close

Вход

Забыли?

вход по аккаунту

?

Cell 9 February 2017

код для вставкиСкачать
Leading Edge
Editorial
Cancer: The Road Ahead
Fighting cancer has been the life mission for generations of biologists and clinicians, and to this effort we dedicate this issue of
Cell. This reviews issue highlights the converging paths in cancer
research that are enabling changes in clinical treatment and
removing obstacles between patients in need and access to
effective care. Empowered by these breakthroughs, it is possible
to imagine major gains against cancer’s relentless advance. In
thinking about the scope of cancer research today, we were
struck by a few themes that transcend the individual topics presented in this issue.
Mapping the Complexity
Cancer is not one disease. For decades, the complexity of its origins, cellular behaviors, and molecular features thwarted researchers’ efforts to grasp its multifaceted biology. It’s different
now. The rapid developments in sequencing technology have
been a game changer for understanding the genetic and transcriptional makeup of cancer at a single-cell resolution. This
granularity permits the reconstruction of the evolutionary path
of cancer genomes, allowing extraction of the underlying principles encoded in the making and breaking of cancer’s DNA.
The complexity in cancer lies not only in its genetics, but also in
its versatile responses to an environment in which it has to
compete to survive and thrive. While we have long known that tumors guzzle fuel pulled from the common reserves, it has been
difficult to directly ask how they acquire and utilize nutrients
in vivo. Studies in model systems are distinguishing the pathways that drive or support tumorigenesis from those that may
be passive bystanders. In parallel, advances in clinical imaging
show how tumors adjust their metabolic profiles to adapt to
local conditions, even exhibiting heterogeneity within the same
tumor mass. Although it remains challenging to effectively target
the heterogeneity in a tumor’s metabolic makeup, a logic is
emerging for how metabolic rewiring might be exploited therapeutically.
Complexity also reigns beyond the boundary of a tumor. There
is an active and dynamic communication between a tumor and
the surrounding microenvironment, including the local immune
components. Probing this interplay led to the identification of
checkpoint blockade approaches that are generating excitement in the clinic. Further teasing out the layered functional interactions between cancer cells and the immune system may open
up new access points that could be targeted. These three—genomics, metabolism, and microenvironment—are only three examples of how our growing capacity to measure, analyze, and
capitalize on the complexity inherent to cancer appears to be
turning the battle to our advantage. Approaches to new treatments are drawing on tumor biology as well as the suite of
host responses.
Translation Goes Both Ways
The concept of ‘‘bench to bedside’’ is widespread in translational
research, but it has mostly been a one-way street: bench science
reveals a vulnerability of cancer cells, small molecules, or antibodies are developed, and then therapeutic value is assessed
in clinical trials. Currently, the outcome of a trial can brand a
translational concept as valid, whereas a failed one can be
seen as nullifying any value of the preceding research efforts at
the early drug development and preclinical stages. However, trials fail for a range of reasons, and the lack of success in one
setting does not necessarily invalidate the promise of a translational idea. Rather, it could mean we don’t know enough of the
underlying biology to effectively design trials for the right patient
populations.
Is there anything we could do differently? The dramatic
success of immunotherapy has propelled discussions of
‘‘reverse translational’’ research. It is clear that even when
we celebrate the achievements of immunotherapy, in many
cases we don’t fully understand why it is effective in some
patients and not others. The large number of ongoing
clinical trials, successful or not, will yield data from which
hypotheses can be generated and tested iteratively in the
laboratory. This cycle doesn’t just move the biology forward,
but also informs trial design and patient selection. Could
there be an opportunity to use the same strategy for other
therapeutic modalities to develop more effective and sustainable cures?
Joining Forces for Patients
The next generation of cures and disease management appears likely to come from combinations of emerging technologies. For example, applying synthetic biology design and
CRISPR-based genome editing to immune cells raises the
hope for a safe and effective class of T-cell-based cancer therapeutics. Likewise, advances in genomic sequencing, high-resolution imaging, and nanotechnology offer new prospects for
early diagnosis. Nowhere is the power of combining technologies more promising than in the design of personalized therapy,
as in cancer immunogenomics, which marries sequencing
technologies with tailored analytical platforms for characterizing neo-antigens and the landscape of infiltrating immune
cells. We are marching into an era of genome-driven oncology
care, and joining together these powerful technologies expands
and advances the gains made through sequencing information
to improve patient outcomes.
Cancer is a research topic, a medical problem, and a social
issue. Barriers still exist between patients in need and effective
care. These go beyond knowledge and technical limitations,
raising many pressing issues. How can academics, commercial
entities, and regulatory agencies work together to make drugs
more affordable? How can we strike a balance between
respecting patient privacy and promoting clinical data sharing
to fulfill the promise of cancer genomics? And, what can we do
to make sure patients with different ethnicities are fairly represented in clinical trials? These may not be the questions that
we ponder daily around the laboratory. Yet, in the fight against
Cell 168, February 9, 2017 ª 2017 Published by Elsevier Inc. 545
cancer, benefiting patients is the ultimate goal, and it is important
to remember that finding targets and developing therapeutics
are not the only tasks we are facing.
In understanding the biology of cancer, we have come a long
way from the ancient concept of a ‘‘creeping ulcer.’’ The ability to
map out the complexities of not only the tumor itself but also the
environment it resides in is expanding the scope of translational
research and drug development from a tumor-centric targeted
view to a system-wide perspective that takes into account
dynamic host responses and tumor-microenvironment communication. Technological innovations are transforming the way we
make biological discoveries, design clinical trials, and deliver
clinical care. Although there is no easy fix when it comes to
cancer, we may be gaining the upper hand in this long battle.
The Cell editorial team is pleased to bring you this special issue
highlighting critical breakthroughs and paths forward in this
moment of cautious optimism. We hope you share our excitement and reasons for hope.
The Cell editorial team
http://dx.doi.org/10.1016/j.cell.2017.01.036
546 Cell 168, February 9, 2017
Leading Edge
Select
Metastasis: Slipping Control
I think that it is our intention to deny cancer any control
over us. —Elizabeth Edwards (interviewed on 60 Minutes)
‘‘And now here it was again, now grown, now in its new
home.’’ In her book Resilience, Elizabeth Edwards recalls
what raced through her mind when she learned that her cancer had come back and had metastasized. After surgery, radiation, and chemotherapy, she thought that the cancer ‘‘had
been chased away.’’
How does cancer find a new home and become metastatic? The short answer is simple: it is complicated. As cancer
forms and develops, some cancer cells manage to be
disseminated from the primary tumor, travel through the
circulatory system, and infiltrate and colonize distant organs,
eventually forming clinically detectable metastasis. Given
the complexity of the metastatic process and technical
and experimental bottlenecks, systematically investigating
metastasis has been challenging, and to this date, metastatic
cancer also remains largely incurable and fatal. Nevertheless,
with the availability of more pathophysiologically relevant
disease models and technological advances in areas such
as imaging and single-cell sequencing, recent studies are
beginning to paint a clearer picture.
Why does cancer come back as metastatic in some patients after the primary tumor has been successfully
removed? The answer might have to do with the timing of
metastatic dissemination. Earlier data from several tumor
types already suggest that, contrary to more conventional
thinking, dissemination can occur rather early during cancer
development, but the mechanism of early dissemination is
unknown. A pair of studies now offers important clues as to
how this happens in breast cancer (Hosseini et al., 2016;
Harper et al., 2016). Using a Her2-driven breast cancer
mouse model, Housseini et al. and Harper et al. provide evi-
Targeting cancer. Image from iStockphoto/Aunt_Spray
dence not only that cancer cells disseminate very early,
before detectable primary tumor formation, but also that
early disseminated cancer cells (eDCCs) have more migration, invasion, and metastasis-initiating capacities (but not
more tumor-initiating capacity) than DCCs from primary tumors at a later stage. Both studies also indicate a critical early
involvement of HER2 in the dissemination of eDCCs, before
its proliferation-promoting role after the primary tumor has
formed. Collectively, these new findings have significant implications and suggest that metastasis might be preventable
if a targetable Achilles’ heel in eDCCs can be identified.
Among the many steps toward clinical manifestation of
metastasis, organ colonization is perhaps the most ratelimiting one given that most arriving cancer cells don’t survive
the new organ’s microenvironment and of the few that do,
some might enter a state of dormancy. By definition, metastasis-initiating cells (MICs) are tumor cells with the capability
to seed secondary tumors in distant organs. Two new studies
dissect the mechanism of metastatic colonization from
different angles. In one study, Pascual et al. set out to ask if
there exists a small population of MICs within tumor-initiating
cells (TICs) that can drive metastasis in oral squamous cell
carcinomas (OSCCs) (Pascual et al., 2017). Using CD44 as
a TIC marker, they identify a subpopulation of CD44bright cells
that are marked by high-level expression of genes associated
with lipid metabolism and metastasis. Among the product of
those genes is CD36, a cell-surface fatty acid receptor
responsible for lipid uptake to provide ATP energy via fatty
acid b-oxidation. Pascual et al. show that CD36 is required
for OSCC metastasis to lymph nodes but has only a slight
effect on primary tumor growth. Feeding mice with a highfat diet or exposing OSCC cells to palmitic acid, a dietary
fatty acid recognized by CD36, also promotes CD36-dependent lymph node metastasis. Importantly, orthotopic inoculation of limiting dilution experiments demonstrates that
CD36+CD44bright cells are metastasis initiating.
In a separate study, van der Weyden et al. use a tour de
force approach to look for contributing factors from the
host microenvironment (van der Weyden et al., 2017). The authors screened 810 mutant mouse lines by using an experimental metastasis assay and identified 23 genes as potential
host regulators of metastatic colonization, many of which
have not been previously implicated in metastasis. Interestingly, mice deficient in the sphingosine-1-phosphate (S1P)
transporter spinster homolog 2 (Spns2) show the greatest
metastatic suppression. S1P is a bioactive lysophospholipid
mediator that has been linked to lymphocyte trafficking. van
der Weyden et al. now report that global or lymphatic endothelial cell-specific deletion of Spns2 reduces melanoma or
colorectal cancer metastasis to the lung by retaining more
anti-tumor effector T cells and natural killer cells in the lung.
As proof of principle, Pascual et al. and van der Weyden
et al. also provide in vivo evidence that CD36 and the SIP1SPNS2 axis could be explored as potential drug targets for
anti-metastasis therapy. In the case of CD36, administration
Cell 168, February 9, 2017 ª 2017 Elsevier Inc. 547
of a CD36-neutralizing antibody in an immunocompetent orthotopic mouse model of OSCC induces strong metastatic
inhibition, and in some cases, complete regression of lymph
node and lung metastases, without toxicity. Although it is still
too early to conclude whether targeting CD36 or SPNS2
might be a clinically viable strategy, these two studies shed
new light on the thus far underexplored roles of lipids in
metastasis. Future studies of lipids, a diversified group of
molecules with a multitude of biological and physiological
functions, might reveal ways in which metastasis will be
losing its control.
REFERENCES
Harper, K.L., Sosa, M.S., Entenberg, D., Hosseini, H., Cheung, J.F., Nobre, R.,
Avivar-Valderas, A., Nagi, C., Girnius, N., Davis, R.J., et al. (2016). Nature. Published online December 14, 2016. http://dx.doi.org/10.1038/nature20609.
, M.M., Hoffmann, M., Harper, K.L., Sosa, M.S.,
Hosseini, H., Obradovic
Werner-Klein, M., Nanduri, L.K., Werno, C., Ehrl, C., Maneck, M., et al.
(2016). Nature. Published online December 14, 2016. http://dx.doi.org/10.
1038/nature20785.
Pascual, G., Avgustinova, A., Mejetta, S., Martı́n, M., Castellanos, A.,
Attolini, C.S., Berenguer, A., Prats, N., Toll, A., Hueto, J.A., et al. (2017). Nature
541, 41–45.
van der Weyden, L., Arends, M.J., Campbell, A.D., Bald, T., Wardle-Jones, H.,
Griggs, N., Velasco-Herrera, M.D., Tüting, T., Sansom, O.J., Karp, N.A., et al;
Sanger Mouse Genetics Project (2017). Nature 541, 233–236.
Xiaohong Helena Yang
Deputy Editor, Cancer Cell
Cell 168, February 9, 2017 549
Leading Edge
Analysis
Getting Data Sharing Right
to Help Fulfill the Promise
of Cancer Genomics
Limited access to the profusion of sequence information derived
from cancer patients worldwide stymies basic research and clinical decisions. Efforts are underway to streamline and safeguard
data use.
Data collected from tens of thousands of
cancer patients have been deposited in
worldwide databases with the hopes that
sharing information will spur development
of new therapies. Despite this collaborative intent, getting access to the data can
sometimes be a struggle, as David Haussler found out while trying to expand his
research in cancer genomics.
‘‘We were rejected on a certain dataset
because we didn’t specify [that] the work
we were doing with it was strictly pediatric
cancer work,’’ says Haussler, who uses
Different studies obtained different sets
of permissions from patients, and these
agreements can impose unexpected
limits on follow-on research. If Haussler
would like to use data deposited in the
database of Genotypes and Phenotypes
(dbGaP) at the National Center for
Biotechnology Information, just writing
the request can be a challenge. ‘‘One of
the big problems with the dbGaP database is that everybody just wrote their
own consent [form] from scratch. Nobody
talked to anybody else about that,’’
‘‘We’ve whittled down from 2,000 patients to ‘here’s the 144
patients that fit your trial.’ Boom. Done.’’
computational techniques to explore the
molecular functions of the human genome
at the University of California, Santa Cruz.
‘‘You couldn’t get the data if you weren’t
restricting your work to pediatric cancer.’’
It’s not that he lacks experience with
sharing data. He’s the director of the
UCSC Cancer Genomics Hub, an online repository that makes available sequencing
data generated by programs of the National Cancer Institute. He co-chairs a
group under the umbrella of the Global Alliance for Genomics and Health (GA4GH) to
standardize genomic and clinical data so
that it can be easily shared among researchers. And way back in 2000, his
team was the first to post publically available human genome sequences on the
Internet.
Yet even Haussler sometimes runs into
one of the typical problems with trying to
gain access to stored data—the fact that
the people collecting the data didn’t take
such sharing into account when they
created their patient consent forms.
Haussler says. ‘‘If I want to analyze 50
different datasets in the dbGaP, I have to
tailor my request to fit the different consent forms. And I have to navigate this
unbelievably byzantine set of different requirements.’’
Getting consent from subjects so that
their information can be used by other researchers is only one of the challenges
involved in sharing data for clinical cancer
research. Data sharing requires guarantees of privacy and data security. There
are also questions of whether data generated by one laboratory are comparable
with those from another. And there’s
concern among some researchers that
making their results freely available will
allow others to get credit for work they
haven’t done. There are even questions
about who should hold patient data and
the practical ways of making it accessible
to others.
These concerns are surfacing because
many people see a lot of potential in
sharing, both for advancing scientific
knowledge and aiding current cancer patients. Understanding rare genetic variations that contribute to cancer requires a
bigger collection of samples than any
one study is likely to generate, and being
able to compare rare variants can help
researchers determine whether or not
they’re disease-causing. Widespread access to information can get patients into
clinical trials faster and help inform doctors’ decisions on how to treat someone.
And having access to other researchers’
results can both aid in validating previous
studies and in generating hypotheses for
new ones.
Indeed, former Vice President Joe Biden
has called for increased data sharing as
part of his Cancer Moonshot, an effort to
double progress in the fight against cancer
over the next five years. His report to the
president, released in mid-October, cites
‘‘a lack of open access and rapid sharing
of research data and results allowing researchers to build on each other’s successes—and failures—to make new discoveries faster,’’ and calls for international
standardization for truly global sharing.
In one attempt at standardization
(Table 1), the GA4GH is developing templates for consent forms. The idea is
to lay out a number of categories and
types of consent that will be broadly applicable to individual research projects and
also allow the data to be shared with other
researchers. Such templates, Haussler
hopes, could make access to and analysis of datasets less complicated.
William Dalton, director of the Moffitt
Cancer Center’s Personalized Medicine
Institute and former CEO of MCC, was
thinking along those lines when he
launched the Total Cancer Care project
11 years ago. When he enrolled patients,
he asked for permission to follow them
throughout their lifetime, collecting their
clinical data, along with blood samples,
tissue from their tumors, and normal tissue. The patients also granted permission
to be contacted again in the future if researchers had questions or felt they could
fit into a new clinical trial. That project
evolved into the Oncology Research Information Exchange Network (ORIEN)
that now includes 13 cancer centers.
The network has gathered data from
more than 130,000 patients, all using the
same standards for acquiring and sharing
data and the same consent form. This,
Cell 168, February 9, 2017 ª 2017 Published by Elsevier Inc. 551
Dalton explains, allows the researchers to
expand far beyond the number of samples that any one cancer center can
acquire, making it easier to find rare genetic variations. The numbers required
are large—a 2014 Broad Institute study
published in Nature estimated that to
find genes that drive cancer in 2% or
more of patients, it is necessary to study
approximately 100,000 tumors.
Armed with this collection of data,
ORIEN researchers can search for patients
who might be eligible for a clinical trial. If a
pharmaceutical company approaches
ORIEN researchers with their criteria for a
trial, it’s easy to search the database for
people who match. ‘‘We’ve done this,’’
says Michael Caligiuri, director of the
Ohio State University Comprehensive
Cancer Center, who cofounded ORIEN
with Dalton. ‘‘We’ve whittled down from
2,000 patients to ‘here’s the 144 patients
that fit your trial.’ Boom. Done.’’
That predictive ability not only makes
the trials more efficient, potentially getting
a new therapy to patients faster; it also
tion, Exchange (GENIE) project contains
clinical and genomic information from
nearly 20,000 patients from eight research
centers. The centers test patients for
known genetic mutations for which there
are existing therapies, but they also
sequence many other genes that may
have mutations for which currently there
is no known treatment, says Charles
Sawyers, chair of the human oncology
and pathogenesis program at Memorial
Sloan Kettering Cancer Center in New
York and chair of the project’s steering
committee. ‘‘We measure them because
we think they’re interesting, but we don’t
know what they are going to tell us.’’ It’s
possible, of course, that the sharing will
also help in clinical decision-making, and
it might help research centers find trial
participants, but GENIE itself does not
have consent to re-contact patients, so
unlike ORIEN, GENIE will not be able to
perform follow-up analyses based on
new clinical data.
An important facet of GENIE is that all
member institutions have agreed to a
‘‘Making data available is more than just taking the data and
throwing it over the wall.’’
helps those patients get into the trial
sooner, which is particularly important
for people with pancreatic or gastric cancer who might only live a few months. ‘‘It
shortened the period of assessment by
weeks,’’ Dalton says.
Each institution in ORIEN is responsible
for taking care of its own patients’ data,
sharing only de-identified information
with other researchers. They can, however, share it with any other researchers
they’d like outside of the network, subject
to a peer-reviewed application process.
Pharmaceutical companies can purchase
subscriptions allowing them to access the
data, and ORIEN is developing a uniform
contract for such subscriptions. When
that’s in place, Caligiuri says, it should
shave a few months off the process of
setting up trials.
By contrast, the focus of the American
Association for Cancer Research’s data
sharing project is aimed not at recruiting
people into clinical trials, but at aiding
research. The first data release from its
Genomics, Evidence, Neoplasia, Informa552 Cell 168, February 9, 2017
common dictionary so that they all refer
to the same conditions and measurements in the same terms to avoid, for
instance, calling the same cancer by two
different names. That kind of effort,
ensuring that data from different sources
lines up so that they can be compared—
knowing, for instance, that one researcher
reporting a variant is talking about the
same thing as a different researcher—is
critical to the success of data sharing.
‘‘Often, people don’t even report the
most trivial things, like what the allele fraction was in a particular sample,’’ says
Louis Staudt, director of the Center for
Cancer Genomics at the National Cancer
Institute. For example, he says, two
studies may identify a tumor as being positive for the KRAS gene, but if one has an
allele frequency of 40 or 50 percent, that
could mean something very different
than one of only 5 or 10 percent.
Staudt is one of the leaders of the
Genomic Data Commons, a $20 million
NCI project at the University of Chicago
designed to store and share genomic
and clinical data from thousands of cancer patients. The GDC, which launched
in June, already contains more than 2 petabytes of data—a petabyte is 1 million
gigabytes—and should easily reach 3 PB
in the coming year, Staudt says. It contains data from large NCI-funded programs such as The Cancer Genome Atlas
and TARGET. In addition to making those
datasets more widely available, Staudt
says, ‘‘We thought, if we made a simpleenough and useful-enough system, people could also upload their own cancer
genetic data that they’ve gotten funded
however from anywhere in the world.’’
To standardize data from those
disparate sources, the GDC runs them
through a common set of analytic pipelines—sets of algorithms used to analyze
the data—so that the data can be
directly compared, and the GDC aligns
them with the latest reference genome.
It also requires users to label their data
in a standard way. ‘‘We do force people
to put all of their clinical fields into a
defined vocabulary,’’ Staudt says.
Databases the size of the GDC bring
logistical challenges. It would take more
than three weeks to download The Cancer Genome Atlas alone. Given that, as
well as concerns about which institution
and which countries have possession,
it’s likely that the data won’t be moved
around in large chunks. For now, Staudt
says, researchers will browse the data
looking for the cancer types or genes or
variants that interest them, then download
only a small fraction of the entire dataset.
Eventually, researchers will probably have
their own analysis software, which they
will send to the database to run computations on the same system that stores the
data, Staudt says.
Agreeing on the significance of data
can also be a challenge. Heidi Rehm,
medical director of the clinical research
sequencing platform at the Broad Institute
in Cambridge, MA, and laboratory director of the Laboratory for Molecular Medicine at Partners Personalized Medicine,
helped develop standard formats for
depositing data into ClinVar, an archive
that collects information about genetic
variation that is clinically relevant in
any disease. Whenever a researcher submits variants to ClinVar, people at the
archive examine the data to make sure it
follows standard nomenclature and can
be mapped to the human genome reference sequence.
The variants in ClinVar and its sister
database, ClinGen, are usually labeled
by the researchers who identify them as
benign or likely benign, likely to cause disease, or of unknown clinical significance.
Rehm and her colleagues found that 11
percent of the nearly 120,000 unique variants in ClinGen had been interpreted by
more than one laboratory, and of those,
nearly 2,300 had their clinical significance interpreted differently by different
labs, sometimes with one researcher labeling the variant benign while another
called it pathenogenic. To deal with such
disagreements, ClinVar implemented a
star-rating system to label how confident
researchers are of a particular variant’s
pathogenicity.
Some of those differences are due to
what was known at the time the variant
was reported. ‘‘The standards to interpret
variants a long time ago were very
different,’’ Rehm says. ‘‘The literature is
rife with false assertions of pathogenicity.’’ Part of the solution is not to simply
accept the claim about the variant, but to
take the evidence gathered from individual experiments and compare it to the
rest of the database. But Rehm believes
that people need to accept that, even if
clinicians are uncertain about the significance of a variant, they can still act on
what they do know. ‘‘It is professional
opinion that guides a lot of medicine,’’
she points out. ‘‘Two different people
looking at the exact same data can have
a different opinion about the certainty.
It’s okay to have disagreements if you’re
all looking at the same evidence.’’
Not only does the biological information
from different sources have to be comparable, but the ways it’s labeled for the
computer have to agree as well. Mark
Musen, a professor of biomedical informatics at Stanford University, runs the
Center for Expanded Data Annotation
and Retrieval (CEDAR), which is supported by the NIH’s Big Data to Knowledge Initiative. CEDAR is building a library
of templates to guide researchers through
the often tedious process of creating
metadata, the set of labels and descriptors that allow computers to search
through a database and find, for example,
all of the data relating to a particular gene
or a particular type of cancer.
Table 1. Acronyms for Data Sharing Projects and Groups
Acronym
Full Name
What is it?
CEDAR
Center for Expanded Data
Annotation and Retrieval
A Stanford University center sponsored by
the National Institutes of Health’s Big Data
to Knowledge Initiative. It aims to improve
the metadata used to label cancer data for
easier retrieval.
ClinGen
Clinical Genome Resource
A National Institutes of Health project to
define the clinical relevance of genes and
variants in various diseases, including
cancer.
ClinVar
Clinical Variant Resource
A partner project of ClinGen that aggregates
information about the relationships between
genetic variation and human health.
dbGaP
The database of Genotypes and
Phenotypes
An archive run by the National Center for
Biotechnology Information that collects
data on the interaction between human
genotypes and phenotypes.
GA4GH
Global Alliance for Genomics and
Health
A coalition of more than 300 health-related
institutions promoting the sharing of clinical
and genomic data.
GENIE
Genomics, Evidence, Neoplasia,
Information, Exchange
An American Association for Cancer
Research project that collects clinical and
genomic information.
ORIEN
Oncology Research Information
Exchange Network
A research center collaborative that shares
a common protocol and provides clinical
trial matching for patients.
TARGET
Therapeutically Applicable
Research
to Generate Effective Treatments
A National Cancer Institute project tracking
the molecular changes driving childhood
cancers.
TCGA
The Cancer Genome Atlas
A joint project of the National Cancer
Institute and National Human Genome
Research Institute that makes maps of key
genetic changes in 33 types of cancer.
The abundance of data sharing projects and associated groups also means an abundance of
acronyms. Here are some of them decoded.
That sounds simple enough, but the
process can be derailed by anything
from typos to using different spellings for
the same units—milliliter versus mL, for
example. Musen says, ‘‘It just boggles
the mind that there are so many different
ways of describing the same thing.’’
CEDAR’s goal is to provide standards
and aids that avoid such confusion.
‘‘Making data available is more than just
taking the data and throwing it over the
wall,’’ Musen says.
While NIH guidelines require researchers to share data, journals are also
encouraging the practice by requiring the
data behind papers to be made available,
sometimes in specific databases. Rehm,
who is an editor at Cold Spring Harbor Molecular Case Studies, persuaded that journal to require that variant data be deposited in ClinVar, and she’s talking to other
journals about doing the same. Like many
journals, MCS also requires sequencing
and phenome information to be deposited
in appropriate databases. In January, the
International Committee of Medical Journal Editors proposed that its members
require researchers to share patient data
within six months of publishing a paper
based on that data. That’s in line with the
GDC’s time frame, although Staudt says
in practice, the process of applying for access to the data will means it will probably
be a year before others can use them. Jeffrey Drazen, editor-in-chief of the New England Journal of Medicine and a member
of ICMJE, says his journal will adopt the
requirement once there is a network in
place for doing that sort of sharing.
Some researchers, though, have expressed resistance to sharing data, at
least quickly, out of fear of what some
Cell 168, February 9, 2017 553
call ‘‘research parasites’’—people who
profit off others’ work by running analyses
of their datasets and writing papers about
them. In August, a group of 282 investigators calling itself the International Consortium of Investigators for Fairness in
Trial Data Sharing published a commentary in NEJM calling on the ICMJE to allow
exclusive use of data for at least two
years. They argue that one of the incentives for researchers to participate in
large, multi-center trials is the opportunity
to write secondary papers based on those
trials and that making the data available
too quickly would put those researchers
in competition with others who didn’t
contribute to the research, thus reducing
the incentive. The requirement might
also lead some people to delay publishing
554 Cell 168, February 9, 2017
their first paper until they had time to write
follow-up papers, the group said.
Drazen agrees that there may need to be
changes in the incentives that drive researchers. That’s something the research
community still needs to work out, he
says. Staudt says that having access to
shared data makes everyone’s work
more productive. ‘‘No one lab can
generate all the gene expression data
they need to fully understand their own
data,’’ Staudt says. ‘‘This kind of sharing
has accelerated research, I would argue,
and it’s the right thing to do because the
primary goal is to help people with cancer.’’
Sawyers doesn’t believe that generating papers from other people’s data in
GENIE will be much of an issue. He also
thinks objections to sharing data are
fading as people come to understand
how much it can advance science. ‘‘The
further along we get, the less resistance
there is to making it available.’’
Moffitt’s Dalton says it’s important to
keep in mind the reasons for sharing
data in the first place. ‘‘Sometimes, we
lose sight and think of data sharing as
the end point. That’s not the end point,
in my opinion,’’ he says. Rather, the point
is to create scientific collaborations that
have the resources to advance science
so that it can aid in the fight against cancer for individual patients and benefit
society as a whole.
Neil Savage
Lowell, MA, USA
http://dx.doi.org/10.1016/j.cell.2017.01.003
Leading Edge
Bench to Bedside
PDGFRA Antibody for Soft Tissue Sarcoma
Lillian R. Klug and Michael C. Heinrich
Portland VA Health Care System and OHSU Knight Cancer Institute, Portland, OR 97239, USA
Correspondence: heinrich@ohsu.edu
http://dx.doi.org/10.1016/j.cell.2017.01.028
NAME
PDGF
A A
A B
Olaratumab (Lartruvo), originally IMC-3G3
C C
B B
APPROVED FOR
Soft tissue sarcoma (for which doxorubicin is appropriate)
Olaratumab
TYPE
Fully humanized monoclonal antibody (IgG1)
MOLECULAR TARGETS
P
α α
P
P
β α
P
α
α
x
PDGFR
Tumor cell
Stromal fibroblast
Vascular endothelial cell
Pericyte
β
PDGFRA, extracellular domains
α
CELLULAR TARGETS
x
Tumor cells, tumor-associated fibroblasts/stromal cells,
endothelial cells, and pericytes
Tumor cell growth
Angiogenesis
EFFECTS ON TARGETS
Inhibits ligand binding and resultant kinase activation
Lartruvo (olaratumab) is a monoclonal antibody against
the extracellular domain of PDGFRA. Olaratumab blocks
ligand binding and thereby inhibits activation of PDGFRA
kinase activity. Pre-clinically, this antibody inhibited
PDGFRA-dependent tumor growth. In a randomized
Phase II study, adding olaratumab to doxorubicin chemotherapy significantly improved overall survival, leading to
FDA approval.
DEVELOPED BY
Eli Lily
Benefit in median
overall survival
Incidence of Soft Tissue Sarcoma
0.5%
of all cancers are
soft tissue sarcomas
With more than
50
1984
Viral oncogene v-sis
from simian sarcoma
virus is PDGF-B
1989
PDGFRA cloned
and mapped
to human
chromosome 4q
1984
12,390 new cases diagnosed in 2016
Most common
forms include:
14.7 months
Undifferentiated
pleomorphic sarcoma
Liposarcoma
subtypes
57% 43%
Male
Leiomyosarcoma
2005
Human PDGFRA
neutralizing antibody,
3G3, has efficacy
in vitro and in vivo
against glioblastoma
and leiomyosarcoma
1993
xenografts
PDGFRA neutralizing
antibody inhibits
PDGF signaling
1992
PDGFRA signaling
promotes cell growth
in human tumor
cells in vitro
1995
Doxorubicin chemotherapy
Adding olaratumab
Extended by
11.8 months
Female
2006
Imatinib mesylate approved
for treatment of a rare
form of soft tissue sarcoma
driven by a secreted
COL1A1-PDGF-B
fusion protein
2010
PDGFRA expression
correlates with poor
survival in widely
resected soft
tissue sarcoma
2009
In Phase II trial, pazopanib
(VEGF/PDGFR inhibitor)
is effective for treatment
of metastatic soft
tissue sarcoma
2005
2010
2016
Open-label Phase IB and
Phase II with olaratumab +
doxirubucin show survival
benefit compared to
doxorubicin alone
2014
Phase 1 trial of
olaratumab in
solid tumors
(did not include
soft tissue sarcoma)
2015
References for further reading are available with this article online: www.cell.com/cell/fulltext/S0092-8674(17)30113-7
Cell 168, February 9, 2017 Published by Elsevier Inc. 555
Leading Edge
Stories
Poisoning the Devil
Zhu Chen and Sai-Juan Chen
Our sight was caught by the subject line of an email we received on March 12, 2016: ‘‘American
Society of Hematology (ASH) Ernest Beutler Lecture and Prize.’’ It came from Charles Abrams,
President of the ASH Society, congratulating Zhu Chen for being the recipient of the 2016 Ernest
Beutler Lecture and Prize together with our close collaborator Hugues de Thé for their contributions to finding a cure for acute promyelocytic leukemia (APL). More than a decade ago, Beutler,
ex-president of ASH and a respectable pioneer in hematology, had visited our institute and was
very encouraging and supportive of our work. Thus, being awarded with this prestigious prize
meant a lot for us. Later, as we prepared for the lecture, our thoughts went back to another
ASH Annual Meeting, one that happened 20 years ago, where Zhu presented our findings on
how arsenic, a notorious poison, can be used to treat APL.
Most people have a hard time thinking of arsenic as a cure of any kind.
‘‘It is tricky to practice
Indeed, the association that comes to mind is the poisoning death of historical
the idea of ‘fighting
and literary characters such as Madame Bovary and Napoléon Bonaparte.
Arsenic acts by its high affinity for the sulfhydryl groups of bioenzymes,
poison with poison,’ but
resulting in various harmful effects through inhibiting the activities. However,
if you do it right and
it is also one of the oldest medicines in the world. It was first mentioned
by Hippocrates (460–370 BC), who used realgar and orpiment pastes to
carefully, you may just
treat ulcers in western medicine. In line with the Chinese philosophy, ‘‘fight
achieve the most thrilling
poison with poison,’’ arsenic was also used for treating diseases in ancient
China. Arsenic pills for the treatment of periodic fever were recorded in
success: saving lives.’’
the Chinese Nei Jing Treaty (263 BC). Hong Ge (284–364 AD), a Taoism alchemist, used arsenic (As4S4) as a disinfector and obtained pure arsenic from
the compounds by heating, documented in his work Baopuzi Inner Chapter. Si-Miao
Sun (581–682 AD) purified a medicine composed of realgar, orpiment, and arsenic in treating
malaria, whereas Shi-Zhen Li (1518–1593 AD) in the Ming Dynasty described the use
of arsenic as a remedy for a variety of diseases in his pharmacopedia. The exploration of
using arsenic to treat leukemia started in the 1970s in China. At the time, Tai-Yun Han, a pharmacist of the First Affiliated Hospital of Harbin Medical University, happened to learn that a
folk remedy composed of arsenic trioxide (ATO), mercury chloride, or toad venom had
some effects on some cancer cases. He then made solutions with the same components
and named them Ailing Solutions. His colleague, Ting-Dong Zhang, explored the effect of
the solution containing ATO and mercury on myeloid leukemia patients and observed some
promising effects. However, as the mode of action of ATO was not clear, the regimen was
not widely applied.
APL, also known as the M3 subtype of acute myeloid leukemia (AML-M3), accounts for 10% of
all AML cases and has a very severe natural course. APL was once one of the most lethal forms
of leukemia because of the heavy burden of leukemia blasts, in which myeloid differentiation is
blocked at the stage of promyelocytes, and because of the presence of hemorrhagic syndrome.
In the 1970s, the combination of anthracyclines and cytosine arabinoside was adopted as the
mainstream treatment, but there was high early mortality due to severe hemorrhage often exacerbated by chemotherapy. APL treatment took a turn for the better in the early 1980s, owing to
the clinical application of all-trans retinoic acid (ATRA) by Zhen-Yi Wang’s team at Shanghai
Institute of Hematology, and the drug could result in a complete remission (CR) rate of over
90%. We were very proud to contribute to this breakthrough while pursuing our Master’s degree
in Wang’s team. In 1980, Professor Zhen-Yi Wang and Zhu Chen set the following objectives:
identify factors promoting leukemic cell differentiation; explore treatment of leukemia with regulatory mechanisms other than chemotherapy. Led by Prof. Wang, our team screened a group
of drugs and finally discovered the effect of ATRA in breaking the differentiation blockade in
leukemic cells. Later, we further explored the pathogenesis of APL and the mechanisms of
ATRA with cytogenetic and molecular biological techniques. While it was exciting to successfully apply the notion of differentiation induction to treat leukemia for the very first time, we
Saijuan and Zhu in Saint Louis Hospital of Paris during their Ph.D. training in the 1980s.
Cell 168, February 9, 2017 ª 2017 Elsevier Inc. 557
Saijuan and Zhu at Shanghai Institute of Hematology in 1997.
soon realized that the problem was not yet solved. Half of the patients receiving ATRA and
chemotherapy showed relapse and became resistant to ATRA treatment.
The devil was resilient, but how should we strike back? At a conference held
‘‘The healing effect of
in 1994, Sai-Juan learned about the earlier attempt of treating leukemia, APL in
poisoning the devil
particular, with ATO from colleagues in Harbin Medical University. This was
intriguing, and we almost immediately thought of the possibility of using ATO to
could go a long way.’’
combat ATRA resistance. The earlier experience of Zhu as a barefoot doctor in
the countryside, during which he made the most of limited, and often not ideal,
resources to treat patients, spoke to him that this was the right direction. Meanwhile, as hematologists received scientific training in both China and Europe, we realized it was a bold idea given
the potential toxicity, and rigorous study of its mechanism and efficacy is the only way to move
forward. Soon, we established collaboration with Ting-Dong Zhang and got pure ATO solution
from a spin-off company of Harbin Medical University. We then launched a systematic study
on the cellular and molecular mechanisms of the effect of arsenic in APL. We found that ATO
potently induced APL cell differentiation and apoptosis by targeting PML-RARa, the oncogenic
driver specific for APL. We also discovered that this effect resulted from a selective action
of arsenic on the PML, but not RARa moiety, of the fusion protein, and our follow-up studies
identified the exact interacting site, structural changes of the oncoprotein upon interacting with
558 Cell 168, February 9, 2017
Saijuan and Zhu in their own laboratory.
arsenic, and the downstream effect triggering protein degradation that leads to APL cell differentiation and apoptosis. These mechanistic studies paved the way for clinical application of arsenic
therapy.
In 1995, we conducted the clinical trial of ATO in a series of APL patients relapsed after ATRA
and chemotherapy. Among the initial 10 cases treated with ATO alone, CR was achieved in
nine. We also demonstrated that the use of arsenic was relatively safe from the pharmacokinetic
aspect. Next, we applied ATO on newly diagnosed APL patients. In the next three years, we
were able to achieve CR in > 70% of treatment-naive patients and > 85% of relapsed patients.
By using PML-RARa transcripts as a biomarker, molecular remission was confirmed, along
with phenotypic remission. Our hard work at both the basic mechanistic and the clinical fronts
paid off. In August 1999, ATO was approved as a new therapeutic reagent for treating APL by
the State Food and Drug Administration of China.
‘‘Looking back to our
Knowing that cancer cells are hard to eradicate, we have been following up on
scientific career, we
the clinical application of ATO over the years. Accumulating clinical data sughave benefited a lot from
gested that ATO monotherapy could achieve a relatively long-term survival in
only a part of cases, whereas a considerable number of patients would relapse
the integration of
again, which showed that arsenic alone, just as ATRA, is not a panacea for APL.
Inspired by our earlier clinical observation back in 1994–1998 that 5 out of 5 cases
western and eastern
treated with both ATO and ATRA obtained CR, we started to explore the underwisdoms.’’
lying mechanisms that might explain and further support the potential synergy
between the two treatments. It is worth mentioning that our long-term collaborator
Hugues de Thé and his team reported that combining ATRA and ATO could specifically target
leukemia-initiating cells. In the meantime, we discovered that ATRA mainly relieves transcriptional
repression, whereas ATO modulates protein network. These two layers of regulation result in cellcycle arrest and differentiation. A striking converging point is the degradation of PML-RARa, with
ATRA targeting the RARa and ATO targeting the PML moieties.
Cell 168, February 9, 2017 559
Encouraged by these insights, in 2000, we launched a clinical trial with the combination of ATRA
and ATO as the front-line therapy—in which ATO was originally used only for relapsed patients—
for newly diagnosed APL, and we found in 2004 that among the 20 patients receiving the combinatory therapy, the durable CR rate with a median follow-up time of 18 months was 100%,
whereas in control groups receiving ATRA or ATO treatment alone, several patients relapsed. In
view of these data, we extended the combination of ATRA+ATO based therapy to all newly diagnosed APL patients in the next 5 years. In 2009, we reported the exciting clinical follow-up results:
5-year relapse-free survival rate of 94.8 ± 2.5% and overall survival rate of 97.4 ± 1.8% in patients
who achieved CR. A subsequent multi-center study on 535 newly diagnosed APL patients led by
our team confirmed an over 90% 5-year disease-free survival rate in APL patients. The dramatic
effect of this synergistic targeted therapy was confirmed by hematologists and oncologists
around the world, including the Australia APML4 protocol led by Iland and the Italian/German trial
led by Prof. Lo Coco, both of which yielded impressive outcomes in ATRA+ATO based therapy for
APL. Indeed, in 2014, ATRA/ATO synergistic targeted therapy was recommended by the USA
National Comprehensive Cancer Network (NCCN) as the first choice for APL treatment. The
Italian/German group very recently found that long-term survival can be achieved in low- and intermediate-risk APL cases without chemotherapy, and the same results were also observed in the
latest APL2012 trial in China.
Looking back to our scientific career, we have benefited a lot from the integration of western and
eastern wisdoms. The Chinese philosophy has taught us how to think in a dialectic way, particularly in exploring the therapeutic value of a poison. It is tricky to practice the idea of ‘‘fighting poison
with poison,’’ but if you do it right and carefully, you may just achieve the most thrilling success:
saving lives. On the other hand, modern analytical thinking and the appreciation on the power of
ever-advancing technology from the western scientific training have enabled us to take a close
look at the molecular movement of the devil so that we could make informed predictions each
step of the way, confidently embracing the positive outcome and breaking new grounds. Indeed,
our recent application of a systems biology approach allowed the discovery of hundreds of potential target proteins for the action of ATO, including key players in the development of solid tumors.
We are optimistic that the healing effect of poisoning the devil could go a long way, and we are just
at the start of the journey.
560 Cell 168, February 9, 2017
Leading Edge
Conversations
A Convergence of Genetics and Epigenetics in Cancer
What is at the forefront of the intersection of genetics and epigenetics in cancer and how do we use
what we’ve learned to devise new cures? These are the questions Cell editor Jiaying Tan posed to
Jan Korbel and Charles Roberts. Annotated excerpts from this conversation are presented below,
and the full conversation is available with the article online.
Jan Korbel
EMBL
Jiaying Tan: I am curious how you see where [the cancer] field
has been and where it is going in terms of epigenetics and
chromatin?
Jan Korbel: For me, this will be more [from] of an outside
perspective, because I’m pursuing more the genetic side than
the epigenetic side, but I do notice in our own data that we often
find interesting links to epigenetics and chromatin modeling,
and Charlie [referring to Charles Roberts] gave a splendid talk
yesterday about how he and others have done that as well, are
following up on very specific mutation pattern or very specific
complexes, how they are mutated, and then have an effect on
the epigenome. I personally see the epigenetic cancer field, if it
exists, fairly nicely coming together with the genetic field by
exploring links between genetic mutations and trying to follow
up mechanistically how they can affect the epigenome .. Given
that a lot of very striking mutations we see in cancer genomes
are mutations affecting epigenetic genes and that when
studying these we can learn more about how arresting a certain
state and certain differentiation state, which then generates an
epigenetic pattern, can contribute to cancer, I see both niches
are coming more closely together right now, which is great.
Charles Roberts: I think it’s, as you say, a drive toward
mechanism. As I look back, when I first started working on
chromatin remodeling in cancer, and I went to my first meeting
that was dedicated to epigenetics, I was very disappointed,
because everything at the meeting was on DNA methylation.
The chromatin hadn’t really been recognized yet, so there was
literally zero other people there working on that type of
chromatin or real-life genetic regulation. Then, as the mutations
Charles Roberts
St. Jude Children’s Research
Hospital
have become known, people more and more started to get into
it, and initially the naive view [was], if it’s ‘‘epigenetic’’ it must be
HDAC inhibitors, and not quite the realization of just how broad
and diverse epigenetic or chromatin contributions are to
cancer. I really think [of a] two-pronged approach. One is
mechanism, really understanding why those mutational things
that are broad transcriptional regulators drive cancer. What are
the mechanisms that are underlying that? Then at the same
time, trying to think about patients. Can we do something for
today’s patients therapeutically?
JT: You mentioned a really interesting point, Charlie. There’s
a lot of questions about how HDAC inhibitor [are] actually
targeting. As one of the speakers mentioned yesterday, this
whole idea of reverse translational research—seeing something
in a patient and going back to try to really figure out how it works.
Do you think that concept has been applied in terms of targeting
epigenetics, or drugs and epigenetics in the cancer field?
CR: I think precisely that has been followed. For us the clue
was in 1997 or 1998, when Olivier Delattre published these
mutations present in this rare pediatric cancer, in an ATPdependent chromatin remodeling complex, and wondering
why an ATP-dependent chromatin remodeling complex looks
like it’s behaving as a tumor suppressor. We jumped in, armed
with also the understanding that often mutations in early onset
pediatric cancers can be a harbinger of important processes in
cancer much more broadly. Then, beginning in 2010, the tidal
wave breaks, and all of these indications of chromatin
regulators begin to be recognized as the human cancer
genome project really came to fruition. Suddenly, the
Cell 168, February 9, 2017 ª 2017 Published by Elsevier Inc. 561
‘‘Suddenly, the realization,
‘‘wow!’’ these things are
prevalent.’’
realization, ‘‘wow!’’ these things are prevalent. Bert Vogelstein
had said at one of the AACR meetings, not much new was
learned from the cancer genome sequencing, except for the
frequencies, and then suddenly the next year pointed out what
we didn’t see was the chromatin regulators. We just didn’t
know about them, so suddenly a new front to the war in
understanding cancer.
JK: I fully agree, and I’m going to take a slightly different
angle in my answer using the question to point out our own
thoughts of how important mechanisms are to understand
genetic and also epigenetic alterations. We’ve been looking
very closely to what extent the [structural] rearrangement of
enhancers can contribute to oncogenic overexpression. What
we and some others (Rick Young’s group for instance) see [is]
very strong epigenetic and chromatin effects, and very local, at
the topologically associating domain where these
rearrangement occur. Studying the mechanism here is
extremely helpful. We don’t immediately have a drug at hand,
but now that we understand, for instance, at the IGF2 locus that
a structure akin to a new topologically associating domain is
found with a super enhancer, this gives us a better angle [for
thinking] about treatment, for instance, inhibiting the super
enhanced gene interaction.
JT: I think it’s very interesting [that] the definition of
epigenetics has been evolving. Initially it was something not in
the sequence, then it was very heavy on modification, and now
it’s seamlessly merged into the whole idea of chromatin
structure. This [reflects] how the field has been moving from
more a granular level to more an overview in terms of how the
genome is really organizing itself and trying to function.
JK: Yes. clearly epigenetics is more than DNA modification,
and it should be.
CR: I think Jan put it very well that really epigenetics is an
interaction of genetic mutations with transcriptional regulation.
If we back up and look where the field has been, in the 1980s
the cutting edge of genomics was looking at chromosomal
translocations in leukemia, and what those essentially are turns
out [to be] oncogenes being translocated into the T cell or B cell
receptor loci. Essentially, that is ‘‘enhancer hijacking’’ that’s
driving those. Those have some of the most potent
transforming abilities. If you put those into mice, and turn on
Myc, [you get] extremely rapid cancer onset, and you get these
genotypes in humans that are very simple. To me, as we look
forward and see, for instance, with the rhabdoid tumors when
you mutate SWI/SNF chromatin subunit, [you] also [get] very
simple genomes. It distills a message that transcriptional
regulation, in my view, is what cancer in many ways is about,
and that the genome instability is an inefficient way to get there.
562 Cell 168, February 9, 2017
It’s a necessary way for many cancers, in order to rack up the
mutations that you need, but it’s really about getting to
transcriptional dysregulation and if you directly dysregulate
transcription in the right ways, you can actually see cancer
evolve in very simple genomes. It’s a fascinating interplay
between the two.
JT: Yesterday there was a talk comparing adult cancer and
pediatric cancer. [There is] a question of whether the ultimate
differences between them is because of the time that you have
to accumulate mutation. Apparently that’s not the full story.
Jan, have you thought about whether there’s some principle
you can learn when you look at those very distinct phenotypes
at different human ages?
JK: These childhood cancers are mostly rising in the context
of cells that are dividing very rapidly, and they’re programmed
to divide very rapidly, because in children cells need to build up
the body, they need to build up our main organs . in my mind
the pediatric cancers are mostly cancers where this program of
rapid expansion in a non-differentiated state is not arrested as it
should be. This appears to require relatively few mutations,
which might be the case that the cells actually have many
properties of what we typically consider to be the purpose of a
tumor, once these mutations act. In adult cancer we see many
more mutations arising. We see that the disease peaks much
later in age, and it is often strongly driven by environmental
influences, like tobacco smoke and UV light in lung and skin
cancer, and hence for cells that are not programmed to expand
that rapidly, may require many more mutations to convince
them to actually start growing again, and to develop the
properties of a tumor, to de-differentiate again.
CR: I think that’s likely to be precisely it . high population,
high frequency of cells, [which] already have many of the
transcriptional pathways ‘‘ON’’ for hyperproliferation, and so
what’s necessary to actually push them over the edge is much
less. Just as Jan said, in the adult cancers, a much lower
frequency, and the cells that are there are often more restricted
in their abilities and likely more mutations, therefore, are
needed to drive them. And yet there are still core themes that
can come through in either case. The transcription factors that
are involved in both pediatric and adults, or for example the
SWI/SNF complex that we work on, is very clearly mutated in
both pediatric and adult, [and there is] a shared theme of certain
pathways or mechanisms, but a difference in what’s needed to
actually get there.
‘‘. transcriptional regulation,
in my view, is what cancer in
many ways is about, and that
the genome instability is an
inefficient way to get there.’’
(L to R) Charles Roberts, Jiaying Tan, and Jan Korbel at the Cell Symposia ‘‘Hallmarks of Cancer’’ in Ghent, Belgium.
JK: I found pediatric cancers very interesting because they
contain so few mutations. We found it easier to essentially look
into certain principles.
JT: It has less noise, so to speak, and bystander phenomena
that accumulate over the time. One more question. From your
own point of view, what’s next big question that you want to
tackle?
JK: I’ll start with the immediate next question we’re
tackling. We’re currently putting this fairly large database of
cancer genomes together .. We’ve collected approximately
2,800 patients with matched tumor and normal whole
genomes. It’s a collaborative work between the International
Cancer Genome Consortium and it’s subprojects including
TCGA, and we [are] specifically exploring principles of how
can we learn from what’s happening in our hereditary genome
and the types of mutations we see in the somatic genome.
This hasn’t been explored so far yet. I think one of the reasons
is that we do carry many germline variants in our genome,
between three and four million, so we need fairly large sample
sizes to make sense of these heritable variances in a
statistical manner. We lack power for this. One of the themes
we’re currently investigating is can we connect the germline
variation world with the somatic variation world. There’s
indeed some fairly interesting links between both that goes
beyond high-risk cancer genes such as BRCA1 and BRCA2,
which predetermine to some extent how the tumor will
mutate. There’s more to come here. That’s one of the things
we are doing in the immediate future.
CR: It’s become clear how frequently chromatin regulators
are mutated in cancer, SWI/SNF complex is in over 20% of all
cancers. [We want to] fundamentally understand the
mechanism of why this is happening, always with the goal
toward, does that enable us to therapeutically target it? At the
first level, we have identified antagonism with the polycomb
complex, and it’s been exciting to see how rapidly that has
moved into the clinic. Now though, I think our understanding is
still relatively rudimentary. If we think about this, p53, Ras, and
Myc, we’ve known about for 3 to 4 decades, the involvement of
chromatin regulators, in many cases, 5, 6, 7 years, so there’s a
lot to be discovered with respect to mechanism that we hope,
yet to be proven, can inform new ways to go about therapy.
Cell 168, February 9, 2017 563
Leading Edge
Voices
Where is the Future of Drug Discovery for Cancer?
With both small molecules and biologics succeeding in trials and in the clinic, the scope of drug
discovery in cancer is changing. We asked a group of researchers to share their visions for how
to identify new targets and how to approach taming them.
Cancer Metabolism Games
Designer Proteins as Cancer
Therapeutics
Giulio Superti-Furga
Growing the Drug Target Space
Craig M. Crews
Research Center for Molecular Medicine of the
Austrian Academy of Sciences
Jennifer Cochran
We have known for a long time that cancer cells
adopt metabolic states that fit their growth
impetus and reflect their relinquishing of tissue
homeostasis, yet the degree and variety of
ways by which metabolic networks are rewired
in tumors continues to surprise us. What has
received less attention is the interplay between
the metabolism of tumors and other cells in
the microenvironment. Metabolism is not cellautonomous; instead, it reflects an obligatory
dialog between tumor cells and the surrounding
tissue. We are increasingly appreciating that
immune cells are also profoundly affected by
metabolism, including nutrient, metal, and
oxygen levels. These insights highlight a potential innovative therapeutic target: the integrated metabolic space of tumor, stromal, and
immune cells, where cells must compete for
nutrients or enter mutually advantageous dependencies. An opportunity now exists to
alter nutrient traffic to draw in and activate the
right immune cell types and to disadvantage
cancer cells. But how to attempt this? The solute carrier and ABC membrane transporters
are responsible for influx and efflux of nutrients
and metabolites. These transporters are differentially expressed in different cell types and
respond to environmental supply and internal
demand. They are also exquisitely druggable.
By targeting transporters, perhaps in combination, we may be able to subtly and safely
turn the tables in the cancer metabolism
game in favor of immune cell well-being and
cancer cell starvation for the re-establishment
of healthy tissue homeostasis.
Monoclonal antibodies dominate the modern
pharmaceutical industry. These agents have
achieved clinical success, led by recent excitement about arming them with chemotherapeutic agents for targeted drug delivery or
interactions with the immune system. Despite
these advances, challenges in the field remain,
including how to best tackle tumor and patient
heterogeneity, rapid drug resistance, and issues with effective tumor penetration and delivery across the blood-brain barrier.
Advances in our understanding of disease
pathophysiology and the development of
rational and combinatorial technologies for
creating protein-based biologics are spawning
drug candidates with improved therapeutic
and safety profiles. We now have ‘‘multi-specific’’ proteins that target and modulate several
key biochemical pathways and ‘‘multi-epitopic’’
proteins that bind different locations of the same
target for improved efficacy. Researchers are
also exploring peptides and so-called ‘‘alternative scaffolds’’ that are modular like antibodies,
but evoke potential benefits such as enhanced
tumor penetration. Along with these elegant approaches come development, manufacturing,
or regulatory hurdles, but also new opportunities for impactful cancer treatments.
Clinical trials are increasingly combining targeted therapies or coupling them with more
traditional modalities such as chemotherapy
or radiation to address multiple facets of cancer. While these approaches bring increased
costs and questions about toxicity, they are
proving highly effective and are poised to offer
new standards of care.
Yale University
Stanford University
564 Cell 168, February 9, 2017 ª 2017 Published by Elsevier Inc.
Greater than 20% of industrial cancer drug
development programs focus on just eight proteins—sadly ironic in this post-genomic era,
when 20,000 possible proteins are known.
While many potential drug targets are enzymes,
it is clear that non-enzymatic proteins also play
key roles in cancer biology. Currently, these
structural and regulatory classes of proteins
appear ‘‘undruggable,’’ since they lack a catalytic site for small-molecule inhibition. This
unsuitability is especially applicable to transcription factors, which regulate gene expression via protein complex formation. Given
these challenges, how can one make these
proteins pharmaceutically vulnerable? RNAi
and CRISPR offer some hope via preventing
oncogene expression. However, their clinical
potential has not been fully realized due to
challenges with cost, delivery, and off-target
effects. Clearly, new approaches are needed
to identify modulators of protein expression
(and thereby, function). Ideally, these approaches should be small molecule based,
should possess favorable pharmaceutical
properties, and should have the potential to
target all proteins, irrespective of protein class.
One emerging approach to target the ‘‘undruggable’’ proteome is the use of small molecule
proteolysis targeting chimera (PROTACs) to
induce the deliberate degradation of specific
proteins by the ubiquitin/proteasome system.
By co-opting the normal cellular quality control
machinery responsible for removing unwanted
proteins, all classes of proteins could be
controlled using small molecules, greatly expanding the number of ‘‘druggable’’ protein
targets.
‘‘Drugging the Unprecedented’’
Probing Epigenetics
Not Undruggable, but #YTBD
Stephen Frye
Gitte Neubauer and Rab Prinjha
Kevan Shokat
University of North Carolina
GSK, Epigenetics DPU and Cellzome, a GSK
company
University of California, San Francisco
The concept of the ‘‘druggable genome’’ has
enumerated the human proteome’s potential
to yield new medicines. ‘‘Drugging the undruggable’’ seeks to expand this target space and
address the many proteins implicated as
therapeutically relevant that fall outside the
druggable genome, e.g., protein-protein interactions (PPIs). While the phrase, drugging the
undruggable, is aspirational, it overstates the
resilience of the original classification of
druggability toward scientific progress. Rather,
potential intervention points for small molecule
ligands are either precedented or unprecedented. Indeed, over the last two decades,
protein kinases have progressed from unprecedented to become a protein family with
28 FDA approvals. While precedent is retrospective, some attempts have also been
made to prospectively analyze the ‘‘ligandability’’ of the proteome by structural and computational methods. These analyses deem unprecedented proteins as either easier (kinases
in 1990) or more difficult (many PPIs today),
but they suffer from an inability to anticipate
induced-fit-binding modes. Given that our current understanding of druggability is strongly
colored by history (‘‘all experience is an arch
wherethrough, Gleams that untraveled world,
whose margin fades, Forever and forever
when I move;’’ ‘‘Ulysses,’’ Alfred, Lord Tennyson) and limitations of computation, efficient
experimental approaches are needed. Fortunately, unbiased assessments of druggability
using the tools of chemical biology and quantitative proteomics are emerging. An experimentally determined druggable genome is within
our grasp.
The discovery of novel medicines is a daunting
task, even more so when entering a new area of
biology, as was the case when GSK started its
early investment in epigenetics. Which of the
epigenetic players that determine or interpret
the histone code in response to environmental
cues offer hope for therapeutic interventions?
In order to tackle these questions, chemical
biology combined with proteomics provided a
vital toolbox. At the outset, a cellular screen
for compounds modulating target gene expression led to biologically active small molecules,
likely affecting epigenetic regulation. We used
such compounds to isolate and identify by
mass spectrometry a new target class, the
BET bromodomain family of epigenetic regulators, that would be tractable for small molecule
inhibition. But this was just the start: in order to
understand function and full therapeutic potential, we characterized the mega-dalton protein
complex surrounding BET proteins, using a
combination of immuno-affinity and chemoproteomic approaches. The unexpected presence
of BET proteins in distinct complexes closely
associated with proteins that are commonly
mutated or trans-located in certain leukemias
pointed to therapeutic indications for BET inhibitors, which are now being tested clinically.
Drug candidates targeting epigenetic regulators hold tremendous promise, but much of
the biology of epigenetics and exciting therapeutic potential in indications beyond cancer
is still to be discovered and is being aided by
modern technologies on the flourishing interface between chemistry and biology.
Insofar as drug discovery is concerned, we are
in the post-cancer-genome era, where we
know the main driver oncogenes. Drugs targeting drivers of common cancers (B-Raf, PIK3CA,
androgen receptor, estrogen receptor, Bcl-2,
etc.) have been approved or are in late-stage
clinical trials. What remain are some wellappreciated oncoproteins like K-Ras and
c-Myc that are crucial in a range of tumors but
lack obvious small molecule binding sites.
These targets are ‘‘yet to be drugged’’ (YTBD),
but that status may be fleeting. Recent success
in discovery of lead compounds for one allele of
K-Ras (G12C) might herald the emergence of
additional drugs for other K-Ras alleles. The
most important ingredient for drugging these
kinds of targets is a long-term commitment by
the community to better understand the key
drivers biochemically, structurally, and functionally and to couple this understanding with
creative chemical approaches to blocking the
targets’ function. BCL-2 was considered undruggable until a decade’s worth of experimentation and application of fragment-based NMR
screening altered and broadened our views
of what drug-like molecules could look like.
However, with major oncogenic drivers known,
we need to ask if we’re out of good targets. I
don’t think so. Components of ‘‘housekeeping’’
cellular machines are underappreciated targets, including those responsible for mRNA
splicing and translation that are hijacked by
cancer cells to selectively promote cancer formation. Our challenge is how to drug these
cellular machines in a manner that blocks their
cancer-specific functions.
Cell 168, February 9, 2017 565
Leading Edge
Commentary
Translating Germline Cancer Risk
into Precision Prevention
Matthew B. Yurgelun,1 Georgia Chenevix-Trench,2 and Scott M. Lippman3,*
1Dana-Farber
Cancer Institute and Harvard Medical School, Boston, MA 02215, USA
Berghofer Medical Research Institute, Brisbane, QLD 4029, Australia
3Moores Cancer Center, University of California San Diego, La Jolla, CA 92093, USA
*Correspondence: slippman@ucsd.edu
http://dx.doi.org/10.1016/j.cell.2017.01.031
2QIMR
Study of the biology of tumors caused by germline mutations has led to recent paradigm-changing therapy and is driving precision prevention efforts, including immune oncology and early
detection research. Here, we explore recent biologic advances that are redefining the spectrum
of cancers linked to various hereditary predisposition syndromes and can be leveraged to
improve personalized risk assessment and develop novel interventions to prevent or intercept
cancer.
Introduction
The concept of precision cancer prevention is inherently reliant on devising
accurate and individualized risk assessment. Germline mutations in highly penetrant cancer susceptibility genes (e.g.,
BRCA1 and BRCA2, DNA mismatch
repair [MMR] genes, APC, TP53, etc.)
are thought to underlie a substantial fraction of all malignancies, and carriers of
such mutations are ideal candidates in
whom to develop and implement precision prevention. Multigene panel testing
has identified families affected by unexpected mutations in high-penetrance
genes (e.g., colon cancer patients with
BRCA1 and BRCA2 mutations and breast
cancer patients with Lynch syndrome),
suggesting that the spectrum of cancer
risk conferred by such germline mutations
is much wider and more complicated
than anticipated (Yurgelun et al., 2017).
Furthermore, even for individuals known
to have inherited cancer susceptibility,
contemporary prevention strategies often
rely on the assumption that all carriers are
at uniformly high cancer risks, resulting in
many healthy mutation carriers being
recommended to undergo prophylactic
radical risk-reducing surgery. Advancing
the field of cancer prevention for individuals with particular germline mutations
will thus require a more advanced understanding of the lifestyle, environmental,
and other factors (including germline
variants) that influence risk of specific malignancies and of novel, precise interventions for reducing cancer risk.
Understanding the Spectrum,
Magnitude, and Modifiers of
Hereditary Cancer Risks
The two most common high-penetrance
inherited cancer syndromes are hereditary
breast and ovarian cancer (caused by
germline mutations in BRCA1 or BRCA2,
associated with homologous recombination and other types of DNA repair) and
Lynch syndrome (caused by germline mutations in the MMR gene family, primarily
MLH1, MSH2, MSH6, or PMS2). These
and other hereditary cancer syndromes
are characterized by high lifetime risks of
particular component cancers (e.g., breast
and ovarian cancer in BRCA1- and
BRCA2-mutation carriers and colorectal
and endometrial cancer in individuals with
Lynch syndrome), which are typically
used to aid in the clinical identification of
families. The fundamental question of why
certain germline mutations in cancer
susceptibility genes with broad functions
(e.g., DNA repair genes) predispose to a
particular spectrum of malignancies, however, rather than generalized systemic cancer risk, remains. Studies of BRCA1 carriers are providing novel insights into the
high penetrance for breast and ovarian
cancer. For example, impaired ovarian hormone regulation and signaling in normal
breast tissue from BRCA1 carriers may
contribute to early stages of breast cancer
development in this setting (Communal
et al., 2016).
Mouse models have begun to provide
extensive insight into hereditary colorectal
cancer syndromes by unraveling some
566 Cell 168, February 9, 2017 ª 2017 Elsevier Inc.
of the key mechanisms of intestinal
neoplastic transformation in these settings.
One recent study in a mouse model of
Lynch syndrome (MMR-deficient intestinal
carcinogenesis) elegantly demonstrated
that the short-chain fatty acid butyrate,
generated by gut microbiota from dietary fibers and complex carbohydrates, can act
as an oncometabolite by promoting the
proliferation of cancer-initiated intestinal
epithelial cells (Belcheva et al., 2014). Interestingly, butyrate can have differential
functions in the colon, including as a tumor-suppressive metabolite, referred to
as the butyrate paradox, which is poorly
understood but may in part reflect the
host’s genetic background. Either alterations to the gut microbiome or reductions
in dietary carbohydrate intake markedly
reduce transformation in MMR-deficient,
but not MMR-proficient, mice. This study
shows how nutrition and microbiota can
contribute to colon polyp and cancer
development and highlights the need to
decipher the interplay between host genetics, microbes, diet, and oncogenesis
(Belcheva et al., 2014). Such mechanistic
data are crucial for understanding why
certain inherited forms of cancer risk predispose to a particular spectrum of cancers
and suggest that there are extrinsic factors
(antibiotic, diet, or microbial reprogramming), which could theoretically be exploited to facilitate cancer prevention.
Building on this concept, another study
demonstrated that celecoxib (a COX-2 inhibitor known to reduce intestinal adenoma
burden) induces alterations in the gut
microbiome and metabolome, reducing intestinal crypt stem cell proliferation and adenoma formation in APCMin/+ mice, a model
of familial adenomatous polyposis (Montrose et al., 2016). In particular, celecoxib
therapy in this study was associated with
increased gut Coriobacteriaceae, which
was felt to suppress production of metabolites (e.g., glycine and serine) that are
known to contribute to oncogenesis.
Further research into the mechanisms of
oncogenesis in various inherited conditions
will most likely facilitate the identification of
novel prevention targets, such as in individuals with Lynch syndrome, where aspirin
and other nonsteroidal anti-inflammatory
drugs (NSAIDs) have already demonstrated significant efficacy in reducing
colorectal cancer risks.
Beyond the question of why certain
germline alterations confer risks of
particular cancers and not others, the
magnitude of risk associated with welldescribed cancer predisposition genes
remains a matter of debate. Most cancer
risk estimates associated with a given syndrome are derived from patients and families ascertained through classic clinical
histories and are thus most likely prone
to significant biases. Furthermore, these
fail to account for other factors that may influence risk and penetrance estimates.
For example, BRCA1 and BRCA2 mutations vary considerably between studies.
In retrospective studies, the cumulative
breast cancer risk estimates to age 70
ranges from 40%–87% and 27%–84%
for BRCA1- and BRCA2-mutation carriers, respectively. For ovarian cancer
risk, the ranges are from 16%–68% and
11%–27% for BRCA1- and BRCA2-mutation carriers, respectively (Milne and Antoniou, 2016). The Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA)
and
the
prospective
International
BRCA1/2 Carrier Cohort Study (IBCCS)
are large-scale, multicenter collaborative
efforts focused on identifying the risks of
breast, ovarian, and other cancers in
BRCA-mutation carriers and assessing
other risk modifiers. Most breast cancer
susceptibility SNPs identified by such efforts are differentially associated with estrogen receptor (ER)-negative or -positive
breast cancer in the general population
and, correspondingly, in BRCA1- or
BRCA2-mutation carriers, respectively.
Similarly, only loci found to be associated
with high-grade serous ovarian cancer in
the general population (e.g., 19p13.11)
modify risk of ovarian cancer in BRCA1and BRCA2-mutation carriers, where
tumors are predominantly of this subtype
(Milne and Antoniou, 2016).
Although, individually, the known cancer-risk-modifying SNPs confer only
moderate increases, their effects can
be combined into polygenic risk scores
(PRSs), which are associated with considerable risks. The influence of PRSs on predicting cancer risk is particularly pronounced in high-risk individuals such as
BRCA1- and BRCA2-mutation carriers
given that even small relative risk modifications to their baseline elevated risk
translate into large differences in their
absolute cancer risk. For example, using
a PRS of 94 breast-cancer- and 18
ovarian-cancer-modifying SNPs, CIMBA
has shown that BRCA1-mutation carriers
at the 90th percentile of the PRS have an
absolute risk of 75% of developing breast
cancer by age 80, while those at the 10th
percentile have a risk of 56%. In BRCA2mutation carriers, the ovarian cancer risk
is 19% by age 80 for women at the 90th
percentile of the PRS compared with
only 6% for those at the 10th percentile
(Kuchenbaecker et al., in press). An immediate challenge is to extend the PRS
studies to risk of prostate and breast cancer in male BRCA carriers and to validate
the PRS in prospective studies. As with
nuclear GWASs, certain inherited mitochondrial DNA alterations modify (lower)
risks of breast cancer in germline BRCA2
mutations. Such studies will be important
in determining how useful personalized
risk prediction will be in altering the
uptake or timing of prophylactic surgery
and other clinical interventions. Whether
a PRS alone would identify BRCA1- and
BRCA2-mutation carriers with a sufficiently low enough risk of ovarian or breast
cancer to obviate (or delay) the need for
risk-reducing surgery is unclear. Ideally,
personalized prediction in BRCA1- and
BRCA2-mutation carriers will also include
information on reproductive and family
history, mammographic density, lifestyle,
and environmental factors, and the position of the mutation within a breast, prostate, or ovarian cancer cluster region—all
validated and implemented through developments in comprehensive risk prediction models, at which point personalized
risk prediction will likely have much
greater clinical utility (Kuchenbaecker
et al., in press; Milne and Antoniou, 2016).
Another critical potential payoff from
cancer risk SNPs, such as those identified
by genome-wide association studies
(GWASs), will be their ability to identify
novel cancer genes and/or pathways
(through functional follow-up studies) underlying the observed risk, which could
be exploited for future drug development
or repositioning for cancer prevention,
possibly aided by modeling in BRCAmutant mouse, organoid, zebrafish, and
other relevant model systems. Notably,
SNPs have only small effects on cancer
risk, although this risk effect size is not
necessarily related to other outcomes or
druggability of a target. GWASs do not
themselves identify the genes targeted
by the associated SNPs. Fine mapping,
in silico predictions, and in vitro, or even
in vivo, functional analyses in follow-up
studies are needed to establish this link.
A comprehensive analysis of the genetic
evidence predicting drug mechanisms
suggested that selecting drug targets
with genetic support from either rare,
high-penetrance Mendelian diseases or
GWASs of common complex diseases
could double the success rates in clinical
trials (Nelson et al., 2015). GWAS data first
revealed the importance of the 1p31.3
locus, which brought attention to the
IL-23/IL-17 axis and ultimately led to
development of several FDA-approved
psoriasis drugs targeting IL-17, which is
a pathway that is of increasing interest in
the development of colorectal and other
cancers (Spira et al., 2016). For BRCAassociated breast cancer, proof of principle for this concept is provided by the
finding that SNPs at 6q25 are associated
with risk of ER-negative and -positive disease in both BRCA1-mutation carriers
and the general population and affect
the expression of ESR1, the target of
tamoxifen (Dunning et al., 2016), and
thus may enhance the chemopreventive
effects of tamoxifen in these settings.
SNPs for novel drug repositioning in prevention could include those at 6p23, a
GWAS-identified
locus
(Michailidou
et al., 2013) particularly associated with
breast cancer risk in BRCA2 carriers,
given that it is located near SIRT5, which
is thought to be a target of histone deacetylase inhibitors such as panobinostat,
Cell 168, February 9, 2017 567
Figure 1. RANK-L Blockade as a Prevention Strategy
A series of studies (e.g., in Brca1-deficient mice and BRCA1-mutant organoids) suggest the following schema: BRCA1mut/+ RANK+ subset of mammary luminal
progenitor (LP) cells (left) give rise to basal-like tumors (far right). Progesterone-dependent RANK signaling in LP cells is responsive to RANK-L inhibition.
Subsequent NF-kB-activated (involving DNA damage [ATM]), progesterone-independent cells may be less sensitive to RANK-L inhibition and more sensitive to
IKKa and PARP inhibitor prevention. Denosumab (human monoclonal antibody [red] targeting RANK-L [pink]) could abrogate progesterone-dependent, mitogenic signaling of RANK+ BRCA1mut/+ LP cells to prevent basal-like tumors. Given these and other compelling data (see text), an international breast cancer
prevention trial of denosumab is planned for BRCA1-mutation carriers (bottom). Abbreviations are as follows: RANK, receptor activator of nuclear factor kappa-B
(NF-kB); RANK-L, RANK ligand; ER, estrogen receptor; PR, progesterone receptor; MaSC, mammary stem cell; NEMO, NF-kB essential modulator; ATM, ataxia
telangiectasia mutated; IKKa, inhibitor of kappaB kinase alpha; PARP, poly (ADP-ribose) polymerase.
which was recently FDA approved for
multiple myeloma.
Precision Cancer Prevention in
Inherited Cancer Risk
Most current evidence-based strategies
for managing patients with known inherited risks of cancer rely on relatively
crude forms of cancer prevention, such
as prophylactic surgeries (e.g., hysterectomy and removal of the fallopian tubes
and ovaries [salpingo-oophorectomy] in
women with Lynch syndrome and salpingo-oophorectomy and mastectomy
in women with BRCA1 and BRCA2 mutations). While mostly effective at reducing
both cancer risk and mortality, they
confer significant long-term morbidity
such as surgical menopause, which
can both limit patient uptake (especially
early in life, when their potential preventive impact is inherently greatest) and
often generate their own medical and
psychological sequelae. Risk-reducing
salpingo-oophorectomy has been a
cornerstone of breast cancer prevention
in BRCA1- and BRCA2-mutation carriers
for decades. However, recent analyses
suggest that this is not as effective for
568 Cell 168, February 9, 2017
premenopausal BRCA1-mutation carriers
(Kotsopoulos et al., 2016), and so there
exists an unmet medical need for other
approaches. Although most breast cancers arising in BRCA1-mutation carriers
are tamoxifen-resistant and ER-negative,
tamoxifen does reduce contralateral
breast cancer in BRCA1 carriers, likely
due to BRCA crosstalk and interactions
with hormonal effects in early ontogeny
or stromal estrogen signaling.
An exciting opportunity in precision
prevention is emerging for BRCA1-mutation carriers based on studies of a highly
proliferative subset of luminal progenitor
cells that give rise to basal-like breast
cancer. These cells constitutively express
RANK (receptor activator of nuclear factor
kappa-B [NF-kB]) and are hyper-responsive to RANK ligand (RANK-L). Interference of this pathway produces significant
preventive activity in breast organoids
and mouse models (Figure 1). Furthermore, in a pilot study of six BRCA1 carriers, a fully human monoclonal antibody
targeted against RANK-L that is FDA
approved with a well-established safety
record for treatment of osteoporosis
(denosumab) inhibited RANK-positive
BRCA1-mutated breast cancer progenitors, with promising effects on proliferation or clonogenic potential in breast
tissue (Nolan et al., 2016). There is
some evidence that germline SNPs in
RANK are associated with breast cancer
risk in these carriers, but larger studies
and demonstration that the SNPs have
regulatory effects on RANK itself will be
required to clarify the relevance of this
finding. RANK-L inhibition in BRCA1-mutation carriers could thus prevent and/or
delay tumor onset for premenopausal
women given that RANK-L is a progesterone-responsive gene and risk-reducing
salpingo-oophorectomy has limited efficacy in this setting (Kotsopoulos et al.,
2016), leaving mastectomy as the primary
preventive approach. Such intriguing
findings support the planned, international trial repurposing denosumab
as a breast cancer prevention agent
for BRCA1-mutation carriers and could
include other BRCA1-associated cancers
(e.g., ovarian and pancreatic) as secondary endpoints.
While not a classic component cancer
of either syndrome, pancreatic cancer
is notoriously lethal with essentially no
effective prevention, early detection, or
therapy, representing an unequivocal unmet medical need. Emerging data indicate a very high (>15% in a clinic-based
cohort) prevalence of germline mutations
in cancer predisposition genes (most
commonly BRCA1 and BRCA2, but also
PALB2, the MMR genes, ATM, CDKN2A,
TP53, and others) among otherwise unselected pancreatic cancer patients and
even higher rates among Ashkenazi
Jewish patients (Salo-Mullen et al.,
2015). Such data recently led to National
Comprehensive Cancer Network (NCCN)
updated guidelines now recommending
germline BRCA1 and BRCA2 testing
for all pancreatic cancer patients of
Ashkenazi ancestry. Data further indicate,
however, that classic high-risk features
(e.g., young age at diagnosis or family
history of cancer) have poor sensitivity for identifying even non-Ashkenazi
pancreatic cancer patients likely to harbor
such germline mutations, suggesting that
performing multigene germline testing
on all newly diagnosed pancreatic cancer
may very well be the most effective
approach for identifying high-risk families.
The first real signal of benefit from
early detection research (imaging people
with high-penetrance mutations) was
recently reported, adding to the enthusiasm for the changing pancreas guidelines for germline testing (Vasen et al.,
2016). Precedent for such a ‘‘test
everyone’’ approach already exists in
ovarian cancer, where NCCN guidelines
recommend that all women with epithelial
ovarian cancer (where screening has
limited benefit) undergo germline analysis
of BRCA1 and BRCA2 given that a
substantial fraction of unselected women
with ovarian cancer will carry such mutations, many of whom lack clinical histories
suggestive of hereditary breast/ovarian
cancer.
Utilization of such genetic screening
approaches is likely to be critical to
devising precision prevention efforts
against particular forms of inherited
cancer. Recent data in BRCA1- and
BRCA2-related pancreatic cancer (Connor et al., 2016) identify somatic mutational patterns and molecular markers of
increased antitumor immunity and tumor
neoantigens, suggesting a role for immune-based mechanisms in treating and
preventing such cancers. Furthermore,
the profound ability of immune checkpoint blockade to successfully treat
MMR-deficient cancers with widespread
accumulation of somatic frameshift mutations, such as those seen in Lynch syndrome, has added to the enthusiasm
for testing tumors for MMR-deficiency
and microsatellite instability (MSI). Such
universal tumor testing is standard practice for all newly diagnosed colorectal
cancers and is becoming commonplace
for all endometrial cancers so as to
comprehensively screen these patients
for Lynch syndrome. Given increasing
interest in identifying patients who could
benefit from immunotherapy, it is likely
that such MMR-deficiency tumor testing
will be expanded to other cancer types,
which may lead to an increase in Lynch
syndrome identification among individuals lacking classic component cancers.
Furthermore, these breakthrough advances regarding the immune system’s
ability to treat and control MMR-deficient cancers have generated interest
in the possibility of using immunebased strategies for cancer prevention,
such as cancer vaccines, targeting
predictable MSI-induced neoantigens in
healthy Lynch syndrome carriers (Spira
et al., 2016).
phylactic surgical approaches to rationally designed personalized strategies
for cancer prevention in individuals with
inherited cancer risk. GWASs illustrate a
relatively untapped treasure trove for
discovering novel genes and/or pathways
underlying cancer risk and immune function, with future possibilities of patient selection and stratification and, perhaps
most importantly, drug development or repositioning. Future GWAS integrating nuclear and mitochondrial studies will provide a more full germline landscape,
including somatic interactions. Leveraging
these many promising advances toward
personalized risk assessment and precision cancer prevention for individuals
with inherited cancer risk will be crucial
for ultimately attaining the goal of more
widespread cancer prevention in the general population.
Conclusion
The full potential for genetic medicine to
facilitate cancer prevention can only be
realized through a comprehensive understanding of the diversity, biology, and
magnitude of risk conferred by given
germline alterations and by the development and implementation of effective
risk-reducing interventions. Major progress is being made in better understanding the phenomenon of inherited cancer
risk, including advances in microbiome
biology and immune oncology, which are
providing tantalizing prevention leads.
Other key advances include the growing
ability to personalize cancer risk estimates for individuals with highly penetrant
inherited syndromes by using genetic
(e.g., PRS), lifestyle, and environmental
factors. Furthermore, compelling studies
of luminal progenitor cell biology in
BRCA1-mutant models have led to a
transformative precision medicine trial.
Additional investigation into these
groundbreaking areas holds the promise
of moving the field beyond empiric, pro-
Communal, L., Vilasco, M., Hugon-Rodin, J.,
Courtin, A., Mourra, N., Lahlou, N., Le Guillou,
M., Perrault de Jotemps, M., Chauvet, M.P.,
Chaouat, M., et al. (2016). Oncotarget 7,
45317–45330.
REFERENCES
Belcheva, A., Irrazabal, T., Robertson, S.J.,
Streutker, C., Maughan, H., Rubino, S., Moriyama,
E.H., Copeland, J.K., Kumar, S., Green, B., et al.
(2014). Cell 158, 288–299.
Connor, A.A., Denroche, R.E., Jang, G.H., Timms,
L., Kalimuthu, S.N., Selander, I., McPherson, T.,
Wilson, G.W., Chan-Seng-Yue, M.A., Borozan, I.,
et al. (2016). JAMA Oncol. Published online
October 20, 2016.
Dunning, A.M., Michailidou, K., Kuchenbaecker,
K.B., Thompson, D., French, J.D., Beesley, J.,
Healey, C.S., Kar, S., Pooley, K.A., LopezKnowles, E., et al.; EMBRACE; GEMO Study Collaborators; HEBON; kConFab Investigators
(2016). Nat. Genet. 48, 374–386.
Kotsopoulos, J., Huzarski, T., Gronwald, J., Singer,
C.F., Moller, P., Lynch, H.T., Armel, S., Karlan, B.,
Foulkes, W.D., Neuhausen, S.L., et al.; Hereditary
Breast Cancer Clinical Study Group (2016).
J. Natl. Cancer Inst. 109.
Kuchenbaecker, K.B., McGuffog, L., Barrowdale,
D., Lee, A., Soucy, P., Healey, S., Dennis, J.,
Lush, M., Robson, M., and Spurdle, A.B., et al. (in
press) J. Natl. Cancer Inst.
Michailidou, K., Hall, P., Gonzalez-Neira, A.,
Ghoussaini, M., Dennis, J., Milne, R.L., Schmidt,
M.K., Chang-Claude, J., Bojesen, S.E., Bolla,
M.K., et al. (2013). Nat. Genet. 45, 353–361.
Milne, R.L., and Antoniou, A.C. (2016). Endocr. Relat. Cancer 23, T69–T84.
Montrose, D.C., Zhou, X.K., McNally, E.M., Sue, E.,
Yantiss, R.K., Gross, S.S., Leve, N.D., Karoly, E.D.,
Cell 168, February 9, 2017 569
Suen, C.S., Ling, L., et al. (2016). Cancer Prev. Res.
(Phila.) 9, 721–731.
milial Breast Cancer (kConFab) (2016). Nat. Med.
22, 933–939.
Nelson, M.R., Tipney, H., Painter, J.L., Shen, J.,
Nicoletti, P., Shen, Y., Floratos, A., Sham, P.C.,
Li, M.J., Wang, J., et al. (2015). Nat. Genet. 47,
856–860.
Salo-Mullen, E.E., O’Reilly, E.M., Kelsen, D.P.,
Ashraf, A.M., Lowery, M.A., Yu, K.H., Reidy, D.L.,
Epstein, A.S., Lincoln, A., Saldia, A., et al. (2015).
Cancer 121, 4382–4388.
Nolan, E., Vaillant, F., Branstetter, D., Pal, B., Giner, G., Whitehead, L., Lok, S.W., Mann, G.B.,
Rohrbach, K., Huang, L.Y., et al.; Kathleen Cuningham Foundation Consortium for Research into Fa-
Spira, A., Disis, M.L., Schiller, J.T., Vilar, E., Rebbeck, T.R., Bejar, R., Ideker, T., Arts, J., Yurgelun,
M.B., Mesirov, J.P., et al. (2016). Proc. Natl. Acad.
Sci. USA 113, 10750–10758.
570 Cell 168, February 9, 2017
Vasen, H., Ibrahim, I., Ponce, C.G., Slater, E.P.,
Matthäi, E., Carrato, A., Earl, J., Robbers, K., van
Mil, A.M., Potjer, T., et al. (2016). J. Clin. Oncol.
34, 2010–2019.
Yurgelun, M.B., Kulke, M.H., Fuchs, C.S., Allen,
B.A., Uno, H., Hornick, J.L., Ukaegbu, C.I.,
Brais, L.K., McNamara, P.G., Mayer, R.J., et al.
(2017). J. Clin. Oncol. Published online January
30, 2017. http://dx.doi.org/10.1200/JCO.2016.
71.0012.
Leading Edge
Commentary
Next-Generation Sequencing
of Circulating Tumor DNA
for Early Cancer Detection
Alexander M. Aravanis,1,2 Mark Lee,1,2 and Richard D. Klausner1,*
1GRAIL,
Menlo Park, CA 94402, USA
author
*Correspondence: klausner.rick@gmail.com
http://dx.doi.org/10.1016/j.cell.2017.01.030
2Co-first
Curative therapies are most successful when cancer is diagnosed and treated at an early stage. We
advocate that technological advances in next-generation sequencing of circulating, tumor-derived
nucleic acids hold promise for addressing the challenge of developing safe and effective cancer
screening tests.
Cancer-specific mortality from most
types of solid tumors has barely
decreased in decades, despite an exponential increase in our knowledge about
cancer pathogenesis and significant investments in the development of effective
treatments. The past few years have
witnessed a dramatic success of immunotherapies in treating a subgroup of
patients with a variety of tumor types,
including lung, bladder, and kidney, as
well as Hodgkin’s lymphoma and melanoma. While such breakthroughs offer
the hope of prolonged survival for some
patients with advanced cancers, finding
cancers earlier would still afford the greatest chance for cure, given that the survival
rates for patients with early diagnoses are
five to ten times higher compared with late
stage disease (Cho et al., 2014). By
enabling diagnosis and localized treatment of early stage invasive cancers
(and, in some cases, pre-invasive states),
screening for cervical and colorectal cancers has contributed to significant declines in mortality from these diseases.
Similarly, effective population screening
paradigms for many other common and
deadly tumor types are clearly needed to
broadly reduce cancer-specific mortality.
Challenges for Cancer Screening
Despite the promise, the field of early cancer detection is filled with cautionary tales
that highlight the challenges. Early stage
solid tumors are small and are thereby
difficult to be non-invasively distinguished
from normal anatomic and biochemical
variation. This ambiguity leads to detec-
tion algorithms that either miss a large
number of invasive cancers or make the
costly trade-off of over-diagnosing and
consequently over treating. For instance,
high false-positive rates from mammography in breast cancer screening, lowdose CT in lung cancer screening, and
prostate-specific antigen (PSA) screening
(Nelson et al., 2016a; Aberle et al., 2011;
Chou et al., 2011) represent a significant
cost to the healthcare system, with resulting mental and physical morbidity, and
even mortality in some cases (Nelson
et al., 2016b).
Even where cancer screening has produced significant stage shifts, as with
breast and prostate cancer screening,
the impact on cancer-specific mortality
has not been a predictable outcome
(Berry, 2014). Multiple explanations may
account for this phenomenon. First, early
stage cancers detected by mammography and PSA are likely to be clinically
insignificant or ‘‘non-lethal,’’ meaning
that, in the absence of screening and
treatment, these tumors would not have
become clinically evident, metastatic, or
contributory to patient mortality. Second,
some lethal tumors diagnosed in an
apparently localized state may already
have disseminated with occult metastatic
disease, making local treatment with
surgery or radiation non-curative. Earlier
diagnosis of these tumors would result in
a ‘‘lead-time bias’’ without impacting cancer-specific mortality. Third, mammography and PSA may be sensitive for indolent tumors, but not sufficiently sensitive
for lethal cancers, particularly when these
cancers are in a pre-metastatic state and
thus still curable. This kinetic aspect of
cancer progression is poorly understood,
but it is essential to informing effective
screening intervals. It is worth noting
that mammography and PSA are only surrogate measures of cancer, which have
poor specificity and provide little insight
into tumor biology. We would argue that
for successful screening, we need a
platform that provides direct, sensitive,
and specific measures of cancer and its
attributes, which have bearing on clinical
behavior.
Circulating Tumor DNA
Profiling of a tumor’s somatic alterations
has become routine, and many clinical
tests are now available that interrogate
anywhere from a few genes to the whole
human genome (Cheng et al., 2015). For
a given tumor, its unique set of somatic alterations creates a biological signature
that is highly specific to the individual tumor and appears to be specific to cancer
in general when compared to normal tissues. Extensive research has demonstrated that tumor-derived somatic alterations in DNA can be detected in the
plasma of cancer patients in the form
of cell-free DNA (cfDNA) (Bettegowda
et al., 2014). This circulating tumorderived DNA (ctDNA) is presumably
shed from tumors, either through necrosis
or apoptosis. Commercial ‘‘liquid biopsy’’
tests performed on patients with metastatic disease interrogate the ctDNA in order to identify clinically actionable tumor
genotypes in lieu of a tissue biopsy.
Cell 168, February 9, 2017 ª 2017 Published by Elsevier Inc. 571
Table 1. Comparison of ctDNA Liquid Biopsy Test to Potential Cancer Screening Test
Indication
Tumor Liquid Biopsy (Genotyping, Monitoring)
Early Cancer Detection
Target population
Patients with known diagnosis of cancer
Asymptomatic individuals
Tissue reference
Can be informed by tissue analyses
No prior knowledge of tissue
Key performance characteristics
Sensitivity and specificity for specific
actionable genotypes
d
d
d
Sensitivity and specificity for clinically
detectable cancer
Premium on specificity in individuals
without detectable cancer
Tissue of origin needed to guide workup
Clinical Endpoint for Utility
Therapeutic benefit with specific therapies
Net outcome improvement with early detection
and local treatment of cancer
Genes Covered
10-50
100-1000s
ctDNA Limit of Detection
0.1%
<0.01%
Importance of Novel Variant Detection
Low
High
Amount of Sequencing
1x
100X
Study Size for Clinical Validity and Utility
100’s
10,000 - 100,000 s
Importantly, the data provided by these
tests indicate that these genotypes are
not common in the plasma of individuals
that are presumably cancer-free (Thress
et al., 2015). It is worth noting that tumor-derived RNA and DNA methylation
patterns can also be detected in the
plasma (Chan et al., 2013) and provide
complementary information to the somatic alterations detected in ctDNA.
At present, relatively little is known
about the biology of cfDNA, such as the
mechanism through which it is produced
and comes into systemic circulation, and
the physiologic variability (e.g., diurnal
variation) and kinetics of clearance.
Further, the quantitative relationship between the relative and absolute level of
ctDNA in the blood and biological features
of a tumor (genotype, size, vascularity,
anatomy) are not known. Basic research
into the fundamental biology of ctDNA is
needed to interpret ctDNA signals and
their associations with clinical manifestations of cancer.
ctDNA Detection
Solid tumors are biologically diverse. Data
from The Cancer Genome Atlas (TCGA)
indicates immense variation between
cancers at the DNA, RNA, and epigenetic
levels, not just between tissues, but
even within a given tissue type (Cancer
Genome Atlas Research Network et al.,
2013). Moreover, the patient-specific tumor microenvironment and immune system impact the evolution and lethality of
a tumor. Compared to infectious disease,
which has well-defined causative agents
572 Cell 168, February 9, 2017
with highly conserved biology, a population of cancer patients behaves as a heterogeneous collection of many diseases,
each of which carries additional heterogeneity in its own right. Therefore, identifying
a finite number of protein or nucleic acid
biomarkers that are highly sensitive and
specific to even a single cancer is difficult,
let alone multiple cancers.
There is evidence for the presence of
ctDNA in early cancers (Bettegowda
et al., 2014). However, the fraction of
tumors that shed detectable levels of
ctDNA, by tumor type and stage, is not
well studied. The studies to date have
small numbers of samples and use a variety of measurement techniques that are
often not comparable. Despite these caveats, ctDNA is a fundamentally different
type of cancer ‘‘biomarker’’ than most
that have been used for cancer detection.
Most importantly, ctDNA detection leverages the hallmarks of cancer as a disease
of genomic alterations and is thus a direct
measurement of the tumor. As such, it has
the potential to be more specific to the
presence of the tumor than other surrogate and downstream measurements of
proteins and metabolites. However, the
implementation of such a test would be
technically challenging, since many genes
would have to be simultaneously queried
for alterations in order to cover enough
of the known diversity in cancer genomes
to see most tumors. While next-generation DNA sequencing technology does
enable high degrees of target multiplexing, the depth of sequencing would also
have to be very high to sample enough
ctDNA molecules to reliably measure
them in a background of mostly non-tumor-derived cfDNA. We estimate that
such a broad and deep sequencing
approach could require orders of magnitude more sequence data than liquid biopsy assays currently use (Table 1). To
generate such a large amount of DNA
sequence data today in the performance
of a routine cancer screening test would
be cost-prohibitive, but it is likely to
become more technically and economically feasible as the cost of computation
and sequencing continues to decrease.
Challenges of Developing a
ctDNA-Based Screening Test
Clinical development of a ctDNA-based
approach for broadly applicable cancer
screening requires very large-scale clinical evidence for both test development
and demonstration of clinical utility. Test
development needs to address the dual
challenges of sensitivity for early stage
disease and the need for exquisite specificity. For sensitivity, the low level of signal
in early stage disease, as well as the heterogeneity of cancer genomes present in
the population, needs to be overcome to
achieve efficacy. For each targeted tumor
type, it is expected that at least several
hundred of examples of cancer patients,
particularly early stage patients, are
required to adequately characterize the
potential cancer-defining variants observable in plasma. To confirm that variants
are indeed cancer-defining, test development must also evaluate specificity in
plasma-cell-free DNA profiles from large
numbers of healthy individuals without a
known diagnosis of cancer and who are
representative controls for the cancer
population. Meanwhile, longitudinal clinical follow-up of these non-cancer subjects for future cancer diagnoses is important for distinguishing false positive
signals from true positives. As true cancer
events are relatively rare within a general
screening population (<<1% per year for
a given tumor type), the demands on
specificity will be particularly high to minimize harm due to false positives and
compatibility with practical application
in the healthcare system. Adding to the
specificity challenge are recent reports of
somatic alterations accumulating in both
solid tissues and the hematopoietic system as a function of age (Genovese
et al., 2014; Alexandrov et al., 2015). To
achieve high clinical specificity, a ctDNAbased screening test must be capable of
distinguishing between the background
signal originating from such non-cancer
or pre-cancerous processes and the invasive malignancy of real interest.
Additional challenges for clinical development arise from the prevailing anatomic
and pathologic paradigm for cancer
diagnosis and staging. First, an ideal
non-invasive screening test would also
provide information on the tissue of origin
to streamline the downstream workup,
including imaging and tissue diagnosis.
This information may be possible through
ctDNA, given the distinct differences in
the patterns of somatic alterations between different tumor types, at least at
a population level (Ciriello et al., 2013).
Second, a proportion of screen-detected
early stage tumors may prove to be clinically insignificant, as has been described
for screen-detected prostate and breast
cancers. These entities may be biologically distinct from lethal variants or potentially held in check by the host immune
system, and finding and treating these
cancers could lead to harm through
over-diagnosis and over-treatment. It is
possible that mutational signatures in
ctDNA could distinguish clinically insignificant biological processes from malignant
and lethal biological processes. In addition, serial measurements of the ctDNA
signal may identify distinct trajectories
with different kinetics for indolent versus
lethal disease, thus functionally stratifying
these patient populations. Third and
finally, early stage tumors, as currently
defined by anatomic staging, are not uniformly cured with local therapy. For some
tumor types such as lung cancer and
pancreatic cancer, the majority of ‘‘localized’’ tumors eventually relapse and thus
could contribute to lead-time bias for
a screening program. These individuals
clearly have occult metastatic disease,
which is not accurately captured with current staging. Post-operative ctDNA levels
have now been shown to predict relapse
(Tie et al., 2016). By identifying ‘‘early
stage’’ patients who have been understaged, ctDNA could minimize the effect
of lead-time bias by directing aggressive
post-operative systemic therapy to a
population where the benefit would be
greatest.
Ultimately, to demonstrate the clinical
utility of a ctDNA cancer screening test,
a prospective clinical trial, comparing the
strategy of early detection based on the
ctDNA test with the standard of care, is
necessary. Given the potential biases in
cancer screening studies, including lead
time bias (Berry, 2014), a randomized
design is likely required to achieve a practice-changing level of evidence. Because
the yearly incidence rate for cancer is
low (1%–2% in aggregate across tumor
types), and because differences in cancer-specific mortality can take years to
manifest, it is anticipated that such a
study would require hundreds of thousands of participants to be appropriately
powered. Early assessment of efficacy
would be possible with large effect sizes
and by evaluating surrogate endpoints,
including cumulative incidence of de
novo and recurrent metastatic disease.
In this regard, GRAIL, a company
recently created by some of the authors,
was formed to develop ctDNA-based
cancer screening tests, is conducting a
study to create a reference library of the
cancer mutations in the blood for the
most common cancers and the background mutations found in matched
healthy subjects. This 10,000-plus subject study, called the Circulating CellFree Genome Atlas (CCGA) (clinicaltrials.
gov identifier: NCT02889978), should
complete enrollment within one year and
will be the largest database on mutations
found in the blood of cancer patients.
The Cancer Moonshot Initiative recently
announced the Blood Profiling Atlas
Project, which also aims to compile
data on cancer signals in the blood.
GRAIL intends to apply a next-generation sequencing approach, combining
sequencing depth and breadth of
genomic coverage, as well as machine
learning, to develop models based on
cell-free DNA for the accurate classification of subjects with and without cancer.
Tests based upon these models, derived
from CCGA and additional studies, will
be prospectively evaluated in observational cohorts and eventually in clinical
utility studies as described above.
Conclusions
The biological, technical, and clinical obstacles to developing a safe and effective
pan-cancer screening test are significant.
However, we have new tools that directly
measure the sin qua non of cancer (its
somatic alterations) in ways not previously possible. Low-cost next-generation
sequencing has enabled the sequencing
of tumor-derived nucleic acids circulating
in the blood. This non-invasive characterization of a tumor’s genome—combined
with new methods of processing complex
data, including machine learning—may
enable sensitive and specific detection
of curable, lethal cancers. Finally, new
adaptive approaches to population-scale
screening trials are emerging that could
be employed to test the clinical utility of
this approach. Given these advances,
the time has come for a new and unprecedented level of effort to develop lifesaving tests that can detect cancer early,
when it can be cured.
ACKNOWLEDGMENTS
A.M.A. and M.L. are employees and shareholders
of GRAIL. R.D.K. is a member of GRAIL’s board
of directors and scientific advisory board and is
also a shareholder.
REFERENCES
Aberle, D.R., Adams, A.M., Berg, C.D., Black,
W.C., Clapp, J.D., Fagerstrom, R.M., Gareen, I.F.,
Gatsonis, C., Marcus, P.M., and Sicks, J.D.;
National Lung Screening Trial Research Team
(2011). N. Engl. J. Med. 365, 395–409.
Alexandrov, L.B., Jones, P.H., Wedge, D.C., Sale,
J.E., Campbell, P.J., Nik-Zainal, S., and Stratton,
M.R. (2015). Nat. Genet. 47, 1402–1407.
Berry, D.A. (2014). Cancer 120, 2784–2791.
Bettegowda, C., Sausen, M., Leary, R.J., Kinde, I.,
Wang, Y., Agrawal, N., Bartlett, B.R., Wang, H.,
Cell 168, February 9, 2017 573
Luber, B., Alani, R.M., et al. (2014). Sci. Transl.
Med. 6, 224ra24-224ra24.
Cancer Genome Atlas Research Network, Weinstein, J.N., Collisson, E.A., Mills, G.B., Shaw,
K.R., Ozenberger, B.A., Ellrott, K., Shmulevich, I.,
Sander, C., and Stuart, J.M. (2013). Nat. Genet.
45, 1113–1120.
Cho, H., Mariotto, A.B., Schwartz, L.M., Luo, J.,
and Woloshin, S. (2014). J. Natl. Cancer Inst.
Monogr. 2014, 187–197.
Nelson, H.D., O’Meara, E.S., Kerlikowske, K.,
Balch, S., and Miglioretti, D. (2016a). Ann. Intern.
Med. 164, 226–235.
Chou, R., Croswell, J.M., Dana, T., Bougatsos, C.,
Blazina, I., Fu, R., Gleitsmann, K., Koenig, H.C.,
Lam, C., Maltz, A., et al. (2011). Ann. Intern. Med.
155, 762–771.
Nelson, H.D., Pappas, M., Cantor, A., Griffin, J.,
Daeges, M., and Humphrey, L. (2016b). Ann.
Intern. Med. 164, 256–267.
Chan, K.C.A., Jiang, P., Chan, C.W.M., Sun, K.,
Wong, J., Hui, E.P., Chan, S.L., Chan, W.C., Hui,
D.S.C., Ng, S.S.M., et al. (2013). Proc. Natl.
Acad. Sci. USA 110, 18761–18768.
Ciriello, G., Miller, M.L., Aksoy, B.A., Senbabaoglu,
Y., Schultz, N., and Sander, C. (2013). Nat. Genet.
45, 1127–1133.
Thress, K.S., Brant, R., Carr, T.H., Dearden, S.,
Jenkins, S., Brown, H., Hammett, T., Cantarini,
M., and Barrett, J.C. (2015). Lung Cancer 90,
509–515.
Cheng, D.T., Mitchell, T.N., Zehir, A., Shah, R.H.,
Benayed, R., Syed, A., Chandramohan, R., Liu,
Z.Y., Won, H.H., Scott, S.N., et al. (2015). J. Mol.
Diagn. 17, 251–264.
Genovese, G., Kähler, A.K., Handsaker, R.E., Lindberg, J., Rose, S.A., Bakhoum, S.F., Chambert, K.,
Mick, E., Neale, B.M., Fromer, M., et al. (2014).
N. Engl. J. Med. 371, 2477–2487.
Tie, J., Wang, Y., Tomasetti, C., Li, L., Springer, S.,
Kinde, I., Silliman, N., Tacey, M., Wong, H.L.,
Christie, M., et al. (2016). Sci. Transl. Med. 8,
346ra92.
574 Cell 168, February 9, 2017
Leading Edge
Commentary
Cancer Clinical Trials:
The Rear-View Mirror and the Crystal Ball
Dave Cescon1 and Lillian L. Siu1,*
1Princess
Margaret Cancer Centre, University of Toronto, Toronto, ON M5G 1Z5, Canada
*Correspondence: lillian.siu@uhn.ca
http://dx.doi.org/10.1016/j.cell.2017.01.027
Clinical trials are key to translating scientific advances into progress in cancer research and care.
Confronted by developments in basic science, the landscape of clinical cancer research is rapidly
evolving. Here, we review recent changes in clinical trials’ design and conduct, and we forecast
future directions toward personalized and global impact.
Introduction
Many of the discoveries in cancer science
are touted as potential avenues toward
improved clinical therapies. Indeed,
revolutionary advancements in technologies for the characterization of
cancer genomes (e.g., next-generation
sequencing), increasing application of
more robust models of clinical disease
(e.g., patient-derived xenografts and organoids), and the development of powerful
tools for the functional evaluation of cancer targets (e.g., CRISPR/Cas9) have had
a major impact on the direction and process of nonclinical anticancer drug development. Simultaneously, the dawn of the
cancer immunotherapy era has seized
the attention of the research community
and is now a major focus for target validation and drug discovery. Despite these
changing landscapes, clinical evaluation
of new agents remains the cornerstone
of drug development, and thus well-designed clinical trials are essential to deliver
new medicines to patients in need.
The scope of clinical investigation in
oncology is enormous, with approximately
830 anticancer agents in clinical trials
or under review by the US Food and
Drug Administration in 2015 (www.phrma.
org/report/medicines-in-developmentfor-cancer-2015-report). While historically
the paradigm of anticancer drug development has taken 10 years or longer from
first-in-human Phase I studies to regulatory
approval, this pace is no longer acceptable
to stakeholders, including industry, investigators, patients, and advocates. In this
Commentary, we will review some of the
important trends in the design and conduct
of clinical trials over the last decade and
anticipate areas for continued innovation
in the coming years. In doing so, we hope
to identify opportunities to address the urgent need to translate emerging science
into therapeutic improvements.
Changes in the Design and Conduct
of Clinical Trials in the Last Decade
The traditional, sequential three-phase
drug development paradigm has reigned
in oncology for more than five decades,
since the development of the Phase II
concept proposed by Gehan at the
National Cancer Institute in 1961 (Gehan,
1961). Therein, the tolerability, pharmacokinetics, and safely administered doses of
new drugs are determined through dose
escalation in Phase I studies, typically in
unselected patients with treatment refractory disease. Agents graduating based on
the above properties move to Phase II
studies in order to estimate antitumor efficacy in a more homogeneous, usually histologically defined population. Finally, the
effectiveness of a treatment compared to
existing standard of care is assessed in
Phase III, typically randomized trials.
However, driven by an urgency to bring
effective therapies to patients, an increasing number of agents with potential
breakthrough anticancer activity have
challenged the sequential three-phase
paradigm. The two major ways that this
has occurred are exemplified by the
development of recently approved immune checkpoint inhibitors (e.g., pembrolizumab) and agents targeting rare and
specific molecular alterations (e.g., crizotinib). Both of these agents took only 3–4
years from first-in-human testing to regulatory approval, a timeline that was unprecedented for solid tumors (Kazandjian
et al., 2014; Poole, 2014).
Seamless Development in Phase I
First-in-Human Trials
In contrast to a typical Phase I trial, which
might enroll a few dozen patients, the
first-in-human study of pembrolizumab,
KEYNOTE-001, evolved into a massive,
multi-armed expansion study enrolling
more than 1,200 patients. A clinical
protocol need not be static and unchanging and can be adapted to incorporate
emerging science and clinical data,
especially for agents demonstrating early
signs of promising antitumor activity
and acceptable toxicity. The flexibility to
amend in-progress studies with regulatory approval must be accompanied by a
clear and strong rationale to alter existent design and justify the expansion.
KEYNOTE-001 is an example of a welltolerated agent with substantial clinical
activity, where serial protocol amendments were deemed to be more expeditious than re-launching later-stage trials.
What began as a standard 3+3 dose
escalation study with three dose levels
and a planned seven-patient expansion
in each of two disease-specific cohorts
(melanoma and renal cell carcinoma) ultimately enrolled 40 times as many patients, including 655 with advanced melanoma and 550 with non-small lung cancer
(NSCLC), and supported the accelerated
approval of pembrolizumab for these
two indications (Garon et al., 2015;
Hamid et al., 2013). In addition, numerous
Phase II and III trials were launched while
KEYNOTE-001 was ongoing.
Accelerated Approval from SingleArm Studies
While immune checkpoint inhibitors such
as anti-programmed death-1 (PD1) and
anti-programmed death ligand-1 (PD-L1)
Cell 168, February 9, 2017 ª 2017 Elsevier Inc. 575
antibodies appear to have broad activity
across several cancer types, agents targeting very specific molecular alterations
represent the other end of the drug development spectrum. Imatinib, the standardbearer of personalized cancer therapy,
initially received accelerated approval for
chronic myelogenous leukemia (CML)
driven by the BCR-ABL translocation on
the basis of open-label, single-arm Phase
II studies (Johnson et al., 2003). The
discovery of EML4-ALK fusions in a
small proportion of NSCLC (Soda et al.,
2007), first reported in 2007, presented a
similar opportunity to demonstrate
exceptional response in a highly selected
patient population. In this case, crizotinib
(PF-02341066) was under investigation
as a c-met/HGF inhibitor in a Phase I
study of patients with unselected solid
tumors. On the basis of its recognized activity against ALK (anaplastic lymphoma
kinase), the discovery of this driver in
NSCLC allowed the repurposing of crizotinib and amending of the trial to include
expansion cohorts enriched for the ALK
fusion. This enabled rapid accrual of this
relatively uncommon molecularly defined
subtype, and in 2010, the initial report
demonstrating a response rate of >50%
in 82-ALK rearranged patients treated on
this open-label Phase I study supported
the accelerated approval of crizotinib (Kazandjian et al., 2014; Kwak et al., 2010).
Next-Generation Targeted Agents
The development of ALK inhibitors also illustrates the phenomenon of rapid and
sequential development of optimized
agents, following demonstration of the
clinical value of an anticancer target.
While there are many historical examples
of iterations on cytotoxic drugs (e.g., taxanes), iterations on molecular targeted
therapies are being developed at an
increasing pace to overcome limitations
of first-generation agents. Both inhibitors
of epidermal growth factor receptor
(EGFR) and ALK have seen the development of new compounds able to surmount common mutational mechanisms
of resistance or designed to improve
penetration into the central nervous
system. With a well-defined population
of interest (e.g., patients resistant to
the first agent), second-generation
studies can be performed with great
efficiency, as evidenced by ceritinib and
alectinib, which entered Phase I testing
576 Cell 168, February 9, 2017
in ALK rearrangement positive crizotinib
failures in 2012, and was approved by
the FDA in 2015. These examples are testament to the impact that clinical proof of
concept can have on the field.
Genomic-Based Precision Oncology
Trials
Increasingly, developers of molecularly
targeted agents are homing in on target
populations linked to putative predictive
biomarkers. Knowledge of the underlying
biology, as well as empiric pharmacogenomic approaches in cell line (Garnett
et al., 2012) or patient-derived model platforms (Gao et al., 2015), may inform patient selection by molecular alterations,
with or without consideration of the underlying disease histology. While the principal goal of Phase I testing is to define
toxicities and pharmacokinetics, incorporating biomarker selection into early
phase testing may be rational for such
agents and could spare the enrolment of
patients who are least likely to derive
any therapeutic benefit. This approach,
however, adds to the complexity of patient recruitment by demanding prescreening, sometimes of large numbers of
patients. Until recently, many genotypespecific trials had molecular prescreening
efforts embedded within the protocols to
recruit eligible patients for therapeutic interventions. These are now being usurped
by more widespread clinical application of
next-generation sequencing, which has
become feasible and more affordable.
Genomic-based trials leveraging this
technology can take multiple forms but
include both umbrella (randomized or
non-randomized clinical trials that are histology specific, investigating different
therapeutic interventions such as different
drugs or drug combinations matched to
molecular aberrations in a single cancer
type) or basket trials (randomized or
non-randomized clinical trials that are histology agnostic, investigating a therapeutic intervention such as a drug or drug
combination targeting a specific molecular aberration across different cancer
types). There are challenges associated with the implementation of these
genomic-based trials. For instance, they
often require a sufficiently broad suite of
agents to enable treatment assignments,
resulting in large collaborative trials
such as the NCI-MATCH program, which
includes drugs from multiple pharma-
ceutical companies (NCT02465060). The
recruitment of rare genotypes into
these trials requires national or international participation to ensure successful
completion.
Correlative Science in Early Phase
Trials
Underscoring the translational interactions between basic discovery and clinical
development teams, correlative science
studies are increasingly incorporated
into early phase trials to assess pharmacodynamics and monitor response or to
detect therapeutic resistance and its underlying mechanisms. While technological advances support the necessary analyses of small biopsy samples or profiling
of circulating tumor markers (e.g., liquid
biopsies), these add an extra layer of
complexity and sometimes demand sophisticated processing, handling, and
tracking of samples at participating sites.
Given the added burden and costs, as
well as the potential risk to participants
associated with invasive procedures, it is
critical that such correlative programs
are scientifically justified, carefully designed, and efficiently implemented.
Future of Clinical Trials
Correlative Studies to Optimize
Clinical Development
While the examples described above
illustrate the potential of rapid translation
of preclinical discoveries into clinical
successes, the underlying principles of
clinical development must be anchored
by thoughtful science and clinical relevance. Though timeliness and efficiency
are desirable for the reasons described
above, the substantial costs—measured
in time, patient participation, and clinical
and financial resources—demand that
study design and implementation not be
rushed. Robust preclinical data to understand therapeutic potential and disease
biology are essential for this pursuit.
However, new and emerging analytical
tools enable correlative studies of clinical
trials, which can increasingly shape and
optimize ongoing clinical development.
Beyond the pharmacokinetic and pharmacodynamic endpoints that are more
traditionally considered in Phase I studies,
the identification and characterization of
exceptional responses and acquired
resistance in these same trials present
an opportunity to hone patient selection
Figure 1. Future Goals for Clinical Trials
(A) Robust and efficient exchange of knowledge between correlative science studies and clinical trials, including basket and umbrella trials. (B) Set clear early
go-no-go criteria and high bars for clinically meaningful and important outcomes. (C) Target micrometastases in adjuvant setting and monitor dynamically for
markers of residual disease. (D) Globalization of clinical trials and data sharing and engagement of all key stakeholders, especially patients.
or design combination treatment strategies. Given the potential impact on subsequent scientific investigation and drug
development, such correlative analyses
of early phase trials should be performed
and disseminated sooner. For example,
acquired resistance develops in 25%
of patients with melanoma achieving
objective responses with PD-1 blockade,
and fairly straightforward paired genomic
analyses in only four pembrolizumabtreated patients permitted the recent
identification of underlying alterations in
interferon signaling and antigen presentation (Zaretsky et al., 2016). Prioritization of
such analyses in the early clinical trials
could likely have resulted in much earlier
delineation of this important biology,
which has implications for ongoing development of immunotherapy strategies.
Striving for Meaningful
Clinical Benefit
The ever-increasing cost of new anti-cancer therapies demands that our efforts be
focused on agents that can deliver meaningful clinical benefit—defined as significant improvements in the efficacy or
substantial reductions in the toxicities of
existing standard of care. In addition to
producing new therapies that are clinically
useful and potentially cost effective, establishing a high bar for meaningful
benefit requires more nimble clinical trials.
Efforts by organizations such as the
American Society of Clinical Oncology to
establish benchmarks for clinical benefit
represent an important step in the right
direction (Ellis et al., 2014). In addition to
defining acceptable magnitude of benefits to justify investment in Phase III trials,
more accurate means to predict success
and avoid futile clinical trials are also
required. Clear go-no-go criteria for
agents to move from early to late-phase
testing are necessary, and the bar should
be set high even at this intermediate
checkpoint. Assuming continued development of biomarker-driven strategies
as described earlier, defining populations
and predictive assays in early phase
expansion cohorts offers an efficient
path to identify signals of activity. However, existing limits on our knowledge of
the prognostic effects of many molecular
alterations makes the interpretation of
time-to-event endpoints (e.g., progres-
sion-free survival) based on historical
data difficult. Unless compelling objective
response rates are observed to support
potential paths for accelerated regulatory
approval, randomized Phase II or III trial
designs will often be most appropriate.
While randomization generally increases
the size of these intermediate or latestage trials, setting a high bar will help
to contain the number of participants
required.
Aiming for Cure with Focus on the
Adjuvant Setting
While immunotherapy strategies are challenging the paradigm that metastatic solid
tumors are incurable, the fact remains
that the best opportunity to improve cancer survival statistics is in the adjuvant or
perioperative setting, where substantially
lower tumor burden could reduce the
probability of drug resistance that prevents disease eradication. Development
of clinical trials in this setting, particularly
among patients at very high risk of lethal
recurrence, needs increased focus and
investment. A major challenge to advances in this area, however, remains
our limited understanding of the biology
Cell 168, February 9, 2017 577
of micrometastastic disease and recurrence and our inability to predict what
makes a curative adjuvant treatment.
Better preclinical models and more basic
and translational research are urgently
needed to address these gaps. In addition, because adjuvant trials are typically
large, long, and expensive to conduct,
improved predictors of risk and validated
surrogates of benefit would be of great
value to accelerate this process. Markers
of minimal residual disease, such as
circulating cell-free (cf)DNA, could have
a place in both patient selection and
dynamic monitoring of response to therapy in this setting.
Collective Wisdom in Clinical Trials
Clinical trials are constantly evolving and
expanding beyond their current geographical, operational, and bioinformatic
boundaries. Driven by many factors such
as lower costs and faster recruitment
rates, the globalization of clinical research
has become an increasingly prevalent
phenomenon, with a rising number of clinical trials being conducted outside of the
US and other developed countries (Viergever and Li, 2015). There are continued efforts to harmonize regulatory standards
and data quality in clinical trials on a global
scale, and some emerging markets have
successfully achieved these goals. In
addition, national and international initiatives to share data obtained from clinical
trials and real-life practice are actively being pursued to enable collective learning
through the power of large sample sizes
(Siu et al., 2016). These efforts will be
578 Cell 168, February 9, 2017
essential to maximize the efficiency of
drug development for increasingly precise
clinical indications.
To summarize, landmark achievements
in cancer science are providing unprecedented opportunities for improved clinical
therapies (Figure 1). By incorporating
such knowledge into clinical trials and implementing non-traditional approaches to
trial design, we have seen more rapid regulatory approval of several effective therapies for specific patient populations over
the last decade. In the future, we expect
that the development of emerging anticancer agents will necessitate increasingly innovative and agile clinical trial
designs, incorporating biomarker selection and correlative science in a dynamic
manner. Importantly, the success of clinical trials in delivering this progress is
contingent on the engagement of key
stakeholders, including academic and
community physicians, basic and translational researchers, regulatory agencies,
the pharmaceutical sector, patients, and
advocates, to ensure that time and resources are expended to address the
most important and meaningful questions
to advance cancer care.
REFERENCES
Ellis, L.M., Bernstein, D.S., Voest, E.E., Berlin, J.D.,
Sargent, D., Cortazar, P., Garrett-Mayer, E.,
Herbst, R.S., Lilenbaum, R.C., Sima, C., et al.
(2014). J. Clin. Oncol. 32, 1277–1280.
Gao, H., Korn, J.M., Ferretti, S., Monahan, J.E.,
Wang, Y., Singh, M., Zhang, C., Schnell, C.,
Yang, G., Zhang, Y., et al. (2015). Nat. Med. 21,
1318–1325.
Garnett, M.J., Edelman, E.J., Heidorn, S.J., Greenman, C.D., Dastur, A., Lau, K.W., Greninger, P.,
Thompson, I.R., Luo, X., Soares, J., et al. (2012).
Nature 483, 570–575.
Garon, E.B., Rizvi, N.A., Hui, R., Leighl, N., Balmanoukian, A.S., Eder, J.P., Patnaik, A., Aggarwal, C.,
Gubens, M., Horn, L., et al.; KEYNOTE-001 Investigators (2015). N. Engl. J. Med. 372, 2018–2028.
Gehan, E.A. (1961). J. Chronic Dis. 13, 346–353.
Hamid, O., Robert, C., Daud, A., Hodi, F.S., Hwu,
W.-J., Kefford, R., Wolchok, J.D., Hersey, P., Joseph, R.W., Weber, J.S., et al. (2013). N. Engl. J.
Med. 369, 134–144.
Johnson, J.R., Bross, P., Cohen, M., Rothmann,
M., Chen, G., Zajicek, A., Gobburu, J., Rahman,
A., Staten, A., and Pazdur, R. (2003). Clin. Cancer
Res. 9, 1972–1979.
Kazandjian, D., Blumenthal, G.M., Chen, H.-Y., He,
K., Patel, M., Justice, R., Keegan, P., and Pazdur,
R. (2014). Oncologist 19, e5–e11.
Kwak, E.L., Bang, Y.-J., Camidge, D.R., Shaw,
A.T., Solomon, B., Maki, R.G., Ou, S.-H.I., Dezube,
B.J., Jänne, P.A., Costa, D.B., et al. (2010). N. Engl.
J. Med. 363, 1693–1703.
Poole, R.M. (2014). Drugs 74, 1973–1981.
Siu, L.L., Lawler, M., Haussler, D., Knoppers, B.M.,
Lewin, J., Vis, D.J., Liao, R.G., Andre, F., Banks, I.,
Barrett, J.C., et al. (2016). Nat. Med. 22, 464–471.
Soda, M., Choi, Y.L., Enomoto, M., Takada, S., Yamashita, Y., Ishikawa, S., Fujiwara, S., Watanabe,
H., Kurashina, K., Hatanaka, H., et al. (2007). Nature 448, 561–566.
Viergever, R.F., and Li, K. (2015). BMJ Open 5,
e008932.
Zaretsky, J.M., Garcia-Diaz, A., Shin, D.S., EscuinOrdinas, H., Hugo, W., Hu-Lieskovan, S., Torrejon,
D.Y., Abril-Rodriguez, G., Sandoval, S., Barthly, L.,
et al. (2016). N. Engl. J. Med. 375, 819–829.
Leading Edge
Commentary
How Much Longer Will We Put Up With
$100,000 Cancer Drugs?
Paul Workman,1 Giulio F. Draetta,2 Jan H.M. Schellens,3,4 and René Bernards5,*
1Cancer
Research UK Cancer Therapeutics Unit, The Institute of Cancer Research, London, UK
of Genomic Medicine and Institute for Applied Cancer Science, The University of Texas MD Anderson Cancer Center, Houston,
Texas, USA
3Division of Clinical Pharmacology, the Netherlands Cancer Institute, 1066 CX Amsterdam, the Netherlands
4Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht University, Utrecht, the Netherlands
5Division of Molecular Carcinogenesis and Cancer Genomics Centre Netherlands, the Netherlands Cancer Institute, 1066 CX Amsterdam,
the Netherlands
*Correspondence: r.bernards@nki.nl
http://dx.doi.org/10.1016/j.cell.2017.01.034
2Department
The spiraling cost of new drugs mandates a fundamentally different approach to keep lifesaving therapies affordable for cancer patients. We call here for the formation of new relationships
between academic drug discovery centers and commercial partners, which can accelerate the
development of truly transformative drugs at sustainable prices.
The Problem
The recently developed targeted drugs
and immunotherapies deliver significant
benefit to cancer patients. However,
the spiraling prices of these new drugs
threaten the financial sustainability of cancer treatment. Healthcare spending has
risen sharply in the United States, reaching 17.1% of the gross domestic product
in 2014. Cancer drugs are of particular
concern, even if they only represent a
fraction of the overall costs. As early as
2012, 12 of the 13 newly-approved cancer
drugs were priced above $100,000 annually, and the situation has only gotten
worse since (Light and Kantarjian, 2013;
Mailankody and Prasad, 2015). Particularly worrisome is the notion that these
drugs often need to be combined for
optimal clinical results. For instance, the
cost of the combination of nivolumab
(anti-PD-1) and ipilimumab (anti-CTLA4)
is priced around $252,000, exceeding
the median cost of a US home ($240,000
in 2016). With a lifetime risk of developing
cancer of close to 40%, the problem
is clear.
The pharmaceutical industry has traditionally defended these high prices by
pointing at the high attrition rate during
clinical drug development and the cost
of large registration studies. But are these
arguments still as sturdy in the new era
of personalized medicine? Targeted cancer drugs are often genotype-selective,
which makes for a higher success rate
due to better upfront patient selection.
Consequently, approval of such drugs
often no longer requires expensive phase
III trials with thousands of patients. As one
example, the registration study of the
ALK inhibitor crizotinib required only
347 patients with ALK-positive lung cancer (Shaw et al., 2013). Furthermore,
in 2016 FDA approved crizotinib for
the treatment of lung cancer patients with mutations in ROS1, which the
drug also inhibits, on the basis of a study
of only 50 patients (Shaw et al., 2014).
So why are drug prices going up instead
of down? One clue can be gleaned
from the pricing of the recent checkpoint
blockade immunotherapeutics nivolumab
and pembrolizumab. Both drugs received
initial FDA approvals in 2014 for metastatic melanoma, but their indication was
widened in 2015 to include certain lung
cancers and renal cancer (nivolumab)
and in 2016 to head and neck cancer
(pembrolizumab) and Hodgkin lymphoma
(nivolumab). If development cost would
be a major factor in the pricing structure,
a simple law of economics would have
mandated a considerable reduction in
price when the eligible patient population
increases, but that has hardly happened.
This is a recurring theme in pharma. For
instance, trastuzumab was first approved
for advanced breast cancer and later
also for early disease (adjuvant) without
a reduction in price. Healthcare payers
should not accept this lack of price-vol-
ume relationship. Moreover, there is very
little relationship between drug price and
clinical benefit (Mailankody and Prasad,
2015). This has sparked widespread criticism, alleging that cancer drug pricing is
primarily based on ‘‘what the market will
bear.’’ There is a clear and urgent necessity to lower cancer drug prices to keep
lifesaving drugs available and affordable
for patients. As one patient advocate
recently put it: ‘‘Innovation is meaningless
if nobody can afford it.’’
Lack of Effective Solutions
Much has been written about the reasons
behind the exorbitant drug prices and
what to do about it. One recurring theme
is the notion that the US federal government is prohibited by law from negotiating drug prices as a result of the 2003
Medicare Prescription Drug, Improvement and Modernization Act. Considering
that Medicare and Medicaid spend $ 140
billion on medicines annually, this represents a serious impediment in driving
down drug prices. Lack of competition
and a general absence of a connection
between drug price, sales volume, and
clinical performance are other arguments
in the drug pricing discussion (Jaffe,
2015). Indeed, lack of competition and
bargaining power made US prices of cancer drugs among the highest in the world,
increasing by 10% annually between
1995 and 2013, far above the average
inflation rate (Howard et al., 2015).
Cell 168, February 9, 2017 ª 2017 Elsevier Inc. 579
While negotiations may bring prices
down, a recent cost comparison in EU
countries shows that the ability of individual nations to negotiate discounts is
limited, most likely due to the modest
market sizes of the EU countries (van
Harten et al., 2016). In the longer term,
it is unlikely that the European Union
will be able to collectively negotiate with
the pharmaceutical industry, given that
some nations in the EU with a large
pharma sector are likely to protect their
national interests. The UK’s pioneering
National Institute for Health and Care
Excellence (NICE) has been able to
restrain prices to £30,000 per ‘‘qualityadjusted life year’’ (QUALY) added or
£50,000 per QUALY for ‘‘end of life’’
treatments that include many cancer
drugs—but this has also resulted in
many oncology agents being turned
down or delayed, leading to complaints
from patient groups and manufacturers.
There is also increasing evidence of pressure on pricing in the US. Researchers at
the Memorial Sloan Kettering Cancer
Center have developed an online tool
called Drug Abacus to help healthcare
providers assess the value of cancer
drugs (www.drugabacus.org/). What is
urgently needed however are mechanisms to encourage scientific and therapeutic innovations that will allow cancer
patients to access new treatments at
affordable and sustainable prices.
Inefficiency in Drug Development
One element that contributes significantly
to the high cost of cancer drugs is the
inefficiency of the overall commercial
enterprise. As one recent example, there
are currently 803 clinical trials testing
checkpoint immune-therapeutics (at least
12 antibodies from a dozen different
pharma companies), which together plan
to enroll over 166,000 patients (Brawley,
2016). There is enormous redundancy in
these studies, as many pharmaceutical
companies perform similar trials with
comparable drugs, but fail to share the
data generated. This herd mentality is
caused in part by the notion that immune
checkpoint therapies can indeed lead to
long-lasting remissions (potentially even
to cures) and that significant numbers of
patients in each clinical indication benefit
from these treatments. While it is in the
short-term good that so many patients
580 Cell 168, February 9, 2017
get access to potentially lifesaving drugs,
in the longer term, patients will have to
pay the price for this inefficiency and
duplication.
Another factor is the frequent absence
of a rigorous biomarker program to identify patients who may benefit from a given
drug. The primary incentive of the pharma
industry is to increase sales, which are
restricted by identifying drug-responsive
subpopulations. Biomarkers are critical,
as they represent a handle to control
drugs costs to society. Regulatory bodies
should set standards for drug approval
employing validated and clinically useful
biomarkers for patient selection. This
will prevent patients being treated with
toxic drugs that do not improve survival
and/or quality of life and will save costs
to society.
Historically, some 90% of all drugs
entering the clinic have failed at some
stage and most biotech start-ups have
a similar fate. This severe attrition rate
in drug discovery and development
has been attributed to various factors,
including failure to validate drug targets
with sufficient rigor; limited predictive
value of animal models; inability to firmly
identify tumor subtypes that might benefit
from treatment; poorly defined clinical
endpoints; organizational pressure to
continue clinical development, such as
stock market considerations for small
one-product biotech firms; and senior
management fear of admiting defeat in
larger pharma. The Tufts Center for the
Study of Drug Development has estimated that the development of a drug on
average costs $2.558 billion, but it is
difficult to determine how this was
calculated—and includes costs of failed
projects and controversially incorporates
selected examples of successful drugs
and the cost of capital (Avorn, 2015).
Some consideration should be given to
the fact that large pharmaceutical companies encounter significant challenges
due to the large size of their operations,
redundant activities at multiple sites,
huge infrastructure costs, heavily matrixed organizations with multiple levels
of decision makers, and endless rounds
of restructuring, mergers, acquisitions
and down-sizing—which all ultimately
contribute to time delays (time is money!),
reduced productivity, and increased
expenses. Smaller biotech start-ups
have significant challenges as well, with
tremendous recent increases in space
and infrastructure build-up costs and
competition for talent (with consequent
increased employee salaries), particularly
in the Boston and San Francisco communities. Furthermore, all commercial entities, large and small, spend very significant sums on hefty salaries for senior
management. We therefore assume that
both organizational and scientific inefficiencies contribute to escalating drug
development costs.
A New Approach
Many of the fundamental discoveries that
formed the basis for new categories of
cancer drugs were made by academia.
Examples of truly disruptive contributions
from academic research that enabled
radically different treatment strategies
include the identification of recurrent
mutations in cancers from the largesale sequencing efforts (enabling targeted therapeutics) and the decadeslong investments in unraveling the basic
biology underlying recognition of cancer
cells by the immune system (enabling
recent immuno-oncology drugs). The National Cancer Institute budget of over
5 billion dollars annually virtually guarantees that this stream of innovations in
oncology drug targets from academia
will not dry up any time soon. It should
also be recognized that academic drug
discovery has already been very successful as several drugs that are part of the
therapeutic armamentarium have been
developed by academic researchers,
e.g., the brain tumor DNA alkylating drug
temozolomide and the prostate cancer
CYP17 inhibitor agent abiraterone, as
well as the biomarker strategy for the
PARP inhibitor olaparib in BRCA mutant
ovarian cancer patients.
To further develop these academic
discoveries, the traditional model of
self-supporting research investigators
who drive their independent research
programs needs to be complemented
by concerted multidisciplinary team
efforts that are adequately financed
and staffed with scientists having all the
required expertise to enable drug discovery (Frye et al., 2011; Schultz Kirkegaard and Valentin, 2014). Indeed, in
recent years there has already been
a steady increase in the number of
academic centers involved in drug discovery. The Cancer Research UK Cancer Therapeutics Unit at The Institute
of Cancer Research, London (www.icr.
ac.uk/our-research/our-research-centres/
cancer-research-uk-cancer-therapeuticsunit) and the Institute for Applied Cancer
Science at MDAnderson cancer center
(www.mdanderson.org/cancermoonshots/
research_platforms/Institute_for_applied_
cancer_science.html) are just two examples of relatively large-scale academic
cancer drug discovery units that have
been established to date. The Academic
Drug Discovery consortium currently
lists 146 drug discovery centers in 15
countries, 80% of which have programs
in oncology drug development (www.
addconsortium.org/).
A key advantage of academic drug
discovery is the freedom and indeed incentivization to tackle major challenges
that would be viewed as too risky by
big pharma and even by many biotech
companies. Currently only a fraction of
the cancer genes listed in the Cancer
Gene Census have drugs or chemical
leads that act on the cognate protein.
This means that there are very large
numbers of cancer genes that remain
to be drugged. For example, we have
no drugs that work directly on mutant
KRAS, mutant p53 or MYC. Hence there
is a huge amount of work to be done to
complete the job of drugging of the cancer genome. Moreover, there are gene
classes that do not fall into the conventional categories of oncogene addiction
and synthetic lethality targets, including
non-oncogene addiction, microenvironmental, drug resistance and immunooncology targets.
Tackling any one of these new targets
carries very high risk—either biological
risk because there is relatively little knowledge of the role of the gene in cancer, or
technical risk because the protein is not
readily druggable by current technology.
These are the targets for which academic
drug discovery can make an enormous
impact in a number of ways. For example:
(1) conducting very rigorous target validation to ensure robustness of the effects;
(2) linkage of the sensitivity to a robust
biomarker that can be potentially used
for patient selection as well as pharmacodynamic biomarkers to demonstrate
target engagement as part of a Pharma-
cologic Audit Trail (Banerji and Workman,
2016); (3) demonstration of druggability
by a small molecule approach; (4) production of chemical probe or biological reagent that demonstrates proof
of concept in a disease-relevant animal
model; (5) progression of a candidate
drug through preclinical development;
and (6) conduct of an early stage clinical trial to show tolerability and proof
of concept in cancer patients (Hoelder
et al., 2012).
Factors that have improved the success of cancer drug discovery efforts
in academia include embedding experienced drug discovery scientists within a
comprehensive cancer center that provides expertise in basic cancer research,
clinical trials and treatment. The recruitment of experienced medicinal chemists
and drug discovery biologists has been
critical. Adequate funding and resources
to support a portfolio of projects as well
the range of expertise and technologies
is important as is experienced leadership
and decision-making.
Once new chemical entities have been
developed and tested in experimental
animals, the (mostly academic) hospitals become involved in the three major
phases of clinical testing of compounds.
Selected Good Laboratory Practice and
Good Manufacturing Practice-certified
academic pharmacies can develop and
manufacture oral and/or parenteral drug
formulations fit for clinical use. Academic clinicians clearly have the skills to
execute large clinical trials, especially
given that some of the recent large clinical
trials were ‘‘investigator initiated,’’ meaning that an academic investigator was
leading the study. Based on the arguments above, it is evident that academia
in principle has all the tools and skill sets
to discover drug targets, to convert these
targets into clinical candidates and to test
these compounds rigorously in clinical
trials.
Bringing Drugs to Patients under
New Assumptions
There are three main reasons why academic drug development typically stalls
at the stage of clinical testing. First, the
stringent quality control over the largescale manufacturing of clinical grade
drugs and their formulation is not a routine
skill of academic groups. Second, the
funds to support the high cost of performing non-clinical regulatory toxicology
studies and clinical trials are hard to raise
by non-profit organizations. Third, even
when these first two steps could be
executed, academic drug discovery and
development units are not equipped to
handle marketing and sales of approved
drugs. Yet, it is at the level of commercialization that the interests of large pharma
to maximize return on investment are
diagonally disparate from the typically
idealistic motivation that drives most
academics to spend countless hours at
modest compensation to solve important
problems in oncology. Nevertheless, academics are driven into the arms of big
pharma after initial proof of concept clinical trials for the reasons listed above.
While it is gratifying for most academic investigators to see their discoveries reach
the clinic, it leaves them unable to influence the pricing of ‘‘their’’ drugs when
they reach the market. There are examples of charities funding later stage trials,
but these are exceptions and academic
drug discovery cannot rely on charity
funding only to bring their candidates to
patients.
How can we break free of this catch 22
situation in academic drug development?
A possible solution may reside in what
happens to cancer drugs when their
patents expire. At this point so-called
‘‘generic’’ drug makers bring generic versions (biosimilars in the case of biological
agents) to the market at greatly reduced
prices. These lower prices are possible
because the companies do not bear the
cost of research and development. Given
that generic drug makers are used to
working with lower profit margins, they
may be one potential partner to develop
highly innovative, but de-risked, drugs
from academic drug discovery and development (Figure 1). Especially when the
drugs have a strong mechanistic rationale
and an associated biomarker of response
(key aspects of academic drug development), the registration trials can be small
and the success rate much higher than
in traditional pharma trials. As a result,
the prices of such drugs can be far lower
than we have witnessed recently. Regulatory bodies are also open to novel
ways for drug approval. The European
Medicines Agency (EMA) has launched
an adaptive licensing program enabling
Cell 168, February 9, 2017 581
Figure 1. The Academic Drug Discovery and Development Continuum and Its Relationships with Stakeholders
Critical components of the cancer drug discovery and development process through commercialization are described. Through comprehensive integration of
expertise, cancer biologists and geneticists, drug discovery scientists and pharmacologists are able to precisely formulate a Clinical Candidate Profile based on
tumor subtype(s) and patient population that might best benefit from treatment. Project financing leverages philanthropic donations and partnerships with CROs
and generic drug makers, allowing not-for profit entities to retain control from the start through commercialization.
CRO, contract research organization, a provider of services to the biopharmaceutical industry; PD, pharmacodynamics, determines a drug mechanism of action
and safety profile; generic drug maker, a high-volume, low profit margin organization devoted to the manufacturing and commercialization of drugs past their
patent expiry; pharmaco-economics, a comprehensive evaluation of the impact of a given program on the health of a population, often leads to decisions on
policies; response biomarker, a biological indicator predicting response to a given treatment.
companies to obtain marketing authorization approval on the basis of a small
well designed, biomarker supported
trial. Post-marketing commitments of the
company forcing them and the community to deliver extended proof of benefit
at acceptable risk is a safeguard to them
and the community that the early market
launch was justified. If such commitments
are not met, the drug will be withdrawn.
Two elements will be mission critical
for this model to succeed. First, academic organizations will need to abide
by their societal responsibility and resist
the temptation to sell their drug candidate to the highest bidder. Second, it will
be imperative that agreements on price
caps are part of the negotiations with
potential investors or with companies
that take forward drugs arising from academic drug development. Ideally, this
approach would also be accompanied
by pricing strategy leading to affordable
drug cost in middle- and low-income
countries, thereby reducing inequality in
global cancer therapy. Given the substantial de-risking achieved prior to commercialization, our model should be attractive
582 Cell 168, February 9, 2017
to these parties and their investors. The
academic drug discovery and development units could be sustained in this
model by receiving royalties on sales of
the drugs they originated.
We need to recognize that building up
a collection of academic centers with
required scale and expertise that would
produce significant numbers of drug candidates will take time and money and will
not be an overnight solution to the global
pricing problem. It is, however, a move
in the right direction. Moreover, the creation of such groups alongside generics
partners or newly created commercial entities will create competition and drive
down prices in conventional pharma and
biotech.
Where to Start with Academic Drug
Discovery and Development?
There is quite a bit of low hanging fruit
to be harvested by academic consortia.
In addition to their main task of discovering mechanistically innovative drugs, a
near term focus should be on the repurposing of existing patent-expired drugs
by finding new indications for these
drugs, linked to effective biomarkers that
are predictive of response. Academic
groups are well equipped to carry out
this research, and it would help if specific
funding mechanisms would be made
available for such projects. We emphasize
that such funds should not be allocated
at the expense of investments in fundamental cancer research. Second, academic consortia should also focus on
finding effective combinations for drugs
that were abandoned for lack of single
agent activity, which appears in one out
of three cases to be the primary reason
new chemical entities are dropped from
early phase clinical trials (Dimasi, 2001).
Lack of single agent activity is often
caused by redundancy or feedback in
signaling pathways, which makes the
inhibition of a single pathway ineffective
without concomitant inhibition of the
redundant or feedback pathway. As one
example, had the BRAF inhibitor vemurafenib been tested initially in BRAF mutant
colon cancer, it would have been discarded as ineffective, whereas it turned
out to be very effective in BRAF mutant
melanoma. We argue that far too many
potentially useful drugs are discarded
early for the wrong reasons.
Besides tackling these initial lower risk
projects, which can serve as a proof of
concept (and cash flow) for the model
proposed here, the academic drug discovery centers should have as a major
emphasis the challenging task of discovering and developing drugs against highly
innovative drug targets emerging from
academic research. Such efforts must
have a sharp focus on mechanism-based
therapies with strong associated biomarkers of response to reduce attrition
rates while allowing small clinical trials to
show efficacy.
By partnering with generic drug makers
or new companies specifically formed
to enable this new model, academic
drug discovery units and interested drug
makers can lead by example and deliver
innovative drugs at sustainable prices.
ACKNOWLEDGMENTS
R.B. is the founder of Qameleon Therapeutics
(www.qamelonrx.com) that seeks to build a new
model of partnership between generic drug
makers and academia as described in this commentary. The authors are involved in multiple
drug discovery and development projects with
actual or potential commercial revenues and act
as advisers to several companies. We welcome
interaction with any drug maker interested in
considering academic partnerships as described
herein.
REFERENCES
Avorn, J. (2015). N. Engl. J. Med. 372, 1877–1879.
Banerji, U., and Workman, P. (2016). Semin. Oncol.
43, 436–445.
Brawley, L. (2016). Cancer Lett. 42, 2–18.
Dimasi, J.A. (2001). Clin. Pharmacol. Ther. 69,
297–307.
Frye, S., Crosby, M., Edwards, T., and Juliano, R.
(2011). Nat. Rev. Drug Discov. 10, 409–410.
Hoelder, S., Clarke, P.A., and Workman, P. (2012).
Mol. Oncol. 6, 155–176.
Howard, D.H., Bach, P.B., Berndt, E.R., and Conti,
R.M. (2015). J. Econ. Perspect. 29, 139–162.
Jaffe, S. (2015). Lancet 386, 2127–2128.
Light, D.W., and Kantarjian, H. (2013). Cancer 119,
3900–3902.
Mailankody, S., and Prasad, V. (2015). JAMA Oncol. 1, 539–540.
Schultz Kirkegaard, H., and Valentin, F. (2014).
Drug Discov. Today 19, 1699–1710.
Shaw, A.T., Kim, D.-W., Nakagawa, K., Seto, T.,
Crinó, L., Ahn, M.-J., De Pas, T., Besse, B., Solomon, B.J., Blackhall, F., et al. (2013). N. Engl. J.
Med. 368, 2385–2394.
Shaw, A.T., Ou, S.H., Bang, Y.J., Camidge, D.R.,
Solomon, B.J., Salgia, R., Riely, G.J., VarellaGarcia, M., Shapiro, G.I., Costa, D.B., et al.
(2014). N. Engl. J. Med. 371, 1963–1971.
van Harten, W.H., Wind, A., de Paoli, P., Saghatchian, M., and Oberst, S. (2016). Lancet Oncol. 17,
18–20.
Cell 168, February 9, 2017 583
Leading Edge
Perspective
Implementing Genome-Driven Oncology
David M. Hyman,1,5 Barry S. Taylor,2,3,4 and José Baselga1,2,5,*
1Department
of Medicine
Oncology and Pathogenesis Program
3Department of Epidemiology and Biostatistics
4Marie-Josée and Henry R. Kravis Center for Molecular Oncology
Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, NY 10065, USA
5Weill Cornell Medical College, 1300 York Avenue, New York, NY 10065, USA
*Correspondence: baselgaj@mskcc.org
http://dx.doi.org/10.1016/j.cell.2016.12.015
2Human
Early successes in identifying and targeting individual oncogenic drivers, together with the
increasing feasibility of sequencing tumor genomes, have brought forth the promise of genomedriven oncology care. As we expand the breadth and depth of genomic analyses, the biological
and clinical complexity of its implementation will be unparalleled. Challenges include target credentialing and validation, implementing drug combinations, clinical trial designs, targeting tumor heterogeneity, and deploying technologies beyond DNA sequencing, among others. We review how
contemporary approaches are tackling these challenges and will ultimately serve as an engine
for biological discovery and increase our insight into cancer and its treatment.
Introduction
Cancer is a disease of the genome, and major strides have been
made in our understanding and treatment of this heterogeneous
collection of diseases, beginning with the initial identification of
oncogenes and tumor suppressor genes to the development of
the first generation of targeted therapies and culminating in the
full annotation of the genomic landscape of the most common
cancer types (Kandoth et al., 2013). Much of this progress can
be traced to technological advances in sequencing, from capillary-based sequencing technologies to the modern massively
parallel sequencing of today, collectively known as next-generation sequencing. These advances have enabled the routine
genomic study of every tumor at the point of care and will redefine clinical management and translational research in transformative ways. Detailed genomic characterization of tumors is
already driving the definition of a new taxonomy of human cancers that will, ultimately, complement current histology-based
classifications (Hoadley et al., 2014). Routine genomic profiling
will also improve prognostication of clinical outcomes, as has
already been achieved with human epidermal growth factor receptor-2 (HER2) amplifications in breast cancer and mutations
in FLT3, NPM1, KIT, and CEBPA in acute myelogenous leukemia. The farthest reaching consequence of routine tumor
profiling, however, will be the identification of genetically driven
tumor dependencies and vulnerabilities that will lead to the
further development of precision therapies and combinatorial
treatment approaches. In fact, as a preview of this concept, there
are already a plethora of genomic alterations for which targeted
therapies have been approved.
Although the promise of such progress is immense, there are
many obstacles to broad implementation of genome-based cancer care. These challenges are both practical and scientific.
Soon, all cancer patients will have the opportunity to obtain
detailed genomic profiles of their tumors, but this is only the first
584 Cell 168, February 9, 2017 ª 2016 Published by Elsevier Inc.
and perhaps easiest step. How do we differentiate between
therapeutically actionable alterations and biologically neutral
passenger changes? How do we manage and prioritize the
biologic credentialing of the large number of novel alterations
now routinely identified through prospective tumor genomicscreening programs? How can we utilize genome-driven clinical
trials to accelerate the biologic investigation of incompletely
characterized alterations now that they are routinely being identified in patients receiving ongoing care? What strategies will be
most effective in engendering prolonged response to targeted
therapy and mitigating the consequences of tumor heterogeneity
and acquired resistance? How do we ensure that our ever-expanding knowledge of the cancer genome and the therapeutic
vulnerabilities encoded therein are shared among the biomedical
community in a manner that maximizes further discovery? What
depth and breadth of genomic characterization of each cancer
type will be required, and how do we incorporate technologies
in the clinic beyond DNA sequencing? How can we improve
the efficiency of genomic hypotheses testing in the clinic, and
how do we ensure we are learning the most we can from each
treated patient? Finally, how do we target mutations that individually occur rarely but, in aggregate, affect a large proportion of
the cancer population? Here, we review how contemporary approaches in precision oncology are beginning to address these
key challenges and, in so doing, serve as an engine for biological
discovery that will ultimately increase our insight into this complex set of diseases.
At the outset, we recognize that as with any new field of science and medicine, a diversity of views on the value of this
approach is inevitable. The emerging field of precision medicine
is no different, and some authoritative voices have raised appropriate concerns (Tannock and Hickman, 2016; Voest and Bernards, 2016). First, it has been pointed out that despite the
immense complexity of the task at hand, there is a lack of
Table 1. Genes Used to Guide FDA-Approved Therapies
Mutations Used to Select Targeted Therapy
ABL1
CML, ALL
EGFR
lung cancer
ALK
lung cancer
ROS1
lung cancer
BRAF
melanoma
ERBB2
breast and gastric cancer
KIT
gastrointestinal stromal tumor
PDGFRA
leukemia, MDS
PDGFRB
dermatofibrosarcoma protuberans
BRCA1 and BRCA2 (germline)
ovarian cancer
Mutations Used to Select against Targeted Therapy
KRAS
colorectal cancer
NRAS
colorectal cancer
BRAF
colorectal cancer
much-needed collaboration among cancer institutions, and
even in those situations in which tumor sequencing takes place,
there is a low rate of patient participation in genomically
matched trials. There is truth in this concern, and later on in
this review, we will touch on some ongoing collaborative initiatives that are precisely aimed at addressing the current
fragmentation of efforts and inefficiency in clinical trials participation. Another far more serious criticism questions whether
this approach will work at all to begin with (Tannock and Hickman, 2016). In support of this view, one recently published randomized trial (the SHIVA study) found equivalent outcomes
when patients with multiple tumor types were randomized to
receive genomically matched versus conventional therapy (Le
Tourneau et al., 2014). This study was designed to explore the
off-label use of marketed drugs in a variety of unvalidated
genomic alterations in multiple tumor types and provides
good evidence of the inadequacy of legacy clinical trial paradigms for evaluating genome-driven medicine. The study was
underpowered, the genomic alterations had not been validated
as optimal targets, and the therapies used were not best in class
but rather commercially available agents. For example, any
alteration present in the PI3K/AKT/mTOR pathway led to treatment with the mTOR inhibitor everolimus despite the fact that
there is strong evidence that PTEN deletions and PI3k/AKT mutations do not confer sensitivity to this agent. Our reaction,
therefore, to this type of study is that the complexity of genomically driven oncology strongly argues that more narrowly
focused studies that ask questions about specific genomic
alterations or drugs rather than large randomized studies attempting to evaluate the entire approach will yield the most
informative data.
In short, we believe that both the underlying science and early
clinical successes of precision oncology support an optimistic
viewpoint, and although we acknowledge the significant
challenges that lay ahead, we have strived here to present a
critically self-reflective but solutions-focused perspective on
the field.
The Current and Emerging Clinical Landscape of
Precision Oncology
Limited genomic data are already being used to guide diagnosis,
inform prognosis, and support treatment decisions in a variety of
cancers. The pioneering example of molecularly driven cancer
medicine was the development and use of the kinase inhibitor
imatinib for the treatment of chronic myelogenous leukemias
(CML) that harbor the BCR-ABL1 balanced chromosomal translocation (Druker et al., 2006). For patients with this previously lethal form of leukemia, imatinib has dramatically improved their
outcomes to the point that the survival of CML patients is now
nearly identical to that of the general population (Bower et al.,
2016). Similarly, the advent of HER2-targeted therapies for the
treatment of women with newly diagnosed metastatic HER2positive breast cancer has radically changed the outcome of
what was until recently the most lethal form of breast cancer. A
woman diagnosed with metastatic HER2-positive breast cancer
can now anticipate a median survival of almost 5 years through
the use of state-of-the-art dual-HER2 blockade compared to
just 20 months prior to the advent of these therapies (Slamon
et al., 2001; Swain et al., 2015). Moreover, the addition of
HER2-directed therapy to routine chemotherapy for patients
with early-stage HER2-positive breast cancer has improved
cure rates by 35%–50% (Piccart-Gebhart et al., 2005).
From these early breakthrough therapies to today, cancer patients are benefiting from a diverse array of therapies targeting
genomic somatic aberrations in their tumors, including amplifications, gain-of-function mutations, and structural rearrangements, as well as germline loss-of-function mutations in at least
11 different genes arising in 10 different cancer types (Table 1).
These therapies have not only transformed the lives of many patients but also provided a powerful validation of the approach of
precision cancer medicine. Individual genomic findings are also
used to forgo therapies that are unlikely to result in clinical
benefit, such as KRAS, NRAS, and BRAF mutations in colorectal
cancers that would otherwise receive anti-epidermal growth
factor (EGFR)-targeted therapies (De Roock et al., 2010). These
approved targeted therapies, administered on the basis of
genomic observations in patients, represent a new era in cancer
therapy. In fact, genomic profiling in some form is now required
for the appropriate clinical management of a variety of tumor
types, including melanoma, gliomas, some sarcomas, as well
as carcinomas of the lung, breast, thyroid, ovary, and colon.
How then will ever-larger-scale prospective sequencing efforts further advance our field? We propose that emerging evidence indicates that a widespread implementation of prospective next-generation sequencing will be transformational at
several levels. Large-scale sequencing studies have demonstrated that nearly all of the genomic alterations currently used
to guide the selection of targeted therapy within specific disease
contexts also occur across a variety of other cancer types, albeit
typically at low frequencies (Chang et al., 2016; Kandoth et al.,
2013). Targeting these clinically validated predictive biomarkers
in a wider array of tumor types, therefore, is an opportunity to
immediately extend the benefits of precision oncology to a
broader population of patients and has been a major focus of
recent clinical development efforts. Evidence is growing that
this approach will work for at least some genomic alterations,
Cell 168, February 9, 2017 585
Table 2. Genes with Clinical Evidence Supporting Them as
Targets for Drug Development
Gene
Alteration(s)
RET
M918, fusions
MET
exon 14 splice, amplifications
AKT1
E17K
ERBB2
activating missense mutations
FGFR1/2/3
fusions, amplifications, activating
missense mutations
FLT3
ITD, D835
IDH1
R132
IDH2
R140
MAP2K1 (MEK1)
activating missense mutations
MTOR
activating missense mutations
BRAF
non-V600-activating mutations,
fusions
NTRK1/2/3
fusions
NRG1
fusions
PIK3CA
activating missense mutations
Homologous recombination
deficiencies (BRCA1, BRCA2,
PALB2, RAD50, ATM, RAD51,
RAD51B/C/D, FANCA,
CHEK1/2)
inactivating alterations
ARAF
S214
EGFR
rare activating missense
mutations, insertions
TSC1/2
inactivating alterations
SMARCA4
inactivating alterations
SMARRB1
inactivating alterations
but in a context- and tumor-dependent fashion. For instance,
BRAF V600 mutations are found in approximately half of cutaneous melanomas, and the use of RAF and MEK inhibitors in
these patients has been shown to improve survival, leading to
the approval of four drugs targeting this pathway (Robert et al.,
2015). However, nearly two-thirds of all BRAF V600 mutations
occur in non-melanoma cancers, suggesting that this therapeutic strategy may be applicable to a much larger number of patients. Clinical trials have been conducted in tumor types in
which the overall incidence of BRAF V600 mutations is sufficient
to conduct disease-specific studies. These efforts have already
identified clinical activity of these inhibitors in lung cancer, hairy
cell leukemia, and thyroid cancer (Brose et al., 2016; Planchard
et al., 2016; Tiacci et al., 2015).
Testing the efficacy of targeting BRAF V600 mutations in the
broad array of cancer types that uncommonly harbor this alteration has been far more challenging and spurred the development of new clinical study paradigms. One increasingly important approach to treating patients with rare genomic variants
has been the use of multiple-tumor-type, genomically selected
‘‘basket’’ studies. These studies revisit the tumor type as the
traditional organizing principle of a clinical trial and instead group
patients by the genomic alterations present in their tumors,
thereby reflecting an increasingly accepted reclassification of
586 Cell 168, February 9, 2017
human cancers not based on the organ of their origin but instead
on the molecular abnormalities that drive their growth and progression (Hoadley et al., 2014). In this way, patients whose tumors harbor the qualifying genomic alterations are eligible for
treatment regardless of cancer type. Tumor types anticipated
to have a sufficient prevalence of the alteration(s) of interest
are enrolled to their own ‘‘basket,’’ whereas all other patients
are grouped in a remaining ‘‘all-comers’’ cohort. These studies
recognize that response may be conditioned by the disease
context in which the genomic alteration arises and therefore
analyze efficacy independently for any tumor type that enrolls
a sufficient number of patients. A variety of statistical designs
have been utilized in these basket studies, and these approaches have been continuously iterated with each successive
generation of studies (Cunanan et al., 2016). In doing so, basket
trials have become a very efficient means of generating clinical
efficacy datasets across a broad variety of tumor types treated
with therapy targeting shared genomic alterations. The early
results of these studies have already begun to change clinical
practice. In the case of BRAF V600, we recently completed a
basket study that is expected to lead to regulatory approval of
the BRAF inhibitor vemurafenib in several additional tumor types,
including a group of rare disorders collectively known as histiocytoses for which no active therapies were previously available
(Hyman et al., 2015a). A similar basket study approach was
undertaken to determine the efficacy of the poly-ADP-ribose
polymerase (PARP) inhibitor olaparib in various solid tumors
harboring loss-of-function germline BRCA1 and BRCA2 alterations. This study led directly to the first regulatory approval of
a PARP inhibitor for women with germline BRCA1/2-associated
ovarian cancer, as well as initial proof of efficacy in prostate and
pancreas cancer (Kaufman et al., 2015).
Expanding the Druggable Genome
Expanding the use of therapy targeting previously validated
genomic biomarkers in a larger set of tumor types is only the first
step in broadening the utility of precision oncology. Maximizing
the potential of genome-driven oncology will require understanding the clinical significance of a much broader set of potentially actionable alterations, both within specific tumor types and
in a tumor-agnostic manner for targets where actionability is not
conditioned by tumor lineage.
The successful targeting of BRAF and PARP in multiple tumor
types suggests that targeting even rare genomic alterations independent of tumor type may be applicable to other driver genes
as well. There are more than 30 promising genomic targets today
(Table 2). This list is only a first iteration of a continuous process,
and we anticipate the number of targets will continue to expand
at a vibrant pace, although the final number of gene targets is
difficult to predict. The cumulative work establishing these targets has already markedly increased the fraction of patients
who are now believed to harbor at least one potentially actionable genomic alteration (Figure 1). Unlike the more common
genomic alterations described above, however, many of these
emerging targets are rare across all tumor types. An example
of this phenomenon is the activating E17K mutation in
AKT1, which leads to its pathologic localization to the plasma
membrane, causing constitutive downstream signaling of the
Figure 1. Druggable Alterations in Oncology Today and in the Near Future
The percent of patients by cancer type harboring a biomarker that, at present, guides the use of either FDA-approved or standard-of-care therapies (open circles)
compared to the fraction of patients in the same tumor type harboring a genomic alteration with compelling clinical evidence that it predicts response to a drug but
neither the genomic biomarker nor the drug are standard-of-care yet in that indication represented by an arrow.
PI3K-AKT-mTOR axis (Carpten et al., 2007). AKT1 E17K is found
in 3% of cases of breast cancer and is even less common
in endometrial, ovarian, cervical, lung, colorectal, and prostate
cancers, as well as several additional tumor types (Bleeker
et al., 2008). No single tumor type has a sufficient prevalence
of AKT1 E17K to allow for a traditional disease-focused drug
development strategy. Instead, and like BRAF before it, this
AKT1 mutation is especially well suited for clinical hypothesis
validation within a basket study. Indeed, AKT inhibitors have
recently shown marked efficacy in AKT1 E17K-mutation-positive
solid tumors in a basket study, and clinical development is now
progressing toward regulatory approval (Hyman et al., 2015b).
Similarly, a basket study of the pan-HER kinase inhibitor neratinib in solid tumors harboring activating mutations in ERBB2 and
ERBB3, which also occur infrequently across a large number of
cancer types, has reported promising preliminary results (Bose
et al., 2013; Hyman et al., 2016).
However, unlike AKT1, where the vast majority of activating
mutations occur at a single allele, activating mutations in
ERBB2 are distributed throughout the extracellular, juxtamembrane, and kinase domains of the gene encoding for the transmembrane HER2 tyrosine kinase receptor. Thus far, more than
10 recurrent missense mutations, in addition to various insertions and even rarer in-frame fusions, have been described.
Moreover, the pattern of ERBB2 mutations varies by tumor
type (Chang et al., 2016). Consequently, responses to HER2-targeted therapy in ERBB2-mutant tumors may be dependent, in
part, on potential differences in function of each mutation, as
well as each variant’s sensitivity to pharmacological inhibition
with a specific inhibitor. Indeed, the unique biochemical properties of individual genomic variants within the same gene have
been a recurrent observation across oncogenes such as BRAF
and KRAS, as well as tumor suppressors such as TP53 (Halevy
et al., 1990; Karnoub and Weinberg, 2008; Yao et al., 2015).
These genomic nuances add significant complexity to the interpretation of genomic data and clinical decision making. Furthermore, as with the basket studies mentioned above, responses
can be conditioned by the tissue of origin, and this must therefore be taken into account. Given the multitude of factors
that can condition response to therapy, larger studies that enroll
sufficient numbers of patients with different mutant alleles and
tumor types are necessary to reach definitive conclusions
regarding such intricacies.
Additional strategies exist to expand the druggable genome
beyond the targeting of individual gain-of-function missense variants. There are increasing efforts to target a growing number of
kinase fusions in various cancer types. As first exemplified by
BCR-ABL1 in leukemia, targeting structural rearrangements resulting in constitutively activated kinases can lead to dramatic efficacy and has become a renewed focus of precision oncology.
Cell 168, February 9, 2017 587
Drugs targeting kinase fusions involving ALK, ROS1, ABL1,
PDGFRA, and PDGFRB have already transformed the care
of patients with lung cancer, leukemia, and sarcoma. The
increasing adoption of DNA- and especially RNA-sequencing
technologies capable of detecting dozens of known and novel kinase fusions have demonstrated that these genomic events are
more common and implicate a larger number of kinases and tumor types than previously recognized (Stransky et al., 2014). As a
result, ongoing studies are already evaluating targeted therapy in
additional sets of fusions involving RET, FGFR1, FGFR2, FGFR3,
NRG1, BRAF, RAF1, NTRK1, NTRK2, NTRK3, and PRKCA
across a diversity of cancer types (Drilon et al., 2016b; Tabernero
et al., 2015). In fact, preliminary data from studies targeting these
novel kinase fusions suggest that the rate and depth of response
as well as the diversity of tumor histologies that will be sensitive
may be greater than for many of the current clinically credentialed missense mutations. As mentioned, we have been particularly struck by the greater magnitude of responses observed
when targeting many of these ‘‘fusion gene targets’’ and speculate that these difficult-to-acquire genomic alterations may indicate a higher level of oncogenic dependency on these complex
lesions. For instance, fusions involving the neurotrophic receptor
tyrosine kinases (NTRK1, 2, and 3) occur in approximately 0.5%
of diverse solid tumors and hematological malignancies, for
which NTRK inhibitors have demonstrated nearly uniform
response (D.S. Hong et al., 2016; AACR Annual Meeting abstract; Vaishnavi et al., 2015). As a consequence of these
impressive early clinical results, the US Food and Drug Administration recently granted Breakthrough Therapy Designation to
one of these agents for the treatment of any solid tumor
harboring an NTRK-fusion transcript, regardless of cancer lineage (Loxo, 2016). We are, therefore, finally on the verge of seeing
a genomically targeted therapy approved and used based solely
on the presence of a genomic alteration, regardless of the tumor
type in which it arises. Similar data have emerged from studies of
selective fibroblast growth factor receptor (FGFR) inhibitors in
FGFR2/3 fusion-positive cancers, including cholangiocarcinomas, bladder cancers, and gliomas (Tabernero et al., 2015).
Although these results targeting kinase fusions are exciting for
the field, they raise a key challenge. Many of these gene fusions
are also present, and sometimes enriched, in pediatric cancers
for which optimal available therapies are lacking. Historically,
targeted drug development in pediatrics has lagged beyond
that for adults, and pediatric patients have become a sadly underserved population when it comes to precision medicine. To
begin to address this deficit, we have recently had success
progressively lowering the minimum permitted age in several
genomically driven studies, a change prompted by the dramatic
responses observed in adults and the enthusiastic support of
health regulators.
Navigating Biological Complexity in Precision Oncology
Credentialing Therapeutically Actionable Mutations
The shift toward larger panel and whole-exome sequencing has
led to the identification of increasing numbers of somatic mutations in potentially actionable cancer genes, the vast majority of
which lack biological or clinical validation. This knowledge gap
significantly impairs our ability to fully utilize data generated by
588 Cell 168, February 9, 2017
prospective profiling to guide patient care and to implement
comprehensive precision-oncology programs. To begin unraveling the biological complexity of both common and rare alleles,
we must create a systematic framework to catalog genomic
alterations and characterize their frequency within and across
cancer types. Large-scale consortia efforts such as The Cancer
Genome Atlas have provided a vital first step but have predominantly studied primary untreated tumors, and of the 33 unique
cancer types profiled to date, only 9 qualify as rare tumors
(http://cancergenome.nih.gov/). On the other hand, prospective
clinical sequencing initiatives reflect a greater diversity and distribution of cancer types seen in patients with advanced disease,
the group most in need of new individualized treatment strategies (Hyman et al., 2015c). Moreover, many of these samples
are collected after the tumors have been exposed to prior therapies and therefore possess mutations that only arise upon selective therapeutic pressure, such as activating mutations in ESR1
(the gene encoding for the estrogen receptor [ER]) in patients
with ER-positive advanced breast cancers that have progressed
after anti-estrogen therapies (Toy et al., 2013).
Assembling large representative databases of clinically
sequenced cancers is only the first step toward saturating the
discovery, and eventual clinical validation, of actionable variants.
The next step is the development of a mutational taxonomy that
classifies each aberration on the basis of its abnormal function
and druggability. To this end, computational frameworks have
been developed for analyzing large-scale sequencing data in order to identify individual positions and genes recurrently
mutated, both within individual tumor types and across cancers,
more often than expected in the absence of selection (Chang
et al., 2016; Lawrence et al., 2014). Such statistically principled
approaches that operationalize different facets of how mutations, both driver and passenger, accrue in cancer genomes
will become increasingly necessary as the community aggregates larger datasets where repeated observations of even passenger alterations are expected. These early efforts have begun
to bear fruit, identifying previously uncharacterized and potentially druggable variants, and have the potential to significantly
expand the proportion of patients that may benefit from precision-oncology approaches. For instance, a recent analysis of
PIK3CA hotspots in more than 11,000 sequenced tumors identified not only the well-characterized helical and kinase domain
mutations E545 and H1047 but also 16 additional statistically
significant recurrent alterations (Chang et al., 2016). These additional hotspots accounted for 32% of all the PIK3CA mutations
observed in the overall cohort, potentially significantly expanding
the number of patients eligible for inhibitors targeting this
pathway. It remains unknown when, as a community, we will
saturate the detection of rare recurrent mutations in PIK3CA
and other clinically actionable genes such as ERBB2, EGFR,
and MET. Such efforts will also help the basic and translational
cancer research community to prioritize biologic investigation
or novel variants and, in doing so, facilitate new target discovery.
Novel allele prioritization is necessary, but not sufficient. The
mutational diversity in already therapeutically actionable genes
is profound and, in many cases, may reflect differences in
phenotype that translate into unique pharmacologic dependencies that in silico approaches alone cannot capture reliably.
For example, BRAF V600 is among the most common somatic
mutations in cancer and results in a constitutively active oncoprotein that signals as a RAF monomer (Davies et al., 2002).
Consequently, BRAF V600 is sensitive to RAF inhibitors that
preferentially bind activated RAF monomers (Bollag et al.,
2010). Conversely, less frequent but recurrent BRAF variants
including K601, L597, and G469 similarly lead to constitutive
BRAF activation but signal primarily as homodimers and are
insensitive to pharmacologic inhibition with first-generation
RAF inhibitors (Poulikakos et al., 2011; Yao et al., 2015). Still
other BRAF mutations such as V599 and G465 impair BRAF kinase activity but appear to activate MEK through RAS-dependent mechanisms (Wan et al., 2004). Finally, the impact of
many other BRAF mutations, including many occurring outside
of the kinase domain, on RAF/RAS signaling are still uncharacterized. Computational approaches alone cannot yet reliably
predict the class to which a previously uncharacterized BRAF
variant belongs and therefore cannot be used to guide selection
of pharmacologic therapy. Moreover, purely genomic approaches do not account for signaling pathway cross-talk and
feedback mechanisms that cloud simple genotype-drug
response relationships but are responsible for clinically important observations such as the lack of response of KRAS mutant
tumors to MEK inhibitors.
How then do we begin to capture at scale these nuances
among an enormous number of as yet to be tested alleles?
Interrogating each of these experimentally is time prohibitive.
Conversely, techniques that facilitate rapid biochemical characterization of large numbers of individual genomic variants simultaneously have been developed to address this need (Kim et al.,
2016; Kitzman et al., 2015). Utilizing recent advances in synthetic
biology, these massively parallel functional genomic approaches
use high-throughput methodologies to characterize the preliminary function of nearly all possible missense mutations within
target genes or pathways of interest. Work is now underway using these methods to build large catalogs of the biochemical
properties of thousands of individual missense mutations. This
approach is exemplified by a recent report on the phenotypic
characterization of a comprehensive set of MAPK1 (ERK2)
missense mutants (Brenan et al., 2016). Employing saturation
mutagenesis using a DNA synthesis-based approach (mutagenesis by integrated TilEs; Melnikov et al., 2014), these investigators were able to screen 6,810 of 6,821 possible (99.84%)
MAPK1 missense mutants for gain and loss of function, as well
as drug sensitivity. This screen confirmed that previously identified recurrently mutated alleles such as MAPK1 E322 were gain
of function but, importantly, also identified other candidate functional alleles that are not yet known to be recurrent in existing
tumor-sequencing databases. Soon, these technologies will
mature such that mutations can be studied in high throughput
in vivo and test a broader panel of molecular and biochemical
phenotypes. This high-throughput strategy for allele prioritization
can then be followed by traditional functional genetic studies
of greater depth using a variety of modern molecular biology
techniques, including CRISPR gene editing.
Although it is clear that a combination of multiple approaches
are necessary to scale up the generation of essential preclinical
data of a large number of candidate alleles, those efforts are pri-
marily focused on exploring individual alleles in isolation. However, this is not typically how these alleles arise in patient tumors.
Indeed, one of the open questions in precision oncology is how
best to intervene therapeutically in patients whose tumors present with multiple actionable mutations. To complicate things
further, the tumor microenvironment may also dictate how a
particular patient will respond to therapy targeting a specific genetic alteration. For example, in colon cancers, but curiously not
in other tumor types, inhibiting BRAF V600 is partially reversed
by abundant EGF in the microenvironment, which activates a
bypass pathway (Prahallad et al., 2012). Ultimately, this dependency framework will be necessary to guide rational drug combinations, as well as to prioritize targeting of one of potentially
multiple alterations in genomically complex tumors.
A New Model for Accelerated Clinical Testing
The scale and potential biological complexity of still uncharacterized genomic variants may not be addressable through traditional clinical testing paradigms. The absence of detailed
biochemical profiling of each genomic alteration is also a principal challenge to physicians trying to incorporate somatic
mutational data into novel treatments for patients with
advanced cancers and in need of such therapies. Biomarker
development has historically proceeded from basic target discovery to biologic validation, culminating in clinical testing.
Although this process has supported significant advancements,
the time necessary to validate each target in this manner is
impractical for many of the aforementioned reasons. In a
reversal of fortunes, the clinical validation of a target is now
frequently occurring prior to its full biological and functional
characterization in the laboratory. We have begun to demonstrate that genome-driven clinical trials designed to evaluate
laboratory-derived hypotheses can also be used to biologically
credential new targets that have not undergone extensive
biologic investigation, thus accelerating drug development
(Figure 2). In this model, clinical studies can become platforms
for exploring the functional consequences of novel genomic
alterations detected in the same patient populations. Such
studies can inform laboratory studies that, in turn, refine our understanding of the role these genomic aberrations play in disease pathogenesis, drug response, and resistance. Although
unconventional in approach, many of the patients currently having their tumors genomically profiled are heavily pretreated and
have no remaining standard therapeutic options. What level of
evidence supporting the actionability of a potential genomic
target should be required in this setting? A proof-of-principle
of this approach was recently undertaken in a heavily pretreated
patient with breast cancer with an ERBB2 L869R mutant tumor
treated on a basket study with an ERBB2 tyrosine kinase inhibitor (neratinib) for solid tumors harboring ERBB2 mutations
(ClinicalTrials.gov, NCT01953926). At the time, this patient’s tumor was identified as harboring an ERBB2 mutation; this particular variant was not known to be recurrent and had not been
functionally characterized. However, it was noted that the
ERBB2 L869 allele is paralogous to EGFR L861, a known activating EGFR mutation that is sensitive to EGFR tyrosine kinase
inhibitors (Yang et al., 2015). Moreover, computational and
three-dimensional modeling of ERBB2, EGFR, and the paralogous residue mutations predicted a similar constitutively
Cell 168, February 9, 2017 589
Figure 2. Approaches to Novel Target Validation
Discovery of a novel target is traditionally followed
by biologic validation before proceeding with
clinical credentialing. This ‘‘monogenic’’ lowthroughput process involves evaluation of one
genomic alteration at a time. Next-generation
sequencing at the point of care now routinely
identifies novel genomic alterations in patients who
are in need of new treatment strategies and cannot
wait for initial biologic validation. In a ‘‘polygenic’’
high-throughput model, novel genomic alterations
observed in patients undergo an initial prequalification using a computational framework that
considers allelic recurrence, paralogy, structure,
expression, and gene dosage to evaluate the
likelihood of clinical relevance. Patients whose
tumors harbor qualifying alterations that are identified as potentially activating can then be enrolled
in genome-driven clinical trials and the responses
observed, potentially providing initial clinical credentialing before biologic characterization. In
this model, clinical studies become platforms for
exploring the functional consequences of novel
genomic alterations detected in the same patient
populations.
activating signaling phenotype. On the basis of these findings,
as well as the absence of standard therapeutic options for the
patient, she was enrolled onto the study and ultimately achieved
a partial response lasting for more than 1 year (Hyman et al.,
2016). In addition to the significant benefit afforded this patient,
this response simultaneously provides valuable insight into the
biologic function of the novel L869R allele.
In the absence of comprehensive biochemical profiling of all
possible mutations in commonly mutated oncogenes, it is
essential that we establish principled approaches for analyzing
novel variants in this way to determine whether and how they
should be acted upon. Under this model, computational frameworks that combine analysis of statistically significant recurrence, sequence paralogy, and protein structure can be useful
if applied judicially and when there are no more-definitive biologic data. Stakeholders, including physicians and their patients,
must then decide whether the predictions generated by these
multifaceted analyses, which cannot replace biologic validation,
provide sufficient preliminary evidence to justify acting on them.
This assessment must necessarily take into account the alternative treatment options available, to the extent any exist, as a
greater degree of speculation may be appropriate in patients
with no alternative evidence-based options. The goal here is to
direct these patients to clinical studies that will collect and report
treatment outcome in a way that contributes to the overall knowledge base while at the same time offering a real possibility of
clinical benefit. Responses in this setting are essential clinical
evidence that, even when anecdotal, can prompt extensive biological exploration of a given allele using traditional molecular
biology techniques, studies that might also be performed in parallel to clinical testing via a co-clinical trial approach. Although
these responses can be extremely informative, the absence of
response cannot similarly be taken as conclusive evidence of
the absence of biologic function and, therefore, should not a priori stifle further biologic evaluation of novel alleles.
590 Cell 168, February 9, 2017
The approach of enrolling patients with candidate but incompletely characterized genomic drivers can be paired with other
study best practices, including generation of advanced models
such as organoids and patient-derived xenografts, that can be
used to further explore basic biology (when paired to the clinical
outcome) from the patient who contributed the sample. Moreover, these patient-derived models permit us to study not only
the initial response phenotype but also the adaptive signaling responses that underlie these outcomes.
Managing Tumor Heterogeneity and Acquired
Resistance
Tumors and the potentially therapeutically actionable mutations
that drive them evolve constantly from an ancestral cell and as
a function of therapy, resulting in both temporal and spatial
genomic heterogeneity (Gerlinger et al., 2012). Next-generation
sequencing has transformed our understanding of intratumoral
heterogeneity, and this phenomenon has been implicated in
both the development of acquired resistance and lesion-specific
differential response to targeted therapy (Diaz et al., 2012; Misale
et al., 2012; Russo et al., 2016). Despite this, the degree to
which this tumor heterogeneity will ultimately impact the utility
of genome-driven oncology remains unclear.
From a therapeutic perspective, the early truncal mutations
that are essential oncogenic drivers are typically shared by all
sites of disease, even in patients with advanced and heavily pretreated cancers, a fact that may mitigate some of the potential
consequences of intratumoral heterogeneity. As an example,
BRAF V600 mutations first arise in dysplastic nevi before they
progress to malignant melanoma and remain as a key tumoral
dependency despite the marked genomic heterogeneity characteristic of melanomas as a result of UV-induced DNA damage.
Nevertheless, therapy targeting even a truncal mutation will still
alter the cellular milieu as the incumbent oncogene-addicted
clone is depleted, facilitating a permissive environment for the
outgrowth of cellular populations that were previously less fit.
Hence, under the pressure of selective targeted therapy, these
initial subclonal genomic mediators of acquired resistance often
become the dominant clone, as exemplified by the emergence of
EGFR T790M in EGFR mutant non-small cell lung cancer patients treated with first-generation inhibitors (Hata et al., 2016).
Even in instances where more complex polyclonal resistance
emerges (different co-existing cell populations driven by genetically distinct mechanisms of resistance), these alterations often
converge on specific genes or pathways, suggesting that even
these scenarios could be managed with drugs or drug combinations that target this evolutionary bottleneck (Juric et al., 2015).
The eventual development of acquired resistance has, therefore, been a near universal observation with targeted cancer
therapy. Although the individual mechanisms underlying this
are varied, two are common—target reactivation due to secondary genomic alterations or activation of upstream effectors and
pathway reactivation mediated by activation of downstream or
bypass effectors. The recognition that resistance to targeted
therapies is often mediated by secondary mutation of the drug
target has already led to the development of drugs that maintain
efficacy despite these tumor adaptations. Resistance to imatinib
in BCR-ABL1 fusion-positive CML is driven almost entirely by
secondary mutations in the drug-binding site, ATP-binding
pocket, catalytic domain, and activation loop (Milojkovic and Apperley, 2009). Identification of these mechanisms has led to the
development and subsequent approval of even more potent
BCR-ABL1 tyrosine kinase inhibitors that maintain activity in
the setting of various secondary mutations (Cortes et al., 2012;
Saglio et al., 2010). Similarly, second- and third-generation inhibitors targeting EGFR mutant and ALK fusion-positive non-small
cell lung cancer have been successfully developed to manage
on-target resistance (Jänne et al., 2015; Shaw et al., 2014, 2016).
Acquired resistance can also be driven by activation of bypass
mechanisms via genomic alterations affecting upstream and
downstream effectors and through pathway-independent
mechanisms whose identification has uncovered biochemically
important facets of key cancer genes and pathways. For
example, the increased expression of the androgen receptor
in castrate-resistant metastatic prostate cancers ultimately led
to the discovery of their continued dependence on androgen
signaling despite apparent resistance to androgen deprivation
and first-generation antagonists. This insight ultimately led to
the successful development of second-generation androgen receptor inhibitors (Chen et al., 2004; Scher et al., 2012). Similarly,
identification of new mutations in the ligand-binding domain of
ESR1 in patients with ER-positive breast cancer treated with
certain classes of antiestrogens led to the discovery that these
mutations induce ligand-independent, ER-dependent gene transcription (Toy et al., 2013). This, in turn, has led to the development of selective ER degraders (Lai et al., 2015). Recently,
deletions of JAK2 and B2M were implicated in resistance to programmed death 1 (PD-1) checkpoint blockade in melanoma,
alluding to the critical role these effectors may play in patients
with regard to tumor-antigen presentation and immune surveillance (Zaretsky et al., 2016).
Undoubtedly, acquired resistance will impact every facet of
precision oncology. This is not an indictment of the therapeutic
approach but will require us to develop strategies to affect
more durable responses in our patients. The identification of
many of the aforementioned mechanisms of resistance has revealed novel biochemical and signaling phenotypes and demonstrated first principles of how multiple signaling pathways
interact via cross-talk and feedback. As a result, drug combinations, when applied in a biologically rational and synergistic
manner, can delay the onset of resistance (Baselga et al.,
2012; Turner et al., 2015). On the other hand, although preclinical
modeling may nominate potential mechanisms of resistance,
few of these mechanisms may manifest clinically in patients.
This further emphasizes the importance of prospectively identifying the genetic alterations acquired with, or selected by, treatment in the clinic and in real time to focus biologic investigation.
In this manner, acquired resistance is teaching us as much about
underlying biological dependencies as is the identification of
sensitizing lesions.
Data and Knowledge Sharing
Accomplishing our goals will necessitate data and knowledge
sharing by the clinical and biomedical community. Indeed, the
mutational heterogeneity and complexity of human cancers is
greater than can be reflected by the data any single center or
commercial lab can generate. Recognizing the critical importance of collaboration, early on our center developed an institution-wide genomics biospecimen protocol (ClinicalTrials.gov,
NCT01775072) that includes a framework for genomic data
sharing that can raise important privacy concerns (Hyman
et al., 2015c). The use of similar genomic biospecimen protocols
has become a common feature of academic centers pursuing
precision-oncology programs, and we view this as critical to
the success of the field. Maximizing the utility of now federated
genomic datasets will require not only consent for data sharing
but also bioinformatic approaches to harmonizing variants
generated by disparate sequencing platforms, each with their
own unique tissue requirements, performance characteristics,
and variant calling pipelines. To address this need, the American
Association of Cancer Research (AACR) has launched Project
GENIE (Genomics, Evidence, Neoplasia, Information, Exchange), which is developing a regulatory-grade registry that
aggregates and links clinical sequencing data from tens of thousands of cancer patients treated at multiple international academic medical centers. Unlike commercial laboratories that
have limited clinical annotation and no follow-up on the samples
they receive, institutional datasets have longitudinal follow-ups
that when aggregated are uniquely powered to answer previously unaddressed questions. For instance, EGFR inhibitors
are currently approved for the treatment of non-small cell lung
cancers with the most common mutational hotspots, including
L858R and exon 19 deletions. Little is known, however, about
less common EGFR variants (such as mutations in exon 18)
and how they might similarly guide evidence-based prescribing
practices for these widely available agents. Efforts to address
this knowledge gap have been unable to aggregate sufficiently
large numbers of patients. Project GENIE and related efforts
are well suited to answer these outstanding questions and immediately expand the proportion of patients who receive and
benefit from precision oncology. Other complementary public
and private efforts, including the NIH Genomic Data Commons
Cell 168, February 9, 2017 591
(Grossman et al., 2016), the Global Alliance for Genomics and
Health (Siu et al., 2016), and Molecular Evidence Development
Consortium (http://med-c.org/), are aggregating such genomic
data to evaluate clinical utility. In fact, data sharing was a cornerstone recommendation of the Blue Ribbon Panel Report issued
as part of Vice President Biden’s Cancer Moonshot initiative.
Sharing data is only one piece of this puzzle. How do we similarly curate, standardize, and aggregate knowledge about the
biological and clinical relevance of individual genomic alterations
to guide principled clinical decision making in prospectively
sequenced patients? The medical genetics community has
long recognized the value of high-quality interpretations of individual genomic variants, which are relied upon to make definitive
recommendations to patients involving complex issues such as
cancer screening, risk-reducing surgery, and reproduction. To
facilitate this decision making, this community has created a
number of resources, most prominently ClinVar (https://www.
ncbi.nlm.nih.gov/clinvar/) (Landrum et al., 2016). To similarly
systematically curate somatic variants, a number of clinical
knowledge bases have been created, including OncoKB
(oncokb.org), MyCancerGenome (mycancergenome.org), CIViC
(civic.genome.wustl.edu), Cancer Genome Interpreter (CGI,
cancergenomeinterpreter.org), CANDL (candl.osu.edu), and
the Personalized Cancer Medicine Knowledge Base (https://
pct.mdanderson.org/). Like ClinVar, these knowledge bases
have established levels of actionability and support community
contribution with expert curation. Centralization, standardization, and timely updates of these critical data resources are
necessary and could be catalyzed by a single national or international effort so that a single principled recommendation per
variant can be democratized to any treating oncologist.
Next-Generation Sequencing in the Clinic
An essential pillar of precision oncology is ensuring that we identify, at the point of care and in every cancer patient, the genomic
alterations on which the growth and progression of their diseases depend. A practical and technical challenge to implementing this paradigm has been the limited quantity and quality of
tumor material typically available for testing. To establish an
initial cancer diagnosis, patients often undergo only a small tumor core biopsy or fine-needle aspiration. A portion of this material is consumed for standard diagnostic evaluation. Moreover,
nearly all surgical and biopsy specimens are fixed in formalin
and may be stored for years in suboptimal conditions before
they are utilized for profiling. Biopsies and surgical specimens
are frequently mixed with stromal tissue, further limiting tumor
content. These limitations in tissue quality and quantity make obtaining comprehensive whole-genome sequencing, whole-transcriptome sequencing, and phosphoproteomics impractical or
even impossible for many cancer patients.
Despite these limitations, DNA enrichment and sequencing
technologies have matured to the point where they can now
generate reliable results on individual tumors within clinically
meaningful time frames using small amounts of paraffinembedded tumor tissue. These approaches have varied from
small gene panels that sequence only recurrently mutated ‘‘hotspot’’ positions (Meric-Bernstam et al., 2015) to targeted
gene panels typically sequencing the entire coding regions of
592 Cell 168, February 9, 2017
50–500 genes (Hyman et al., 2015c) to whole-exome
and -genome sequencing. Smaller hotspot-based panels, which
typically utilize amplicon-based sequencing, miss many of the
less common but potentially actionable alterations that in aggregate constitute the majority of genomic variants in cancer and, in
addition, are not well suited for detecting copy-number alterations and structural rearrangements. Conversely, larger targeted panels, which typically utilize hybridization enrichment,
can sequence the entire coding region of targeted genes, are
capable of detecting copy-number alterations, as well as select
structural rearrangements, and cover most, if not all, of the
genomic targets for which there are currently drugs in clinical
development. Importantly, these hybrid capture platforms
perform well with small tumor specimens and provide sufficient
sensitivity for the detection of actionable alterations in samples
with low tumor purity. In addition, these panels can be easily
expanded to include new genomic content as the actionable
genome expands. Consequently, targeted gene panels using
this technology have been favored by many of the high-volume
academic institutions and commercial laboratories pioneering
the use of this testing (Cheng et al., 2015; Frampton et al.,
2013). Although broader whole-exome sequencing provides
the opportunity to detect recurrent alterations in genes not previously implicated in cancer, and is therefore useful for discovery, it requires more input DNA, typically offers lower coverage,
and often has less sensitivity for structural rearrangements due
to limited sequencing of intronic regions (Beltran et al., 2015).
Whole-genome sequencing similarly offers lower coverage and
also remains cost prohibitive and therefore impractical for
most patients. In the near term, these different approaches will
continue to represent tradeoffs in terms of sequencing breadth,
depth, and practicality of overall adoption. In weighing these
tradeoffs, we believe that the rarity of promising genomic alterations currently dictates that clinical genomic profiling strategies
be optimized to screen more patients using sufficiently broad
targeted gene panels rather than fewer patients with even
more comprehensive assays. Although even further from clinical
implementation, single-cell DNA and RNA sequencing are also
actively being investigated as a means of providing unprecedented resolution into the heterogeneity of each tumor, as well
as its microenvironment (Tirosh et al., 2016)
The increased sensitivity afforded by targeted next-generation
sequencing has substantially improved our ability to identify
known actionable alterations in patients and has also led to the
discovery of new potentially druggable alterations. One recent
example is MET splice site mutations, which promote MET
exon 14 skipping that results in augmented MET kinase activity
through loss of ubiquitin-mediated degradation and appear sensitive to MET inhibitors such as crizotinib (A. Drilon et al., 2016a;
ASCO Annual Meeting abstract; Peschard et al., 2001). These
heterogeneous alterations, which are present in 3% of lung adenocarcinomas and rarely in other cancer types, can span 200
bases and are not easily detected by hotspot-based sequencing
approaches (Frampton et al., 2015).
Technical advances are now accelerating the development of
genomic profiling strategies suitable for sequencing tumorderived cell free DNA (cfDNA) from plasma with high sensitivity
and specificity. These assays combine deep targeted hybrid
capture sequencing with the use of unique molecular identifiers
that allow for the differentiation of sequencing artifacts from very
low-allele-frequency mutations (Kivioja et al., 2011; Lanman
et al., 2015). The potential advantages of cfDNA, also known
as a ‘‘liquid biopsy,’’ compared to tumor sequencing are many.
Tumor biopsies are complex, invasive, and expensive, which
precludes their widespread implementation and repeated use
in patients. Conversely, cfDNA only requires a blood draw and
can be repeated as frequently as clinically indicated with little
risk to the patient. Furthermore, cfDNA sequencing may better
capture the genomic heterogeneity of patient disease by detecting mutations that are both shared and private to individual
tumor sites. cfDNA sequencing can also be used to monitor
response to targeted therapy. In fact, real-time analysis of cfDNA
may determine response to therapy within days of treatment initiation, as opposed to weeks with conventional imaging studies.
As the sensitivity and specificity of cfDNA profiling improves,
as does our ability to interpret the presence and biological significance of rare mutations in circulation, we envision the power of
cfDNA being utilized as a screening tool to detect early-stage
cancers, when the likelihood of cure would be far higher.
The Reach of Precision Oncology into the Germline: The
Case for Integrated Testing
The growing compendium of germline polymorphisms established by large-scale projects such as the EXaC consortium
(http://exac.broadinstitute.org/) has improved the ability to
distinguish between somatic and germline variants in tumoronly prospective sequencing. Despite this progress, these databases are under-represented by individuals of non-European
ancestry and do not entirely address private germline variations
that are unlikely to be sufficiently captured without orders of
magnitude of additional germline data, if ever. As a result, clinical tumor sequencing, where inaccurate variant classification
could potentially lead to improper treatment selection, is
increasingly utilizing patient-matched normal controls with the
primary objective of distinguishing between germline and somatic variants. However, matched germline sequencing also
facilitates simultaneous diagnosis of cancer predisposition syndromes that are themselves therapeutically actionable. Early
evidence from large-scale clinical sequencing initiatives suggests a higher rate of pathologic germline variants in both adult
and pediatric cancer patients than would be predicted from
conventional germline screening guidelines based on personal
and family history (Schrader et al., 2016; Zhang et al., 2015).
The utility of detecting germline mutations has been recently
highlighted in prostate cancer, for which putative loss-of-function germline mutations in BRCA2, ATM, CHEK2, BRCA1, and
PALB2 are found in nearly 12% of patients, a rate higher than
expected (Pritchard et al., 2016). Importantly, these genomic
alterations are all potentially targetable with PARP inhibitors,
as well as other inhibitors of DNA-damage repair or DNAdamaging agents. Similarly, defects in mismatch repair through
germline, somatic, and epigenetic mechanisms may be targetable with immune checkpoint blockade and are readily detectable through tumor and matched germline sequencing (Le
et al., 2015). These findings demonstrate that combined tumor
and matched germline sequencing not only improves the detec-
tion of somatic mutations but also simplifies comprehensive
testing for cancer patients while further expanding the scope
of actionable alterations. Moreover, although variants of unknown significance plague the clinical interpretation of most
germline alleles, combined analysis with the corresponding tumor that reveals somatic loss of heterozygosity can implicate
a germline variant of uncertain significance in the absence of
functional data and perhaps inform therapy in advanced patients who otherwise lack treatment options. Therefore, such
integrative analysis may provide insights that individual testing
cannot capture.
A Future for Genome-Driven Oncology
We have described a framework that can begin to successfully
navigate the scientific challenges and broaden the scope and
utility of genome-driven oncology. Equally important will be optimizing the way we evaluate the resulting genomic hypotheses in
the clinic. Maximizing progress will require us to improve every
step in the precision medicine ecosystem, beginning with how
we sequence patients, identifying the best targets from this
testing, notifying stakeholders, improving access to relevant
clinical studies, and finally, making sure these studies are appropriately designed to advance the field.
The Hallmarks of a Precision-Oncology Study: Learning
More from Each Patient
Clinical studies evaluating a genomic-driven hypothesis should
be designed to learn from each case in an unprecedented way.
To capture maximal information from each enrolled patient, precision-oncology studies should include, when possible, not only
traditional clinical response evaluation but also the systematic
analysis of patient-derived biospecimens and even potential
clinical strategies to overcome adaptive or acquired resistance
in real time (Figure 3). We propose that the hallmarks of a modern
precision-oncology study include four primary scientific objectives: identification of the target, confirmation of target inhibition,
biologic credentialing of the target, and description of the mechanisms underlying acquired resistance. Collection and analysis
of biospecimens should be organized around, and driven by,
these key objectives. For example, pretreatment tumor and
liquid biopsies should be used to confirm the presence of the
target and define the broader genomic context in which it arises.
Confirmation of target engagement and early adaptive responses to target inhibition can be evaluated by tumor biopsies
obtained shortly after the initiation of treatment, when early
compensatory feedback mechanisms that may modify treatment efficacy can be observed, potentially nominating rational
combinatorial strategies. Target engagement can also be assessed through functional imaging studies, such as 18F-fluoroestradiol positron emission tomography (PET), or indirectly
via analysis of circulating cfDNA, exosomes, and tumor cells.
Biologic target validation can be greatly facilitated through generation of patient-derived xenografts or organoids at the time of
tumor biopsy. As described earlier, the identification of mechanisms that result in de novo and acquired resistance in patients
is critical and often requires comparative analysis of pre- and
post-treatment tumor using a variety of experimental approaches, including DNA and RNA sequencing, phosphoprotein
Cell 168, February 9, 2017 593
Figure 3. The Hallmarks of a PrecisionOncology Study
Shown are multiple facets of a modern oncology
trial that not only refines a biomarker hypothesis in
a scientifically principled manner but also can
serve as an engine to drive new scientific discoveries. The hallmarks of a modern precisiononcology study include four primary scientific
objectives: identification of the target, confirmation
of target inhibition, biologic credentialing of the
target, and description of the mechanisms underlying acquired resistance. Collection and analysis
of biospecimens should be organized around, and
driven by, these key objectives.
analysis, and immune profiling. Interrogating tumor-derived
cfDNA from plasma can also identify the emergence of resistance at an early stage and identify clonal evolution as a
response to therapy that may facilitate the addition of other
agents or lead to a switch to alternative targeted agents entirely
in order to prevent progression of disease (Dawson et al.,
2013; Thress et al., 2015). This approach can provide rapid clinical confirmation of underlying resistance mechanisms nominated by preclinical models even when treating small numbers
of patients. The study of circulating tumor cells can be considered when key study-related biological questions require intact
cells, and it can be visualized as a ‘‘live liquid biopsy’’ where cells
can be interrogated in an unprecedented fashion (Yu et al.,
2013). Similarly, tumor-derived exosomes appear to contain
not only DNA but also RNA, microRNA, proteins, and lipids
from cancer cells and provide an opportunity for the non-invasive, multidimensional profiling of cancer (Hoshino et al., 2015).
The viability of this approach has been facilitated by techniques
that selectively enrich for tumor-derived exosomes (Melo et al.,
2015). Finally, for those patients who ultimately succumb to their
disease, rapid autopsy programs can provide invaluable insights
as well as sustaining models to further accelerate drug development (Juric et al., 2015).
Unexplored Facets of Sensitizing Biomarkers
Optimizing genome-driven oncology will likely demand that we
move beyond simple classifications of biomarkers as being
594 Cell 168, February 9, 2017
present or absent in a given patient.
Additional facets of a genomic alteration,
including absolute copy number, clonality,
and zygosity, may condition its function or
modify response to targeted therapy.
Advanced analytical techniques, originally
developed to generate precise estimates
of subclonal heterogeneity, allele-specific
absolute copy number, and zygosity from
whole-exome sequencing, are now being
scaled down to targeted sequencing
data to address this unmet need (Carter
et al., 2012; Shen and Seshan, 2016).
An important and still unresolved clinical question is the degree to which subclonal heterogeneity of a sensitizing
genomic biomarker affects the likelihood
of response. With modern analytic techniques, it is now possible
to catalog the clonality of the individual genomic variants within a
single sequenced site and incorporate evolutionary inference to
identify the order in which they were acquired within ‘‘molecular’’
time. These analyses will provide unique insights into how
tumors evolve and permit an additional dimension of genotype-to-response-phenotype correlation. Similarly, copy-number estimates made by most clinically implemented next-generation sequencing bioinformatic pipelines do not currently correct
for tumor purity or provide estimates of allele-specific absolute
copy number. An increasing body of evidence, however, suggests that high absolute copy number of gene amplifications,
including MET and FGFR2, are necessary to condition response
(D.R. Camidge et al., 2014; ASCO Annual Meeting abstract;
Pearson et al., 2016). As a result, standardized reporting of
allele-specific absolute copy number will enhance the optimal
genomic selection of patients when targeting oncogene
amplifications.
Understanding the zygosity of genomic alterations involving
tumor-suppressor genes may also be important for targeted
therapies that exploit synthetic lethality. For instance, PARP inhibitors have shown efficacy in treating homologous recombination-deficient (HRD) tumors with deleterious germline BRCA1/2
mutations. One clinically important question is whether PARP inhibitors may also be effective in patients whose tumors harbor
homologous recombination deficiency driven by somatic mutations. Preliminary data suggest that somatic alterations in key
HRD genes may require biallelic inactivation to sensitize tumors
to PARP inhibitors (Mateo et al., 2015). Computational techniques capable of determining whether somatic truncating variants in HRD genes are accompanied by loss of heterozygosity of
the remaining wild-type allele may therefore be necessary to
optimally select patients for these therapies. Similarly, recent
data demonstrate that biallelic loss of SMARCA4 (BRG1) and
SMARCB1 (INI1) in rhabdoid carcinomas and epithelioid sarcomas, respectively, may be targetable with EZH2 inhibitors
(Kim et al., 2015). Efforts are currently underway to extend this
therapeutic strategy to a wider range of epigenetic modifiers,
including BAP1 and ARID1A (Bitler et al., 2015; LaFave
et al., 2015).
DNA-based signatures are also emerging as an important tool
with which to select targeted therapy. For example, several
DNA-based gene signatures have been developed to detect
the genomic instability associated with homologous recombination deficiency caused by inactivation of BRCA1, BRCA2, and
potentially a variety of other related genes (Telli et al., 2016).
Demonstrating the potential value of this approach, a recent
phase III study in ovarian cancer utilizing a DNA-based homologous recombination signature demonstrated that signature-positive patients treated with the PARP inhibitor niraparib versus placebo had a near tripling of their progression-free survival (12.9
versus 3.8 months) compared to signature-negative patients
(Mirza et al., 2016). These homologous recombination-deficiency signatures, also sometimes referred to as ‘‘BRCA-ness’’
or ‘‘genome scar’’ assays, also demonstrate how the field will
begin to extend precision-oncology approaches into tumor
types, such as prostate and pancreatic cancer, that otherwise
do not harbor frequent actionable genomic variants.
Moving beyond DNA
Although DNA sequencing provides a wealth of information,
genomic alterations are only one of several important biologic
drivers of cancer. As a consequence, it is understood that DNA
sequencing will not be sufficient to optimally select patients
for all classes of targeted therapy. In the laboratory, highthroughput technologies, including RNA sequencing, genomewide DNA methylation profiling, microRNA profiling, and phosphoprotein arrays, have been extensively used to further improve
our understanding of the biologic dependencies of cancer. Of
these various technologies, the one that is closest to the clinic
is RNA sequencing, also sometimes referred to as transcriptome
sequencing. The incremental value of RNA sequencing in
addition to existing DNA sequencing is two-fold. First, when
compared to targeted or whole-exome DNA sequencing, RNA
sequencing is better suited to detecting structural rearrangements resulting in fusion gene products. Along these lines, we
have recently instituted in our center a ‘‘cancer of unknown
driver’’ initiative whereby cancers for which prospective targeted
DNA sequencing fails to identify a genomic driver are reflexed to
targeted RNA sequencing with the goal of identifying cryptogenic
kinase fusions. This approach has already identified novel and
potentially actionable kinase fusions involving NTRK1/2/3,
FGFR2/3, BRAF, NRG1, RET, ERBB2, and AKT1. Similar approaches have been adopted by some high-volume commercial
laboratories in cancer types such as sarcomas and hematologic
malignancies in which the diversity of potential gene fusions
cannot be efficiency interrogated by targeted DNA-based
methods (He et al., 2016). In addition to enhanced detecting
sensitivity for structural rearrangements in the genome, RNA
sequencing offers the ability to measure the transcription of
both mutant and wild-type proteins. In breast cancer, earlier approaches to measuring gene-expression profiling are already being used to provide a more refined prognosis than that based on
standard clinical risk factors alone and, in doing so, guide treatment decisions (Cardoso et al., 2016). The understanding of the
transcriptome at the point of care in each patient will provide an
invaluable additional dimension of information, and we believe
it is the next frontier for precision oncology. Pilot programs
have already begun evaluating the feasibility of incorporating
broader RNA sequencing into prospective precision-medicine
programs (Roychowdhury et al., 2011). The preliminary results
have been encouraging but are limited by the quantity and type
of tumor material required, as well as challenges associated
with scaling this approach to larger patient volumes. Over time,
we expect these technical, scientific, and financial obstacles to
be solved, and we must begin to consider now how to begin to
incorporate these technologies into the clinic.
Expanding Accessibility and Enrollment to PrecisionOncology Studies
The timely and efficient execution of precision-oncology studies,
like those we detail here, represents a true engineering challenge
on top of an already inefficient clinical-trials model. In fact, less
than 5% of cancer patients are enrolled into clinical trials. Therefore, lack of execution represents a real threat that needs to be
addressed. The components of this engineering problem include
routine tumor sequencing as a central component of cancer
care, the input and annotation of genomic findings into the
medical records, and both patient and physician education.
Furthermore, robust methods for connecting patients with rare
actionable alterations to clinical trials targeting these genomic
variants are needed. Within our own institution, we have developed a system that automatically identifies, tracks, and recruits
patients with qualifying genomic alterations to the appropriate
genome-driven study (Eubank et al., 2016). This system has allowed treating physicians to rely on domain experts to interpret
the actionability of individual genomic alterations and identify
relevant and immediately available treatment opportunities for
their patients. The broader field urgently needs a similar centralized genomic ‘‘clearing house’’ wherein the results of clinical
genomic sequencing can be shared and matched in real time
against the qualifying genomic alterations being targeted by individual studies nationwide. Unlike current static study registries
such as ClinicalTrials.gov, such a system would ensure a privacy-compliant two-way exchange between patients and study
sponsors who could indicate the specific genomic variants of interest and provide a path to enrollment for qualifying patients.
Several groups, including the Global Alliance for Genomics and
Health, are working to develop such a resource (Siu et al., 2016).
Facilitating the identification of highly relevant and immediately available genomically matched studies is only the first
step; patients must be able to readily access these studies.
Geographical considerations remain an important barrier to
Cell 168, February 9, 2017 595
study access. The current generation of precision-oncology
studies are beginning to address this problem in a number of
ways, from supporting patient travel to study centers to
bringing the study to the patient once they are identified. These
efforts have significantly improved access to genomically
matched therapy studies in the community, where the majority
of cancer patients continue to be treated. Access to genomically matched therapy has also been improved by the consolidation of precision-oncology trials into larger protocols that
have alternatively been referred to as ‘‘master,’’ ‘‘umbrella,’’
or ‘‘molecular allocation’’ studies. These protocols generally
offer multiple therapeutic options matched to the patient’s
individual tumor genome. Several of these studies, such
as the iSpy2 (NCT01042379), Lung-MAP (NCT02154490),
ALCHEMIST (NCT02194738), and BATTLE-2 (NCT01248247)
trials, have explored genomically defined subtypes of specific
cancers. Master protocols can also offer treatment across a variety of tumor types and, in this way, essentially become a
collection of individual basket studies; these include NCIMATCH (NCT02465060), MyPathway (NCT02091141), and the
American Society of Clinical Oncology (ASCO) Targeted Agent
and Profiling Utilization Registry (TAPUR) (NCT02693535). The
most ambitious effort to date is the NCI-MATCH study, which is
anticipated to have more than 30 unique treatment arms assigned primarily on the basis of genomic selection criteria.
Unlike many related efforts, NCI-MATCH incorporates centralized high-multiplexed tumor genomic screening, with the plan
to biopsy and sequence 6,000 patients. This feature has
made NCI-MATCH particularly attractive to community oncologists who might otherwise not have access to high-quality tumor genetic screening.
Conclusion
This is a transformative time for cancer therapy. The feasibility
of establishing the detailed molecular portraits of individual
cancers, even at the point of care, is no longer the primary
obstacle to progress. Similarly, new highly potent and selective
purpose-built inhibitors are being developed at a time when
our understanding of actionable mutations in cancer genomes
is improving steadily, resulting in continuous erosion of the market share of what has been called the undruggable genome.
What is most lacking, therefore, is the knowledge of how best
to use these new and powerful tools that are already available
to us. In short, we have an engineering problem. The strategies
we propose here establish a framework to begin to address
this critical knowledge and implementation gap. We cannot
achieve the progress needed with conventional approaches.
Rapid progress demands a new degree of collaboration and information exchange between basic and translational laboratory
scientists and clinical investigators acting as equal partners.
Evaluating the genomic hypotheses that emerge from this
collaboration will require us to improve the efficiency with which
we sequence patients, annotate results, identify accountable alterations, notify involved parties, enroll into relevant studies, and
finally, interpret the outcome and iterate as necessary. If we do
so, we will finally have the tools to make truly transformative insights into the basic biology of cancer and its treatment.
596 Cell 168, February 9, 2017
ACKNOWLEDGMENTS
We thank Alison M. Schram and Maurizio Scaltriti for their critical reading of
this manuscript, Jianjiong Gao and Nikolaus Schultz for clinical annotation of
sequenced specimens, and Scott Johnson for assistance with medical
graphics. The authors acknowledge support from the Prostate Cancer Foundation, the Sontag Foundation, the Josie Robertson Foundation, the Breast
Cancer Research Foundation, the Geoffrey Beene Cancer Research Center,
the Marie-Josée and Henry R. Kravis Center for Molecular Oncology, Cycle
for Survival, and National Institutes of Health awards R01 CA190642-01A1
(J.B.), P50 CA092629 (B.S.T.), R01 CA207244 (D.M.H. and B.S.T.), U54
OD202355 (B.S.T.), and P30 CA008748. J.B. is a member of the scientific advisory board of GRAIL.
REFERENCES
Baselga, J., Campone, M., Piccart, M., Burris, H.A., 3rd, Rugo, H.S.,
Sahmoud, T., Noguchi, S., Gnant, M., Pritchard, K.I., Lebrun, F., et al.
(2012). Everolimus in postmenopausal hormone-receptor-positive advanced
breast cancer. N. Engl. J. Med. 366, 520–529.
Beltran, H., Eng, K., Mosquera, J.M., Sigaras, A., Romanel, A., Rennert, H.,
Kossai, M., Pauli, C., Faltas, B., Fontugne, J., et al. (2015). Whole-exome
sequencing of metastatic cancer and biomarkers of treatment response.
JAMA Oncol. 1, 466–474.
Bitler, B.G., Aird, K.M., Garipov, A., Li, H., Amatangelo, M., Kossenkov, A.V.,
Schultz, D.C., Liu, Q., Shih, IeM., Conejo-Garcia, J.R., et al. (2015). Synthetic
lethality by targeting EZH2 methyltransferase activity in ARID1A-mutated cancers. Nat. Med. 21, 231–238.
Bleeker, F.E., Felicioni, L., Buttitta, F., Lamba, S., Cardone, L., Rodolfo, M.,
Scarpa, A., Leenstra, S., Frattini, M., Barbareschi, M., et al. (2008). AKT1(E17K)
in human solid tumours. Oncogene 27, 5648–5650.
Bollag, G., Hirth, P., Tsai, J., Zhang, J., Ibrahim, P.N., Cho, H., Spevak, W.,
Zhang, C., Zhang, Y., Habets, G., et al. (2010). Clinical efficacy of a RAF inhibitor needs broad target blockade in BRAF-mutant melanoma. Nature 467,
596–599.
Bose, R., Kavuri, S.M., Searleman, A.C., Shen, W., Shen, D., Koboldt, D.C.,
Monsey, J., Goel, N., Aronson, A.B., Li, S., et al. (2013). Activating HER2 mutations in HER2 gene amplification negative breast cancer. Cancer Discov. 3,
224–237.
Bower, H., Björkholm, M., Dickman, P.W., Höglund, M., Lambert, P.C., and
Andersson, T.M. (2016). Life expectancy of patients with chronic myeloid leukemia approaches the life expectancy of the general population. J. Clin. Oncol.
34, 2851–2857.
Brenan, L., Andreev, A., Cohen, O., Pantel, S., Kamburov, A., Cacchiarelli, D.,
Persky, N.S., Zhu, C., Bagul, M., Goetz, E.M., et al. (2016). Phenotypic characterization of a comprehensive set of MAPK1/ERK2 missense mutants. Cell
Rep. 17, 1171–1183.
Brose, M.S., Cabanillas, M.E., Cohen, E.E., Wirth, L.J., Riehl, T., Yue, H.,
Sherman, S.I., and Sherman, E.J. (2016). Vemurafenib in patients with
BRAF(V600E)-positive metastatic or unresectable papillary thyroid cancer refractory to radioactive iodine: a non-randomised, multicentre, open-label,
phase 2 trial. Lancet Oncol. 17, 1272–1282.
Cardoso, F., van’t Veer, L.J., Bogaerts, J., Slaets, L., Viale, G., Delaloge, S.,
Pierga, J.Y., Brain, E., Causeret, S., DeLorenzi, M., et al.; MINDACT Investigators (2016). 70-gene signature as an aid to treatment decisions in early-stage
breast cancer. N. Engl. J. Med. 375, 717–729.
Carpten, J.D., Faber, A.L., Horn, C., Donoho, G.P., Briggs, S.L., Robbins,
C.M., Hostetter, G., Boguslawski, S., Moses, T.Y., Savage, S., et al. (2007).
A transforming mutation in the pleckstrin homology domain of AKT1 in cancer.
Nature 448, 439–444.
Carter, S.L., Cibulskis, K., Helman, E., McKenna, A., Shen, H., Zack, T., Laird,
P.W., Onofrio, R.C., Winckler, W., Weir, B.A., et al. (2012). Absolute quantification of somatic DNA alterations in human cancer. Nat. Biotechnol. 30,
413–421.
Chang, M.T., Asthana, S., Gao, S.P., Lee, B.H., Chapman, J.S., Kandoth, C.,
Gao, J., Socci, N.D., Solit, D.B., Olshen, A.B., et al. (2016). Identifying recurrent
mutations in cancer reveals widespread lineage diversity and mutational specificity. Nat. Biotechnol. 34, 155–163.
Chen, C.D., Welsbie, D.S., Tran, C., Baek, S.H., Chen, R., Vessella, R., Rosenfeld, M.G., and Sawyers, C.L. (2004). Molecular determinants of resistance to
antiandrogen therapy. Nat. Med. 10, 33–39.
Cheng, D.T., Mitchell, T.N., Zehir, A., Shah, R.H., Benayed, R., Syed, A.,
Chandramohan, R., Liu, Z.Y., Won, H.H., Scott, S.N., et al. (2015). Memorial
Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets
(MSK-IMPACT): A hybridization capture-based next-generation sequencing
clinical assay for solid tumor molecular oncology. J. Mol. Diagn. 17, 251–264.
Cortes, J.E., Kantarjian, H., Shah, N.P., Bixby, D., Mauro, M.J., Flinn, I.,
O’Hare, T., Hu, S., Narasimhan, N.I., Rivera, V.M., et al. (2012). Ponatinib in
refractory Philadelphia chromosome-positive leukemias. N. Engl. J. Med.
367, 2075–2088.
Cunanan, K.M., Gonen, M., Shen, R., Hyman, D.M., Riely, G.J., Begg, C.B.,
and Iasonos, A. (2016). Basket trials in oncology: a trade-off between
complexity and efficiency. J. Clin. Oncol. 35, 271–273.
Davies, H., Bignell, G.R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague,
J., Woffendin, H., Garnett, M.J., Bottomley, W., et al. (2002). Mutations of the
BRAF gene in human cancer. Nature 417, 949–954.
Dawson, S.J., Tsui, D.W., Murtaza, M., Biggs, H., Rueda, O.M., Chin, S.F.,
Dunning, M.J., Gale, D., Forshew, T., Mahler-Araujo, B., et al. (2013). Analysis
of circulating tumor DNA to monitor metastatic breast cancer. N. Engl. J. Med.
368, 1199–1209.
De Roock, W., Claes, B., Bernasconi, D., De Schutter, J., Biesmans, B., Fountzilas, G., Kalogeras, K.T., Kotoula, V., Papamichael, D., Laurent-Puig, P., et al.
(2010). Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy
of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol. 11, 753–762.
Diaz, L.A., Jr., Williams, R.T., Wu, J., Kinde, I., Hecht, J.R., Berlin, J., Allen, B.,
Bozic, I., Reiter, J.G., Nowak, M.A., et al. (2012). The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature
486, 537–540.
Drilon, A., Rekhtman, N., Arcila, M., Wang, L., Ni, A., Albano, M., Van Voorthuysen, M., Somwar, R., Smith, R.S., Montecalvo, J., et al. (2016b). Cabozantinib
in patients with advanced RET-rearranged non-small-cell lung cancer:
an open-label, single-centre, phase 2, single-arm trial. Lancet Oncol. 17,
1653–1660.
Druker, B.J., Guilhot, F., O’Brien, S.G., Gathmann, I., Kantarjian, H., Gattermann, N., Deininger, M.W., Silver, R.T., Goldman, J.M., Stone, R.M., et al.;
IRIS Investigators (2006). Five-year follow-up of patients receiving imatinib
for chronic myeloid leukemia. N. Engl. J. Med. 355, 2408–2417.
Eubank, M.H., Hyman, D.M., Kanakamedala, A.D., Gardos, S.M., Wills, J.M.,
and Stetson, P.D. (2016). Automated eligibility screening and monitoring for
genotype-driven precision oncology trials. J. Am. Med. Inform. Assoc. 23,
777–781.
Frampton, G.M., Fichtenholtz, A., Otto, G.A., Wang, K., Downing, S.R., He, J.,
Schnall-Levin, M., White, J., Sanford, E.M., An, P., et al. (2013). Development
and validation of a clinical cancer genomic profiling test based on massively
parallel DNA sequencing. Nat. Biotechnol. 31, 1023–1031.
Halevy, O., Michalovitz, D., and Oren, M. (1990). Different tumor-derived p53
mutants exhibit distinct biological activities. Science 250, 113–116.
Hata, A.N., Niederst, M.J., Archibald, H.L., Gomez-Caraballo, M., Siddiqui,
F.M., Mulvey, H.E., Maruvka, Y.E., Ji, F., Bhang, H.E., Krishnamurthy Radhakrishna, V., et al. (2016). Tumor cells can follow distinct evolutionary paths to
become resistant to epidermal growth factor receptor inhibition. Nat. Med.
22, 262–269.
He, J., Abdel-Wahab, O., Nahas, M.K., Wang, K., Rampal, R.K., Intlekofer,
A.M., Patel, J., Krivstov, A., Frampton, G.M., Young, L.E., et al. (2016). Integrated genomic DNA/RNA profiling of hematologic malignancies in the clinical
setting. Blood 127, 3004–3014.
Hoadley, K.A., Yau, C., Wolf, D.M., Cherniack, A.D., Tamborero, D., Ng, S.,
Leiserson, M.D., Niu, B., McLellan, M.D., Uzunangelov, V., et al.; Cancer
Genome Atlas Research Network (2014). Multiplatform analysis of 12 cancer
types reveals molecular classification within and across tissues of origin.
Cell 158, 929–944.
Hoshino, A., Costa-Silva, B., Shen, T.L., Rodrigues, G., Hashimoto, A., Tesic
Mark, M., Molina, H., Kohsaka, S., Di Giannatale, A., Ceder, S., et al. (2015).
Tumour exosome integrins determine organotropic metastasis. Nature 527,
329–335.
Hyman, D.M., Puzanov, I., Subbiah, V., Faris, J.E., Chau, I., Blay, J.Y., Wolf, J.,
Raje, N.S., Diamond, E.L., Hollebecque, A., et al. (2015a). Vemurafenib in multiple nonmelanoma cancers with BRAF V600 mutations. N. Engl. J. Med. 373,
726–736.
Hyman, D.M., Smyth, L., Bedard, P.L., Oza, A., Dean, E., Armstrong, A., Lima,
J., Bando, H., Kabos, P., Perez-Fidalgo, J.A., et al. (2015b). AZD5363, a catalytic pan-Akt inhibitor, in Akt1 E17K mutation positive advanced solid tumors.
Mol. Cancer Ther. 14, Abstract B109.
Hyman, D.M., Solit, D.B., Arcila, M.E., Cheng, D.T., Sabbatini, P., Baselga, J.,
Berger, M.F., and Ladanyi, M. (2015c). Precision medicine at Memorial Sloan
Kettering Cancer Center: clinical next-generation sequencing enabling nextgeneration targeted therapy trials. Drug Discov. Today 20, 1422–1428.
Hyman, D.M., Piha-Paul, S., Rodon, J., Saura, C., Puzanov, I., Shapiro, G., Loi,
S., Joensuu, H., Hanrahan, A., Modi, S., et al. (2016). Neratinib for ERBB2
mutant, HER2 non-amplified, metastatic breast cancer: Preliminary analysis
from a multicenter, open-label, multi-histology phase II basket trial. Cancer
Res. 76, Abstract PD5–05.
Jänne, P.A., Yang, J.C., Kim, D.W., Planchard, D., Ohe, Y., Ramalingam, S.S.,
Ahn, M.J., Kim, S.W., Su, W.C., Horn, L., et al. (2015). AZD9291 in EGFR inhibitor-resistant non-small-cell lung cancer. N. Engl. J. Med. 372, 1689–1699.
Juric, D., Castel, P., Griffith, M., Griffith, O.L., Won, H.H., Ellis, H., Ebbesen,
S.H., Ainscough, B.J., Ramu, A., Iyer, G., et al. (2015). Convergent loss of
PTEN leads to clinical resistance to a PI(3)Ka inhibitor. Nature 518, 240–244.
Kandoth, C., McLellan, M.D., Vandin, F., Ye, K., Niu, B., Lu, C., Xie, M., Zhang,
Q., McMichael, J.F., Wyczalkowski, M.A., et al. (2013). Mutational landscape
and significance across 12 major cancer types. Nature 502, 333–339.
Karnoub, A.E., and Weinberg, R.A. (2008). Ras oncogenes: split personalities.
Nat. Rev. Mol. Cell Biol. 9, 517–531.
Kaufman, B., Shapira-Frommer, R., Schmutzler, R.K., Audeh, M.W., Friedlander, M., Balmaña, J., Mitchell, G., Fried, G., Stemmer, S.M., Hubert, A.,
et al. (2015). Olaparib monotherapy in patients with advanced cancer and a
germline BRCA1/2 mutation. J. Clin. Oncol. 33, 244–250.
Frampton, G.M., Ali, S.M., Rosenzweig, M., Chmielecki, J., Lu, X., Bauer, T.M.,
Akimov, M., Bufill, J.A., Lee, C., Jentz, D., et al. (2015). Activation of MET via
diverse exon 14 splicing alterations occurs in multiple tumor types and confers
clinical sensitivity to MET inhibitors. Cancer Discov. 5, 850–859.
Kim, E., Ilic, N., Shrestha, Y., Zou, L., Kamburov, A., Zhu, C., Yang, X., Lubonja,
R., Tran, N., Nguyen, C., et al. (2016). Systematic functional interrogation of
rare cancer variants identifies oncogenic alleles. Cancer Discov. 6, 714–726.
Gerlinger, M., Rowan, A.J., Horswell, S., Larkin, J., Endesfelder, D., Gronroos,
E., Martinez, P., Matthews, N., Stewart, A., Tarpey, P., et al. (2012). Intratumor
heterogeneity and branched evolution revealed by multiregion sequencing.
N. Engl. J. Med. 366, 883–892.
Kim, K.H., Kim, W., Howard, T.P., Vazquez, F., Tsherniak, A., Wu, J.N., Wang,
W., Haswell, J.R., Walensky, L.D., Hahn, W.C., et al. (2015). SWI/SNF-mutant
cancers depend on catalytic and non-catalytic activity of EZH2. Nat. Med. 21,
1491–1496.
Grossman, R.L., Heath, A.P., Ferretti, V., Varmus, H.E., Lowy, D.R., Kibbe,
W.A., and Staudt, L.M. (2016). Toward a shared vision for cancer genomic
data. N. Engl. J. Med. 375, 1109–1112.
Kitzman, J.O., Starita, L.M., Lo, R.S., Fields, S., and Shendure, J. (2015).
Massively parallel single-amino-acid mutagenesis. Nat. Methods 12, 203–
206, 4, 206.
Cell 168, February 9, 2017 597
Kivioja, T., Vähärautio, A., Karlsson, K., Bonke, M., Enge, M., Linnarsson, S.,
and Taipale, J. (2011). Counting absolute numbers of molecules using unique
molecular identifiers. Nat. Methods 9, 72–74.
LaFave, L.M., Béguelin, W., Koche, R., Teater, M., Spitzer, B., Chramiec, A.,
Papalexi, E., Keller, M.D., Hricik, T., Konstantinoff, K., et al. (2015). Loss of
BAP1 function leads to EZH2-dependent transformation. Nat. Med. 21,
1344–1349.
Pearson, A., Smyth, E., Babina, I.S., Herrera-Abreu, M.T., Tarazona, N., Peckitt, C., Kilgour, E., Smith, N.R., Geh, C., Rooney, C., et al. (2016). High-level
clonal FGFR amplification and response to FGFR inhibition in a translational
clinical trial. Cancer Discov. 6, 838–851.
Peschard, P., Fournier, T.M., Lamorte, L., Naujokas, M.A., Band, H., Langdon,
W.Y., and Park, M. (2001). Mutation of the c-Cbl TKB domain binding site on
the Met receptor tyrosine kinase converts it into a transforming protein. Mol.
Cell 8, 995–1004.
Lai, A., Kahraman, M., Govek, S., Nagasawa, J., Bonnefous, C., Julien, J.,
Douglas, K., Sensintaffar, J., Lu, N., Lee, K.J., et al. (2015). Identification of
GDC-0810 (ARN-810), an orally bioavailable selective estrogen receptor
degrader (SERD) that demonstrates robust activity in tamoxifen-resistant
breast cancer xenografts. J. Med. Chem. 58, 4888–4904.
Piccart-Gebhart, M.J., Procter, M., Leyland-Jones, B., Goldhirsch, A., Untch,
M., Smith, I., Gianni, L., Baselga, J., Bell, R., Jackisch, C., et al.; Herceptin
Adjuvant (HERA) Trial Study Team (2005). Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N. Engl. J. Med. 353, 1659–1672.
Landrum, M.J., Lee, J.M., Benson, M., Brown, G., Chao, C., Chitipiralla, S., Gu,
B., Hart, J., Hoffman, D., Hoover, J., et al. (2016). ClinVar: public archive
of interpretations of clinically relevant variants. Nucleic Acids Res. 44(D1),
D862–D868.
Planchard, D., Kim, T.M., Mazieres, J., Quoix, E., Riely, G., Barlesi, F.,
Souquet, P.J., Smit, E.F., Groen, H.J., Kelly, R.J., et al. (2016). Dabrafenib in
patients with BRAF(V600E)-positive advanced non-small-cell lung cancer: a
single-arm, multicentre, open-label, phase 2 trial. Lancet Oncol. 17, 642–650.
Lanman, R.B., Mortimer, S.A., Zill, O.A., Sebisanovic, D., Lopez, R., Blau, S.,
Collisson, E.A., Divers, S.G., Hoon, D.S., Kopetz, E.S., et al. (2015). Analytical
and clinical validation of a digital sequencing panel for quantitative, highly accurate evaluation of cell-free circulating tumor DNA. PLoS ONE 10, e0140712.
Poulikakos, P.I., Persaud, Y., Janakiraman, M., Kong, X., Ng, C., Moriceau, G.,
Shi, H., Atefi, M., Titz, B., Gabay, M.T., et al. (2011). RAF inhibitor resistance is
mediated by dimerization of aberrantly spliced BRAF(V600E). Nature 480,
387–390.
Lawrence, M.S., Stojanov, P., Mermel, C.H., Robinson, J.T., Garraway, L.A.,
Golub, T.R., Meyerson, M., Gabriel, S.B., Lander, E.S., and Getz, G. (2014).
Discovery and saturation analysis of cancer genes across 21 tumour types.
Nature 505, 495–501.
Prahallad, A., Sun, C., Huang, S., Di Nicolantonio, F., Salazar, R., Zecchin, D.,
Beijersbergen, R.L., Bardelli, A., and Bernards, R. (2012). Unresponsiveness of
colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR.
Nature 483, 100–103.
Le, D.T., Uram, J.N., Wang, H., Bartlett, B.R., Kemberling, H., Eyring, A.D.,
Skora, A.D., Luber, B.S., Azad, N.S., Laheru, D., et al. (2015). PD-1 blockade
in tumors with mismatch-repair deficiency. N. Engl. J. Med. 372, 2509–2520.
Le Tourneau, C., Paoletti, X., Servant, N., Bièche, I., Gentien, D., Rio Frio, T.,
Vincent-Salomon, A., Servois, V., Romejon, J., Mariani, O., et al. (2014). Randomised proof-of-concept phase II trial comparing targeted therapy based on
tumour molecular profiling vs conventional therapy in patients with refractory
cancer: results of the feasibility part of the SHIVA trial. Br. J. Cancer 111,
17–24.
Loxo (2016). Press release: Loxo Oncology receives breakthrough therapy
designation from U.S. Food and Drug Administration for LOXO-101. https://
ir.loxooncology.com/press-releases/loxo-oncology-receives-breakthroughtherapy-designation-from-u.s.-food-and-drug-administration-for-loxo-101.
Mateo, J., Carreira, S., Sandhu, S., Miranda, S., Mossop, H., Perez-Lopez, R.,
Nava Rodrigues, D., Robinson, D., Omlin, A., Tunariu, N., et al. (2015). DNArepair defects and olaparib in metastatic prostate cancer. N. Engl. J. Med.
373, 1697–1708.
Melnikov, A., Rogov, P., Wang, L., Gnirke, A., and Mikkelsen, T.S. (2014).
Comprehensive mutational scanning of a kinase in vivo reveals substratedependent fitness landscapes. Nucleic Acids Res. 42, e112.
Melo, S.A., Luecke, L.B., Kahlert, C., Fernandez, A.F., Gammon, S.T., Kaye, J.,
LeBleu, V.S., Mittendorf, E.A., Weitz, J., Rahbari, N., et al. (2015). Glypican-1
identifies cancer exosomes and detects early pancreatic cancer. Nature
523, 177–182.
Meric-Bernstam, F., Brusco, L., Shaw, K., Horombe, C., Kopetz, S., Davies,
M.A., Routbort, M., Piha-Paul, S.A., Janku, F., Ueno, N., et al. (2015). Feasibility of large-scale genomic testing to facilitate enrollment onto genomically
matched clinical trials. J. Clin. Oncol. 33, 2753–2762.
Milojkovic, D., and Apperley, J. (2009). Mechanisms of resistance to imatinib
and second-generation tyrosine inhibitors in chronic myeloid leukemia. Clin.
Cancer Res. 15, 7519–7527.
Mirza, M.R., Monk, B.J., Herrstedt, J., Oza, A.M., Mahner, S., Redondo, A.,
Fabbro, M., Ledermann, J.A., Lorusso, D., Vergote, I., et al.; ENGOT-OV16/
NOVA Investigators (2016). Niraparib maintenance therapy in platinum-sensitive, recurrent ovarian cancer. N. Engl. J. Med. 375, 2154–2164.
Misale, S., Yaeger, R., Hobor, S., Scala, E., Janakiraman, M., Liska, D., Valtorta, E., Schiavo, R., Buscarino, M., Siravegna, G., et al. (2012). Emergence
of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal
cancer. Nature 486, 532–536.
598 Cell 168, February 9, 2017
Pritchard, C.C., Mateo, J., Walsh, M.F., De Sarkar, N., Abida, W., Beltran, H.,
Garofalo, A., Gulati, R., Carreira, S., Eeles, R., et al. (2016). Inherited DNArepair gene mutations in men with metastatic prostate cancer. N. Engl. J.
Med. 375, 443–453.
Robert, C., Karaszewska, B., Schachter, J., Rutkowski, P., Mackiewicz, A.,
Stroiakovski, D., Lichinitser, M., Dummer, R., Grange, F., Mortier, L., et al.
(2015). Improved overall survival in melanoma with combined dabrafenib
and trametinib. N. Engl. J. Med. 372, 30–39.
Roychowdhury, S., Iyer, M.K., Robinson, D.R., Lonigro, R.J., Wu, Y.M., Cao,
X., Kalyana-Sundaram, S., Sam, L., Balbin, O.A., Quist, M.J., et al. (2011).
Personalized oncology through integrative high-throughput sequencing: a pilot study. Sci. Transl. Med. 3, 111ra121.
Russo, M., Siravegna, G., Blaszkowsky, L.S., Corti, G., Crisafulli, G., Ahronian,
L.G., Mussolin, B., Kwak, E.L., Buscarino, M., Lazzari, L., et al. (2016). Tumor
heterogeneity and lesion-specific response to targeted therapy in colorectal
cancer. Cancer Discov. 6, 147–153.
Saglio, G., Kim, D.W., Issaragrisil, S., le Coutre, P., Etienne, G., Lobo, C., Pasquini, R., Clark, R.E., Hochhaus, A., Hughes, T.P., et al.; ENESTnd Investigators (2010). Nilotinib versus imatinib for newly diagnosed chronic myeloid leukemia. N. Engl. J. Med. 362, 2251–2259.
Scher, H.I., Fizazi, K., Saad, F., Taplin, M.E., Sternberg, C.N., Miller, K., de Wit,
R., Mulders, P., Chi, K.N., Shore, N.D., et al.; AFFIRM Investigators (2012).
Increased survival with enzalutamide in prostate cancer after chemotherapy.
N. Engl. J. Med. 367, 1187–1197.
Schrader, K.A., Cheng, D.T., Joseph, V., Prasad, M., Walsh, M., Zehir, A., Ni,
A., Thomas, T., Benayed, R., Ashraf, A., et al. (2016). Germline variants in
targeted tumor sequencing using matched normal DNA. JAMA Oncol. 2,
104–111.
Shaw, A.T., Kim, D.W., Mehra, R., Tan, D.S., Felip, E., Chow, L.Q., Camidge,
D.R., Vansteenkiste, J., Sharma, S., De Pas, T., et al. (2014). Ceritinib in
ALK-rearranged non-small-cell lung cancer. N. Engl. J. Med. 370, 1189–1197.
Shaw, A.T., Friboulet, L., Leshchiner, I., Gainor, J.F., Bergqvist, S., Brooun, A.,
Burke, B.J., Deng, Y.L., Liu, W., Dardaei, L., et al. (2016). Resensitization to crizotinib by the lorlatinib ALK resistance mutation L1198F. N. Engl. J. Med. 374,
54–61.
Shen, R., and Seshan, V.E. (2016). FACETS: allele-specific copy number and
clonal heterogeneity analysis tool for high-throughput DNA sequencing. Nucleic Acids Res. 44, e131.
Siu, L.L., Lawler, M., Haussler, D., Knoppers, B.M., Lewin, J., Vis, D.J., Liao,
R.G., Andre, F., Banks, I., Barrett, J.C., et al. (2016). Facilitating a culture of
responsible and effective sharing of cancer genome data. Nat. Med. 22,
464–471.
Slamon, D.J., Leyland-Jones, B., Shak, S., Fuchs, H., Paton, V., Bajamonde,
A., Fleming, T., Eiermann, W., Wolter, J., Pegram, M., et al. (2001). Use of
chemotherapy plus a monoclonal antibody against HER2 for metastatic breast
cancer that overexpresses HER2. N. Engl. J. Med. 344, 783–792.
Stransky, N., Cerami, E., Schalm, S., Kim, J.L., and Lengauer, C. (2014). The
landscape of kinase fusions in cancer. Nat. Commun. 5, 4846.
Swain, S.M., Baselga, J., Kim, S.B., Ro, J., Semiglazov, V., Campone, M., Ciruelos, E., Ferrero, J.M., Schneeweiss, A., Heeson, S., et al.; CLEOPATRA
Study Group (2015). Pertuzumab, trastuzumab, and docetaxel in HER2-positive metastatic breast cancer. N. Engl. J. Med. 372, 724–734.
Tabernero, J., Bahleda, R., Dienstmann, R., Infante, J.R., Mita, A., Italiano, A.,
Calvo, E., Moreno, V., Adamo, B., Gazzah, A., et al. (2015). Phase I dose-escalation study of JNJ-42756493, an oral pan-fibroblast growth factor receptor inhibitor, in patients with advanced solid tumors. J. Clin. Oncol. 33, 3401–3408.
Tannock, I.F., and Hickman, J.A. (2016). Limits to personalized cancer medicine. N. Engl. J. Med. 375, 1289–1294.
Telli, M.L., Timms, K.M., Reid, J., Hennessy, B., Mills, G.B., Jensen, K.C., Szallasi, Z., Barry, W.T., Winer, E.P., Tung, N.M., et al. (2016). Homologous recombination deficiency (HRD) score predicts response to platinum-containing
neoadjuvant chemotherapy in patients with triple-negative breast cancer.
Clin. Cancer Res. 22, 3764–3773.
Toy, W., Shen, Y., Won, H., Green, B., Sakr, R.A., Will, M., Li, Z., Gala, K., Fanning, S., King, T.A., et al. (2013). ESR1 ligand-binding domain mutations in hormone-resistant breast cancer. Nat. Genet. 45, 1439–1445.
Turner, N.C., Ro, J., André, F., Loi, S., Verma, S., Iwata, H., Harbeck, N., Loibl,
S., Huang Bartlett, C., Zhang, K., et al.; PALOMA3 Study Group (2015). Palbociclib in hormone-receptor-positive advanced breast cancer. N. Engl. J. Med.
373, 209–219.
Vaishnavi, A., Le, A.T., and Doebele, R.C. (2015). TRKing down an old oncogene in a new era of targeted therapy. Cancer Discov. 5, 25–34.
Voest, E.E., and Bernards, R. (2016). DNA-guided precision medicine for cancer: a case of irrational exuberance? Cancer Discov. 6, 130–132.
Wan, P.T., Garnett, M.J., Roe, S.M., Lee, S., Niculescu-Duvaz, D., Good, V.M.,
Jones, C.M., Marshall, C.J., Springer, C.J., Barford, D., and Marais, R.; Cancer
Genome Project (2004). Mechanism of activation of the RAF-ERK signaling
pathway by oncogenic mutations of B-RAF. Cell 116, 855–867.
Yang, J.C., Sequist, L.V., Geater, S.L., Tsai, C.M., Mok, T.S., Schuler, M.,
Yamamoto, N., Yu, C.J., Ou, S.H., Zhou, C., et al. (2015). Clinical activity of
afatinib in patients with advanced non-small-cell lung cancer harbouring uncommon EGFR mutations: a combined post-hoc analysis of LUX-Lung 2,
LUX-Lung 3, and LUX-Lung 6. Lancet Oncol. 16, 830–838.
Yao, Z., Torres, N.M., Tao, A., Gao, Y., Luo, L., Li, Q., de Stanchina, E., AbdelWahab, O., Solit, D.B., Poulikakos, P.I., and Rosen, N. (2015). BRAF mutants
evade ERK-dependent feedback by different mechanisms that determine their
sensitivity to pharmacologic inhibition. Cancer Cell 28, 370–383.
Thress, K.S., Paweletz, C.P., Felip, E., Cho, B.C., Stetson, D., Dougherty, B.,
Lai, Z., Markovets, A., Vivancos, A., Kuang, Y., et al. (2015). Acquired EGFR
C797S mutation mediates resistance to AZD9291 in non-small cell lung cancer
harboring EGFR T790M. Nat. Med. 21, 560–562.
Yu, M., Bardia, A., Wittner, B.S., Stott, S.L., Smas, M.E., Ting, D.T., Isakoff,
S.J., Ciciliano, J.C., Wells, M.N., Shah, A.M., et al. (2013). Circulating breast
tumor cells exhibit dynamic changes in epithelial and mesenchymal composition. Science 339, 580–584.
Tiacci, E., Park, J.H., De Carolis, L., Chung, S.S., Broccoli, A., Scott, S., Zaja,
F., Devlin, S., Pulsoni, A., Chung, Y.R., et al. (2015). Targeting mutant BRAF in
relapsed or refractory hairy-cell leukemia. N. Engl. J. Med. 373, 1733–1747.
Zaretsky, J.M., Garcia-Diaz, A., Shin, D.S., Escuin-Ordinas, H., Hugo, W., HuLieskovan, S., Torrejon, D.Y., Abril-Rodriguez, G., Sandoval, S., Barthly, L.,
et al. (2016). Mutations associated with acquired resistance to PD-1 blockade
in melanoma. N. Engl. J. Med. 375, 819–829.
Tirosh, I., Izar, B., Prakadan, S.M., Wadsworth, M.H., 2nd, Treacy, D.,
Trombetta, J.J., Rotem, A., Rodman, C., Lian, C., Murphy, G., et al. (2016). Dissecting the multicellular ecosystem of metastatic melanoma by single-cell
RNA-seq. Science 352, 189–196.
Zhang, J., Walsh, M.F., Wu, G., Edmonson, M.N., Gruber, T.A., Easton, J.,
Hedges, D., Ma, X., Zhou, X., Yergeau, D.A., et al. (2015). Germline mutations
in predisposition genes in pediatric cancer. N. Engl. J. Med. 373, 2336–2346.
Cell 168, February 9, 2017 599
Leading Edge
Primer
Applications of Immunogenomics to Cancer
X. Shirley Liu1,* and Elaine R. Mardis2,*
1Department
of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, Harvard T.H. Chan School of Public Health,
450 Brookline Ave, Boston MA 02215, USA
2Institute for Genomic Medicine, Nationwide Children’s Hospital, and The Ohio State University College of Medicine, 575 Children’s
Crossroad, Columbus OH 43205, USA
*Correspondence: xsliu@jimmy.harvard.edu (X.S.L.), elaine.mardis@nationwidechildrens.org (E.R.M.)
http://dx.doi.org/10.1016/j.cell.2017.01.014
Cancer immunogenomics originally was framed by research supporting the hypothesis that cancer
mutations generated novel peptides seen as ‘‘non-self’’ by the immune system. The search for
these ‘‘neoantigens’’ has been facilitated by the combination of new sequencing technologies,
specialized computational analyses, and HLA binding predictions that evaluate somatic alterations
in a cancer genome and interpret their ability to produce an immune-stimulatory peptide. The resulting information can characterize a tumor’s neoantigen load, its cadre of infiltrating immune
cell types, the T or B cell receptor repertoire, and direct the design of a personalized therapeutic.
Brief History of Tumor-Specific Mutant Antigens and
Immunogenomics
The underpinnings of modern immunogenomics resulted from
hypotheses generated and tested by visionaries in cancer immunology during the late 1980s through the 1990s. Their central
hypothesis was that cancer cells presented novel, tumor-specific (i.e., mutated) peptides on the cancer cell surface bound
by the patient’s HLA molecules. By virtue of this cell surface presentation, specific T cell immunity might be elicited to these
‘‘neoantigens.’’ Supporting evidence for this hypothesis was
demonstrated in cancers of non-viral origin (Old and Boyse,
1964; Foley, 1953; Prehn and Main, 1957). This foundational
work led to the identification and characterization of the role
of MHC proteins in antigen presentation (Babbitt et al., 1985;
Bjorkman et al., 1987). Concomitantly, methods to grow antigen-specific cytolytic T lymphocytes (CTLs) in culture were
also developed (Cerottini et al., 1974; Gillis and Smith, 1977),
as were the molecular biology procedures to clone and express
gene products. Thierry Boon’s laboratory combined these new
methods to identify the first tumor specific antigen (TSA), a point
mutation in a protein called P91A (De Plaen et al., 1988). Subsequently, Hans Schreiber’s laboratory demonstrated that TSAs
also function as neoantigens using primary UV-induced mouse
tumors (Monach et al., 1995). Similarly, groups studying human
melanomas showed they could identify T cells in the peripheral
circulation that bind melanoma cells preferentially over normal
cells from the same patient (Dubey et al., 1997; Knuth et al.,
1984; Robbins et al., 1996; Van den Eynde et al., 1989). Shortly
thereafter, Boon’s laboratory cloned the first human TSA, called
MAGEA1 (van der Bruggen et al., 1991), and Sahin’s group
demonstrated an autologous antibody-based method to clone
and identify different human TSAs (Sahin et al., 1995). While
these foundational studies established supporting evidence for
the existence of tumor-specific peptide neoantigens, the lengthy
and painstaking nature of these processes was unlikely to scale
to clinical application for cancer patients.
600 Cell 168, February 9, 2017 ª 2017 Elsevier Inc.
More recently, these limitations have been alleviated by the
application of new sequencing technologies and associated
computational data analysis approaches. These methods,
collectively referred to as ‘‘immunogenomics,’’ have improved
the facility with which individual cancers can be studied to predict their neoantigens for prognostic purposes or to inform
immunotherapeutic interventions. Complementary methods
have been developed to study the changes in the T cell repertoire, to characterize the gene expression signatures of the immune cell types present in the tumor mass, and to design
personalized vaccines or adoptive cell transfer (ACT) therapies.
The now scalable nature of immunogenomic methods should
permit their widespread clinical application, although there
remain issues and challenges to be resolved. This primer will
highlight the specific methods and describe the known strengths
and weaknesses in modern immunogenomics.
Somatic Mutations Generate Neoantigens
It has long been known that cancer is caused by alterations to
genomic DNA that impact protein functions, ultimately disrupting cellular control of pathways and resulting in the outgrowth
of a tumor mass. Methods using next generation sequencing
platforms generate data from tumor and normal DNA isolates
that, once aligned to the Human Reference Genome sequence,
can be interpreted to identify somatic alterations (Ley et al.,
2008). In practice, such analyses aim to identify DNA alterations
in known cancer genes, both oncogenes and tumor suppressors
that combine to transform the founder cell. For certain oncogenes, identified mutations indicate therapeutic interventions
that may successfully halt the tumor cell growth. By contrast,
immunogenomic approaches aim to identify tumor-specific
DNA alterations that predict amino acid sequence changes in
all encoded proteins, and then evaluate their potential as neoantigens. In practice, most TSAs identified to-date are highly
unique to each patient and generally do not involve known cancer genes.
Hence, the widespread use of next-generation sequencing
(NGS) instrumentation has enabled immunogenomics, providing
a facile way to generate data to predict tumor-specific neoantigens in a rapid, inexpensive and comprehensive manner (Gubin
et al., 2015). NGS technologies have rapidly evolved over
the past 10 years, resulting in dramatically increased amounts
of sequencing data produced per instrument run at everdecreasing costs (Mardis, 2017). In immunogenomics, since
the focus is protein-coding genes, solution hybridization-based
methods are used to select these sequences (‘‘exome’’) prior
to sequencing (Bainbridge et al., 2010; Gnirke et al., 2009;
Hodges et al., 2009). Importantly, the concomitant development
of advanced variant detection algorithms that identify different
classes of mutations from NGS data has enabled the identification of all classes of somatic variation. Accurate detection of variants in this setting is influenced by multiple factors, which are
presented here in detail.
One important consideration for somatic variant detection is
depth of coverage by NGS sequencing reads from the tumor.
In principle, since tumor samples include variable percentages
of normal cells, adjustments to the depth of NGS data generated
must be flexible to ensure that a sufficient representation of tumor-derived sequence reads are obtained. Isolating DNA from
selected, tumor-rich areas of a biopsy or resection sample is
ideal, but not always possible, so average read depths of 300to 500-fold exome coverage are typically attempted to compensate for the normal cell DNA-derived reads. A second reason for
high coverage of the tumor-derived DNA is to enable the evaluation of founder clone versus subclonal mutations in the resulting
data. Here, we define founder mutations as the original set of
mutations present in the cell that transformed from normal to
neoplastic, whereas subclonal mutations occur as the daughter
cells of this founder acquire additional mutations during growth
of the tumor mass. Based on this definition, founder clone
mutations in diploid regions of the exome have a proportional
fraction of variant-containing sequencing reads (variant allele
fraction or VAF) that is around 50% (adjusted for normal DNA
contribution), since most somatic mutations are heterozygous.
In theory, neoantigens that result from founder clone mutations
should elicit a T cell response that targets all cancer cells rather
than the subset of tumor cells that would be targeted by T cell
response to subclonal neoantigens in the vaccine.
Equally important to appropriate coverage depth for accurate
prediction of variants is the algorithm or set of algorithms
used to identify variants from the NGS exome data. The factors
to consider here include the types of variants one wishes to
evaluate in neoantigen discovery. For example, single nucleotide variants (point mutations) are easiest to predict with high accuracy because reads containing a single variant are readily
aligned to their reference genome ‘‘match,’’ and because there
are a variety of different algorithms that also can detect low
VAF variants. Variant detection from NGS reads has been an
area of rapid development and there are many algorithms to
choose from, with variable performance, as has been evaluated
(Cornish and Guda, 2015; Ghoneim et al., 2014; Krøigård et al.,
2016). By contrast, variants resulting from insertion or deletion
of one or a few nucleotides (‘‘indels’’) are significantly more difficult to identify due to issues of read alignment by standard align-
ment algorithms, that lead often to lower coverage in these
regions for the variant-containing sequencing reads (Jiang
et al., 2012, 2015; Ratan et al., 2015). However, indels may be
important to immunogenomics efforts because they can introduce frameshift mutations that result in highly divergent amino
acid sequences in the resulting protein and hence may produce
strong predicted neoantigens. Increased read lengths on NGS
platforms have improved indel detection, as has the use of
gapped alignment or split-read algorithms that are computationally intensive but better able to align the indel-containing reads
to the reference genome. Assembly-based realignment approaches also have been developed to improve the precision
of indel variant detection (Mose et al., 2014; Narzisi et al., 2014).
Another type of somatic variation that can lead to highly
altered amino acid sequences, and as a result create a neoantigenic peptide, is a structural variant which fuses two proteincoding sequences. These can result from inversion or deletion
of a chromosomal segment or from chromosomal translocations. Detecting these alterations from exome sequencing data
is quite challenging and error-prone, but RNA-based analysis
can identify the resulting fusion transcript (Li et al., 2011; Scolnick et al., 2015; Zhang et al., 2016a; Kumar et al., 2016) and
compare the predicted fusion sequence to NGS data from
DNA (whole genome or exome sequencing) to identify supporting evidence of the genomic event causing the fusion. Recently,
we adapted this approach for neoantigen prediction with a process called IntegrateNEO, using the TMPRSS2-ERG fusions
common in prostate cancer to evaluate its ability to identify
fusion peptide neoantigens (Zhang et al., 2016b). RNaseq data
bring added value to immunogenomics efforts beyond the
detection of fusion peptides, as will be described later.
Once variant detection is completed, each variant is annotated
to predict the resulting amino acid change(s) that result from the
altered DNA sequence (if any). There are widely utilized computational tools such as Annovar and VEP available to produce the
translated peptides from the DNA data. The translated peptides
constitute one type of input data for the neoantigen prediction
software to calculate the class I or class II predicted binding affinities.
The second data input for neoantigen prediction are the HLA
haplotypes of the patient, also derived from exome data, since
these reagents capture the HLA gene loci. Heretofore, HLA
typing was performed using a PCR-based and Sanger
sequencing-based clinical assay. The repetitive nature of the
HLA genes requires a high-stringency assembly of these genes,
which can be achieved using the >500 bp read lengths from
Sanger data. Sequence analysis of these regions based on
hybrid capture-derived NGS reads, which are relatively short
(100 bp), requires a stringent alignment of the read data to
the IMGT/HLA database (Robinson et al., 2001) using a haplotype-resolved algorithm to interpret the HLA class I and II haplotypes. There now exist multiple algorithms for accomplishing
these data interpretations, including Polysolver (Shukla et al.,
2015), HLAMiner (Warren et al., 2012), and OptiType (Szolek
et al., 2014). Typically, one interprets the normal tissue-derived
exome data to obtain the HLA haplotypes. Clinical analysis of
these genes also should include repeating the alignment of the
tumor-derived exome data and identification of mutations in
Cell 168, February 9, 2017 601
order to identify HLA alleles that are impacted by nonsense mutations, deletions, or other similarly deleterious types of somatic
alterations that may influence the presence of that allele (Shukla
et al., 2015). Some algorithms also can use RNA-derived data to
interpret the HLA haplotypes (Warren et al., 2012).
Another critical component of identifying neoantigens is the
in silico prediction of HLA class I and II binding affinities for
specific peptides. These predictions are quite computationally
complex and require machine learning-based approaches to
establish models for the different types of binding site interactions. In particular, each peptide interacts with the binding
pocket residues of the many different HLA proteins through the
amino acid side chains of specific residues. Therefore, the
binding affinity of any peptide is sequence-specific relative to
that patient’s HLA proteins, some of which may be common
and some rare. There also are differences in the binding of
peptides by class I or class II HLA that impact the precision
of neoantigen prediction, as described later. Finally, there is
considerable debate about an appropriate cutoff value for binding affinity in terms of what does or does not constitute a strong
neoantigen candidate (Duan et al., 2014)(Bassani-Sternberg
et al., 2016)
The initial approach to computational HLA binding predictions
utilized a neural network-based learning method developed from
a training set of experimentally derived binding affinities for class
I HLA proteins and different peptides. This effort resulted in an
HLA class I binding prediction software known as netMHC,
devised by researchers in the Center for Biological Sequence
Analysis at the Technical University of Denmark (Lundegaard
et al., 2008a, 2008b; Nielsen et al., 2003). The predictor has
improved over time with the availability of training datasets for
HLA proteins that are more rare in the population, although
calculated binding affinities for the most rare HLA alleles in humans remain less certain (Wang et al., 2010). An interim
approach to address rare HLA class I binding calculations was
PickPocket, which extrapolated from variants with known binding specificity to those without existing experimental data (Zhang
et al., 2009). The most recent version is netMHCstabpan (Rasmussen et al., 2016), which uses a neural network approach
based on a dataset of stability values calculated for different
peptide-MHC-1 complexes, rather than their binding affinity
values, since the stability of their interaction has experimentally
been shown to be more strongly correlated to T cell immunogenicity. Another early method developed to generate class I binding predictions was based on a stabilized matrix method (SMM)
algorithm developed by Peters and Sette (Peters and Sette,
2005). This approach models the sequence specificity of binding
processes as a means of predicting outcomes for untested sequences. SMM not only predicts HLA binding but also evaluates
peptide transport as a function of antigen presentation and proteasomal cleavage with the TAP algorithm. Subsequent efforts
to develop new class I binding affinity prediction software have
included the use of combined support vector machine-based
(SVM) and random forest machine-learning approaches (Srivastava et al., 2013), or combined the information obtained from
amino acid pairwise contact potentials and quantum topology
molecular similarity descriptors (Saethang et al., 2013) to better
model HLA class I peptide interactions.
602 Cell 168, February 9, 2017
With the requisite information generated by NGS to call somatic variants and interpret their impact on protein sequences,
and to identify the HLA haplotypes specific to the patient, neoantigen prediction software can be used to predict both the class I
and class II HLA binding affinities for each tumor-unique set of
peptides. Considerations and specifics for these prediction approaches are described in detail below. There are a number of
binding prediction software and associated immunogenomics
algorithms available at the Immune Epitope DataBase (IEDB)
analysis resource (http://tools.immuneepitope.org/main/) (Robinson et al., 2013). The IEDB web interface permits the input
of peptide sequences for sequential evaluation by user-configured steps using the software of choice to predict neoantigens.
Publicly available software pipelines also are available for
local download and computing of neoantigen predictions by
end-users, including pVAC-seq (https://github.com/griffithlab/
pVAC-Seq) and epidisco (https://github.com/hammerlab). An
overall workflow for the processes described above is shown
in Figure 1.
Class I Predictions
Approaches to predict HLA class I neoantigens typically begin
by parsing the tumor-specific peptides predicted from variant
calling as 21-mer peptides that encompass the variant amino
acid(s) placed as near to the center of the 21-mer as possible.
This is easiest to envisage for simple non-synonymous amino
acid substitutions, shown in Figure 2A, which then are tiled
across the variant-containing peptides to define a set of 8-mer
to 11mers to input for binding calculations, based on HLA
class I binding characteristics (Figure 2B). These peptide sets
are parsed along with their corresponding wild-type peptide
sequences as input data for consideration by neoantigen
prediction software, along with information about the HLA
class I haplotypes determined for the patient. The resulting list
of neoantigens can be quite extensive, depending upon the
numbers and types of input peptide sequences and the diversity
of the HLA haplotypes. Applying several criteria, if desired,
can winnow the numbers of neoantigens. One conventional
approach is to only consider variant peptides with a strongto intermediate-binding affinity (typically lower than 500 nM)
but this arbitrary cut-off is controversial because strong neoantigens can have lower calculated affinities than actual. This
sometimes is due to the presence of a rare HLA haplotype, for
which the neural net software provides an inaccurate binding affinity prediction. Thus, for each altered locus, one can select the
candidate peptide with the single best binding affinity to each
corresponding HLA allele across all peptide lengths considered,
or proceed with all candidates for all HLA alleles to additional
filtering steps, as follows.
Three important additional filters should be applied to remove false positives, (1) RNA-based filtering to remove genes
with no evidence of expression, (2) filtering based on exome
data coverage depth at the variant loci, and (3) filtering based
on variant allele fraction (VAF)-based metrics. The RNA expression filter ensures that each peptide is supported by evidence
of RNA expression, wherein evidence of RNA expression is
considered a reasonable, but not absolute, proxy that the
gene is expressed in the tumor cell proteome. For the NGS
coverage filter, a minimum level of normal read coverage depth
Figure 1. An Overall Workflow for Neoantigen Discovery and Personalized Cancer
Vaccine Design
Starting from next-generation sequencing of
DNA exomes to compare tumor to normal DNA,
and of tumor RNA to evaluate gene expression,
this figure illustrates the steps outlined in the
primer to identify tumor-specific mutant antigens
(neoantigens) from NGS data, to evaluate the
neoantigens, and to design a personalized neoantigen vaccine.
is required to ensure there is sufficient sequencing data
coverage from the normal tissue (i.e., supports a true positive
somatic variant call). Finally, both DNA and RNA data should
be evaluated to ascertain the percentage of variant-containing
reads or variant allele fraction (VAF). As described earlier, this
criterion helps to inform the final list of neoantigen candidates
by providing information on whether a specific alteration is
shared across all tumor cells (i.e., in the founder clone) or is
subclonal, based on DNA sequencing data, and ensures that
a variant is expressed in the tumor RNA. The latter is especially
important in tumor types with a high mutation load such as
those with chemical or UV damage to DNA, since upward of
50% of mutations are typically not expressed in RNA (or protein
by inference) for these tumors. With these filtering steps
completed, a list of high confidence, predicted neoantigenic
peptides and the HLA class I proteins predicted to bind
them, their calculated binding affinity value(s), and the binding
affinity of the cognate wild-type peptide values can be parsed
for further consideration in vaccine design or other immunological evaluations such as neoantigen burden. In the former case,
neoantigen predictions have been tested in clinical trials of
personalized vaccines, with demonstrated ability to elicit spe-
cific T cell responses (Schumacher
et al., 2014; Carreno et al., 2015; Tran
et al., 2014). In the latter approach, there
are demonstrated correlations between
neoantigen burden and the likelihood of
response to checkpoint blockade inhibition therapies (Le et al., 2015; Rizvi et al.,
2015; Snyder et al., 2014; Van Allen
et al., 2015), and a demonstration that
predicted neoantigens also are the epitopes targeted by checkpoint blockade
immunotherapies (Gubin et al., 2014).
Class II Predictions
HLA class II predictions are significantly
more difficult to generate with precision
due to the nature of the HLA class II
proteins. First, class II HLA proteins are
heterodimers of alpha and beta peptides
encoded by four different loci in the
human genome. Only one of these four
loci is not highly polymorphic (Robinson
et al., 2003), meaning there is extensive
HLA class II polymorphism in the general
population. This becomes somewhat less
complex if neoantigen predictions focus on the most frequently
expressed class II molecules (McKinney et al., 2013). Second,
certain peptides bind to multiple different HLA class II molecules
and are responsible for the majority of antigen-specific T cell responses. These so-called ‘‘promiscuous peptides’’ are difficult
to predict using computational approaches. Third, the HLA
class II binding groove is open on both ends, and although the
core binding motif is a 9-mer amino acid, variable length peptides are allowed to bind. Many of the HLA-II polymorphic sites
comprise other regions of the binding groove outside the core
motif binding region, which allows the flanking amino acid sequences on either side of the motif sequence to influence its
binding affinity. As a result, binding affinities are difficult to
predict with a high degree of precision. Input data for MHC
class II binding predictions consist of 15-mer representatives of each somatic neoantigen candidate peptide, along
with the patient’s HLA class II haplotypes. A cutoff binding
of <1,000 nM may be utilized to distinguish strong binders but
given the vagaries of binding affinity predictions described
above, this cutoff may not be appropriate. RNA expression
level has been identified as a critical filtering parameter for predicted class II neoantigen candidates, whereby those peptides
Cell 168, February 9, 2017 603
Figure 2. Idealized Selection of MutantContaining Peptides for Neoantigen Prediction
(A) The localized peptides that tile across and
contain the mutated amino acid substitution are
identified and parsed into the neoantigen prediction pipeline. Each peptide is considered for HLA
binding strength relative to its non-mutant (wildtype) counterpart.
(B) Shown is the top scoring candidate peptide
that was selected across all specified k-mers and
between all HLA types that were input to the neoantigen prediction pipeline.
to and properly presented by HLA
molecules on the cell surface. This
critical component of T cell activation
must occur for the neoantigen to stimulate a specific immune response, yet it
is presently not possible to computationally predict the processing and
presentation of peptides by HLA. One
way to inform neoantigen prediction
methods is using experimental measurements of T-cell-based immune responses to the predicted peptide epitopes. There are conventional methods
such as EliSpot (IFN-gamma release)
assays (Cole, 2005), flow cytometrybased dextramer assays (Carreno
et al., 2015), and mass spectrometrybased evaluation of HLA-bound peptides (Gubin et al., 2014). However,
scalable, high-throughput methods are
in development at present and will
require time and testing.
corresponding to genes with higher relative expression values
from RNaseq data analysis are considered to be the strongest
candidates (Kreiter et al., 2015).
Computational predictions, considering the aforementioned
caveats for both class I and II, therefore only offer putative
neoantigen candidates that may be subject to a variety of errors or sources of inaccuracy. In addition to what we already
have described, there are other challenges to accurate neoantigen prediction. First, even though RNA evidence supports
a variant as being expressed, the most accurate evidence
of a peptide’s presence in the cell is identifying that peptide from mass spectrometry-based proteomic data derived
from the specific tumor under study. Second, binding affinity
calculations are more accurate for the common class I HLA
haplotypes, less so for rarer haplotypes. Third, a significant
biological confounder of neoantigen discovery is our inability
to predict precisely which of the putative neoantigen peptides
will be processed in the tumor cell degradasome, then bound
604 Cell 168, February 9, 2017
Immune Repertoire Profiling
Cellular immune responses from T cells
and humoral immune responses from
B cells are stimulated by exposures to antigens, including pathogens, allergens, and neoantigens. V(D)J recombination in the
primary lymphoid organs creates the incredibly diverse and
unique repertoire of the hypervariable regions of B cell receptors
(BCR) and T cell receptors (TCR), and somatic hypermutations
contribute to additional BCR diversity (Figure 3). During B and
T cell development, self-antigens are presented to B and
T cells to select out self-reacting types, and to ensure only B
and T cells that recognize and attack foreign antigens are in
the circulation. T cells only recognize foreign proteins presented
on MHC, while B cells can also target foreign DNA, lipids, or carbohydrates. Upon recognition of foreign antigens and with the
presence of co-stimulatory molecules, B and T cells express
cell surface activation markers, attack foreign antigens, secrete
cytokines, stimulate each other, and proliferate (Pasternack,
1994). One goal of immunogenomic studies is to characterize
the repertoire of B and T cells in patients with cancer, especially
before and after immunotherapy-based interventions.
Figure 3. Structure and Diversity in the T
Cell Receptor
(A) The mature T cell heterodimer, consisting of
a- and b-subunit chains. The a subunit chains
consist of variable (V), joining (J), and constant
(C) regions, whereas the b subunit includes an
additional diversity (D) region.
(B) V-D-J recombination and post-transcriptional
processing of a TCR-b subunit chain.
DNA sequencing approaches have enabled the characterization of immune repertoires (Pasternack 1994; Robins 2013). After
a pioneering study introduced the technique (Freeman et al.,
2009), a plethora of immune repertoire methods have been published and commercial solutions are also available. Several
studies (Calis and Rosenberg, 2014; Hou et al., 2016; Yaari
and Kleinstein, 2015) have evaluated the experimental techniques and practical advice needed for immune repertoire
profiling. Basically, multiplex PCR can amplify the recombined
V(D)J regions from either mRNA or DNA in the B or T cells. The
V(D)J, and most importantly, the variable complementaritydetermining region CDR3 sequences, and their respective
abundance can be resolved by high-throughput sequencing.
Paired-end sequencing with additional PCR primers in the
middle of the fragment permits full-length TCR repertoire
sequencing with short read NGS technology to resolve the V/J
pairing (Cole et al., 2016). One caveat to this approach is that
PCR biases and sequencing errors can falsely increase the total
repertoire with deeper sequencing coverage, so unique molecular identifier barcodes should be used to eliminate such artifacts
(Cole et al., 2016), although such an approach is presently only
available for RNA-based repertoire profiling.
Computational methods, as summarized in (Greiff et al.,
2015a, 2015b) are important components for the analysis, annotation, and visualization of immune repertoires. To this end,
IMGT (Giudicelli et al., 1997) is the most widely cited immunogenetics database and provides many useful tools such as
V-QUEST and HighV-QUEST (Alamyar et al., 2012) as well as
statistical metrics (Aouinti et al., 2015) for the analysis and annotation of immune repertoire data. VDJtools (Shugay et al., 2015)
is a comprehensive analysis framework for T cell and B cell
repertoire sequencing data. It includes MIXCR for fast alignment
and clonal type assembly (Bolotin et al.,
2015), MIGEC for removing duplicates
and combining barcodes (Shugay et al.,
2014), and VDJviz for visualization (Bagaev et al., 2016), and provides basic statistical analyses for characterizing and
comparing different immune repertoires.
The initial output from a repertoire
profiling analysis is a list of BCR/TCR
CDR3 sequences, sometimes including
the adjoining V and J sequences, each
followed by an abundance estimate.
This output allows samples to be
compared and clustered, if desired. For
example, common CDR3 sequences
that are shared among individuals indicate BCR/TCR clones that recognize common antigens such
as herpes or common cold viruses. In comparison, CDR3 clones
that are rare among patients but are abundant within a tumor,
and more importantly for BCR lineage-related CDR3s with small
numbers of mutations, indicate T/B cell recognition of the patient
tumor-specific antigens (Saul et al., 2016). Repertoire profiles
from individuals with similar ethnic backgrounds, lifestyles, or
environmental exposures are often clustered. Two independent
metrics, diversity (often measured by the Shannon entropy), and
evenness (indicative of the degree of clonal expansion), have
been proposed as important characteristics of immune repertoires (Greiff et al., 2015b). Since V(D)J recombination in TCR
only occurs in children, TCR diversity generally declines with
age. In contrast, V(D)J recombination in BCR occurs throughout
life although at reduced levels in adults, and activated BCR undergoes somatic hypermutation to improve the antibody affinity
to the recognized antigen, so the BCR diversity distributions assume more complex patterns. Although immune-stimulating
events such as allergy or vaccination could shift the abundance
of some clones, the immune repertoire has been suggested as a
means to monitor an individual’s immune health (Johnson et al.,
2014). The utility of this metric depends on the accurate measure
of clonal abundance, which requires linear amplification from
multiplex PCR products and additional normalization of TCR/
BCR expression levels for RNA-based profiles. Furthermore,
the method and time-span of sample storage can also influence
sample quality for repertoire profiling.
While immune repertoires are informative, profiling them over
large sample cohorts can be expensive. Computational methods
have been developed to directly infer immune repertoires from
unselected bulk tumor RNaseq data, such as TRUST for TCR
(Li et al., 2016a) and V’DJer for BCR (Mose et al., 2016). The
Cell 168, February 9, 2017 605
hypervariability of the CDR3 regions of TCR and BCR renders the
RNaseq reads from these regions unmappable to the human
reference genome sequence, and somatic hypermutation adds
additional challenges to BCR mapping and alignment. Both of
the aforementioned methods select unmappable RNaseq reads,
align these unmapped reads to each other with de Brujin graphing methods, de novo assemble these alignments into contigs,
and use IMGT (Giudicelli et al., 1997) to annotate those containing CDR3 motifs as potential BCR or TCR. Although these
approaches only recover the most abundant of the immune repertoires, they were used to analyze RNA-seq data across tumor
samples profiled by The Cancer Genome Atlas (TCGA) and resulted in novel findings. For example, TRUST revealed increased
T cell clonal diversity in tumor types with higher mutational loads
and potential neoantigens based on their co-occurrence with
CDR3-containing sequences in the tumors (Li et al., 2016a),
while V’DJer reported higher somatic hypermutation in IgG and
IgA than in IgM (Mose et al., 2016).
Published studies have made fascinating observations on how
immune repertoires can reflect an individual patient’s immune
health and predict their response to therapy. The ability to reconstruct a more diverse TCR repertoire after autologous hematopoietic stem cell transplantation has been observed to predict
better transplant outcomes in multiple sclerosis patients (Johnson et al., 2014; Muraro et al., 2014). Another study used TCR
repertoire sequencing to compare each patient’s TCR before
and after dendritic cell-based neoantigen vaccine dosing, illustrating expanded TCRs for the vaccine peptides that elicited a
T cell response (Carreno et al., 2015). For metastatic melanoma
patients, the anti-CTLA4 antibody ipilimumab has been shown to
increase peripheral blood TCR diversity (Robert et al., 2014), and
those patients with higher peripheral TCR diversity before treatment were reported to respond better to ipilimumab (Postow
et al., 2015). In contrast, the anti-PD-1 antibody pembrolizumab
showed better efficacy in melanoma patients whose pre-treatment tumor-infiltrating T cells were less diverse and more clonal
(Tumeh et al., 2014). This study also demonstrated that more tumor-infiltrating T cell clones expanded after treatment in the
therapy responsive group than in the (non-responding) disease
progression group. Although these pioneering studies were conducted on a limited number of patients, they do suggest TCR
repertoire as a universal cancer immunotherapy biomarker
(McNeel, 2016). Potentially overall patient immune health from
the peripheral TCR and signs of neoantigen recognition and
clonal expansion from the tumor TCR before treatment could
predict better patient response to cancer immunotherapies. As
an example, one bioinformatics study using a Potential Support
Vector Machine-based approach reported the ability to predict
an individual’s age, health, transplantation status, and development of lymphoid cancer based on repertoire profiles (Greiff
et al., 2015b).
Distribution of Tumor Infiltrating Lymphocytes
Large-scale molecular tumor profiling often selects samples
with high tumor purity to best characterize the molecular signatures of the tumor. While most cancer genomics studies are
focused on the cancerous cells in the tumor tissue, the impurities, such as stromal cells, endothelial cells, and immune cells,
606 Cell 168, February 9, 2017
could have major impact on the development and progression of
cancer. With genomic profiling, tumor purity could be estimated
from DNA copy number (Carter et al., 2012), SNP allele frequency (Li and Li, 2014), RNA-seq (Yoshihara et al., 2013), or
DNA methylation (Zhang et al., 2015; Zheng et al., 2014)
data. Interestingly, these methods using orthogonal tumor
profiling modalities yield very consistent tumor purity estimates,
in distinct contrast to the estimates provided by pathologists,
suggesting that molecular and morphological changes in the tumor do not appear simultaneously.
Pertinent to immunogenomic studies of cancer is the evaluation of tumor-infiltrating lymphocytes (TILs), which can involve
traditional approaches such as flow cytometry and multiplex
immunohistochemistry. Flow cytometry uses antibodies against
proteins uniquely expressed on different subpopulations of immune cells to isolate specific subsets of these cells from blood
or tissues. The resulting cell counts characterize the relative
abundance of different subpopulations in individual cancer samples and can reveal changes following treatment. Flow-cytometry requires relatively large fresh tissue samples for study, but the
resulting isolated cells, once sorted, can be cultured and profiled. Multiplex immunohistochemistry (IHC) can simultaneously
capture the expression levels of multiple proteins in formalinfixed paraffin-embedded (FFPE) tissue, with the advantage of
capturing their spatial organization and co-expression patterns,
although the number of proteins that can be differentially stained
on each tissue slide is limited.
In addition to these conventional approaches, recent computational methods have also advanced our understanding of TILs.
In a seminal study (Rooney et al., 2015), Rooney and colleagues
used Granzyme A and perforin expression levels to model the
immune cytolytic activities in tumors studied in TCGA, observing
increased cytolytic activities in tumors with higher mutation load,
copy number aberration, viral infection, and lower tumor stage.
This signature-gene based approach has been employed by
lu et al.,
two recent studies (Angelova et al., 2015; Sxenbabaog
2016) to estimate immune subset abundance based on a collection of pre-selected markers. CIBERSORT (Newman et al., 2015)
used an expert-selected signature of about 500 genes to infer
the abundance of 22 different tumor infiltrating immune components. In contrast, TIMER (Li et al., 2016b) selected cancer-specific signature genes to eliminate the bias from highly expressed
genes in cancer cells and deconvolved only six immune components to ensure that colinear expression between closely related
immune cells did not affect the deconvolution accuracy. These
studies confirmed previous observations (Bindea et al., 2013;
Rooney et al., 2015) and reported that CD8+ T cells are associated with better overall survival and fewer relapses, whereas
macrophages are associated with worse clinical outcome in
many cancer types (Li et al., 2016b).
There have been inconsistent observations on whether the
abundance of B cells is associated with improved cancer survival (DiLillo et al., 2010; Perricone et al., 2004; Qin et al., 1998;
Schultz et al., 1990). One potential reason is that B cells
with different activation statuses may either inhibit or promote
T cell functions (Nelson, 2010). Another possible reason is that
B cells are sometimes enriched at the margins of tumor capsules instead of evenly distributed throughout the tumor tissue
(Kroeger et al., 2016; Lao et al., 2016; Nelson, 2010; Shi et al.,
2013). Therefore, abundance estimates of B cells may be variable due to the specific tumor section under assay. By contrast,
TCR-seq of different sections of a large ovarian tumor (Emerson
et al., 2013) revealed that T cells are spatially homogeneous
within the tumor, similar to peripheral blood. Therefore, it is
possible that the correlation of TIL abundance with patient
outcome will depend on the homogeneity of TIL distribution for
different cancer types.
Applications
The culmination of our renewed understanding of the immune
system and its interaction potential with cancer cells has
been a decades-long effort to develop therapeutic approaches
that boost existing immune responses against neoplastic
cells. These efforts span widely variable approaches, and a
comprehensive review has been recently published that explores the broad landscape of cancer immunotherapies (Galluzzi
et al., 2014).
Certain types of cancer immunotherapies act to re-invigorate
existing immunity that has been suppressed in the tumor microenvironment. These so-called ‘‘checkpoint blockade’’ therapies
were devised to address our fundamental understanding of
immunosuppression and T cell exhaustion, and provide a relatively tumor-specific immune response. However, there often
are attendant side effects of variable severity, because their action targets native immune molecules such as CTLA-4, PD-1 and
PD-L1. Potentially, more specific targeting could result from using putative neoantigens predicted by NGS-based analysis,
described above, delivered as patient-specific vaccines meant
to stimulate an immune response that is highly specific for the
tumor cells. In this paradigm, several different vaccine types
(or ‘‘platforms’’) have emerged and are actively being tested in
pre-clinical and clinical settings, as follows (Hirayama and Nishimura, 2016; Overwijk et al., 2013; Vormehr et al., 2015; Zhang
et al., 2016c).
DNA minicassette vaccines: One vaccine platform is based
on piecing together the individual coding sequences for
each predicted neoantigen peptide into a DNA construct
that contains a specific human promoter element to drive
peptide production, once introduced into the patient. The
sequence-verified vaccine construct can be electroporated
into patient-derived dendritic cells and the DCs then reinfused into the patient. Synthetic DNA is relatively cheaply
and quickly obtained, even with the attendant GMP requirements for sequence verification prior to use in a human vaccine. Hence, concerns about cost and scalability of this
approach are minimal. One design consideration is ensuring
that no self-antigens are potentially encoded by the junctions
between each neoantigen sequence, but this is relatively
easy to confirm computationally once the proposed vaccine
design is in-hand.
Peptide vaccines: Synthetic peptides representing computationally identified neoantigens can be combined and solubilized in the presence of one or more immune-stimulatory
adjuvants to create patient-specific peptide vaccines.
These can be directly injected intramuscularly, intradermally,
or subcutaneously as a means of presenting the neoantigenic
peptides during maturation of native dendritic cells, which
then can prime a robust and specific immune response. Short
neoantigen peptides of 8–12 amino acids can directly bind
to HLAs expressed on the surface of antigen-presenting dendritic cells, thereby priming a T cell specific response. Peptide
vaccines also can be comprised of synthetic long peptides
(25–30 amino acids), which require uptake, processing, and
presentation by antigen-presenting cells in order to elicit an
immune response. While GMP-grade peptides are expensive
to manufacture, this is a scalable enterprise and, when
coupled with the simplicity of the peptide vaccine design, is
being applied in clinical trials of patient-specific vaccines
(W. Gillanders, personal communication).
RNA vaccines: Conceptually similar to DNA and peptide vaccines are RNA-based neoantigen vaccines, wherein the RNA
encodes the various predicted neoantigens that are unique to
each patient’s tumor. As with all RNA-based therapeutics, the
lability of RNA invokes a need to stabilize the RNA molecules
and to provide for appropriate uptake by antigen presenting
cells so the encoded peptides can be processed and presented. Cost and scalability of RNA synthesis are similarly
straightforward as for DNA, so the packaging and stabilization are the challenging puzzles for this platform, which is being actively pursued in the research setting.
Autologous dendritic cell vaccines: Dendritic cells (DCs)
occupy a unique position at the interface of innate and
adaptive immunity, and have been shown to effect a robust,
therapeutically relevant anti-neoplastic immune response.
In particular, autologous dendritic cells can be isolated
from patients and conditioned ex-vivo to mature, thereby
providing immune-stimulatory functions. When coupled
with neoantigenic peptides from patient-specific analyses,
the resulting dendritic cell vaccine can be re-infused and
has been shown to elicit neoantigen-specific T cell immunity
and an attendant expansion of the neoantigen-specific TCR
(Carreno et al., 2015; Galluzzi et al., 2014). Emphasizing their
specificity for tumor cells, no severe adverse events were recorded in this initial trial of patient-specific DC vaccines.
However, not all of the predicted neoantigens elicited a
T cell response, indicating that our ability to predict even
class I neoantigens will require additional precision, as discussed herein. While these early first-in-human results are
exciting, the preparation of dendritic cell vaccines requires
significant amounts of peripheral blood mononuclear cells
for dendritic cell isolation, as well as time- and effort-intensive
laboratory work to culture and mature the DCs ex vivo.
Hence, their scalability may be in question for broad-based
clinical use.
Besides cancer vaccines, another type of genomics-driven
patient-specific cancer immunotherapy is adoptive cell transfer
(ACT), as pioneered by Rosenberg and colleagues (Rosenberg
and Restifo, 2015). Basically, T cells extracted from a cancer patient, either from peripheral blood or the resected tumor, can be
activated and expanded ex vivo by IL2 treatment, before infusing
them back to the same patient to kill the cancer cells. Preparatory lymphodepletion either by chemotherapy or radiation of
Cell 168, February 9, 2017 607
the patient is an important step done prior to infusion, to improve
the engraftment and persistence of the adoptively transferred
T cells, thus increasing durability of tumor regression (Dudley
et al., 2002). ACT cells not only persist months after infusion,
but also expand in the patient. Two additional genomic approaches below have been shown to further enhance tumorspecific killing and broaden the applicable cancer types suitable
for ACT. Despite the cost and the technical and logistical challenges of ACT, this personalized immunotherapy has demonstrated promising rates and duration of response.
Genetically engineering T cells: T cells extracted from patients can be genetically engineered to express TCRs that
specifically recognize proteins expressed only in the patient’s
cancer cells, such as the melanoma/melanocyte specific
MART-1 antigen (Morgan et al., 2006) or the cancer-testis antigen NY-ESO-1 (Robbins et al., 2011). T cells also can be engineered by viral transduction to express a chimeric antigen
receptor (CAR) that uniquely recognizes the B cell specific
CD19 (Kalos et al., 2011; Kochenderfer et al., 2010) on the
cell surface. Linking the CAR with a co-stimulatory domain
such as CD137 (Imai et al., 2004; Milone et al., 2009) or engineering the cells to express another chimeric costimulatory
receptor recognizing a second antigen (Kloss et al., 2013)
have both improved T cell antitumor activity. Recently a
new clinical trial has been proposed, where CRISPR technology is applied to further engineer the NY-ESO-1-targeting
CAR T cells. Using a small number of CRISPR guide RNAs
to knock out the PD-1 gene and the cells’ intrinsic TCR,
this approach aims to eliminate immune suppression and
improve the NY-ESO-1 receptor response. If proven effective, genome engineering technology could provide new opportunities to manipulate other genes in immune cells ex vivo
using the CRISPR technology to achieve desired cancer
killing phenotypes.
Expanding tumor-specific T cells: Instead of engineering the
autologous T cells ex vivo, this approach separately cultures
tumor-infiltrating T cell clones or subpopulations, then
selects those reacting against tumor cells for massive
expansion before patient infusion. With the emergence of
exome-sequencing, scientists can call somatic mutations
from the tumor and computationally predict immunogenic
neoantigens. Testing the immunogenicity of these mutations
in parallel uses minigene constructs (described above) encoding the mutated peptides into expression vectors and
the in vitro transcribed RNA from the vectors can be electroporated into antigen presenting cells (APC). Culturing the
tumor-infiltrating T cells for reactivity against these APCs
selects the tumor-specific T cells and identifies the immunogenic mutant minigenes (Robbins et al., 2013). Compared to
genetically engineered T cells, the final T cells infused into
patients using this approach are comprised of populations
of different T cells recognizing different neoantigens. While
most of these neoantigens are hypothesized to be passenger
mutations, recently the Rosenberg group identified four
T cell clones that specifically react to the KRAS G12D mutation in colon cancer (Tran et al., 2016), thereby drugging the
undruggable.
608 Cell 168, February 9, 2017
Future Perspectives
As high-throughput technologies improve and our immunology
knowledge grows, the future of immunogenomics-based application to cancer appears quite promising and likely will continue
to broaden. Technological and computational innovations will be
instrumental to overcome existing challenges and move the field
forward. First, despite the advances offered by algorithms such
as NetMHC-pan, both the accuracy of MHC presentation prediction, especially for rarer alleles, and of MHC class II presentation
await improvements. In addition, most studies use MHC presentation of somatic mutations as a proxy to predict immunogenicity, although it is unclear which presented somatic mutations
will elicit immune responses. Experimental assays such as
EliSpot are currently used to validate the predicted neoantigens,
although such assays are still conducted in a low throughput
fashion (Cole, 2005).
Second, TIL deconvolution methods such as CIBERSORT
and TIMER use reference expression profiling data on sorted
immune components from peripheral blood. These methods
could be combined with Nanostring-based measures of immune
marker genes in addition to bulk tissue RNA-seq data for inexpensive profiling of large archival tumor cohorts. However,
expression of immune cells in tumors might differ significantly
from that in peripheral blood, which could influence the accuracy
of these inference methods. Recent developments in single
cell analyses techniques, such as CyTOF (Newell et al., 2012)
and single-cell RNA-seq (Klein et al., 2015; Macosko et al.,
2015; Tirosh et al., 2016), might offer more quantitative alternatives. However, for very detailed TIL deconvolution on large
sample cohorts, the required starting tumor material and cost
of single cell experiments need to decrease significantly for
widespread use.
Third, monitoring an individual patient’s immune repertoire in
peripheral blood or tumors provides insights into their immune
health as well as their response to allergens, vaccines or therapies (Robins, 2013). However, there are still many challenges
ahead, such as how to identify the specific TCR / BCR that
recognizes each specific somatic mutation and how accurate
the immune repertoire is at predicting patient response to immunotherapy. Other challenges include how to robustly estimate
the total immune repertoire in different samples from the same
individual, normalize bias from minor immune events, and distinguish immune repertoire signals from normal versus pathogenic
immune events.
Last but not least, predicting response to immunotherapies,
including tumor killing effects and autoimmune side effects, is still
an open question. So far, higher T cell infiltration (Taube et al.,
2012; Tumeh et al., 2014), higher PD-1 or PD-L1/L2 expression
(Garon et al., 2015; Herbst et al., 2014; Taube et al., 2012), higher
neoantigen load from BRCA or somatic mutations in DNA repair
pathway genes (Hugo et al., 2016; Snyder et al., 2014; Van Allen
et al., 2015), or microsatellite instability (Le et al., 2015), higher peripheral baseline TCR diversity (Postow et al., 2015), lower tumor
infiltrating TCR diversity (Tumeh et al., 2014), lack of mutations in
interferon gamma (INFG) (Gao et al., 2016), beta-2-microglobulin
(B2M) (Zaretsky et al., 2016), or JAK1/JAK2 (Zaretsky et al., 2016)
have been associated with better response to immunotherapies
in various cancer types. A comprehensive model that integrates
all of these factors to accurately predict patient response to
immunotherapy is still lacking, and likely will require much more
data to train and refine. In addition, methods to predict the
optimal combination of immunotherapies or with other targeted,
chemo, or radiation therapies for individual patients still await
development. Despite all the aforementioned challenges, the
exciting results obtained to-date from cancer immunotherapies
will continue to motivate the biomedical research community to
overcome these challenges and explore new frontiers.
ACKNOWLEDGEMENTS
This work was partially supported by the National Institutes of Health (1U01
CA180980 to X.S.L.).
breadth and diversity of melanoma neoantigen-specific T cells. Science 348,
803–808.
Carter, S.L., Cibulskis, K., Helman, E., McKenna, A., Shen, H., Zack, T., Laird,
P.W., Onofrio, R.C., Winckler, W., Weir, B.A., et al. (2012). Absolute quantification of somatic DNA alterations in human cancer. Nat. Biotechnol. 30,
413–421.
Cerottini, J.C., Engers, H.D., Macdonald, H.R., and Brunner, T. (1974). Generation of cytotoxic T lymphocytes in vitro. I. Response of normal and immune
mouse spleen cells in mixed leukocyte cultures. J. Exp. Med. 140, 703–717.
Cole, G.A. (2005). Interferon-g ELISPOT Assay for the Quantitative Measurement of Antigen-Specific Murine CD8 T-Cells. In Handbook of ELISPOT
(Springer), pp. 191–204.
Cole, C., Volden, R., Dharmadhikari, S., Scelfo-Dalbey, C., and Vollmers, C.
(2016). Highly Accurate Sequencing of Full-Length Immune Repertoire Amplicons Using Tn5-Enabled and Molecular Identifier-Guided Amplicon Assembly.
J. Immunol. 196, 2902–2907.
REFERENCES
Cornish, A., and Guda, C. (2015). A Comparison of Variant Calling Pipelines
Using Genome in a Bottle as a Reference. BioMed Res. Int. 2015, 456479.
Alamyar, E., Duroux, P., Lefranc, M.-P., and Giudicelli, V. (2012). IMGT()
tools for the nucleotide analysis of immunoglobulin (IG) and T cell receptor
(TR) V-(D)-J repertoires, polymorphisms, and IG mutations: IMGT/VQUEST and IMGT/HighV-QUEST for NGS. Methods Mol. Biol. 882,
569–604.
De Plaen, E., Lurquin, C., Van Pel, A., Mariamé, B., Szikora, J.P., Wölfel, T., Sibille, C., Chomez, P., and Boon, T. (1988). Immunogenic (tum-) variants of
mouse tumor P815: cloning of the gene of tum- antigen P91A and identification
of the tum- mutation. Proc. Natl. Acad. Sci. USA 85, 2274–2278.
Angelova, M., Charoentong, P., Hackl, H., Fischer, M.L., Snajder, R., Krogsdam, A.M., Waldner, M.J., Bindea, G., Mlecnik, B., Galon, J., and Trajanoski,
Z. (2015). Characterization of the immunophenotypes and antigenomes of
colorectal cancers reveals distinct tumor escape mechanisms and novel targets for immunotherapy. Genome Biol. 16, 64.
Aouinti, S., Malouche, D., Giudicelli, V., Kossida, S., and Lefranc, M.-P. (2015).
IMGT/HighV-QUEST Statistical Significance of IMGT Clonotype (AA) Diversity
per Gene for Standardized Comparisons of Next Generation Sequencing Immunoprofiles of Immunoglobulins and T Cell Receptors. PLoS ONE 10,
e0142353.
Babbitt, B.P., Allen, P.M., Matsueda, G., Haber, E., and Unanue, E.R. (1985).
Binding of immunogenic peptides to Ia histocompatibility molecules. Nature
317, 359–361.
Bagaev, D.V., Zvyagin, I.V., Putintseva, E.V., Izraelson, M., Britanova, O.V.,
Chudakov, D.M., and Shugay, M. (2016). VDJviz: a versatile browser for immunogenomics data. BMC Genomics 17, 453.
Bainbridge, M.N., Wang, M., Burgess, D.L., Kovar, C., Rodesch, M.J., D’Ascenzo, M., Kitzman, J., Wu, Y.-Q., Newsham, I., Richmond, T.A., et al.
(2010). Whole exome capture in solution with 3 Gbp of data. Genome Biol.
11, R62.
Bassani-Sternberg, M., Bräunlein, E., Klar, R., Engleitner, T., Sinitcyn, P., Audehm, S., Straub, M., Weber, J., Slotta-Huspenina, J., Specht, K., et al. (2016).
Direct identification of clinically relevant neoepitopes presented on native human melanoma tissue by mass spectrometry. Nat. Commun. 7, 13404.
Bindea, G., Mlecnik, B., Tosolini, M., Kirilovsky, A., Waldner, M., Obenauf,
A.C., Angell, H., Fredriksen, T., Lafontaine, L., Berger, A., et al. (2013). Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape
in human cancer. Immunity 39, 782–795.
Bjorkman, P.J., Saper, M.A., Samraoui, B., Bennett, W.S., Strominger, J.L.,
and Wiley, D.C. (1987). Structure of the human class I histocompatibility antigen, HLA-A2. Nature 329, 506–512.
Bolotin, D.A., Poslavsky, S., Mitrophanov, I., Shugay, M., Mamedov, I.Z., Putintseva, E.V., and Chudakov, D.M. (2015). MiXCR: software for comprehensive adaptive immunity profiling. Nat. Methods 12, 380–381.
DiLillo, D.J., Yanaba, K., and Tedder, T.F. (2010). B cells are required for
optimal CD4+ and CD8+ T cell tumor immunity: therapeutic B cell depletion
enhances B16 melanoma growth in mice. J. Immunol. 184, 4006–4016.
Duan, F., Duitama, J., Al Seesi, S., Ayres, C.M., Corcelli, S.A., Pawashe, A.P.,
Blanchard, T., McMahon, D., Sidney, J., Sette, A., et al. (2014). Genomic and
bioinformatic profiling of mutational neoepitopes reveals new rules to predict
anticancer immunogenicity. J. Exp. Med. 211, 2231–2248.
Dubey, P., Hendrickson, R.C., Meredith, S.C., Siegel, C.T., Shabanowitz, J.,
Skipper, J.C., Engelhard, V.H., Hunt, D.F., and Schreiber, H. (1997). The immunodominant antigen of an ultraviolet-induced regressor tumor is generated by
a somatic point mutation in the DEAD box helicase p68. J. Exp. Med. 185,
695–705.
Dudley, M.E., Wunderlich, J.R., Robbins, P.F., Yang, J.C., Hwu, P., Schwartzentruber, D.J., Topalian, S.L., Sherry, R., Restifo, N.P., Hubicki, A.M., et al.
(2002). Cancer regression and autoimmunity in patients after clonal repopulation with antitumor lymphocytes. Science 298, 850–854.
Emerson, R.O., Sherwood, A.M., Rieder, M.J., Guenthoer, J., Williamson,
D.W., Carlson, C.S., Drescher, C.W., Tewari, M., Bielas, J.H., and Robins,
H.S. (2013). High-throughput sequencing of T-cell receptors reveals a homogeneous repertoire of tumour-infiltrating lymphocytes in ovarian cancer.
J. Pathol. 231, 433–440.
Foley, E.J. (1953). Antigenic properties of methylcholanthrene-induced tumors
in mice of the strain of origin. Cancer Res. 13, 835–837.
Freeman, J.D., Warren, R.L., Webb, J.R., Nelson, B.H., and Holt, R.A. (2009).
Profiling the T-cell receptor beta-chain repertoire by massively parallel
sequencing. Genome Res. 19, 1817–1824.
Galluzzi, L., Vacchelli, E., Bravo-San Pedro, J.-M., Buqué, A., Senovilla, L.,
Baracco, E.E., Bloy, N., Castoldi, F., Abastado, J.-P., Agostinis, P., et al.
(2014). Classification of current anticancer immunotherapies. Oncotarget 5,
12472–12508.
Gao, J., Shi, L.Z., Zhao, H., Chen, J., Xiong, L., He, Q., Chen, T., Roszik, J.,
Bernatchez, C., Woodman, S.E., et al. (2016). Loss of IFN-g Pathway Genes
in Tumor Cells as a Mechanism of Resistance to Anti-CTLA-4 Therapy. Cell
167, 397–404.e9.
Calis, J.J.A., and Rosenberg, B.R. (2014). Characterizing immune repertoires
by high throughput sequencing: strategies and applications. Trends Immunol.
35, 581–590.
Garon, E.B., Rizvi, N.A., Hui, R., Leighl, N., Balmanoukian, A.S., Eder, J.P., Patnaik, A., Aggarwal, C., Gubens, M., Horn, L., et al.; KEYNOTE-001 Investigators (2015). Pembrolizumab for the treatment of non-small-cell lung cancer.
N. Engl. J. Med. 372, 2018–2028.
Carreno, B.M., Magrini, V., Becker-Hapak, M., Kaabinejadian, S., Hundal, J.,
Petti, A.A., Ly, A., Lie, W.-R., Hildebrand, W.H., Mardis, E.R., and Linette,
G.P. (2015). Cancer immunotherapy. A dendritic cell vaccine increases the
Ghoneim, D.H., Myers, J.R., Tuttle, E., and Paciorkowski, A.R. (2014). Comparison of insertion/deletion calling algorithms on human next-generation
sequencing data. BMC Res. Notes 7, 864.
Cell 168, February 9, 2017 609
Gillis, S., and Smith, K.A. (1977). Long term culture of tumour-specific cytotoxic T cells. Nature 268, 154–156.
Giudicelli, V., Chaume, D., Bodmer, J., Müller, W., Busin, C., Marsh, S., Bontrop, R., Marc, L., Malik, A., and Lefranc, M.P. (1997). IMGT, the international
ImMunoGeneTics database. Nucleic Acids Res. 25, 206–211.
Gnirke, A., Melnikov, A., Maguire, J., Rogov, P., LeProust, E.M., Brockman,
W., Fennell, T., Giannoukos, G., Fisher, S., Russ, C., et al. (2009). Solution
hybrid selection with ultra-long oligonucleotides for massively parallel targeted
sequencing. Nat. Biotechnol. 27, 182–189.
Greiff, V., Miho, E., Menzel, U., and Reddy, S.T. (2015a). Bioinformatic and Statistical Analysis of Adaptive Immune Repertoires. Trends Immunol. 36,
738–749.
Greiff, V., Bhat, P., Cook, S.C., Menzel, U., Kang, W., and Reddy, S.T. (2015b).
A bioinformatic framework for immune repertoire diversity profiling enables
detection of immunological status. Genome Med. 7, 49.
Gubin, M.M., Zhang, X., Schuster, H., Caron, E., Ward, J.P., Noguchi, T., Ivanova, Y., Hundal, J., Arthur, C.D., Krebber, W.-J., et al. (2014). Checkpoint
blockade cancer immunotherapy targets tumour-specific mutant antigens.
Nature 515, 577–581.
Gubin, M.M., Artyomov, M.N., Mardis, E.R., and Schreiber, R.D. (2015). Tumor
neoantigens: building a framework for personalized cancer immunotherapy.
J. Clin. Invest. 125, 3413–3421.
Herbst, R.S., Soria, J.-C., Kowanetz, M., Fine, G.D., Hamid, O., Gordon, M.S.,
Sosman, J.A., McDermott, D.F., Powderly, J.D., Gettinger, S.N., et al. (2014).
Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in
cancer patients. Nature 515, 563–567.
Hirayama, M., and Nishimura, Y. (2016). The present status and future prospects of peptide-based cancer vaccines. Int. Immunol. 28, 319–328.
Hodges, E., Rooks, M., Xuan, Z., Bhattacharjee, A., Benjamin Gordon, D., Brizuela, L., Richard McCombie, W., and Hannon, G.J. (2009). Hybrid selection of
discrete genomic intervals on custom-designed microarrays for massively
parallel sequencing. Nat. Protoc. 4, 960–974.
Hou, X.-L., Wang, L., Ding, Y.-L., Xie, Q., and Diao, H.-Y. (2016). Current status
and recent advances of next generation sequencing techniques in immunological repertoire. Genes Immun. 17, 153–164.
Hugo, W., Zaretsky, J.M., Sun, L., Song, C., Moreno, B.H., Hu-Lieskovan, S.,
Berent-Maoz, B., Pang, J., Chmielowski, B., Cherry, G., et al. (2016). Genomic
and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic
Melanoma. Cell 165, 35–44.
Imai, C., Mihara, K., Andreansky, M., Nicholson, I.C., Pui, C.-H., Geiger, T.L.,
and Campana, D. (2004). Chimeric receptors with 4-1BB signaling capacity
provoke potent cytotoxicity against acute lymphoblastic leukemia. Leukemia
18, 676–684.
Jiang, Y., Wang, Y., and Brudno, M. (2012). PRISM: pair-read informed splitread mapping for base-pair level detection of insertion, deletion and structural
variants. Bioinformatics 28, 2576–2583.
Jiang, Y., Turinsky, A.L., and Brudno, M. (2015). The missing indels: an estimate of indel variation in a human genome and analysis of factors that impede
detection. Nucleic Acids Res. 43, 7217–7228.
Johnson, P.L.F., Goronzy, J.J., and Antia, R. (2014). A population biological
approach to understanding the maintenance and loss of the T-cell repertoire
during aging. Immunology 142, 167–175.
Kalos, M., Levine, B.L., Porter, D.L., Katz, S., Grupp, S.A., Bagg, A., and June,
C.H. (2011). T cells with chimeric antigen receptors have potent antitumor effects and can establish memory in patients with advanced leukemia. Sci.
Transl. Med. 3, 95ra73.
Klein, A.M., Mazutis, L., Akartuna, I., Tallapragada, N., Veres, A., Li, V., Peshkin, L., Weitz, D.A., and Kirschner, M.W. (2015). Droplet barcoding for singlecell transcriptomics applied to embryonic stem cells. Cell 161, 1187–1201.
Kloss, C.C., Condomines, M., Cartellieri, M., Bachmann, M., and Sadelain, M.
(2013). Combinatorial antigen recognition with balanced signaling promotes
selective tumor eradication by engineered T cells. Nat. Biotechnol. 31, 71–75.
610 Cell 168, February 9, 2017
Knuth, A., Danowski, B., Oettgen, H.F., and Old, L.J. (1984). T-cell-mediated
cytotoxicity against autologous malignant melanoma: analysis with interleukin
2-dependent T-cell cultures. Proc. Natl. Acad. Sci. USA 81, 3511–3515.
Kochenderfer, J.N., Wilson, W.H., Janik, J.E., Dudley, M.E., Stetler-Stevenson, M., Feldman, S.A., Maric, I., Raffeld, M., Nathan, D.-A.N., Lanier, B.J.,
et al. (2010). Eradication of B-lineage cells and regression of lymphoma in a patient treated with autologous T cells genetically engineered to recognize CD19.
Blood 116, 4099–4102.
Kreiter, S., Vormehr, M., van de Roemer, N., Diken, M., Löwer, M., Diekmann,
J., Boegel, S., Schrörs, B., Vascotto, F., Castle, J.C., et al. (2015). Mutant MHC
class II epitopes drive therapeutic immune responses to cancer. Nature 520,
692–696.
Kroeger, D.R., Milne, K., and Nelson, B.H. (2016). Tumor-Infiltrating Plasma
Cells Are Associated with Tertiary Lymphoid Structures, Cytolytic T-Cell Responses, and Superior Prognosis in Ovarian Cancer. Clin. Cancer Res. 22,
3005–3015.
Krøigård, A.B., Thomassen, M., Lænkholm, A.-V., Kruse, T.A., and Larsen,
M.J. (2016). Evaluation of Nine Somatic Variant Callers for Detection of Somatic Mutations in Exome and Targeted Deep Sequencing Data. PLoS ONE
11, e0151664.
Kumar, S., Razzaq, S.K., Vo, A.D., Gautam, M., and Li, H. (2016). Identifying
fusion transcripts using next generation sequencing. Wiley Interdiscip. Rev.
RNA 7, 811–823.
Lao, X.-M., Liang, Y.-J., Su, Y.-X., Zhang, S.-E., Zhou, X.I., and Liao, G.-Q.
(2016). Distribution and significance of interstitial fibrosis and stroma-infiltrating B cells in tongue squamous cell carcinoma. Oncol. Lett. 11, 2027–2034.
Le, D.T., Uram, J.N., Wang, H., Bartlett, B.R., Kemberling, H., Eyring, A.D.,
Skora, A.D., Luber, B.S., Azad, N.S., Laheru, D., et al. (2015). PD-1 Blockade
in Tumors with Mismatch-Repair Deficiency. N. Engl. J. Med. 372, 2509–2520.
Ley, T.J., Mardis, E.R., Ding, L., Fulton, B., McLellan, M.D., Chen, K., Dooling,
D., Dunford-Shore, B.H., McGrath, S., Hickenbotham, M., et al. (2008). DNA
sequencing of a cytogenetically normal acute myeloid leukaemia genome. Nature 456, 66–72.
Li, B., and Li, J.Z. (2014). A general framework for analyzing tumor subclonality
using SNP array and DNA sequencing data. Genome Biol. 15, 473.
Li, Y., Chien, J., Smith, D.I., and Ma, J. (2011). FusionHunter: identifying fusion
transcripts in cancer using paired-end RNA-seq. Bioinformatics 27,
1708–1710.
Li, B., Li, T., Pignon, J.-C., Wang, B., Wang, J., Shukla, S.A., Dou, R., Chen, Q.,
Hodi, F.S., Choueiri, T.K., et al. (2016a). Landscape of tumor-infiltrating T cell
repertoire of human cancers. Nat. Genet. 48, 725–732.
Li, B., Severson, E., Pignon, J.-C., Zhao, H., Li, T., Novak, J., Jiang, P., Shen,
H., Aster, J.C., Rodig, S., et al. (2016b). Comprehensive analyses of tumor immunity: implications for cancer immunotherapy. Genome Biol. 17, 174.
Lundegaard, C., Lund, O., and Nielsen, M. (2008a). Accurate approximation
method for prediction of class I MHC affinities for peptides of length 8, 10
and 11 using prediction tools trained on 9mers. Bioinformatics 24, 1397–1398.
Lundegaard, C., Lamberth, K., Harndahl, M., Buus, S., Lund, O., and Nielsen,
M. (2008b). NetMHC-3.0: accurate web accessible predictions of human,
mouse and monkey MHC class I affinities for peptides of length 8-11. Nucleic
Acids Res. 36, W509–W512.
McNeel, D.G. (2016). TCR diversity - a universal cancer immunotherapy
biomarker? J. Immunother. Cancer 4, 69.
Macosko, E.Z., Basu, A., Satija, R., Nemesh, J., Shekhar, K., Goldman, M., Tirosh, I., Bialas, A.R., Kamitaki, N., Martersteck, E.M., et al. (2015). Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter
Droplets. Cell 161, 1202–1214.
Mardis, E.R. (2017). DNA sequencing technologies: 2006-2016. Nat. Protoc.
12, 213–218.
McKinney, D.M., Southwood, S., Hinz, D., Oseroff, C., Arlehamn, C.S.L.,
Schulten, V., Taplitz, R., Broide, D., Hanekom, W.A., Scriba, T.J., et al.
(2013). A strategy to determine HLA class II restriction broadly covering the
DR, DP, and DQ allelic variants most commonly expressed in the general population. Immunogenetics 65, 357–370.
Prehn, R.T., and Main, J.M. (1957). Immunity to methylcholanthrene-induced
sarcomas. J. Natl. Cancer Inst. 18, 769–778.
Milone, M.C., Fish, J.D., Carpenito, C., Carroll, R.G., Binder, G.K., Teachey, D.,
Samanta, M., Lakhal, M., Gloss, B., Danet-Desnoyers, G., et al. (2009).
Chimeric receptors containing CD137 signal transduction domains mediate
enhanced survival of T cells and increased antileukemic efficacy in vivo.
Mol. Ther. 17, 1453–1464.
Qin, Z., Richter, G., Schüler, T., Ibe, S., Cao, X., and Blankenstein, T. (1998). B
cells inhibit induction of T cell-dependent tumor immunity. Nat. Med. 4,
627–630.
Monach, P.A., Meredith, S.C., Siegel, C.T., and Schreiber, H. (1995). A unique
tumor antigen produced by a single amino acid substitution. Immunity 2,
45–59.
Morgan, R.A., Dudley, M.E., Wunderlich, J.R., Hughes, M.S., Yang, J.C.,
Sherry, R.M., Royal, R.E., Topalian, S.L., Kammula, U.S., Restifo, N.P., et al.
(2006). Cancer regression in patients after transfer of genetically engineered
lymphocytes. Science 314, 126–129.
Mose, L.E., Wilkerson, M.D., Hayes, D.N., Perou, C.M., and Parker, J.S. (2014).
ABRA: improved coding indel detection via assembly-based realignment. Bioinformatics 30, 2813–2815.
Mose, L.E., Selitsky, S.R., Bixby, L.M., Marron, D.L., Iglesia, M.D., Serody,
J.S., Perou, C.M., Vincent, B.G., and Parker, J.S. (2016). Assembly-based
inference of B-cell receptor repertoires from short read RNA sequencing
data with V’DJer. Bioinformatics 32, 3729–3734.
Muraro, P.A., Robins, H., Malhotra, S., Howell, M., Phippard, D., Desmarais,
C., de Paula Alves Sousa, A., Griffith, L.M., Lim, N., Nash, R.A., and Turka,
L.A. (2014). T cell repertoire following autologous stem cell transplantation
for multiple sclerosis. J. Clin. Invest. 124, 1168–1172.
Narzisi, G., O’Rawe, J.A., Iossifov, I., Fang, H., Lee, Y.-H., Wang, Z., Wu, Y.,
Lyon, G.J., Wigler, M., and Schatz, M.C. (2014). Accurate de novo and transmitted indel detection in exome-capture data using microassembly. Nat.
Methods 11, 1033–1036.
Nelson, B.H. (2010). CD20+ B cells: the other tumor-infiltrating lymphocytes.
J. Immunol. 185, 4977–4982.
Newell, E.W., Sigal, N., Bendall, S.C., Nolan, G.P., and Davis, M.M. (2012). Cytometry by time-of-flight shows combinatorial cytokine expression and virusspecific cell niches within a continuum of CD8+ T cell phenotypes. Immunity
36, 142–152.
Newman, A.M., Liu, C.L., Green, M.R., Gentles, A.J., Feng, W., Xu, Y., Hoang,
C.D., Diehn, M., and Alizadeh, A.A. (2015). Robust enumeration of cell subsets
from tissue expression profiles. Nat. Methods 12, 453–457.
Rasmussen, M., Fenoy, E., Harndahl, M., Kristensen, A.B., Nielsen, I.K., Nielsen, M., and Buus, S. (2016). Pan-Specific Prediction of Peptide-MHC Class I
Complex Stability, a Correlate of T Cell Immunogenicity. J. Immunol. 197,
1517–1524.
Ratan, A., Olson, T.L., Loughran, T.P., Jr., and Miller, W. (2015). Identification
of indels in next-generation sequencing data. BMC Bioinformatics 16, 42.
Rizvi, N.A., Hellmann, M.D., Snyder, A., Kvistborg, P., Makarov, V., Havel, J.J.,
Lee, W., Yuan, J., Wong, P., Ho, T.S., et al. (2015). Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung
cancer. Science 348, 124–128.
Robbins, P.F., El-Gamil, M., Li, Y.F., Kawakami, Y., Loftus, D., Appella, E., and
Rosenberg, S.A. (1996). A mutated beta-catenin gene encodes a melanomaspecific antigen recognized by tumor infiltrating lymphocytes. J. Exp. Med.
183, 1185–1192.
Robbins, P.F., Morgan, R.A., Feldman, S.A., Yang, J.C., Sherry, R.M., Dudley,
M.E., Wunderlich, J.R., Nahvi, A.V., Helman, L.J., Mackall, C.L., et al. (2011).
Tumor regression in patients with metastatic synovial cell sarcoma and melanoma using genetically engineered lymphocytes reactive with NY-ESO-1.
J. Clin. Oncol. 29, 917–924.
Robbins, P.F., Lu, Y.-C., El-Gamil, M., Li, Y.F., Gross, C., Gartner, J., Lin, J.C.,
Teer, J.K., Cliften, P., Tycksen, E., et al. (2013). Mining exomic sequencing
data to identify mutated antigens recognized by adoptively transferred tumor-reactive T cells. Nat. Med. 19, 747–752.
Robert, L., Tsoi, J., Wang, X., Emerson, R., Homet, B., Chodon, T., Mok, S.,
Huang, R.R., Cochran, A.J., Comin-Anduix, B., et al. (2014). CTLA4 blockade
broadens the peripheral T-cell receptor repertoire. Clin. Cancer Res. 20,
2424–2432.
Robins, H. (2013). Immunosequencing: applications of immune repertoire
deep sequencing. Curr. Opin. Immunol. 25, 646–652.
Robinson, J., Waller, M.J., Parham, P., Bodmer, J.G., and Marsh, S.G. (2001).
IMGT/HLA Database–a sequence database for the human major histocompatibility complex. Nucleic Acids Res. 29, 210–213.
Nielsen, M., Lundegaard, C., Worning, P., Lauemøller, S.L., Lamberth, K.,
Buus, S., Brunak, S., and Lund, O. (2003). Reliable prediction of T-cell epitopes
using neural networks with novel sequence representations. Protein Sci. 12,
1007–1017.
Robinson, J., Waller, M.J., Parham, P., de Groot, N., Bontrop, R., Kennedy,
L.J., Stoehr, P., and Marsh, S.G. (2003). IMGT/HLA and IMGT/MHC: sequence
databases for the study of the major histocompatibility complex. Nucleic Acids
Res. 31, 311–314.
Old, L.J., and Boyse, E.A. (1964). IMMUNOLOGY OF EXPERIMENTAL
TUMORS. Annu. Rev. Med. 15, 167–186.
Robinson, J., Halliwell, J.A., McWilliam, H., Lopez, R., Parham, P., and Marsh,
S.G.E. (2013). The IMGT/HLA database. Nucleic Acids Res. 41, D1222–D1227.
Overwijk, W.W., Wang, E., Marincola, F.M., Rammensee, H.-G., and Restifo,
N.P. (2013). Mining the mutanome: developing highly personalized Immunotherapies based on mutational analysis of tumors. J. Immunother. Cancer
1, 11.
Rooney, M.S., Shukla, S.A., Wu, C.J., Getz, G., and Hacohen, N. (2015). Molecular and genetic properties of tumors associated with local immune cytolytic activity. Cell 160, 48–61.
Pasternack, M.S. (1994). Review of Fundamental Immunology Edited by William E. Paul, 3rd edition (New York: Raven Press), 1993. 1,490 pp., illustrated.
$95. Clin. Infect. Dis. 19, 996–997.
Perricone, M.A., Smith, K.A., Claussen, K.A., Plog, M.S., Hempel, D.M., Roberts, B.L., St George, J.A., and Kaplan, J.M. (2004). Enhanced efficacy of melanoma vaccines in the absence of B lymphocytes. J. Immunother. 27,
273–281.
Rosenberg, S.A., and Restifo, N.P. (2015). Adoptive cell transfer as personalized immunotherapy for human cancer. Science 348, 62–68.
Saethang, T., Hirose, O., Kimkong, I., Tran, V.A., Dang, X.T., Nguyen, L.A.T.,
Le, T.K.T., Kubo, M., Yamada, Y., and Satou, K. (2013). PAAQD: Predicting
immunogenicity of MHC class I binding peptides using amino acid pairwise
contact potentials and quantum topological molecular similarity descriptors.
J. Immunol. Methods 387, 293–302.
Peters, B., and Sette, A. (2005). Generating quantitative models describing the
sequence specificity of biological processes with the stabilized matrix method.
BMC Bioinformatics 6, 132.
Sahin, U., Türeci, O., Schmitt, H., Cochlovius, B., Johannes, T., Schmits, R.,
Stenner, F., Luo, G., Schobert, I., and Pfreundschuh, M. (1995). Human neoplasms elicit multiple specific immune responses in the autologous host.
Proc. Natl. Acad. Sci. USA 92, 11810–11813.
Postow, M.A., Manuel, M., Wong, P., Yuan, J., Dong, Z., Liu, C., Perez, S., Tanneau, I., Noel, M., Courtier, A., et al. (2015). Peripheral T cell receptor diversity
is associated with clinical outcomes following ipilimumab treatment in metastatic melanoma. J. Immunother. Cancer 3, 23.
Saul, L., Ilieva, K.M., Bax, H.J., Karagiannis, P., Correa, I., Rodriguez-Hernandez, I., Josephs, D.H., Tosi, I., Egbuniwe, I.U., Lombardi, S., et al. (2016). IgG
subclass switching and clonal expansion in cutaneous melanoma and normal
skin. Sci. Rep. 6, 29736.
Cell 168, February 9, 2017 611
Schultz, K.R., Klarnet, J.P., Gieni, R.S., HayGlass, K.T., and Greenberg, P.D.
(1990). The role of B cells for in vivo T cell responses to a Friend virus-induced
leukemia. Science 249, 921–923.
Schumacher, T., Bunse, L., Pusch, S., Sahm, F., Wiestler, B., Quandt, J.,
Menn, O., Osswald, M., Oezen, I., Ott, M., et al. (2014). A vaccine targeting
mutant IDH1 induces antitumour immunity. Nature 512, 324–327.
Scolnick, J.A., Dimon, M., Wang, I.-C., Huelga, S.C., and Amorese, D.A.
(2015). An Efficient Method for Identifying Gene Fusions by Targeted RNA
Sequencing from Fresh Frozen and FFPE Samples. PLoS ONE 10, e0128916.
lu, Y., Gejman, R.S., Winer, A.G., Liu, M., Van Allen, E.M., de VeSxenbabaog
lasco, G., Miao, D., Ostrovnaya, I., Drill, E., Luna, A., et al. (2016). Tumor immune microenvironment characterization in clear cell renal cell carcinoma
identifies prognostic and immunotherapeutically relevant messenger RNA signatures. Genome Biol. 17, 231.
Shi, J.-Y., Gao, Q., Wang, Z.-C., Zhou, J., Wang, X.-Y., Min, Z.-H., Shi, Y.-H.,
Shi, G.-M., Ding, Z.-B., Ke, A.-W., et al. (2013). Margin-infiltrating CD20(+)
B cells display an atypical memory phenotype and correlate with favorable
prognosis in hepatocellular carcinoma. Clin. Cancer Res. 19, 5994–6005.
Shugay, M., Britanova, O.V., Merzlyak, E.M., Turchaninova, M.A., Mamedov,
I.Z., Tuganbaev, T.R., Bolotin, D.A., Staroverov, D.B., Putintseva, E.V., Plevova, K., et al. (2014). Towards error-free profiling of immune repertoires.
Nat. Methods 11, 653–655.
Shugay, M., Bagaev, D.V., Turchaninova, M.A., Bolotin, D.A., Britanova, O.V.,
Putintseva, E.V., Pogorelyy, M.V., Nazarov, V.I., Zvyagin, I.V., Kirgizova, V.I.,
et al. (2015). VDJtools: Unifying Post-analysis of T Cell Receptor Repertoires.
PLoS Comput. Biol. 11, e1004503.
Shukla, S.A., Rooney, M.S., Rajasagi, M., Tiao, G., Dixon, P.M., Lawrence,
M.S., Stevens, J., Lane, W.J., Dellagatta, J.L., Steelman, S., et al. (2015).
Comprehensive analysis of cancer-associated somatic mutations in class I
HLA genes. Nat. Biotechnol. 33, 1152–1158.
Snyder, A., Makarov, V., Merghoub, T., Yuan, J., Zaretsky, J.M., Desrichard,
A., Walsh, L.A., Postow, M.A., Wong, P., Ho, T.S., et al. (2014). Genetic basis
for clinical response to CTLA-4 blockade in melanoma. N. Engl. J. Med. 371,
2189–2199.
Srivastava, A., Ghosh, S., Anantharaman, N., and Jayaraman, V.K. (2013).
Hybrid biogeography based simultaneous feature selection and MHC class I
peptide binding prediction using support vector machines and random forests.
J. Immunol. Methods 387, 284–292.
Szolek, A., Schubert, B., Mohr, C., Sturm, M., Feldhahn, M., and Kohlbacher,
O. (2014). OptiType: precision HLA typing from next-generation sequencing
data. Bioinformatics 30, 3310–3316.
Taube, J.M., Anders, R.A., Young, G.D., Xu, H., Sharma, R., McMiller, T.L.,
Chen, S., Klein, A.P., Pardoll, D.M., Topalian, S.L., and Chen, L. (2012). Colocalization of inflammatory response with B7-h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape.
Sci. Transl. Med. 4, 127ra37.
Tirosh, I., Izar, B., Prakadan, S.M., Wadsworth, M.H., 2nd, Treacy, D., Trombetta, J.J., Rotem, A., Rodman, C., Lian, C., Murphy, G., et al. (2016). Dissecting the multicellular ecosystem of metastatic melanoma by single-cell
RNA-seq. Science 352, 189–196.
Tran, E., Turcotte, S., Gros, A., Robbins, P.F., Lu, Y.-C., Dudley, M.E., Wunderlich, J.R., Somerville, R.P., Hogan, K., Hinrichs, C.S., et al. (2014). Cancer
immunotherapy based on mutation-specific CD4+ T cells in a patient with
epithelial cancer. Science 344, 641–645.
Tran, E., Robbins, P.F., Lu, Y.-C., Prickett, T.D., Gartner, J.J., Jia, L., Pasetto,
A., Zheng, Z., Ray, S., Groh, E.M., et al. (2016). T-Cell Transfer Therapy Targeting Mutant KRAS in Cancer. N. Engl. J. Med. 375, 2255–2262.
612 Cell 168, February 9, 2017
Tumeh, P.C., Harview, C.L., Yearley, J.H., Shintaku, I.P., Taylor, E.J.M., Robert, L., Chmielowski, B., Spasic, M., Henry, G., Ciobanu, V., et al. (2014). PD-1
blockade induces responses by inhibiting adaptive immune resistance. Nature
515, 568–571.
Van Allen, E.M., Miao, D., Schilling, B., Shukla, S.A., Blank, C., Zimmer, L.,
Sucker, A., Hillen, U., Geukes Foppen, M.H., Goldinger, S.M., et al. (2015).
Genomic correlates of response to CTLA-4 blockade in metastatic melanoma.
Science 350, 207–211.
Van den Eynde, B., Hainaut, P., Hérin, M., Knuth, A., Lemoine, C., Weynants,
P., van der Bruggen, P., Fauchet, R., and Boon, T. (1989). Presence on a human melanoma of multiple antigens recognized by autologous CTL. Int. J.
Cancer 44, 634–640.
van der Bruggen, P., Traversari, C., Chomez, P., Lurquin, C., De Plaen, E., Van
den Eynde, B., Knuth, A., and Boon, T. (1991). A gene encoding an antigen
recognized by cytolytic T lymphocytes on a human melanoma. Science 254,
1643–1647.
Vormehr, M., Schrörs, B., Boegel, S., Löwer, M., Türeci, Ö., and Sahin, U.
(2015). Mutanome Engineered RNA Immunotherapy: Towards PatientCentered Tumor Vaccination. J. Immunol. Res. 2015, 595363.
Wang, P., Sidney, J., Kim, Y., Sette, A., Lund, O., Nielsen, M., and Peters, B.
(2010). Peptide binding predictions for HLA DR, DP and DQ molecules. BMC
Bioinformatics 11, 568.
Warren, R.L., Choe, G., Freeman, D.J., Castellarin, M., Munro, S., Moore, R.,
and Holt, R.A. (2012). Derivation of HLA types from shotgun sequence datasets. Genome Med. 4, 95.
Yaari, G., and Kleinstein, S.H. (2015). Practical guidelines for B-cell receptor
repertoire sequencing analysis. Genome Med. 7, 121.
Yoshihara, K., Shahmoradgoli, M., Martı́nez, E., Vegesna, R., Kim, H., TorresGarcia, W., Treviño, V., Shen, H., Laird, P.W., Levine, D.A., et al. (2013). Inferring tumour purity and stromal and immune cell admixture from expression
data. Nat. Commun. 4, 2612.
Zaretsky, J.M., Garcia-Diaz, A., Shin, D.S., Escuin-Ordinas, H., Hugo, W., HuLieskovan, S., Torrejon, D.Y., Abril-Rodriguez, G., Sandoval, S., Barthly, L.,
et al. (2016). Mutations Associated with Acquired Resistance to PD-1
Blockade in Melanoma. N. Engl. J. Med. 375, 819–829.
Zhang, H., Lund, O., and Nielsen, M. (2009). The PickPocket method for predicting binding specificities for receptors based on receptor pocket similarities: application to MHC-peptide binding. Bioinformatics 25, 1293–1299.
Zhang, N., Wu, H.-J., Zhang, W., Wang, J., Wu, H., and Zheng, X. (2015). Predicting tumor purity from methylation microarray data. Bioinformatics 31,
3401–3405.
Zhang, J., White, N.M., Schmidt, H.K., Fulton, R.S., Tomlinson, C., Warren,
W.C., Wilson, R.K., and Maher, C.A. (2016a). INTEGRATE: gene fusion discovery using whole genome and transcriptome data. Genome Res. 26, 108–118.
Zhang, J., Mardis, E.R., and Maher, C.A. (2016b). INTEGRATE-neo: a pipeline
for personalized gene fusion neoantigen discovery. Bioinformatics, btw674.
Zhang, X., Sharma, P.K., Peter Goedegebuure, S., and Gillanders, W.E.
(2016c). Personalized cancer vaccines: Targeting the cancer mutanome. Vaccine. Published online July 20, 2016. http://dx.doi.org/10.1016/j.vaccine.2016.
05.073.
Zheng, X., Zhao, Q., Wu, H.-J., Li, W., Wang, H., Meyer, C.A., Qin, Q.A., Xu, H.,
Zang, C., Jiang, P., et al. (2014). MethylPurify: tumor purity deconvolution and
differential methylation detection from single tumor DNA methylomes.
Genome Biol. 15, 419.
Leading Edge
Review
Clonal Heterogeneity and Tumor Evolution:
Past, Present, and the Future
Nicholas McGranahan1,2 and Charles Swanton1,2,3,*
1Cancer Research UK Lung Cancer Centre of Excellence, University College London Cancer Institute, Paul O’Gorman Building,
72 Huntley Street, London WC1E 6BT, UK
2Translational Cancer Therapeutics Laboratory, The Francis Crick Institute, 1 Midland Rd, London NW1 1AT, UK
3Department of Medical Oncology, University College London Hospitals, 235 Euston Rd, Fitzrovia, London NW1 2BU, UK
*Correspondence: charles.swanton@crick.ac.uk
http://dx.doi.org/10.1016/j.cell.2017.01.018
Intratumor heterogeneity, which fosters tumor evolution, is a key challenge in cancer medicine.
Here, we review data and technologies that have revealed intra-tumor heterogeneity across cancer
types and the dynamics, constraints, and contingencies inherent to tumor evolution. We emphasize
the importance of macro-evolutionary leaps, often involving large-scale chromosomal alterations,
in driving tumor evolution and metastasis and consider the role of the tumor microenvironment in
engendering heterogeneity and drug resistance. We suggest that bold approaches to drug development, harnessing the adaptive properties of the immune-microenvironment while limiting those
of the tumor, combined with advances in clinical trial-design, will improve patient outcome.
Introduction
In a famous thought experiment, evolutionary biologist Stephen
Jay Gould asked, if the ‘‘tape of life’’ could be turned back to the
very beginning, would the same outcome prevail (Gould, 2000)?
While re-playing the tape of life remains a hypothetical experiment, the ‘‘tape of cancer’’ is being played and re-played with
increasing regularity and too frequently with lethal results. Cancer is one of the leading causes of morbidity and mortality worldwide, with approximately 14 million new cases and 8.2 million
cancer-related deaths occurring in 2012 alone (Siegel et al.,
2013). Alarmingly, the number of new cases is predicted to rise
by 70% over the next two decades.
A key factor contributing to the lethal outcome of cancer, therapeutic failure, and drug resistance is intratumor heterogeneity
(ITH) (Greaves, 2015). ITH provides diverse genetic and epigenetic material upon which selection and Darwinian evolution
can act. However, this diversity also permits the tape of each
cancer’s life to be deciphered, revealing the temporal order of
genomic events and shedding light on constraints and contingencies to cancer evolutionary trajectories. The advent of nextgeneration sequencing has enabled more powerful analysis of
tumor evolution and has improved our understanding of tumor
initiation and development, as well as the interaction between
cancer cells and the immune-microenvironment. Despite these
advances, knowledge of ITH and clonal evolution and the
potential for competitive release of resistant subclones is infrequently considered in the therapeutic setting to inform clinical
trial design.
In this review, we explore the extent and clinical implications of
ITH and discuss how profiling tumor genetic diversity has been
used to trace tumors’ life histories and their patterns of evolution.
We emphasize the importance of viewing tumor development in
the context of an evolutionary framework, which includes largescale genomic alterations, and consider the dynamic evolution of
tumor cells and their interaction with the microenvironment.
Finally, we outline approaches to address cancer systemically,
taking into account ITH and tumor evolution.
Intra-tumor Heterogeneity: Patterns and Prevalence
How Much Heterogeneity Is There?
Genomic diversity within single tumors has long been recognized. Indeed, as early as 1958, evolutionary biologist Julian
Huxley commented on ‘‘genetic inhomogeneity’’ in cancer,
noting, ‘‘it will be of great interest to discover the extent of
such new variance and the rate at which it occurs’’ (Huxley,
1958). However, it is only with the advent of next-generation
sequencing studies that the full extent of genomic ITH is
becoming apparent. Sequencing of spatially or temporally
distinct tumor regions has begun to uncover the bewildering
extent of diversity within tumors (Figure 1).
These studies have revealed that the degree of ITH can be
highly variable, with between 0 and over 8,000 thousand coding
mutations found to be heterogeneous within primary tumors or
between primary and metastatic or recurrence sites (Johnson
et al., 2014). Despite caveats regarding differences in sampling
procedure, tumor stage, and sequencing depth, it is evident
that certain tumor types, such as melanoma and lung cancer,
harbor a significantly larger homogeneous coding mutational
burden than other types. The involvement of powerful exogenous mutagens, such as ultraviolet light and tobacco carcinogens, which a stem cell niche may be exposed to for years prior
to the first invasive step, likely explains the elevated clonal mutational burden in these cancers. Accordingly, the prevalence of
different types of base-substitutions (within a trinucleotide
context) observed in these tumors reflects their exogenous exposures (Alexandrov et al., 2013).
It is also evident that a large clonal burden does not equate to a
large subclonal burden, or vice versa. Predominantly, this likely
Cell 168, February 9, 2017 Crown Copyright ª 2017 Published by Elsevier Inc. 613
Figure 1. Heterogeneity of Non-silent Mutations from Multiple-Sample Sequencing
across a Range of Cancer Types
For each tumor type, each point represents one
tumor, with the proportion of heterogeneous mutations (ITH proportion), as well as the absolute
numbers of heterogeneous and homogeneous
non-silent mutations, shown. Black circles represent treatment naive tumors, with red triangles
indicating tumors that have received treatment.
Notably, these data are restricted to non-silent
mutations and does not include copy-number alterations. The data are extracted from the following
primary publications: Diffuse intrinsic pontine glioma (Nikbakht et al., 2016); Neuroblastoma (Eleveld et al., 2015); low-grade gliomba/glioblastoma
multiforme (Johnson et al., 2014; Kim et al., 2015;
Wang et al., 2016); breast cancer (Yates et al.,
2015); ccRCC (Gerlinger et al., 2014a); multiple
myeloma (Bolli et al., 2014); lung adenocarcinomas
(de Bruin et al., 2014; Zhang et al., 2014); prostate
(Gundem et al., 2015; Ju et al., 2014); bladder
(Lamy et al., 2016); colorectal adenocarcinomas
(Uchi et al., 2016); liver hepatocellular carcinoma
(Xue et al., 2016); esophageal squamous cell carcinoma (Hao et al., 2016); ovarian (Bashashati
et al., 2013; Eckert et al., 2016); esophageal
adenocarcinoma (Murugaesu et al., 2015); melanoma (Harbst et al., 2016).
reflects the fact that distinct mutational processes may operate
at different times in tumor evolution (de Bruin et al., 2014;
McGranahan et al., 2015; Nik-Zainal et al., 2012). Indeed, the
most notable outliers with regard to high subclonal mutational
burden, but low clonal burden, are low-grade gliomas that recur
as glioblastomas after treatment with the alkylating agent temozolomide (Johnson et al., 2014; Kim et al., 2015). In this case,
the abundance of subclonal mutations can be directly linked
to therapy-induced mutations that are compounded by loss
of mismatch repair machinery. As more data accumulates, the
impact of therapy on mutational load and ITH will likely become
clearer.
In other cancer types, such as non-small cell lung cancer
(NSCLC) and bladder cancer, the presence of a large subclonal
mutation burden can be attributed to the action of the APOBEC
family of cytidine deaminases (de Bruin et al., 2014; McGranahan
et al., 2015; Zhang et al., 2014). In colorectal and prostate cancers, alterations to mismatch repair or proofreading machinery
can occasionally play a key role in generating both clonal and
subclonal mutations (Kumar et al., 2016; Uchi et al., 2016).
Consistent with results from mathematical modeling (Tomasetti et al., 2013), these studies also suggest that the clonal mutation burden in certain cancer types, including lung cancer and
melanoma, predominantly reflects mutations accumulated prior
to carcinogenesis, while in others, mutation burden may be more
614 Cell 168, February 9, 2017
reflective of somatic events occurring after tumorigenesis. As such, the mutation
rate of a tumor cannot necessarily be inferred from a single biopsy, and simply
considering the proportion of heterogeneous mutations represents a poor surrogate for diversity (Figure 1).
Importantly, heterogeneity does not simply affect coding mutations. A multitude of epigenetic mechanisms, including DNA
methylation, chromatin remodeling, and post-translational modification of histones, can contribute to diversity within tumors (for
a review, see Mazor et al. [2016]). Analysis of ITH in gliomas (Mazor et al., 2015), as well as prostate cancers (Aryee et al., 2013)
and esophageal squamous cell carcinomas (Hao et al., 2016),
has suggested that the extent of ITH calculated from DNA
methylation mirrors ITH measures captured at the genomic level.
Genomic copy-number heterogeneity can also be extensive
within tumors, to the extent that copy-number ITH in clear cell
renal cell carcinoma (ccRCC) mirrors the extent of copy-number
diversity between tumors (Martinez et al., 2013). Large-scale
chromosomal alterations may have profound impact upon the
genome, disrupting hundreds of genes, and can be considered
macro-evolutionary events (see below), which may contribute
to tumor progression (Notta et al., 2016). Loss of genomic material through chromosomal instability may also contribute to
mutational ITH (McPherson et al., 2016; Murugaesu et al.,
2015), highlighting the importance of considering both copy
number and mutation data when inferring the evolutionary history of tumors.
Heterogeneity Can Reveal a Tumor’s Life History
A tumor’s mutational catalog represents a historical record of
alterations that have accumulated during its life history. The
Figure 2. Evolutionary Trees Illustrating Intratumor Heterogeneity across Cancer Types
For each cancer type, the mean number of clonal and subclonal non-silent mutations are depicted as trunks (blue) and branches (yellow and red), respectively.
The data is extracted from the same publications as Figure 1.
heterogeneity present between cancer cells can be used to illuminate the temporal order of these events. Alterations identified
in every sequenced cancer cell can be considered to form the
trunk of a cancer’s somatic evolutionary tree, while subclonal
mutations, present in only a subset of cancer cells, make up
the branches (Figure 2). Further, the prevalence of subclonal mutations in different cancer cells can be used to infer the subclonal
hierarchy of the tumor’s phylogeny.
A plethora of bioinformatics tools have been developed to help
decipher the temporal order of mutations and determine which
are clonal or subclonal. The majority of tools focus on somatic
point mutations and either restrict the analysis to copy neutral regions of the genome or make the assumption that a single mutation will be present at the same copy-number state in every cancer cell (Carter et al., 2012; Miller et al., 2014; Roth et al., 2014).
While tools to infer copy-number heterogeneity have also been
developed (Ha et al., 2014; Shen and Seshan, 2016), more recent
tools seek to combine copy number and mutational data (Fischer
et al., 2014; Jiang et al., 2016) and, furthermore, attempt to
infer evolutionary relationships between subclonal populations
(Deshwar et al., 2015; Jiang et al., 2016). Relatedly, orthogonal
tools to dissect heterogeneity by using data from single-cell
sequencing have also been developed (Roth et al., 2016).
However, despite considerable advances, it is self-evident
that even cutting-edge tools can only dissect the ITH within the
sample(s) subject to sequencing. As such, our ability to distinguish truly clonal from pseudo-clonal mutations is largely dependent on the number of tumor regions sequenced (de Bruin et al.,
2014; Yates et al., 2015), the depth and purity of what is
sequenced, and, further, whether single-cell sequencing is also
implemented (Roth et al., 2016).
Driver Alterations and Heterogeneity
Studies exploring ITH within solid tumors have demonstrated
a tendency for established cancer genes to harbor clonal muta-
tions (Kumar et al., 2016; Makohon-Moore et al., 2017;
McGranahan et al., 2015). However, despite this tendency,
even mutations in established cancer genes can be present in
only a subset of cancer cells within a tumor.
Subclonal driver mutations can give an illusion of clonality due
to sampling bias. Whole-genome sequencing of 33 pairs of medulloblastomas, pre- and post-therapy, found that the majority of
putative drug targets that were identified pre-treatment appeared clonal but were revealed to be subclonal or absent at
recurrence (Morrissy et al., 2016). Equally, ITH may lead to significant underestimates of the number of driver alterations present
in a tumor. Analysis of 86 cases of diverse primary tumors and
brain metastases revealed that, in 53% of cases, putative drug
targets, including PTEN and PIK3CA, were exclusively identified
in the brain lesions and not in the primary tumor (Brastianos
et al., 2015).
Accumulating evidence suggests that certain driver alterations may be more likely to be subclonal than others. Subclonal mutations in PIK3CA have been found in NSCLC (de Bruin
et al., 2014), breast (Yates et al., 2015), colorectal (Uchi et al.,
2016), melanoma (Harbst et al., 2016), esophageal squamous
cell carcinoma (Hao et al., 2016), ccRCC (Gerlinger et al.,
2014a), and ovarian cancers (Bashashati et al., 2013). In keeping with these results, across nine cancer types, mutations in
the PI3K-AKT-mTOR pathway were found to harbor a higher
proportion of subclonal mutations compared to genes associated with RAS-MAPK pathway (McGranahan et al., 2015). However, other driver mutations exhibit a tendency to be clonal in
certain cancer types, but not others. Mutations in TP53 appear
almost exclusively clonal in NSCLC (de Bruin et al., 2014; Zhang
et al., 2014), esophageal adenocarcinomas (Murugaesu et al.,
2015), and ovarian cancers (Bashashati et al., 2013), yet are
often subclonal in ccRCC (Gerlinger et al., 2014a) and chronic
lymphocytic leukemia (CLL) (Landau et al., 2013). Such
Cell 168, February 9, 2017 615
differences may reflect the importance of epistasis in cancer
evolution and are in agreement with findings that co-occurrence
and mutual exclusivity relationships between cancer driver alterations can vary extensively in different cancer types (Park
and Lehner, 2015).
The subclonal nature of genomic driver alterations can have
important clinical implications. A recent report demonstrated
that two gastric cancers with high clonal amplification of
FGFR2 responded to the FGFR inhibitor AZD4547. Conversely,
the six tumors with low or subclonal amplification of FGFR2
did not respond (Pearson et al., 2016). In CLL, the presence
of subclonal driver alterations is associated with decreased
relapse-free survival (Landau et al., 2013).
Processes of Cancer Genome Evolution, and
Evolutionary Debates
Selection and Neutral Evolution in Cancer
‘‘.the fittest will survive, and a race will be eventually produced
adapted to the conditions in which it lives’’ (Wallace, 1867).
Although originally framed in relation to the evolution of individual organisms within a population, the fundamental principles
of Darwinian evolution, involving variation with differential fitness
that is heritable, can be applied in the context of tumor evolution
(Nowell, 1976). In this setting, the population of cancer cells are
subject to selection, and the genetic variation between these
cells, influenced by endogenous and exogenous mutational processes, provides the fuel for selection to act (Figure 3).
Although heterogeneity is required for Darwinian evolution,
positive selection does not necessarily lead to heterogeneity
(Waclaw et al., 2015). As such, the extent to which positive selection can account for the degree of ITH in tumors has been called
into question (Ling et al., 2015; Williams et al., 2016). Specifically,
while evidence for selection of driver events in cancer development, as well as the selection pressures imposed by therapy,
are undisputed, following a ‘‘big bang’’ of diversity early in tumor
evolution, ITH development can follow the laws of neutral growth
(Sottoriva et al., 2015). In support of this, Williams et al. (2016)
noted that, for a subset of tumors, the relationship between the
number of subclonal mutations and their relative abundance
was consistent with a neutral growth pattern rather than subclonal expansions. Likewise, in an extensively sampled colorectal
cancer, the degree of heterogeneity appeared more consistent
with neutral growth (Ling et al., 2015), and lineage-tracing
studies in mice have suggested that ITH may emerge from a
stem cell hierarchy of cancer cells evolving under neutral evolution (Driessens et al., 2012).
Further work is warranted to explore the temporal and spatial
dynamics of clones in human tumors. Longitudinal sequencing
data from CLL has demonstrated clonal dynamics and shifts
in selection pressures, even in the absence of therapy (Nadeu
et al., 2016). Conceivably, during Darwinian evolution, both selection and neutral growth may operate simultaneously within
the same tumor, and this may alter dynamically over time. The
observation that multiple different diversity measures in Barrett’s
esophagus predict progression to esophageal adenocarcinoma
(Maley et al., 2006; Merlo et al., 2010) is indicative that diversity
may lead to selection of aggressive subclones, even without
therapeutic selection pressures. Relatedly, the fact that cancer
616 Cell 168, February 9, 2017
genes can harbor a statistically significant enrichment of subclonal mutations suggests that a signal of selection can be present
throughout tumor evolution (McGranahan et al., 2015). It remains
an open question whether distinct clinical behaviors can be
observed depending on the mode of tumor evolution or whether
the occurrence of neutral evolution and drift may limit the ability
to predict a tumor’s next step (Lipinski et al., 2016).
Contingency and Convergence
Using the tape of life metaphor, Gould emphasized the importance of chance and unpredictability in the evolution of life on
earth, suggesting that the end result is causally dependent on
antecedent steps, or ‘‘historical contingency’’ (Gould, 2000).
Examples of genetic contingency impacting upon the clinical
course of the disease can be found in cancer evolution. The order in which the two driver events in JAK2 and TET2 are acquired
in myeloproliferative disorders affects the clinical course of
the disease (Ortmann et al., 2015). If a TET2 mutation is acquired
first, expansion of hematopoietic stem and progenitor cells
occurs, blocking expansion of erythroid progenitors until cells
acquire a JAK2 mutation. Conversely, if a JAK2 mutation is acquired first, megakaryocyte number increases, with no expansion of the hematopoietic stem and progenitor pool until a
TET2 mutation is acquired. Patients acquiring a JAK2 mutation
first are younger at disease onset and are more likely to present
with polycythemia rubra vera and develop thrombosis than they
are to develop essential thrombocythemia. Relatedly, the cell of
origin may also have important consequences on the impact of
identical somatic events. Despite the fact that it is possible to
induce pancreatic adenocarcinoma and non-small cell cancers
from the same initiating events (TP53 inactivation coupled with
KRAS activation), these tumors have been found to exhibit
distinct metabolic requirements, making use of branched-chain
amino acids in different ways (Mayers et al., 2016).
An alternative (and likely complementary) view of evolution
emphasizes the importance of convergence and was first
encapsulated by Darwin discussing analogical variation, noting,
‘‘the common rule throughout nature is infinite diversity of
structure for gaining the same end’’ (Darwin, 1859), and
subsequently echoed nearly 150 years later by Conway Morris,
when he observed that there is a ‘‘recurrent tendency of
biological organization to arrive at the same solution’’ (Conway
Morris, 2003).
Evidence supporting both historical contingency and convergence toward the same solution is apparent from cancer evolutionary studies. Indeed, in germline VHL mutant carriers with
synchronous renal cell carcinomas developing in the same patient, evidence for both contingency and convergence can be
found; despite distinct secondary 3p loss of heterozygosity
events and driver mutations in different cancers from the same
patient, there was evidence for convergent PI3K signal transduction pathway activation (Fisher et al., 2014).
There is also extensive evidence for convergence of both genotype and phenotype in cancer evolution. Indeed, the notion of
cancer hallmarks supports the occurrence of convergence in
cancer evolution. Further, Conway Morris (2003) assertion that
‘‘it matters little what our starting point may have been: the
different routes will not prevent a convergence to similar ends’’
could be used to describe the tendency for a cancer stem cell
Figure 3. Clonal Heterogeneity and Tumor Evolution: Modes, Mechanisms, Ecosystems, and Evolutionary Therapeutics
The first three panels depict different aspects of cancer genome evolution, which all need to be understood to develop improved evolutionary therapeutics
(panel four).
transcriptome to derive on multiple occasions across multiple
distinct tumor types (Chen and He, 2016).
Moreover, convergence is frequently seen within individual
tumors, termed parallel evolution, which, in the context of cancer, refers to the independent evolution of similar traits starting
from a single ancestral clone. Campbell et al. (2010) reported
two overlapping out of frame deletions in exon 6 of PARK2 in
distinct pancreatic cancer metastases from the same patient.
Similarly, recurrent and independent acquisition of copy-number
events such as deletions in PAX5, ETV6, and CDKN2A have
been described within distinct subclones in acute lymphoblastic
leukemia (ALL) (Anderson et al., 2011). Recurrent disruption of
the SWI/SNF complex or activation of the PI3K pathway through
distinct mutations in mTOR, TSC1, PTEN, and PIK3CA are
frequently observed in different subclones from the same renal
cancer (Gerlinger et al., 2012, 2014a; Voss et al., 2014).
Cell 168, February 9, 2017 617
As the resolution of cancer evolutionary analyses improves,
the number of examples of parallel evolution increases, including
but not limited to events involving EGFR in glioblastoma (Francis
et al., 2014; Kim et al., 2015), TP53 and ATRX in glioma (Johnson
et al., 2014), activation of the MAPK pathway in multiple
myeloma (Bolli et al., 2014; Melchor et al., 2014), NOTCH1 and
GNPTAB recurrent mutations in esophageal adenocarcinoma
(Murugaesu et al., 2015), SMO mutations in medulloblastoma
(Morrissy et al., 2016), distinct AR amplification events in prostate cancer (Gundem et al., 2015), KMTD2D and CREBBP mutations in follicular lymphoma (Okosun et al., 2014), PTEN and
TP53 mutations, FGFR2 amplifications, and RUNX1 deletions
in primary breast cancer (Yates et al., 2015), as well as distinct
CCNE1 amplifications in ovarian cancers (McPherson et al.,
2016). In ccRCC, an early clonal event is 3p loss of heterozygosity, which appears to prime the tumor for second hits in SETD2,
PBRM1, and BAP1 (all of which are encoded on chromosome
3p) later in tumor evolution (Gerlinger et al., 2014a). Similarly, in
breast cancer, three of the four parallel evolutionary events
(TP53, PTEN, and RUNX1) documented by Yates et al. (2015)
occurred as the second hit, following a clonal event.
During the selection pressures of targeted therapies, parallel
evolution driving polyclonal-acquired drug resistance has been
frequently documented. For example, in 13 out of 16 patients
with BRAF mutant melanomas with resistance to RAF inhibition,
multiple parallel mechanisms of resistance were observed (Shi
et al., 2014). Likewise, following EGFR monoclonal antibody
therapy, multiple KRAS mutations have been observed in circulating free DNA (Bettegowda et al., 2014; Misale et al., 2012).
One patient acquired a codon 12 KRAS, codon 61 KRAS, and
a codon 61 NRAS mutation together with a BRAF codon 600
mutation following acquired resistance to EGFR monoclonal
antibody therapy that were not detectable prior to therapy (Bettegowda et al., 2014). Following acquired resistance to a PI3K
alpha inhibitor, Juric et al. (2015) found parallel evolution of six
distinct PTEN aberrations across 10 metastatic sites on the
background of a clonal single copy PTEN deletion, reminiscent
of second hit tumor suppressor gene loss following an early
clonal event witnessed in breast and renal cancers.
These observations suggest that despite the stochastic nature
of genomic change, microenvironmental, epistatic, and lineage
constraints operate that might allow the prediction of a limited
set of subsequent evolutionary moves.
Gradualism versus Punctuated Evolution
Another long-standing evolutionary debate that has reemerged
in the context of tumor development centers on whether tumor
evolution occurs gradually through the sequential accumulation
of mutations and clonal expansions or whether it is characterized
by punctuated bursts (Figure 3). Such a dichotomy has been
framed in the context of micro- versus macro-evolution, with
gradual accumulation of point mutations (micro-evolution) presented in opposition to a saltationist view, which emphasizes
the importance of large-scale chromosomal alterations and
bursts of mutations (macro-evolution) (Gerlinger et al., 2014b).
An incremental, gradual accumulation of mutations during the
life history of a tumor is evidenced by the presence of clock-like
mutational signatures that correlate with the chronological age of
the patient (Alexandrov et al., 2015). However, not all mutations
618 Cell 168, February 9, 2017
accumulate in a clock-like manner. In several cancer types (Alexandrov et al., 2013), a phenomenon termed kataegis has been
observed, describing a small localized mutational process that
results in hyper-mutation (a few to several hundred C>T and/or
C>G substitutions, enriched at TpC sites) on the same DNA
strand. A punctuated mode of tumor evolution is also supported
from lineage tracing studies. Graham and colleagues used lineage-tracing techniques based on nuclear and mitochondrial
DNA lesions in human colon adenomas to identify stem cell populations within adenoma crypts with multipotent potential and
map their evolution over time. A punctuated model of rare clonal
expansions interspersed with prolonged periods of stasis was
suggested (Humphries et al., 2013).
Cancer evolution is conceptually similar to evolution in asexually reproducing organisms, and in yeast, it was recently demonstrated that tetraploid strains showed faster adaptation to a
poor carbon source and accumulated more genomic diversity
compared to diploid counterparts (Selmecki et al., 2015). In cancers, genome doublings have been estimated to occur at high
frequencies (Zack et al., 2013) and are also associated with
elevated rates of chromosomal aberrations (Dewhurst et al.,
2014; Fujiwara et al., 2005; Zack et al., 2013). Genome doublings
may serve to reduce the impact of Muller’s ratchet, a process by
which asexual genomes accumulate deleterious mutations in an
irreversible manner. Specifically, although the impact of a deleterious mutation cannot be removed through sexual reproduction, it can be mitigated by the presence of additional, doubled,
wild-type (WT) alleles.
Chromothripsis, characterized as a single catastrophic event
resulting in tens of hundreds of locally clustered rearrangements
affecting one or a few chromosomes, has also been documented
to be widespread in cancers, occurring in over 30% of bladder
cancers (Morrison et al., 2014), lung adenocarcinomas (Malhotra
et al., 2013), esophageal adenocarcinomas (Nones et al., 2014),
glioblastomas (Malhotra et al., 2013), uterine leiomyomas (Mehine et al., 2013), and pancreatic cancers (Notta et al., 2016).
These events occurred both clonally and subclonally during tumor evolution. Single-nucleus sequencing of 1,000 cancer cells
from 12 triple-negative breast cancers found evidence for
copy-number alterations that had accumulated in short punctuated bursts early in tumor evolution, but not late (Gao et al.,
2016). The progression of esophageal adenocarcinomas from
Barrett’s esophagus is thought to involve a punctuated path
whereby a TP53 mutant cell undergoes a whole-genomedoubling event followed by the acquisition of oncogenic
amplifications (Stachler et al., 2015), conceivably through chromothripsis (Nones et al., 2014).
Chromothripsis and large-scale genomic rearrangements
have parallels with evolutionary biologist Richard Goldschmidt’s
notion of ‘‘hopeful monsters’’ and ‘‘macromutations’’ (Goldschmidt, 1982). Such mutations were described as ‘‘of the
most extraordinary rarity to provide the world with the important
material for evolution’’ and appear analogous to the simultaneous disruption of multiple pre-neoplastic driver events
(CDKN2A, TP53, and SMAD4) in single chromothriptic events
in prostate cancer (Notta et al., 2016).
Perhaps unsurprisingly, while large-scale genomic rearrangements and chromothriptic events are often associated with
aggressive cancers, its common occurrence in uterine leiomyomas highlights that it can also be involved in the development of
benign tumors. In fact, consistent with ‘‘hopeful’’ and ‘‘hopeless
monster’’ evolutionary thought, chromothripsis can even occasionally have a direct positive impact on patient outcome.
McDermott et al. (2015) reported a case study in which a chromothriptic event resulted in a cure for a patient with an inherited
immunodeficiency disease caused by over-activity of a mutated
chemokine receptor CXCR4. Specifically, the chromothriptic
event led to deletion of the aberrant allele in a single hematopoietic stem cell, which subsequently repopulated the bone marrow
and restored normal immune function.
Taken together, these data highlight the frequent occurrence
of macro-evolutionary events in cancers. However, temporal
and multi-regional tumor analyses will be required to reveal
the true extent to which chromosomal alterations occur dynamically throughout tumor evolution. The observation that tumors
with an extreme level of chromosomal instability appear associated with improved prognosis, compared to intermediate levels
(Andor et al., 2016; Birkbak et al., 2011), supports the hypothesis that there may be a delicate balance between too much and
too little instability and that there may be potent selection pressures in cancer evolution for a ‘‘just-right’’ level of cell-to-cell
variation.
Speciation and the Metastatic Process in Cancer
Tumor metastasis is frequently cited to be responsible for 90%
of all cancer-related deaths. The process has been likened to
a speciation event with macro-evolutionary leaps required to
endow a tumor cell with metastatic potential (for reviews, see
Gerlinger et al., 2014b; Turajlic and Swanton, 2016).
In certain tumors, metastatic spread has been found to be
monophyletic, with a single subclone in the primary tumor appearing to seed multiple metastases at different sites, resulting
in low inter-metastatic ITH (McPherson et al., 2016; Schwarz
et al., 2015). However, in other tumors, subclones at distinct
metastatic sites are more closely related to subclones within
the primary tumor than they are to each other, indicative of a
polyphyletic metastatic process. Importantly, polyphyletic
metastatic spread suggests that multiple distinct evolutionary
trajectories within a single tumor can result in metastatic
dissemination. A study of seven patients with ovarian cancer
found that five patients exhibited monoclonal and uni-directional seeding from the ovary to intraperitoneal sites, while the
remaining two patients exhibited polyphyletic spread and
reseeding (McPherson et al., 2016). However, convergent selection pressures in the metastatic setting, even in the context
of polyphyletic spread, are evidenced by the occurrence of parallel evolution at distinct metastatic sites (Campbell et al., 2010).
Finally, multiple rounds of metastasis involving re-seeding may
also occur, highlighting the diverse patterns of metastatic
spread that can occur, even within single tumors (for review,
see Turajlic and Swanton, 2016).
Lineage tracing studies have informed our understanding of
the patterns of tumor metastatic seeding. In an autochthonous
model of mouse pancreatic cancer, Maddipati and Stanger
(2015) used multi-color lineage tracing strategies to track early
development of KRAS/p53 mutant pancreatic pre-invasive
lesions through to metastatic disease. Each pancreatic mass
contained an average of four single-color lesions, indicating
the presence of distinct tumors originating from independent
genetic events in the pancreas. A quarter of pre-malignant precursor pancreatic lesions, acinar-to-ductal metaplasias (ADMs),
displayed heterogeneous colors, indicative of their evolution
from multiple acinar cells. However, pancreatic intra-epithelial
neoplasia (PanIN) lesions displayed single colors, indicative of
a bottlenecking event in the evolution of the pre-malignant disease from ADMs to PanIN lesions. Analysis of metastatic lesions
in the lung, liver, and peritoneum revealed a high frequency of
polyclonal metastasis, suggesting potential cooperativity between cancer subclones facilitating metastatic colonization.
Evidence for polyclonal seeding of metastases was also
observed in a common mouse model of breast cancer (Cheung
et al., 2016). Supporting a clonal cooperativity model of tumor
metastases, the authors found evidence for collective invasion
and migration of polyclonal clusters of cells within the circulation
seeding polyclonal disease at metastatic sites. These data
reflect reports of circulating tumor cell clusters associated
with poor prognosis in breast and prostate cancer (Aceto
et al., 2014) and highlight the need to view cancer as an
ecosystem of subclones that may act cooperatively or antagonistically.
The Cancer Ecosystem
Functional Cooperativity
The common occurrence of ITH challenges the view that tumor
phenotypes are entirely driven by the dominant tumor clone in
a cell-autonomous manner, in which driver mutations only confer
a benefit to the cancer cell in which they occur. If cancer cells act
in a non-cell-autonomous way, whereby driver mutations confer
benefits to neighboring cells, it will result in ITH and cooperative
or social networks governing tumor behavior.
Anton Berns and colleagues studied metastatic potential in a
mouse model of small cell lung carcinoma (Calbo et al., 2011).
Mesenchymal and neuroendocrine cells derived from a common
progenitor, when engrafted into mice as a heterogeneous population, triggered metastatic behavior of the neuro-endocrine
cells. In adult GBM, Inda and colleagues noted that EGFRvIII
deletion mutants account for a minority of the total population
in some tumors. The authors found a paracrine mechanism
sustaining growth of the dominant WT-EGFR clones through
IL-6 and LIF from the EGFRvIII deletion mutants, leading to
WT-EGFR activation in neighboring clones, sustaining tumor
heterogeneity (Inda et al., 2010).
To investigate subclonal cooperativity further, Polyak and
colleagues studied subclonal interactions in mouse xenograft
models. Sub-populations of tumor cells could sustain the survival and growth of all tumor cells through IL-11-mediated microenvironment change. Notably, if minor subclones, sustaining the
growth of the majority, were outcompeted by tumor subclones
with greater proliferative capacity, tumor collapse resulted, suggesting that non-cell autonomous drivers may be required for tumor development (Marusyk et al., 2014). Similarly, using a mouse
model of breast cancer, Cleary et al. (2014) demonstrated that
clonal cooperation can be essential for tumor maintenance.
Bi-clonal mouse tumors containing genetically distinct luminal
and basal subclones were separated into their component
Cell 168, February 9, 2017 619
subclonal populations and subsequently transplanted into WT
host animals, either separately or as a 1:1 admixture. Whereas
the bi-clonal cell mixture was highly tumorigenic, mono-clonal
populations failed to elicit tumor formation.
Such cooperativity also extends to the field of drug resistance.
Hobor and Bardelli noted that only a fraction of some cetuximabresistant colorectal cancer samples harbored KRAS mutations,
commonly described to result in acquired resistance to EGFR
directed therapies. The authors found evidence that TGF alpha
and amphiregulin secretion from EGFR inhibitor resistant cells
were capable of sustaining the growth of KRAS WT drug-sensitive cells in a paracrine manner (Hobor et al., 2014).
The Tumor Microenvironment
The tumor microenvironment likely imposes profound constraints upon cancer evolution, both at primary and distant sites.
Such constraints arise through resource limitations, immune
predation, and adverse growth conditions in the form of tissue
hypoxia, acidosis, and cancer therapeutics, among others.
Increasing evidence supports the ability of tumor cells to shape
their own advantageous growth environment and the ability of
the microenvironment to protect tumor cells from the deleterious
impact of exogenous sources of microenvironment change
derived from systemic therapy. Therefore, cancer evolution
cannot be fully understood without a detailed understanding
of the source and impact of micro-environmental selection
pressures.
Computational, pathological, and tumor-imaging approaches
are increasingly being used to describe the complex tumor
microenvironment in a relatively unbiased manner. Aerts et al.
(2014) applied radiomics, which refers to the quantification of tumor phenotype using multiple imaging features, to head and
neck and lung cancers. The authors found that multiple radiomic
features associated with heterogeneity were linked to poorer
survival outcome in both tumor types. Through the integration
of gene expression and somatic copy number data with automated microenvironment analysis from standard hematoxylin
and eosin slides, Yuan et al. (2012) demonstrated that survival
predictions in ER negative breast cancer can be optimized.
The authors found that the spatial distribution of stromal cells
was an independent prognostic factor for survival outcome.
Similarly, in ovarian cancers, the percentage of stromal cells
(assessed from hematoxylin and eosin slides) was significantly
associated with poor overall and progression-free survival,
even after controlling for clinical parameters including surgical
debulking status and age (Natrajan et al., 2016). In prostate cancers, a measure of genomic instability, coupled with intratumoral
hypoxia, was found to be able to significantly improve prognostic
accuracy, beyond conventional clinical parameters (Lalonde
et al., 2014). A landmark study in melanoma, which sequenced
over 4,500 single cells from 19 patients (including malignant,
immune, stromal, and epithelial cells), identified therapy resistance tumor subpopulations present prior to treatment, which
may have been missed with bulk-sequencing, as well as a relationship between cancer-associated fibroblasts and preferential
expression of an AXL-high and MITF-low transcriptional program (Tirosh et al., 2016).
Recently, approaches to analyze the tumor microenvironment
in 3D have been developed, allowing a quantitative measure of
620 Cell 168, February 9, 2017
micro-environmental heterogeneity to be assessed (ecosystem
diversity index) (Natrajan et al., 2016). In grade 3 breast cancers,
high micro-environmental diversity was associated with poor
prognosis, independent of tumor size or genomic features. The
authors suggest that these data are indicative of cooperation between tumor cells and the microenvironment. Further, the spatial
diversity of resources inherent in a heterogeneous tumor microenvironment may select for a metastatic phenotype. A complementary explanation is that a heterogeneous tumor microenvironment may also contribute to unequal drug penetration
brought about by disordered blood vessel development that
might contribute to resistant cell populations emerging through
therapy (Fu et al., 2015; Junttila and de Sauvage, 2013) or to
the development of diverse niches, including hypoxic or perivascular regions that might support cancer stem cell phenotypes
and chemo-resistance (Mao et al., 2013).
Immune-Mediated Editing
Seminal work from the Schreiber laboratory in mouse models
demonstrated the capacity of the immune system to maintain tumors in a state of equilibrium, where clonal expansions are attenuated by adaptive immunity (Koebel et al., 2007). These observations begin to shed light on tumor dormancy and how patients
with early stage breast cancer may have disseminated tumor
cells in bone marrow, which never give rise to metastatic disease
(Hartkopf et al., 2014).
One substrate for immune-mediated disease control of tumor
growth can be patient-specific neo-antigens that arise as a
consequence of tumor-somatic mutations. A number of studies
have revealed an association of tumor mutational burden with
response to immune checkpoint blockade. In both melanoma
and NSCLC, evidence is building that the mutational load and
neo-antigen burden correlates with benefit to anti-CTLA4 therapy in melanoma (Snyder et al., 2014; Van Allen et al., 2015)
and anti-PD1 therapy in NSCLC (Rizvi et al., 2015). Likewise, hyper-mutated mismatch-repair-deficient tumors are significantly
more responsive to immune-checkpoint blockade than their
mismatch-repair-proficient counterparts (Le et al., 2015).
How tumor cells evade such hostile immune predation is
becoming an active research area. Hacohen and colleagues
devised an RNA sequencing (RNA-seq)-based signature of cytolytic activity, incorporating Granzyme A and Perforin, genes
which are upregulated following CD8+ T cell activation (Rooney
et al., 2015). Application of the signature to the TCGA dataset revealed that certain tumors such as ccRCC exhibit high cytolytic
activity, while others such as glioma and prostate cancer tend
to display low cytolytic activity. Somatic mutations in specific
genes, such as inactivating mutations in Caspase 8, were associated with higher cytolytic activity. These data are consistent
with previous work revealing that Caspase 8 blockade results
in tumor T cell escape in two murine tumor models (Medema
et al., 1999). Cytotoxic T cell (CTL) activity was associated with
both the rate of mutations and the rate of mutations resulting in
predicted neo-antigens across multiple tumor types (Rooney
et al., 2015). Intriguingly, colorectal and kidney cancer harbored
significantly fewer putative neo-antigens per non-silent than
expected, consistent with immune-editing, where immune activity likely results in the depletion of emerging tumor clones with
productive neo-antigens.
The evolving somatic mutational landscape may also influence
immune surveillance and response to checkpoint blockade. Evidence is emerging that the clonal status of a neo-antigen might
influence immune checkpoint benefit. Tumors with a high clonal
neo-antigen burden and low subclonal neo-antigenic heterogeneity appeared to be enriched in patients benefiting from antiPD1 therapy in NSCLC or anti-CTLA4 therapy in melanoma
(McGranahan et al., 2016). Whether subclonal neo-antigens
developing in a rapidly evolving tumor actively distract the immune response from the effective targeting of clonal neo-antigens is unclear. We have recently shown that APOBEC-induced
mutagenesis contributes to branched evolution and the acquisition of subclonal mutations in adenocarcinoma of the lung, estrogen receptor (ER)-negative breast cancer, head and neck squamous carcinoma, and esophageal adenocarcinomas (de Bruin
et al., 2014; McGranahan et al., 2015; Rosenthal et al., 2016).
In this regard, the parallels with HIV-based evolution of diversity
mediated by APOBEC activity are intriguing. Evidence suggests
that APOBEC 3G/3F-induced mutations in HIV are less immunogenic and reduce CD8+ T cell responses against common HIV
epitopes ex vivo (Monajemi et al., 2014). Whether APOBECinduced mutagenesis provides the tumor with a similar immune
evasion escape warrants further investigation.
The dynamic nature of the ‘‘predator-prey’’ relationship between the immune microenvironment and the tumor has recently
been highlighted by work demonstrating the ability of tumors to
lose the expression of neo-antigens (Anagnostou et al., 2016;
Verdegaal et al., 2016). Thus, therapeutic efforts may have to
be oriented toward the targeting of multiple clonal neo-antigens
to optimize disease control and minimize the potential for immune escape.
Safe Havens
Recent studies have suggested that mechanisms of tumor resistance need not always be mediated by the selection for resistant
populations of tumor cells (Hirata et al., 2015).
Using detailed intravital imaging in a mouse model of cancer,
Sahai and colleagues have demonstrated that resistance to a
BRAF inhibitor PLX4720 is mediated through melanoma-associated fibroblast-induced matrix remodeling (Figure 3). This promotes integrin Beta 1-FAK-Src signaling within melanoma cells
and reactivation of ERK signal transduction and BRAF inhibitor
resistance that can be circumvented by combined BRAF/FAK inhibition. Consistent with these data, fibronectin matrices (Hirata
et al., 2015) or fibronectin induction (Fedorenko et al., 2016) are
sufficient to circumvent the impact of BRAF inhibition on tumor
cells. Co-cultures of fibroblasts, stromal cells, and tumor cells
revealed stromal-derived HGF as a mediator of RAF inhibitor
resistance in BRAF mutant melanoma cells via activation of its
cognate receptor cMET (Straussman et al., 2012). Moreover,
BRAF inhibition results in TGF beta release from melanoma cells,
which promotes the differentiation of fibroblasts, fibronectin
expression, and HGF secretion, which together triggers PI3K/
AKT pathway activity (Fedorenko et al., 2016). Similar evidence
in the field of anti-angiogenic therapies implicates the stroma
in resistance to therapy through release of PDGF-C by cancerassociated fibroblasts (CAFs) (Crawford et al., 2009).
Just as genotoxic damage can foster the emergence of drug
resistant cells and histological transformation of the tumor into
a high-grade recurrence (Johnson et al., 2014; Kim et al.,
2015), the stroma can also be adversely affected by genotoxic
agents to support the survival of cancer cells in the face of
such selection pressures. Genotoxic therapy can promote the
microenvironmental secretion of WNT16B that reduces the
impact of cytotoxic therapy upon prostate cancer cells in vivo
(Sun et al., 2012). Genotoxic chemotherapy can also result in
secretion of IL-6 and TIMP-1 from thymic endothelium in a
mouse model of Burkitt’s lymphoma, which promotes the survival of minimal residual disease in the thymus (Gilbert and Hemann, 2010). Exploring this chemoresistant perivascular niche
phenomenon further, Hodivala-Dilke and colleagues demonstrated that following DNA damage, FAK loss from endothelial
cells sensitizes cancer cells to doxorubicin through the suppression of NF-kB-induced cytokine production from endothelial
cells and concomitant reduced phospho-STAT3 in tumor cells
following doxorubicin exposure, without having any measurable
impact on endothelial function (Tavora et al., 2014).
Finally, increasing evidence supports the ability of the immune
microenvironment to support the survival of tumor cells. MAPK
pathway inhibition increases macrophage infiltration into the
tumor microenvironment, promoting resistance to BRAF and
MEK inhibition through TNF alpha release from myeloid cells,
mediating the induction of the melanocytic-specific transcription
and survival factor, MITF (Smith et al., 2014).
Taken together, these data are consistent with the concept
that the microenvironment can foster a ‘‘safe haven’’ for the evolution of drug tolerant cells, providing an explanation as to how
cells survive in the period between initial tumor response and
disease progression. Conceivably, through this permissive
microenvironment, tumor cells may proceed to acquire genetically encoded drug-resistant mechanisms that might dominate
specific lesions at progression. Targeting the microenvironment
itself may therefore present an effective strategy to manage cancer clonal evolution. Using a humanized mouse model Bcl2/Myc
driven lymphoma, treated with alemtuzumab (anti-CD52 antibody), Hemann and colleagues demonstrated that infiltration of
resistant leukemia cells into the bone marrow rewires the tumor
microenvironment to inhibit engulfment of antibody-targeted
tumor cells. Combination therapy of cyclophosphamide with
the antibody eliminated residual disease by inducing a secretion
phenotype that increased infiltration of macrophages into the
bone marrow and enhanced phagocytic activity (Pallasch
et al., 2014).
Managing Clonal Evolution
Despite extensive clinical data documenting ITH and its clinical
relevance, drug development and novel clinical trial designs to
account for the dynamic evolution of tumors have lagged behind.
Longitudinal Sampling Strategies
Given the evolutionary capabilities of tumors, monitoring disease
evolution to guide therapeutic interventions and to understand
evolutionary trajectories of individual tumors has become a vital
research area. Serial sampling of tumor genomes from plasma
and circulating tumor cells is now increasingly implemented in
both clinical research settings to monitor cancer clonal evolution
and drug-resistance mechanisms over time (Haber and Velculescu, 2014), as well as the evolution of metastatic disease
Cell 168, February 9, 2017 621
(Carreira et al., 2014). Sequencing of circulating tumor DNA
(ctDNA) through therapy can reveal somatic mutations acquired
at resistance following cytotoxic and targeted therapies (Murtaza et al., 2013), the detection of which has higher sensitivity
and dynamic range than conventional blood based markers
(Dawson et al., 2013). Furthermore, an increase in ctDNA heralds
progressive disease in advance of conventional imaging approaches (Dawson et al., 2013).
Problems inherent to tumor sampling bias due to ITH may
be mitigated with ctDNA sampling (Jamal-Hanjani et al., 2016;
Russo et al., 2016; Siravegna et al., 2015). Moreover, such
methods can track mechanisms of resistance to targeted
agents, which may emerge during the course of therapy. Multiple
mutations in KRAS and NRAS were identified in the same patient
through BEAMing analysis of ctDNA, acquired during cetuximab
or panitumumab exposure in advanced colorectal cancer,
converging upon the reactivation of ERK signaling. These results
suggest that a rational strategy to limit acquired therapy resistance may be through dual blockade of both EGFR and MEK
signaling (Misale et al., 2012). Studies are also revealing the dynamic clonal evolution that occurs subsequent to the acquisition
of drug resistance. Bardelli and colleagues (Siravegna et al.,
2015) demonstrated the waning of KRAS mutant subclones
upon EGFR monoclonal antibody drug withdrawal, suggesting
fitness costs following the acquisition of KRAS mutations later
in tumor evolution and providing an explanation for further tumor
responses following drug re-challenge. However, a drawback
of ctDNA sampling strategies is that they cannot necessarily
provide an accurate portrayal of the copy-number state of the
cancer genome nor a detailed phylogeny, and there may be
over-representation of DNA from dying cells. In this respect,
circulating tumor cells may provide additional information.
Targeting Clonal Events
Targeting clonal events, present in every tumor cell, may present
an attractive model for drug development. Indeed, it is likely that
many targeted therapies that have successfully passed through
the drug development process, demonstrating robust progression free survival benefits in clinical trials, are targeting early
clonal events present at all sites of disease. However, even in
the context of a clonal driver, resistance to such therapies is
frequent in the advanced disease setting and may be driven
by the selection of resistance cancer cells present at low frequencies prior to therapy (Bhang et al., 2015; Su et al., 2012;
Turke et al., 2010) or may evolve through de novo mutations
that are acquired during therapy (Hata et al., 2016).
Modeling approaches have estimated that most lesions
identifiable through radiographic techniques harbor ten or
more resistant subclones (Bozic and Nowak, 2014). Combination therapy approaches that act through distinct pathways
may help circumvent this problem (Bozic et al., 2013). However,
in practice, the feasibility of such approaches may be complicated by the toxicity of combination therapy, as well as the
occurrence of mutations that confer resistance to multiple drugs.
Conceivably, vaccine or adoptive T cell therapy approaches
targeting multiple clonal neo-antigens may provide the specificity required to minimize normal tissue toxicity and maximize
tumor cell kill while minimizing the possibility for acquired drug
resistance to occur. We have recently found evidence for
622 Cell 168, February 9, 2017
CD8+PD1+ T cell populations recognizing clonal neo-antigens
present in all cells of a tumor (McGranahan et al., 2016). However, the extent to which tumors could circumvent such
strategies through chromosomal instability-driven loss of neoantigens remains unclear.
Attenuating or Exploiting Genome Instability
Genome instability acts as a fuel for cell-to-cell variation and,
hence, selection and evolution. The clinical relevance of genomic
instability is evidenced by the association of chromosomal instability with outcomes across multiple cancer types (McGranahan
et al., 2012) and accumulating evidence linking chromosomal
chaos with metastasis (for review, see Turajlic and Swanton,
2016). Targeting specific genome instability mechanisms may
provide a means to arrest tumor evolution and limit disease progression, particularly in the early disease.
The success of PARP inhibitors in the treatment of BRCA
mutant cancers exemplifies how genomically unstable cancers
can be targeted by elevating instability to lethal levels and exploiting synthetic lethality (for review, see Lord and Ashworth,
2016). However, even in this context, resistance can occur, for
example, through BRCA reversion, either directly through additional mutations to BRCA or indirectly through, for example,
inactivation of 53BP1 (Lord and Ashworth, 2016).
Maley and colleagues have explored the impact of non-steroidal anti-inflammatory agents (NSAIDs) upon the evolution of preinvasive Barrett’s esophagus to invasive esophageal cancer. In a
longitudinal analysis of 13 patients who had been exposed to
NSAIDs over a period of several years, the authors found
evidence that NSAID use was associated with a reduced rate
of somatic genomic abnormalities, as defined by SNP array analyses (Kostadinov et al., 2013).
Evolutionary studies are revealing distinct mutagenic processes that occur through the disease course. Evidence is
emerging in lung adenocarcinoma, bladder cancer, estrogen
receptor negative breast cancer, head and neck squamous carcinoma, and esophageal squamous carcinoma that APOBECinduced mutagenesis is enriched later in tumor evolution,
suggesting that efforts to target the cytidine deaminase family
may demonstrate utility to limit ongoing mutagenesis (de Bruin
et al., 2014; McGranahan et al., 2015). A recent study, using
clinical data and xenograft experiments, found evidence that
APOBEC3B can facilitate tamoxifen resistance in ER-positive
breast cancer (Law et al., 2016). Similarly, chemotherapy-resistant urothelial carcinomas exhibited an enrichment of
APOBEC3B mutations following therapy (Faltas et al., 2016). Inhibiting APOBEC3B may provide an effective strategy to
improve efficacies of cancer therapies by limiting the evolutionary potential of cancer cells.
Competitive Release and Adaptive Therapy
While benefits in progression-free survival times are commonly
reported in clinical trials, these rarely translate to equivalent clinically relevant overall survival benefits (Fojo et al., 2014). Notwithstanding the complexities of clinical trial design, the mismatch
between progression-free and overall survival times may reflect
clonal competition and competitive release. Conceivably, elimination of a dominant drug-sensitive clone in the investigational
arm might allow the competitive release of resistant subclones
to undergo accelerated growth in a resource-rich environment,
Figure 4. Competitive Release of Resistance Subclones
Similar overall survival times, yet divergent progression-free survival times,
between treated and un-treated patients may reflect competitive release of
aggressive subclones.
resulting in more rapid disease progression compared to the
control arm after the observed progression-free survival benefit
(Figure 4).
Thus, new trial concepts accounting for competitive release of
resistant subclones may maximize the overall survival benefits
with current therapies. Gatenby and colleagues have devised
approaches in animal models to exploit the fitness cost of resistant subclones by maintaining a stable population of sensitive
subclones, thereby restricting the growth of resistant cells (Enriquez-Navas et al., 2016). In contrast to standard clinical practice,
where the goals of therapy are to maximally reduce tumor
burden, the focus of adaptive therapy is to maximize time to progression by stabilizing tumor size (Enriquez-Navas et al., 2016;
Gatenby et al., 2009). Adaptive therapy requires variable drug
dosing and schedules in two phases: an induction phase, to control tumor progression from exponential growth, and a maintenance phase that might require progressively lower dosing or
even omitted schedules in order to achieve better progressionfree survival times compared to standard fixed dosing. In keeping with the benefits of an adaptive therapy approach, in the
context of patient-derived melanoma xenografts, Stuart and colleagues demonstrated how vemurafenib-resistant melanomas
can exhibit drug dependency, such that an intermittent rather
than continuous dosing of the drug can forestall the onset of lethal drug resistance (Das Thakur et al., 2013).
Exploiting Evolutionary Constraints
Traditional approaches to cancer management are primarily
reactive, focusing on the management of drug-resistant disease.
Pro-active management of cancers through attempts to predict
or attenuate a cancer’s next evolutionary move, exploiting evolutionary constraints or synthetic lethality, might be feasible as
knowledge of evolution across cancer types increases.
Emerging evidence in ccRCC suggests that the constraints
upon activation of the PI3K/mTOR pathway, manifested as
recurrent deleterious or activating mutations in PTEN, PIK3CA,
TSC1, or mTOR, might be exploitable for therapeutic benefit.
Voss and colleagues examined renal tumors from five patients
who had experienced a prolonged benefit from mTOR pathway
inhibition with everolimus or temsirolimus. Multi-region tumor
sampling revealed parallel evolution with distinct somatic mutations predicted to lead to activation of the mTOR pathway in
different tumor regions in three of the five cases (Voss et al.,
2014). These data suggest that targeting constraints to tumor
evolution might be practical if appropriate biomarker assays
could be developed that could detect parallel evolution leading
to signal transduction pathway convergence.
A further tractable approach may be derived from exploiting
iatrogenic evolutionary selection pressures. An approach termed
‘‘collateral sensitivity’’ leverages the phenomenon where resistance acquired to one drug comes at the expense of sensitivity
to another (Hill, 1986; Jensen et al., 1997). Hemann and colleagues have exploited such evolutionary constraints in a murine
model of Philadelphia chromosome-positive ALL (Zhao et al.,
2016). The authors described collateral sensitivity induced by
treatment with dasatinib, which resulted in the selection of the
acquired resistance BCR-ABL1 V299L mutation at intermediate
stages of evolution of Ph+ ALL cells. This rendered the cells sensitive to non-classical BCR-ABL inhibitors such as cabozantinib
and vandetanib.
Similar methods have been applied to the targeting of aneuploid populations. Rong Li and colleagues used an ‘‘evolutionary
trap’’ by reducing karyotypic heterogeneity to a defined predictable state through initial drug exposure, which can then by targeted by a secondary drug. Specifically, exposure of aneuploid
budding yeast to radicicol, an HSP90 inhibitor, results in the
selection of multiple copies of chromosome XV. Amplification
of chromosome XV results in resistance to radicicol; however,
it also engenders sensitivity to hygromycin B (Chen et al., 2015).
Evidence is emerging for the potential of similar strategies in
the clinical setting. Engelman and colleagues explored resistance mechanisms in a patient with ALK-rearranged NSCLC,
which harbored a subclonal C1156Y mutation in the kinase
domain, acquired following progression on crizotinib (Shaw
et al., 2016). Although the tumor did not benefit from a secondgeneration ALK inhibitor, it did respond to the third-generation
ALK inhibitor, lorlatinib. Following progression on lorlatinib, the
tumor acquired a L1198F mutation, which, together with the
pre-existing C1156Y alteration, prevented drug interaction with
the kinase. However, the L1198F mutation promoted re-sensitization to crizotinib, thereby resulting in improvement in the patient’s symptoms. This case study illustrates how evolutionary
constraints and collateral sensitivity can be exploited for patient
benefit.
Conclusions
Although our understanding of cancer genome evolution, and
the dynamic interplay between tumor cells and the microenvironment, has dramatically increased, the field of cancer evolutionary therapeutics is still in its infancy. As we attempt to forecast evolution and proactively manage a dynamic tumor
genome and its microenvironment, it is worth remembering
that Darwin recognized that such challenges ‘‘throw up a handful
of feathers, and all fall to the ground according to definite laws;
Cell 168, February 9, 2017 623
but how simple is the problem where each shall fall compared to
that of the action and reaction of innumerable plants and animals
which have determined, in the course of centuries, the proportional numbers and kinds of trees now growing’’ (Darwin,
1859). Predicting the innumerable interactions of cancer subclones with each other and the microenvironment is an equally
formidable task. Computational and technological advances,
coupled with prospective longitudinal studies exploring the cancer genome and the immune microenvironment, will be needed
to gain a deeper understanding of the evolutionary trajectories
of tumors and the extent to which a tumor’s next step may be
predicted. Such studies may also allow new insights into the
processes generating diversity and how constraints to tumor
evolution may be exploited, permitting proactive management
of cancers, which leverage an adaptive immune response.
ACKNOWLEDGMENTS
We thank Erik Sahai, Samra Turajlic and Gareth Wilson for their input and discussions regarding the manuscript. We thank Poppy Swanton and Kip Lyall for
artistic help with the figures. C.S. is a Royal Society Napier Research Professor. This work was supported by the Francis Crick Institute, which receives its
core funding from Cancer Research UK (FC001169), the UK Medical Research
Council (FC001169), and the Wellcome Trust (FC001169) and by the UK Medical Research Council (grant reference MR/FC001169/1). C.S. is funded by
Cancer Research UK (TRACERx), the CRUK Lung Cancer Centre of Excellence, Stand Up 2 Cancer (SU2C), the Rosetrees Trust, NovoNordisk Foundation (ID 16584), the Prostate Cancer Foundation, the Breast Cancer Research
Foundation, and the European Research Council (THESEUS). Support was
provided to C.S. by the National Institute for Health Research, the University
College London (UCL) Hospitals Biomedical Research Centre, and the Cancer
Research UK UCL Experimental Cancer Medicine Centre. C.S. reports
personal fees from Pfizer, Boehringer Ingelheim, Novartis, Celgene, Servier,
Eli Lily and Glazo Smithkline outside the submitted work, and stock options:
1. Achilles Therapeutics, 2. Epic Biosciences, 3. GRAIL, 4. Apogen Biotech.
REFERENCES
Aceto, N., Bardia, A., Miyamoto, D.T., Donaldson, M.C., Wittner, B.S.,
Spencer, J.A., Yu, M., Pely, A., Engstrom, A., Zhu, H., et al. (2014). Circulating
tumor cell clusters are oligoclonal precursors of breast cancer metastasis. Cell
158, 1110–1122.
Aerts, H.J., Velazquez, E.R., Leijenaar, R.T., Parmar, C., Grossmann, P.,
Carvalho, S., Bussink, J., Monshouwer, R., Haibe-Kains, B., Rietveld, D.,
et al. (2014). Decoding tumour phenotype by noninvasive imaging using a
quantitative radiomics approach. Nat. Commun. 5, 4006.
Alexandrov, L.B., Nik-Zainal, S., Wedge, D.C., Aparicio, S.A., Behjati, S., Biankin, A.V., Bignell, G.R., Bolli, N., Borg, A., Børresen-Dale, A.L., et al.; Australian
Pancreatic Cancer Genome Initiative; ICGC Breast Cancer Consortium; ICGC
MMML-Seq Consortium; ICGC PedBrain (2013). Signatures of mutational processes in human cancer. Nature 500, 415–421.
Alexandrov, L.B., Jones, P.H., Wedge, D.C., Sale, J.E., Campbell, P.J., NikZainal, S., and Stratton, M.R. (2015). Clock-like mutational processes in human somatic cells. Nat. Genet. 47, 1402–1407.
Anagnostou, V., Smith, K.N., Forde, P.M., Niknafs, N., Bhattacharya, R., White,
J., Zhang, T., Adleff, V., Phallen, J., Wali, N., et al. (2016). Evolution of Neoantigen Landscape During Immune Checkpoint Blockade in Non-Small Cell Lung
Cancer. Cancer Discov. Published online December 28, 2016. http://dx.doi.
org/10.1158/2159-8290.CD-16-0828.
Anderson, K., Lutz, C., van Delft, F.W., Bateman, C.M., Guo, Y., Colman, S.M.,
Kempski, H., Moorman, A.V., Titley, I., Swansbury, J., et al. (2011). Genetic
variegation of clonal architecture and propagating cells in leukaemia. Nature
469, 356–361.
624 Cell 168, February 9, 2017
Andor, N., Graham, T.A., Jansen, M., Xia, L.C., Aktipis, C.A., Petritsch, C., Ji,
H.P., and Maley, C.C. (2016). Pan-cancer analysis of the extent and consequences of intratumor heterogeneity. Nat. Med. 22, 105–113.
Aryee, M.J., Liu, W., Engelmann, J.C., Nuhn, P., Gurel, M., Haffner, M.C.,
Esopi, D., Irizarry, R.A., Getzenberg, R.H., Nelson, W.G., et al. (2013). DNA
methylation alterations exhibit intraindividual stability and interindividual heterogeneity in prostate cancer metastases. Sci. Transl. Med. 5, 169ra10.
Bashashati, A., Ha, G., Tone, A., Ding, J., Prentice, L.M., Roth, A., Rosner, J.,
Shumansky, K., Kalloger, S., Senz, J., et al. (2013). Distinct evolutionary trajectories of primary high-grade serous ovarian cancers revealed through spatial
mutational profiling. J. Pathol. 231, 21–34.
Bettegowda, C., Sausen, M., Leary, R.J., Kinde, I., Wang, Y., Agrawal, N.,
Bartlett, B.R., Wang, H., Luber, B., Alani, R.M., et al. (2014). Detection of
circulating tumor DNA in early- and late-stage human malignancies. Sci.
Transl. Med. 6, 224ra24.
Bhang, H.E., Ruddy, D.A., Krishnamurthy Radhakrishna, V., Caushi, J.X.,
Zhao, R., Hims, M.M., Singh, A.P., Kao, I., Rakiec, D., Shaw, P., et al.
(2015). Studying clonal dynamics in response to cancer therapy using highcomplexity barcoding. Nat. Med. 21, 440–448.
Birkbak, N.J., Eklund, A.C., Li, Q., McClelland, S.E., Endesfelder, D., Tan, P.,
Tan, I.B., Richardson, A.L., Szallasi, Z., and Swanton, C. (2011). Paradoxical
relationship between chromosomal instability and survival outcome in cancer.
Cancer Res. 71, 3447–3452.
Bolli, N., Avet-Loiseau, H., Wedge, D.C., Van Loo, P., Alexandrov, L.B.,
Martincorena, I., Dawson, K.J., Iorio, F., Nik-Zainal, S., Bignell, G.R., et al.
(2014). Heterogeneity of genomic evolution and mutational profiles in multiple
myeloma. Nat. Commun. 5, 2997.
Bozic, I., and Nowak, M.A. (2014). Timing and heterogeneity of mutations
associated with drug resistance in metastatic cancers. Proc. Natl. Acad. Sci.
USA 111, 15964–15968.
Bozic, I., Reiter, J.G., Allen, B., Antal, T., Chatterjee, K., Shah, P., Moon, Y.S.,
Yaqubie, A., Kelly, N., Le, D.T., et al. (2013). Evolutionary dynamics of cancer in
response to targeted combination therapy. eLife 2, e00747.
Brastianos, P.K., Carter, S.L., Santagata, S., Cahill, D.P., Taylor-Weiner, A.,
Jones, R.T., Van Allen, E.M., Lawrence, M.S., Horowitz, P.M., Cibulskis, K.,
et al. (2015). Genomic Characterization of Brain Metastases Reveals Branched
Evolution and Potential Therapeutic Targets. Cancer Discov. 5, 1164–1177.
Calbo, J., van Montfort, E., Proost, N., van Drunen, E., Beverloo, H.B., Meuwissen, R., and Berns, A. (2011). A functional role for tumor cell heterogeneity in a
mouse model of small cell lung cancer. Cancer Cell 19, 244–256.
Campbell, P.J., Yachida, S., Mudie, L.J., Stephens, P.J., Pleasance, E.D.,
Stebbings, L.A., Morsberger, L.A., Latimer, C., McLaren, S., Lin, M.L., et al.
(2010). The patterns and dynamics of genomic instability in metastatic pancreatic cancer. Nature 467, 1109–1113.
Carreira, S., Romanel, A., Goodall, J., Grist, E., Ferraldeschi, R., Miranda, S.,
Prandi, D., Lorente, D., Frenel, J.S., Pezaro, C., et al. (2014). Tumor clone dynamics in lethal prostate cancer. Sci. Transl. Med. 6, 254ra125.
Carter, S.L., Cibulskis, K., Helman, E., McKenna, A., Shen, H., Zack, T., Laird,
P.W., Onofrio, R.C., Winckler, W., Weir, B.A., et al. (2012). Absolute quantification of somatic DNA alterations in human cancer. Nat. Biotechnol. 30,
413–421.
Chen, H., and He, X. (2016). The Convergent Cancer Evolution toward a Single
Cellular Destination. Mol. Biol. Evol. 33, 4–12.
Chen, G., Mulla, W.A., Kucharavy, A., Tsai, H.J., Rubinstein, B., Conkright, J.,
McCroskey, S., Bradford, W.D., Weems, L., Haug, J.S., et al. (2015). Targeting
the adaptability of heterogeneous aneuploids. Cell 160, 771–784.
Cheung, K.J., Padmanaban, V., Silvestri, V., Schipper, K., Cohen, J.D., Fairchild, A.N., Gorin, M.A., Verdone, J.E., Pienta, K.J., Bader, J.S., and Ewald,
A.J. (2016). Polyclonal breast cancer metastases arise from collective dissemination of keratin 14-expressing tumor cell clusters. Proc. Natl. Acad. Sci. USA
113, E854–E863.
Cleary, A.S., Leonard, T.L., Gestl, S.A., and Gunther, E.J. (2014). Tumour cell
heterogeneity maintained by cooperating subclones in Wnt-driven mammary
cancers. Nature 508, 113–117.
Conway Morris, S. (2003). Life’s solution: inevtiable humans in a lonely universe (New York: Cambridge University Press).
Crawford, Y., Kasman, I., Yu, L., Zhong, C., Wu, X., Modrusan, Z., Kaminker,
J., and Ferrara, N. (2009). PDGF-C mediates the angiogenic and tumorigenic
properties of fibroblasts associated with tumors refractory to anti-VEGF treatment. Cancer Cell 15, 21–34.
Darwin, C. (1859). On the Origin of Species by Means of Natural Selection, Or,
The Preservation of Favoured Races in the Struggle for Life (J. Murray).
Das Thakur, M., Salangsang, F., Landman, A.S., Sellers, W.R., Pryer, N.K.,
Levesque, M.P., Dummer, R., McMahon, M., and Stuart, D.D. (2013). Modelling vemurafenib resistance in melanoma reveals a strategy to forestall drug
resistance. Nature 494, 251–255.
Dawson, S.J., Tsui, D.W., Murtaza, M., Biggs, H., Rueda, O.M., Chin, S.F.,
Dunning, M.J., Gale, D., Forshew, T., Mahler-Araujo, B., et al. (2013). Analysis
of circulating tumor DNA to monitor metastatic breast cancer. N. Engl. J. Med.
368, 1199–1209.
de Bruin, E.C., McGranahan, N., Mitter, R., Salm, M., Wedge, D.C., Yates, L.,
Jamal-Hanjani, M., Shafi, S., Murugaesu, N., Rowan, A.J., et al. (2014). Spatial
and temporal diversity in genomic instability processes defines lung cancer
evolution. Science 346, 251–256.
Deshwar, A.G., Vembu, S., Yung, C.K., Jang, G.H., Stein, L., and Morris, Q.
(2015). PhyloWGS: reconstructing subclonal composition and evolution from
whole-genome sequencing of tumors. Genome Biol. 16, 35.
Dewhurst, S.M., McGranahan, N., Burrell, R.A., Rowan, A.J., Grönroos, E.,
Endesfelder, D., Joshi, T., Mouradov, D., Gibbs, P., Ward, R.L., et al. (2014).
Tolerance of whole-genome doubling propagates chromosomal instability
and accelerates cancer genome evolution. Cancer Discov. 4, 175–185.
Driessens, G., Beck, B., Caauwe, A., Simons, B.D., and Blanpain, C. (2012).
Defining the mode of tumour growth by clonal analysis. Nature 488, 527–530.
Eckert, M.A., Pan, S., Hernandez, K.M., Loth, R.M., Andrade, J., Volchenboum, S.L., Faber, P., Montag, A., Lastra, R., Peter, M.E., et al. (2016). Genomics of Ovarian Cancer Progression Reveals Diverse Metastatic Trajectories
Including Intraepithelial Metastasis to the Fallopian Tube. Cancer Discov. 6,
1342–1351.
Eleveld, T.F., Oldridge, D.A., Bernard, V., Koster, J., Daage, L.C., Diskin, S.J.,
Schild, L., Bentahar, N.B., Bellini, A., Chicard, M., et al. (2015). Relapsed
neuroblastomas show frequent RAS-MAPK pathway mutations. Nat. Genet.
47, 864–871.
Enriquez-Navas, P.M., Kam, Y., Das, T., Hassan, S., Silva, A., Foroutan, P.,
Ruiz, E., Martinez, G., Minton, S., Gillies, R.J., and Gatenby, R.A. (2016). Exploiting evolutionary principles to prolong tumor control in preclinical models
of breast cancer. Sci. Transl. Med. 8, 327ra24.
Faltas, B.M., Prandi, D., Tagawa, S.T., Molina, A.M., Nanus, D.M., Sternberg,
C., Rosenberg, J., Mosquera, J.M., Robinson, B., Elemento, O., et al. (2016).
Clonal evolution of chemotherapy-resistant urothelial carcinoma. Nat. Genet.
48, 1490–1499.
Fedorenko, I.V., Abel, E.V., Koomen, J.M., Fang, B., Wood, E.R., Chen, Y.A.,
Fisher, K.J., Iyengar, S., Dahlman, K.B., Wargo, J.A., et al. (2016). Fibronectin
induction abrogates the BRAF inhibitor response of BRAF V600E/PTEN-null
melanoma cells. Oncogene 35, 1225–1235.
Fischer, A., Vázquez-Garcı́a, I., Illingworth, C.J., and Mustonen, V. (2014).
High-definition reconstruction of clonal composition in cancer. Cell Rep. 7,
1740–1752.
Fisher, R., Horswell, S., Rowan, A., Salm, M.P., de Bruin, E.C., Gulati, S.,
McGranahan, N., Stares, M., Gerlinger, M., Varela, I., et al. (2014). Development of synchronous VHL syndrome tumors reveals contingencies and constraints to tumor evolution. Genome Biol. 15, 433.
Fojo, T., Mailankody, S., and Lo, A. (2014). Unintended consequences of
expensive cancer therapeutics—the pursuit of marginal indications and a
me-too mentality that stifles innovation and creativity: the John Conley Lecture. JAMA Otolaryngol. Head Neck Surg. 140, 1225–1236.
Francis, J.M., Zhang, C.Z., Maire, C.L., Jung, J., Manzo, V.E., Adalsteinsson,
V.A., Homer, H., Haidar, S., Blumenstiel, B., Pedamallu, C.S., et al. (2014).
EGFR variant heterogeneity in glioblastoma resolved through single-nucleus
sequencing. Cancer Discov. 4, 956–971.
Fu, F., Nowak, M.A., and Bonhoeffer, S. (2015). Spatial heterogeneity in drug
concentrations can facilitate the emergence of resistance to cancer therapy.
PLoS Comput. Biol. 11, e1004142.
Fujiwara, T., Bandi, M., Nitta, M., Ivanova, E.V., Bronson, R.T., and Pellman, D.
(2005). Cytokinesis failure generating tetraploids promotes tumorigenesis in
p53-null cells. Nature 437, 1043–1047.
Gao, R., Davis, A., McDonald, T.O., Sei, E., Shi, X., Wang, Y., Tsai, P.C., Casasent, A., Waters, J., Zhang, H., et al. (2016). Punctuated copy number evolution
and clonal stasis in triple-negative breast cancer. Nat. Genet. 48, 1119–1130.
Gatenby, R.A., Silva, A.S., Gillies, R.J., and Frieden, B.R. (2009). Adaptive therapy. Cancer Res. 69, 4894–4903.
Gerlinger, M., Rowan, A.J., Horswell, S., Larkin, J., Endesfelder, D., Gronroos,
E., Martinez, P., Matthews, N., Stewart, A., Tarpey, P., et al. (2012). Intratumor
heterogeneity and branched evolution revealed by multiregion sequencing.
N. Engl. J. Med. 366, 883–892.
Gerlinger, M., Horswell, S., Larkin, J., Rowan, A.J., Salm, M.P., Varela, I.,
Fisher, R., McGranahan, N., Matthews, N., Santos, C.R., et al. (2014a).
Genomic architecture and evolution of clear cell renal cell carcinomas defined
by multiregion sequencing. Nat. Genet. 46, 225–233.
Gerlinger, M., McGranahan, N., Dewhurst, S.M., Burrell, R.A., Tomlinson, I.,
and Swanton, C. (2014b). Cancer: evolution within a lifetime. Annu. Rev.
Genet. 48, 215–236.
Gilbert, L.A., and Hemann, M.T. (2010). DNA damage-mediated induction of a
chemoresistant niche. Cell 143, 355–366.
Goldschmidt, R. (1982). The Material Basis of Evolution: Reissued, Second
Edition (Yale University).
Gould, S.J. (2000). Wonderful life: the Burgess Shale and the nature of history
(Random House).
Greaves, M. (2015). Evolutionary determinants of cancer. Cancer Discov. 5,
806–820.
Gundem, G., Van Loo, P., Kremeyer, B., Alexandrov, L.B., Tubio, J.M.,
Papaemmanuil, E., Brewer, D.S., Kallio, H.M., Högnäs, G., Annala, M., et al.;
ICGC Prostate UK Group (2015). The evolutionary history of lethal metastatic
prostate cancer. Nature 520, 353–357.
Ha, G., Roth, A., Khattra, J., Ho, J., Yap, D., Prentice, L.M., Melnyk, N.,
McPherson, A., Bashashati, A., Laks, E., et al. (2014). TITAN: inference of
copy number architectures in clonal cell populations from tumor wholegenome sequence data. Genome Res. 24, 1881–1893.
Haber, D.A., and Velculescu, V.E. (2014). Blood-based analyses of cancer:
circulating tumor cells and circulating tumor DNA. Cancer Discov. 4, 650–661.
Hao, J.J., Lin, D.C., Dinh, H.Q., Mayakonda, A., Jiang, Y.Y., Chang, C., Jiang,
Y., Lu, C.C., Shi, Z.Z., Xu, X., et al. (2016). Spatial intratumoral heterogeneity
and temporal clonal evolution in esophageal squamous cell carcinoma. Nat.
Genet. 48, 1500–1507.
Harbst, K., Lauss, M., Cirenajwis, H., Isaksson, K., Rosengren, F., Törngren, T.,
Kvist, A., Johansson, M.C., Vallon-Christersson, J., Baldetorp, B., et al. (2016).
Multiregion Whole-Exome Sequencing Uncovers the Genetic Evolution and
Mutational Heterogeneity of Early-Stage Metastatic Melanoma. Cancer Res.
76, 4765–4774.
Hartkopf, A.D., Stefanescu, D., Wallwiener, M., Hahn, M., Becker, S.,
Solomayer, E.F., Fehm, T.N., Brucker, S.Y., and Taran, F.A. (2014). Tumor
cell dissemination to the bone marrow and blood is associated with poor
outcome in patients with metastatic breast cancer. Breast Cancer Res. Treat.
147, 345–351.
Hata, A.N., Niederst, M.J., Archibald, H.L., Gomez-Caraballo, M., Siddiqui,
F.M., Mulvey, H.E., Maruvka, Y.E., Ji, F., Bhang, H.E., Krishnamurthy Radhakrishna, V., et al. (2016). Tumor cells can follow distinct evolutionary paths to
Cell 168, February 9, 2017 625
become resistant to epidermal growth factor receptor inhibition. Nat. Med. 22,
262–269.
vidual and limited intraindividual genomic diversity among tumors from men
with metastatic prostate cancer. Nat. Med. 22, 369–378.
Hill, B.T. (1986). Potential of continuous tumour cell lines for establishing patterns of cross-resistance and collateral sensitivity in vitro. Drugs Exp. Clin.
Res. 12, 293–298.
Lalonde, E., Ishkanian, A.S., Sykes, J., Fraser, M., Ross-Adams, H., Erho, N.,
Dunning, M.J., Halim, S., Lamb, A.D., Moon, N.C., et al. (2014). Tumour
genomic and microenvironmental heterogeneity for integrated prediction of
5-year biochemical recurrence of prostate cancer: a retrospective cohort
study. Lancet Oncol. 15, 1521–1532.
Hirata, E., Girotti, M.R., Viros, A., Hooper, S., Spencer-Dene, B., Matsuda, M.,
Larkin, J., Marais, R., and Sahai, E. (2015). Intravital imaging reveals how BRAF
inhibition generates drug-tolerant microenvironments with high integrin b1/
FAK signaling. Cancer Cell 27, 574–588.
Hobor, S., Van Emburgh, B.O., Crowley, E., Misale, S., Di Nicolantonio, F., and
Bardelli, A. (2014). TGFa and amphiregulin paracrine network promotes
resistance to EGFR blockade in colorectal cancer cells. Clin. Cancer Res.
20, 6429–6438.
Humphries, A., Cereser, B., Gay, L.J., Miller, D.S., Das, B., Gutteridge, A., Elia,
G., Nye, E., Jeffery, R., Poulsom, R., et al. (2013). Lineage tracing reveals
multipotent stem cells maintain human adenomas and the pattern of clonal
expansion in tumor evolution. Proc. Natl. Acad. Sci. USA 110, E2490–E2499.
Huxley, J. (1958). The Biological Aspects of Cancer (New York: Harcourt, Brace).
Inda, M.M., Bonavia, R., Mukasa, A., Narita, Y., Sah, D.W., Vandenberg, S.,
Brennan, C., Johns, T.G., Bachoo, R., Hadwiger, P., et al. (2010). Tumor
heterogeneity is an active process maintained by a mutant EGFR-induced
cytokine circuit in glioblastoma. Genes Dev. 24, 1731–1745.
Jamal-Hanjani, M., Wilson, G.A., Horswell, S., Mitter, R., Sakarya, O.,
Constantin, T., Salari, R., Kirkizlar, E., Sigurjonsson, S., Pelham, R., et al.
(2016). Detection of Ubiquitous and Heterogeneous Mutations in Cell-Free
DNA from Patients with Early-Stage Non-Small-Cell Lung Cancer. Ann. Oncol.
27, 862–867.
Jensen, P.B., Holm, B., Sorensen, M., Christensen, I.J., and Sehested, M.
(1997). In vitro cross-resistance and collateral sensitivity in seven resistant
small-cell lung cancer cell lines: preclinical identification of suitable drug partners to taxotere, taxol, topotecan and gemcitabin. Br. J. Cancer 75, 869–877.
Jiang, Y., Qiu, Y., Minn, A.J., and Zhang, N.R. (2016). Assessing intratumor
heterogeneity and tracking longitudinal and spatial clonal evolutionary
history by next-generation sequencing. Proc. Natl. Acad. Sci. USA 113,
E5528–E5537.
Johnson, B.E., Mazor, T., Hong, C., Barnes, M., Aihara, K., McLean, C.Y.,
Fouse, S.D., Yamamoto, S., Ueda, H., Tatsuno, K., et al. (2014). Mutational
analysis reveals the origin and therapy-driven evolution of recurrent glioma.
Science 343, 189–193.
Ju, Y.S., Alexandrov, L.B., Gerstung, M., Martincorena, I., Nik-Zainal, S., Ramakrishna, M., Davies, H.R., Papaemmanuil, E., Gundem, G., Shlien, A., et al.;
ICGC Breast Cancer Group; ICGC Chronic Myeloid Disorders Group; ICGC
Prostate Cancer Group (2014). Origins and functional consequences of somatic mitochondrial DNA mutations in human cancer. eLife 3, e02935.
Lamy, P., Nordentoft, I., Birkenkamp-Demtröder, K., Thomsen, M.B., Villesen,
P., Vang, S., Hedegaard, J., Borre, M., Jensen, J.B., Høyer, S., et al. (2016).
Paired Exome Analysis Reveals Clonal Evolution and Potential Therapeutic
Targets in Urothelial Carcinoma. Cancer Res. 76, 5894–5906.
Landau, D.A., Carter, S.L., Stojanov, P., McKenna, A., Stevenson, K.,
Lawrence, M.S., Sougnez, C., Stewart, C., Sivachenko, A., Wang, L., et al.
(2013). Evolution and impact of subclonal mutations in chronic lymphocytic
leukemia. Cell 152, 714–726.
Law, E.K., Sieuwerts, A.M., LaPara, K., Leonard, B., Starrett, G.J., Molan,
A.M., Temiz, N.A., Vogel, R.I., Meijer-van Gelder, M.E., Sweep, F.C., et al.
(2016). The DNA cytosine deaminase APOBEC3B promotes tamoxifen resistance in ER-positive breast cancer. Sci. Adv. 2, e1601737.
Le, D.T., Uram, J.N., Wang, H., Bartlett, B.R., Kemberling, H., Eyring, A.D.,
Skora, A.D., Luber, B.S., Azad, N.S., Laheru, D., et al. (2015). PD-1 Blockade
in Tumors with Mismatch-Repair Deficiency. N. Engl. J. Med. 372, 2509–2520.
Ling, S., Hu, Z., Yang, Z., Yang, F., Li, Y., Lin, P., Chen, K., Dong, L., Cao, L.,
Tao, Y., et al. (2015). Extremely high genetic diversity in a single tumor points
to prevalence of non-Darwinian cell evolution. Proc. Natl. Acad. Sci. USA 112,
E6496–E6505.
Lipinski, K.A., Barber, L.J., Davies, M.N., Ashenden, M., Sottoriva, A., and Gerlinger, M. (2016). Cancer Evolution and the Limits of Predictability in Precision
Cancer Medicine. Trends Cancer 2, 49–63.
Lord, C.J., and Ashworth, A. (2016). BRCAness revisited. Nat. Rev. Cancer 16,
110–120.
Maddipati, R., and Stanger, B.Z. (2015). Pancreatic Cancer Metastases Harbor Evidence of Polyclonality. Cancer Discov. 5, 1086–1097.
Makohon-Moore, A.P., Zhang, M., Reiter, J.G., Bozic, I., Allen, B., Kundu, D.,
Chatterjee, K., Wong, F., Jiao, Y., Kohutek, Z.A., et al. (2017). Limited heterogeneity of known driver gene mutations among the metastases of individual
patients with pancreatic cancer. Nat. Genet. Published online January 16,
2017. http://dx.doi.org/10.1038/ng.3764.
Maley, C.C., Galipeau, P.C., Finley, J.C., Wongsurawat, V.J., Li, X., Sanchez,
C.A., Paulson, T.G., Blount, P.L., Risques, R.A., Rabinovitch, P.S., and Reid,
B.J. (2006). Genetic clonal diversity predicts progression to esophageal
adenocarcinoma. Nat. Genet. 38, 468–473.
Junttila, M.R., and de Sauvage, F.J. (2013). Influence of tumour micro-environment heterogeneity on therapeutic response. Nature 501, 346–354.
Malhotra, A., Lindberg, M., Faust, G.G., Leibowitz, M.L., Clark, R.A., Layer,
R.M., Quinlan, A.R., and Hall, I.M. (2013). Breakpoint profiling of 64 cancer genomes reveals numerous complex rearrangements spawned by homology-independent mechanisms. Genome Res. 23, 762–776.
Juric, D., Castel, P., Griffith, M., Griffith, O.L., Won, H.H., Ellis, H., Ebbesen,
S.H., Ainscough, B.J., Ramu, A., Iyer, G., et al. (2015). Convergent loss of
PTEN leads to clinical resistance to a PI(3)Ka inhibitor. Nature 518, 240–244.
Mao, Q., Zhang, Y., Fu, X., Xue, J., Guo, W., Meng, M., Zhou, Z., Mo, X., and
Lu, Y. (2013). A tumor hypoxic niche protects human colon cancer stem cells
from chemotherapy. J. Cancer Res. Clin. Oncol. 139, 211–222.
Kim, H., Zheng, S., Amini, S.S., Virk, S.M., Mikkelsen, T., Brat, D.J., Grimsby,
J., Sougnez, C., Muller, F., Hu, J., et al. (2015). Whole-genome and multisector
exome sequencing of primary and post-treatment glioblastoma reveals patterns of tumor evolution. Genome Res. 25, 316–327.
Martinez, P., Birkbak, N.J., Gerlinger, M., McGranahan, N., Burrell, R.A.,
Rowan, A.J., Joshi, T., Fisher, R., Larkin, J., Szallasi, Z., and Swanton, C.
(2013). Parallel evolution of tumour subclones mimics diversity between tumours. J. Pathol. 230, 356–364.
Koebel, C.M., Vermi, W., Swann, J.B., Zerafa, N., Rodig, S.J., Old, L.J., Smyth,
M.J., and Schreiber, R.D. (2007). Adaptive immunity maintains occult cancer in
an equilibrium state. Nature 450, 903–907.
Marusyk, A., Tabassum, D.P., Altrock, P.M., Almendro, V., Michor, F., and
Polyak, K. (2014). Non-cell-autonomous driving of tumour growth supports
sub-clonal heterogeneity. Nature 514, 54–58.
Kostadinov, R.L., Kuhner, M.K., Li, X., Sanchez, C.A., Galipeau, P.C., Paulson,
T.G., Sather, C.L., Srivastava, A., Odze, R.D., Blount, P.L., et al. (2013).
NSAIDs modulate clonal evolution in Barrett’s esophagus. PLoS Genet. 9,
e1003553.
Mayers, J.R., Torrence, M.E., Danai, L.V., Papagiannakopoulos, T., Davidson,
S.M., Bauer, M.R., Lau, A.N., Ji, B.W., Dixit, P.D., Hosios, A.M., et al. (2016).
Tissue of origin dictates branched-chain amino acid metabolism in mutant
Kras-driven cancers. Science 353, 1161–1165.
Kumar, A., Coleman, I., Morrissey, C., Zhang, X., True, L.D., Gulati, R., Etzioni,
R., Bolouri, H., Montgomery, B., White, T., et al. (2016). Substantial interindi-
Mazor, T., Pankov, A., Johnson, B.E., Hong, C., Hamilton, E.G., Bell, R.J.,
Smirnov, I.V., Reis, G.F., Phillips, J.J., Barnes, M.J., et al. (2015). DNA
626 Cell 168, February 9, 2017
Methylation and Somatic Mutations Converge on the Cell Cycle and Define
Similar Evolutionary Histories in Brain Tumors. Cancer Cell 28, 307–317.
(2015). Tracking the genomic evolution of esophageal adenocarcinoma
through neoadjuvant chemotherapy. Cancer Discov. 5, 821–831.
Mazor, T., Pankov, A., Song, J.S., and Costello, J.F. (2016). Intratumoral Heterogeneity of the Epigenome. Cancer Cell 29, 440–451.
Nadeu, F., Delgado, J., Royo, C., Baumann, T., Stankovic, T., Pinyol, M., Jares,
P., Navarro, A., Martı́n-Garcı́a, D., Beà, S., et al. (2016). Clinical impact of
clonal and subclonal TP53, SF3B1, BIRC3, NOTCH1, and ATM mutations in
chronic lymphocytic leukemia. Blood 127, 2122–2130.
McDermott, D.H., Gao, J.L., Liu, Q., Siwicki, M., Martens, C., Jacobs, P.,
Velez, D., Yim, E., Bryke, C.R., Hsu, N., et al. (2015). Chromothriptic cure of
WHIM syndrome. Cell 160, 686–699.
McGranahan, N., Burrell, R.A., Endesfelder, D., Novelli, M.R., and Swanton, C.
(2012). Cancer chromosomal instability: therapeutic and diagnostic challenges. EMBO Rep. 13, 528–538.
McGranahan, N., Favero, F., de Bruin, E.C., Birkbak, N.J., Szallasi, Z., and
Swanton, C. (2015). Clonal status of actionable driver events and the timing
of mutational processes in cancer evolution. Sci. Transl. Med. 7, 283ra54.
McGranahan, N., Furness, A.J., Rosenthal, R., Ramskov, S., Lyngaa, R., Saini,
S.K., Jamal-Hanjani, M., Wilson, G.A., Birkbak, N.J., Hiley, C.T., et al. (2016).
Clonal neoantigens elicit T cell immunoreactivity and sensitivity to immune
checkpoint blockade. Science 351, 1463–1469.
McPherson, A., Roth, A., Laks, E., Masud, T., Bashashati, A., Zhang, A.W., Ha,
G., Biele, J., Yap, D., Wan, A., et al. (2016). Divergent modes of clonal spread
and intraperitoneal mixing in high-grade serous ovarian cancer. Nat. Genet.
48, 758–767.
Medema, J.P., de Jong, J., van Hall, T., Melief, C.J., and Offringa, R. (1999).
Immune escape of tumors in vivo by expression of cellular FLICE-inhibitory
protein. J. Exp. Med. 190, 1033–1038.
Mehine, M., Kaasinen, E., Mäkinen, N., Katainen, R., Kämpjärvi, K., Pitkänen,
E., Heinonen, H.R., Bützow, R., Kilpivaara, O., Kuosmanen, A., et al. (2013).
Characterization of uterine leiomyomas by whole-genome sequencing.
N. Engl. J. Med. 369, 43–53.
Melchor, L., Brioli, A., Wardell, C.P., Murison, A., Potter, N.E., Kaiser, M.F.,
Fryer, R.A., Johnson, D.C., Begum, D.B., Hulkki Wilson, S., et al. (2014). Single-cell genetic analysis reveals the composition of initiating clones and phylogenetic patterns of branching and parallel evolution in myeloma. Leukemia 28,
1705–1715.
Merlo, L.M., Shah, N.A., Li, X., Blount, P.L., Vaughan, T.L., Reid, B.J., and
Maley, C.C. (2010). A comprehensive survey of clonal diversity measures in
Barrett’s esophagus as biomarkers of progression to esophageal adenocarcinoma. Cancer Prev. Res. (Phila.) 3, 1388–1397.
Miller, C.A., White, B.S., Dees, N.D., Griffith, M., Welch, J.S., Griffith, O.L., Vij,
R., Tomasson, M.H., Graubert, T.A., Walter, M.J., et al. (2014). SciClone: inferring clonal architecture and tracking the spatial and temporal patterns of tumor
evolution. PLoS Comput. Biol. 10, e1003665.
Misale, S., Yaeger, R., Hobor, S., Scala, E., Janakiraman, M., Liska, D., Valtorta, E., Schiavo, R., Buscarino, M., Siravegna, G., et al. (2012). Emergence
of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal
cancer. Nature 486, 532–536.
Natrajan, R., Sailem, H., Mardakheh, F.K., Arias Garcia, M., Tape, C.J., Dowsett, M., Bakal, C., and Yuan, Y. (2016). Microenvironmental Heterogeneity Parallels Breast Cancer Progression: A Histology-Genomic Integration Analysis.
PLoS Med. 13, e1001961.
Nik-Zainal, S., Alexandrov, L.B., Wedge, D.C., Van Loo, P., Greenman, C.D.,
Raine, K., Jones, D., Hinton, J., Marshall, J., Stebbings, L.A., et al.; Breast Cancer Working Group of the International Cancer Genome Consortium (2012).
Mutational processes molding the genomes of 21 breast cancers. Cell 149,
979–993.
Nikbakht, H., Panditharatna, E., Mikael, L.G., Li, R., Gayden, T., Osmond, M.,
Ho, C.Y., Kambhampati, M., Hwang, E.I., Faury, D., et al. (2016). Spatial and
temporal homogeneity of driver mutations in diffuse intrinsic pontine glioma.
Nat. Commun. 7, 11185.
Nones, K., Waddell, N., Wayte, N., Patch, A.M., Bailey, P., Newell, F., Holmes,
O., Fink, J.L., Quinn, M.C., Tang, Y.H., et al. (2014). Genomic catastrophes
frequently arise in esophageal adenocarcinoma and drive tumorigenesis.
Nat. Commun. 5, 5224.
Notta, F., Chan-Seng-Yue, M., Lemire, M., Li, Y., Wilson, G.W., Connor, A.A.,
Denroche, R.E., Liang, S.B., Brown, A.M., Kim, J.C., et al. (2016). A renewed
model of pancreatic cancer evolution based on genomic rearrangement patterns. Nature 538, 378–382.
Nowell, P.C. (1976). The clonal evolution of tumor cell populations. Science
194, 23–28.
Okosun, J., Bödör, C., Wang, J., Araf, S., Yang, C.Y., Pan, C., Boller, S.,
Cittaro, D., Bozek, M., Iqbal, S., et al. (2014). Integrated genomic analysis identifies recurrent mutations and evolution patterns driving the initiation and progression of follicular lymphoma. Nat. Genet. 46, 176–181.
Ortmann, C.A., Kent, D.G., Nangalia, J., Silber, Y., Wedge, D.C., Grinfeld, J.,
Baxter, E.J., Massie, C.E., Papaemmanuil, E., Menon, S., et al. (2015). Effect
of mutation order on myeloproliferative neoplasms. N. Engl. J. Med. 372,
601–612.
Pallasch, C.P., Leskov, I., Braun, C.J., Vorholt, D., Drake, A., Soto-Feliciano,
Y.M., Bent, E.H., Schwamb, J., Iliopoulou, B., Kutsch, N., et al. (2014). Sensitizing protective tumor microenvironments to antibody-mediated therapy. Cell
156, 590–602.
Park, S., and Lehner, B. (2015). Cancer type-dependent genetic interactions
between cancer driver alterations indicate plasticity of epistasis across cell
types. Mol. Syst. Biol. 11, 824.
Monajemi, M., Woodworth, C.F., Zipperlen, K., Gallant, M., Grant, M.D., and
Larijani, M. (2014). Positioning of APOBEC3G/F mutational hotspots in the
human immunodeficiency virus genome favors reduced recognition by
CD8+ T cells. PLoS ONE 9, e93428.
Pearson, A., Smyth, E., Babina, I.S., Herrera-Abreu, M.T., Tarazona, N., Peckitt, C., Kilgour, E., Smith, N.R., Geh, C., Rooney, C., et al. (2016). High-Level
Clonal FGFR Amplification and Response to FGFR Inhibition in a Translational
Clinical Trial. Cancer Discov. 6, 838–851.
Morrison, C.D., Liu, P., Woloszynska-Read, A., Zhang, J., Luo, W., Qin, M.,
Bshara, W., Conroy, J.M., Sabatini, L., Vedell, P., et al. (2014). Whole-genome
sequencing identifies genomic heterogeneity at a nucleotide and chromosomal level in bladder cancer. Proc. Natl. Acad. Sci. USA 111, E672–E681.
Rizvi, N.A., Hellmann, M.D., Snyder, A., Kvistborg, P., Makarov, V., Havel, J.J.,
Lee, W., Yuan, J., Wong, P., Ho, T.S., et al. (2015). Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung
cancer. Science 348, 124–128.
Morrissy, A.S., Garzia, L., Shih, D.J., Zuyderduyn, S., Huang, X., Skowron, P.,
Remke, M., Cavalli, F.M., Ramaswamy, V., Lindsay, P.E., et al. (2016). Divergent clonal selection dominates medulloblastoma at recurrence. Nature 529,
351–357.
Rooney, M.S., Shukla, S.A., Wu, C.J., Getz, G., and Hacohen, N. (2015).
Molecular and genetic properties of tumors associated with local immune
cytolytic activity. Cell 160, 48–61.
Murtaza, M., Dawson, S.J., Tsui, D.W., Gale, D., Forshew, T., Piskorz, A.M.,
Parkinson, C., Chin, S.F., Kingsbury, Z., Wong, A.S., et al. (2013). Non-invasive
analysis of acquired resistance to cancer therapy by sequencing of plasma
DNA. Nature 497, 108–112.
Murugaesu, N., Wilson, G.A., Birkbak, N.J., Watkins, T.B., McGranahan, N.,
Kumar, S., Abbassi-Ghadi, N., Salm, M., Mitter, R., Horswell, S., et al.
Rosenthal, R., McGranahan, N., Herrero, J., Taylor, B.S., and Swanton, C.
(2016). DeconstructSigs: delineating mutational processes in single tumors
distinguishes DNA repair deficiencies and patterns of carcinoma evolution.
Genome Biol. 17, 31.
Roth, A., Khattra, J., Yap, D., Wan, A., Laks, E., Biele, J., Ha, G., Aparicio, S.,
Bouchard-Côté, A., and Shah, S.P. (2014). PyClone: statistical inference of
clonal population structure in cancer. Nat. Methods 11, 396–398.
Cell 168, February 9, 2017 627
Roth, A., McPherson, A., Laks, E., Biele, J., Yap, D., Wan, A., Smith, M.A., Nielsen, C.B., McAlpine, J.N., Aparicio, S., et al. (2016). Clonal genotype and population structure inference from single-cell tumor sequencing. Nat. Methods
13, 573–576.
Russo, M., Siravegna, G., Blaszkowsky, L.S., Corti, G., Crisafulli, G., Ahronian,
L.G., Mussolin, B., Kwak, E.L., Buscarino, M., Lazzari, L., et al. (2016). Tumor
Heterogeneity and Lesion-Specific Response to Targeted Therapy in Colorectal Cancer. Cancer Discov. 6, 147–153.
Schwarz, R.F., Ng, C.K., Cooke, S.L., Newman, S., Temple, J., Piskorz, A.M.,
Gale, D., Sayal, K., Murtaza, M., Baldwin, P.J., et al. (2015). Spatial and temporal heterogeneity in high-grade serous ovarian cancer: a phylogenetic analysis. PLoS Med. 12, e1001789.
lial-cell FAK targeting sensitizes tumours to DNA-damaging therapy. Nature
514, 112–116.
Tirosh, I., Izar, B., Prakadan, S.M., Wadsworth, M.H., 2nd, Treacy, D., Trombetta, J.J., Rotem, A., Rodman, C., Lian, C., Murphy, G., et al. (2016). Dissecting the multicellular ecosystem of metastatic melanoma by single-cell RNAseq. Science 352, 189–196.
Tomasetti, C., Vogelstein, B., and Parmigiani, G. (2013). Half or more of the somatic mutations in cancers of self-renewing tissues originate prior to tumor
initiation. Proc. Natl. Acad. Sci. USA 110, 1999–2004.
Turajlic, S., and Swanton, C. (2016). Metastasis as an evolutionary process.
Science 352, 169–175.
Selmecki, A.M., Maruvka, Y.E., Richmond, P.A., Guillet, M., Shoresh, N., Sorenson, A.L., De, S., Kishony, R., Michor, F., Dowell, R., and Pellman, D. (2015).
Polyploidy can drive rapid adaptation in yeast. Nature 519, 349–352.
Turke, A.B., Zejnullahu, K., Wu, Y.L., Song, Y., Dias-Santagata, D., Lifshits, E.,
Toschi, L., Rogers, A., Mok, T., Sequist, L., et al. (2010). Preexistence and
clonal selection of MET amplification in EGFR mutant NSCLC. Cancer Cell
17, 77–88.
Shaw, A.T., Friboulet, L., Leshchiner, I., Gainor, J.F., Bergqvist, S., Brooun, A.,
Burke, B.J., Deng, Y.L., Liu, W., Dardaei, L., et al. (2016). Resensitization to Crizotinib by the Lorlatinib ALK Resistance Mutation L1198F. N. Engl. J. Med.
374, 54–61.
Uchi, R., Takahashi, Y., Niida, A., Shimamura, T., Hirata, H., Sugimachi, K.,
Sawada, G., Iwaya, T., Kurashige, J., Shinden, Y., et al. (2016). Integrated
Multiregional Analysis Proposing a New Model of Colorectal Cancer Evolution.
PLoS Genet. 12, e1005778.
Shen, R., and Seshan, V.E. (2016). FACETS: allele-specific copy number and
clonal heterogeneity analysis tool for high-throughput DNA sequencing. Nucleic Acids Res. 44, e131.
Van Allen, E.M., Miao, D., Schilling, B., Shukla, S.A., Blank, C., Zimmer, L.,
Sucker, A., Hillen, U., Geukes Foppen, M.H., Goldinger, S.M., et al. (2015).
Genomic correlates of response to CTLA-4 blockade in metastatic melanoma.
Science 350, 207–211.
Shi, H., Hugo, W., Kong, X., Hong, A., Koya, R.C., Moriceau, G., Chodon, T.,
Guo, R., Johnson, D.B., Dahlman, K.B., et al. (2014). Acquired resistance
and clonal evolution in melanoma during BRAF inhibitor therapy. Cancer Discov. 4, 80–93.
Siegel, R., Naishadham, D., and Jemal, A. (2013). Cancer statistics, 2013. CA
Cancer J. Clin. 63, 11–30.
Siravegna, G., Mussolin, B., Buscarino, M., Corti, G., Cassingena, A., Crisafulli,
G., Ponzetti, A., Cremolini, C., Amatu, A., Lauricella, C., et al. (2015). Clonal
evolution and resistance to EGFR blockade in the blood of colorectal cancer
patients. Nat. Med. 21, 795–801.
Smith, M.P., Sanchez-Laorden, B., O’Brien, K., Brunton, H., Ferguson, J.,
Young, H., Dhomen, N., Flaherty, K.T., Frederick, D.T., Cooper, Z.A., et al.
(2014). The immune microenvironment confers resistance to MAPK pathway
inhibitors through macrophage-derived TNFa. Cancer Discov. 4, 1214–1229.
Snyder, A., Makarov, V., Merghoub, T., Yuan, J., Zaretsky, J.M., Desrichard,
A., Walsh, L.A., Postow, M.A., Wong, P., Ho, T.S., et al. (2014). Genetic basis
for clinical response to CTLA-4 blockade in melanoma. N. Engl. J. Med. 371,
2189–2199.
Sottoriva, A., Kang, H., Ma, Z., Graham, T.A., Salomon, M.P., Zhao, J., Marjoram, P., Siegmund, K., Press, M.F., Shibata, D., and Curtis, C. (2015). A Big
Bang model of human colorectal tumor growth. Nat. Genet. 47, 209–216.
Stachler, M.D., Taylor-Weiner, A., Peng, S., McKenna, A., Agoston, A.T., Odze,
R.D., Davison, J.M., Nason, K.S., Loda, M., Leshchiner, I., et al. (2015). Paired
exome analysis of Barrett’s esophagus and adenocarcinoma. Nat. Genet. 47,
1047–1055.
Straussman, R., Morikawa, T., Shee, K., Barzily-Rokni, M., Qian, Z.R., Du, J.,
Davis, A., Mongare, M.M., Gould, J., Frederick, D.T., et al. (2012). Tumour
micro-environment elicits innate resistance to RAF inhibitors through HGF
secretion. Nature 487, 500–504.
Su, K.Y., Chen, H.Y., Li, K.C., Kuo, M.L., Yang, J.C., Chan, W.K., Ho, B.C.,
Chang, G.C., Shih, J.Y., Yu, S.L., and Yang, P.C. (2012). Pretreatment
epidermal growth factor receptor (EGFR) T790M mutation predicts shorter
EGFR tyrosine kinase inhibitor response duration in patients with non-smallcell lung cancer. J. Clin. Oncol. 30, 433–440.
Sun, Y., Campisi, J., Higano, C., Beer, T.M., Porter, P., Coleman, I., True, L.,
and Nelson, P.S. (2012). Treatment-induced damage to the tumor microenvironment promotes prostate cancer therapy resistance through WNT16B. Nat.
Med. 18, 1359–1368.
Tavora, B., Reynolds, L.E., Batista, S., Demircioglu, F., Fernandez, I., Lechertier, T., Lees, D.M., Wong, P.P., Alexopoulou, A., Elia, G., et al. (2014). Endothe-
628 Cell 168, February 9, 2017
Verdegaal, E.M., de Miranda, N.F., Visser, M., Harryvan, T., van Buuren, M.M.,
Andersen, R.S., Hadrup, S.R., van der Minne, C.E., Schotte, R., Spits, H., et al.
(2016). Neoantigen landscape dynamics during human melanoma-T cell interactions. Nature 536, 91–95.
Voss, M.H., Hakimi, A.A., Pham, C.G., Brannon, A.R., Chen, Y.B., Cunha, L.F.,
Akin, O., Liu, H., Takeda, S., Scott, S.N., et al. (2014). Tumor genetic analyses
of patients with metastatic renal cell carcinoma and extended benefit from
mTOR inhibitor therapy. Clin. Cancer Res. 20, 1955–1964.
Waclaw, B., Bozic, I., Pittman, M.E., Hruban, R.H., Vogelstein, B., and Nowak,
M.A. (2015). A spatial model predicts that dispersal and cell turnover limit intratumour heterogeneity. Nature 525, 261–264.
Wallace, A.R. (1867). Mimicry, and Other Protective Resemblances Among
Animals. Westminster Review.
Wang, J., Cazzato, E., Ladewig, E., Frattini, V., Rosenbloom, D.I., Zairis, S.,
Abate, F., Liu, Z., Elliott, O., Shin, Y.J., et al. (2016). Clonal evolution of glioblastoma under therapy. Nat. Genet. 48, 768–776.
Williams, M.J., Werner, B., Barnes, C.P., Graham, T.A., and Sottoriva, A.
(2016). Identification of neutral tumor evolution across cancer types. Nat.
Genet. 48, 238–244.
Xue, R., Li, R., Guo, H., Guo, L., Su, Z., Ni, X., Qi, L., Zhang, T., Li, Q., Zhang, Z.,
et al. (2016). Variable Intra-Tumor Genomic Heterogeneity of Multiple Lesions
in Patients With Hepatocellular Carcinoma. Gastroenterology 150, 998–1008.
Yates, L.R., Gerstung, M., Knappskog, S., Desmedt, C., Gundem, G., Van Loo,
P., Aas, T., Alexandrov, L.B., Larsimont, D., Davies, H., et al. (2015). Subclonal
diversification of primary breast cancer revealed by multiregion sequencing.
Nat. Med. 21, 751–759.
Yuan, Y., Failmezger, H., Rueda, O.M., Ali, H.R., Gräf, S., Chin, S.F., Schwarz,
R.F., Curtis, C., Dunning, M.J., Bardwell, H., et al. (2012). Quantitative image
analysis of cellular heterogeneity in breast tumors complements genomic
profiling. Sci. Transl. Med. 4, 157ra143.
Zack, T.I., Schumacher, S.E., Carter, S.L., Cherniack, A.D., Saksena, G., Tabak,
B., Lawrence, M.S., Zhsng, C.Z., Wala, J., Mermel, C.H., et al. (2013). Pan-cancer
patterns of somatic copy number alteration. Nat. Genet. 45, 1134–1140.
Zhang, J., Fujimoto, J., Zhang, J., Wedge, D.C., Song, X., Zhang, J., Seth, S.,
Chow, C.W., Cao, Y., Gumbs, C., et al. (2014). Intratumor heterogeneity in
localized lung adenocarcinomas delineated by multiregion sequencing. Science 346, 256–259.
Zhao, B., Sedlak, J.C., Srinivas, R., Creixell, P., Pritchard, J.R., Tidor, B., Lauffenburger, D.A., and Hemann, M.T. (2016). Exploiting Temporal Collateral
Sensitivity in Tumor Clonal Evolution. Cell 165, 234–246.
Leading Edge
Review
Transcriptional Addiction in Cancer
James E. Bradner,1 Denes Hnisz,2 and Richard A. Young2,3,*
1Novartis
Institutes for Biomedical Research, 181 Massachusetts Avenue, Cambridge, MA 02139, USA
Institute for Biomedical Research, 9 Cambridge Center, Cambridge, MA 02142, USA
3Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
*Correspondence: young@wi.mit.edu
http://dx.doi.org/10.1016/j.cell.2016.12.013
2Whitehead
Cancer arises from genetic alterations that invariably lead to dysregulated transcriptional programs. These dysregulated programs can cause cancer cells to become highly dependent on
certain regulators of gene expression. Here, we discuss how transcriptional control is disrupted
by genetic alterations in cancer cells, why transcriptional dependencies can develop as a consequence of dysregulated programs, and how these dependencies provide opportunities for novel
therapeutic interventions in cancer.
Introduction
Gene dysregulation is a hallmark of cancer. Recent progress in
our understanding of transcription and its role in cancer pathogenesis suggests that many new insights will soon be leveraged
for patient benefit. Thus, the bulk of the phenotypes, including
those affecting their clinical progression and therapeutic responsiveness, are likely to be strongly regulated by the dysregulated versions of transcriptional programs operating within
cancer cells. Increasingly, the molecular regulators of these programs—notably, proteins involved in transcriptional control—
are coming into view as attractive targets of a new generation
of drugs that perturb their functions and thus the transcriptional
programs that they govern.
By now, extensive cancer genome sequencing studies have
revealed recurrent somatic mutations in tumor cells that affect
nearly every DNA, RNA, and protein component of normal transcriptional control. These findings provide insights into the genes
whose alterations influence the cancer state and identify potential therapeutic targets. A number of excellent reviews describe
these alterations that affect cell signaling, transcription factors,
enhancer elements, chromatin regulators, and chromosome
structure (Garraway and Lander, 2013; Kandoth et al., 2013;
Lawrence et al., 2014; Stratton et al., 2009; Sur and Taipale,
2016; Vogelstein et al., 2013; Watson et al., 2013).
An alternative approach to understanding cancer and identifying therapeutic targets is to discover the key components on
which the dysregulated transcriptional programs depend in cancer cells (Figure 1). Such transcriptional dependencies are not
typically identified by cancer genome sequencing, but are discovered through focused mechanistic studies of gene control programs operating in both normal and neoplastic cells. We describe
our current views of the transcriptional programs operating in
normal cells, explain how these programs are altered in tumor
cells, and discuss recent insights into components of transcriptional control on which certain cancer cells become dependent.
Transcriptional Programs in Normal Cells
Cell identity—more specifically, the identity of one or another
differentiated cell type—is controlled in large part by the action
of transcription factors (TFs) that recognize and bind specific sequences in the genome and thereby regulate gene expression.
While nearly half of all of the TFs encoded in the human genome
are expressed in any one cell type (Vaquerizas et al., 2009), a
small number of master TFs, sometimes called lineage regulators, are sufficient to establish control of the gene expression
programs that define cell identity (Buganim et al., 2013; Graf
and Enver, 2009; Lee and Young, 2013; Morris and Daley,
2013; Sancho-Martinez et al., 2012; Vierbuchen and Wernig,
2012; Yamanaka, 2012). Thus, the control of transcriptional programs that characterize normal differentiated cell states is dominated by these master TFs, which are expressed at high levels in
selected cell types, tend to co-occupy most enhancers together
with other master TFs, and typically regulate their own genes
through an autoregulatory loop that forms the core transcriptional regulatory circuitry of a cell (Figure 2A) (Lee and Young,
2013). The master TFs of any one cell type can be found at the
enhancers of a majority of the active cell-type-specific genes
and may thus account for much of the organization of celltype-specific gene expression programs.
The master TFs bind cooperatively to enhancer DNA elements
and recruit coactivators and the transcription apparatus (Bulger
and Groudine, 2011; Levine et al., 2014; Long et al., 2016; Malik
and Roeder, 2010; Ong and Corces, 2011; Spitz and Furlong,
2012). These TFs can activate transcription from the enhancer
elements themselves (Figure 2B), producing enhancer RNAs
(eRNAs) that bind certain TFs and cofactors and contribute to
enhancer maintenance and dynamics (Lai et al., 2013; Li et al.,
2016). Enhancers, which tend to be cell type specific because
they are generally established by cell-type-specific master TFs,
have been mapped in a broad spectrum of human tissue types
by using epigenetic marks associated with enhancer activity
(ENCODE Project Consortium et al., 2012; Roadmap Epigenomics Consortium et al., 2015).
Bound by master TFs, clusters of enhancers known as superenhancers (SEs) regulate genes that play prominent roles in cell
identity or specialized cellular function (Chapuy et al., 2013;
Hnisz et al., 2013, 2015; Whyte et al., 2013). Enhancer-associated proteins and RNAs, including TFs, cofactors such as
Cell 168, February 9, 2017 ª 2016 Elsevier Inc. 629
Figure 1. Genetic Changes and Dysregulated Gene Expression Programs Lead to
Cancer Cell State
The path to a cancer cell involves genetic alterations that lead to changes in the gene expression
program. The dysregulated program can create
dependencies on transcriptional regulators that
make the tumor cells more sensitive to inhibition of
these regulators than normal cells.
Mediator, chromatin regulators, signaling factors, RNA polymerase II (RNAPII), enhancer-associated chromatin marks
(H3K27Ac), and eRNAs, are all found at especially high density
at the constituent enhancers of SEs. The constituent enhancers
of SEs physically associate with one another (Figure 2C) (Dowen
et al., 2014; Ji et al., 2016; Kieffer-Kwon et al., 2013) and can
function as independent or interdependent components of these
large transcription-regulating complexes to drive high-level
expression of their associated genes (Hah et al., 2015; Hay
et al., 2016; Hnisz et al., 2015; Jiang et al., 2016; Shin et al.,
2016).
Enhancers and super-enhancers become physically juxtaposed to target gene promoters by looping of the chromatin
and, having become so, stimulate transcription from these promoters. Although enhancers can activate any gene, they are
physically and functionally constrained to act within insulated
neighborhoods (Figure 2D) (Hnisz et al., 2016a). Insulated neighborhoods are chromosomal loop structures formed by the interaction of two DNA sites bound by the CTCF protein and occupied
by the cohesin complex. These chromosomal neighborhoods
engender specific enhancer-gene interactions and are essential
for normal gene activation and repression. The CTCF-CTCF loops
that form insulated neighborhoods are the mechanistic basis of
higher-order chromosome structures, such as topologically associating domains (TADs), and form a chromosome scaffold that is
largely preserved throughout development (Gibcus and Dekker,
2013; Gorkin et al., 2014; Phillips-Cremins and Corces, 2013).
Normal cell states depend on signals received from the tissue
microenvironments. Much of this contextual information is delivered by signaling pathways to SEs and, to a lesser extent, to
typical enhancers (Figure 2E). SEs have been shown to integrate
input from Wnt, TGFb, and LIF signaling pathways operating
within embryonic stem cells (ESCs) (Hnisz et al., 2015) and
BDNF and KCl signaling at c-Fos in neurons (Joo et al., 2016).
This signal integration is thought to be a consequence of the ability of master TFs to recruit signal-activated TFs to enhancer sites
previously established by the master TFs (Mullen et al., 2011;
Trompouki et al., 2011). The extent to which extracellular and
intracellular information is delivered to enhancers is underappreciated, due in part to the paucity of literature on nuclear signaling
relative to cytoplasmic signaling.
Enhancers that are responsive to various types of afferent
signaling are thought to be highly dynamic because the
activity of signaling TFs is linked to their destruction
(Figure 2F). For example, activated TGFb and BMP receptor kinases phosphorylate Smad TFs, which become fully functional
transcriptional activators after being further phosphorylated by
the transcriptional cyclin-dependent kinases (CDKs) CDK8 and
CDK9. However, following their initial activation, these phos630 Cell 168, February 9, 2017
phorylated Smads are recognized by specific ubiquitin ligases
operating in the nucleus, leading to their proteasome-mediated
destruction (Alarcón et al., 2009). Remarkably, ubiquitylation
and degradation of many transcriptional activators occurs at
enhancer/promoter sites and can be required for efficient transcription (Geng et al., 2012; Thomas and Tyers, 2000).
Multiple cyclin-dependent kinases (CDKs), including CDK7, 8,
9, 12, and 13, are dynamic effectors of gene control and function,
in part, by phosphorylating serine residues in the C-terminal
domain of RNA polymerase II (Eick and Geyer, 2013). For
example, CDK7 and CDK9 contribute to control of transcription
initiation and elongation, respectively (Jonkers and Lis, 2015).
TFs may thus control initiation or elongation by their interactions
with these CDKs.
DNA is packaged into nucleosomes, which consist of histones
that are substrates for chromatin regulators that can modify
various amino acid residues or bind in a modification-dependent
manner to these histones (Figure 2G). The roles of these diverse
chromatin regulators in gene control have been reviewed extensively elsewhere (Campos and Reinberg, 2009; Kouzarides,
2007; Piunti and Shilatifard, 2016; Soshnev et al., 2016; Tessarz
and Kouzarides, 2014). The positioning of nucleosomes on DNA
can also influence gene control, for example, by limiting access
of TFs to regulatory sequences, and ATP-dependent remodeling
complexes influence transcriptional states by mobilizing nucleosomes (Kadoch and Crabtree, 2015). The common functional
theme of these regulators is that they facilitate maintenance of
positive or negative gene expression states.
DNA methylation contributes to gene control at three levels.
Methylation of enhancer and promoter sites can prevent TF binding and thus silence genes (Figure 2H) (Ziller et al., 2013). Methylation of CTCF loop anchor sites prevents CTCF binding and can
thus alter insulated neighborhood structure (Ghirlando and Felsenfeld, 2016; Liu et al., 2016b). Cytosine methylation and hydroxymethylation present spatially positioned chemical motifs
that can be recognized by chromatin-associated proteins (e.g.,
MECP2), thereby influencing transcriptional regulation (Baubec
et al., 2013; Liu et al., 2016a; Mellén et al., 2012).
In summary, normal cells have transcriptional programs that
are established and maintained by master TFs that regulate
genes by binding specific enhancer elements, which in turn
interact with genes within insulated neighborhoods. The maintenance of normal cell states depends on the tissue environment,
and such information is delivered by signaling pathways ultimately to enhancers. Maintenance of cell identity and dynamics
of cell states also depend on a large number of histone readers,
writers, and erasers, as well as regulators of DNA methylation,
that together ensure chromatin states are appropriate for positive and negative gene regulation.
Figure 2. Key Features of Transcriptional Regulation of Gene Expression Programs
(A) In this model of cell-type-specific core regulatory circuitry, master TFs co-regulate their own genes, forming an interconnected autoregulatory loop (left), as
well as those of those of many other cell-type-specific genes (right).
(B) Enhancers are DNA elements bound by multiple TFs that recruit coactivators (such as Mediator) and RNA polymerase II, which can initiate transcription within
enhancer sequences to produce eRNAs.
(C) SE constituents are physically connected. Interactions among SE constituent enhancers, and between SEs and a target gene, are indicated by red arcs.
(D) The human genome contains over 10,000 insulated neighborhoods, which are produced by multimerization of CTCF bound to two sites and reinforced with
cohesin; enhancer-gene interactions occur predominantly within these neighborhoods.
(E) SEs are enriched for signaling TFs, and their associated genes, which tend to play prominent roles in cell identity, are thus especially responsive to signaling.
(F) Signaling TFs and other TFs that play important regulatory roles are themselves regulated by ubiquitin and proteasome-mediated destruction.
(G) Chromatin regulators that act through ‘‘writing,’’ ‘‘erasing,’’ or ‘‘reading’’ histone modifications.
(H) DNA methylation can contribute to gene control by causing loss of TF binding at enhancers or loss of CTCF at insulated neighborhood loop anchors.
Transcriptional Dysregulation in Cancer
The complements of genetic alterations that collaborate to
transform normal cells into neoplastic derivatives exhibit a high
degree of tissue specificity. Tissue-specific enhancers are struc-
turally altered to drive expression of oncogenes (e.g., TMPRSS2ERG in prostate cancer, IgH-locus alterations in B cell malignancies, TCR-locus alterations in T cell malignancies) (Sur and
Taipale, 2016). In addition, many oncogenic signaling pathways
Cell 168, February 9, 2017 631
Figure 3. Components of Gene Control Altered in Cancer
Examples of three types of cis-factors (enhancers, promoters, and insulators) and various trans-factors (TFs, cofactors, chromatin regulators, RNA polymerase II,
and histones) that acquire recurrent somatic mutations in cancer cells.
require cell-specific chromatin contexts (e.g., NOTCH1 activation in T cell, but not B cell, leukemia and EZH2 activation in B
cell, but not T cell, lymphoma). Because many tissue-specific
differentiation programs (specifically, those that define cell identity) persist in cancer cells, it is clear that cancer arises from the
collaborative interplay of oncogenic events acquired during
multi-step tumor formation with the tissue-specifying gene
expression programs that survive neoplastic progression and
continue to influence cancer cell behavior.
This transcriptional dysregulation arises in cancer from disease-defining genetic alterations either indirectly, via mutation
of signaling factors converging on transcriptional control, or
directly, via genetic alterations in gene control factors themselves. Cancer-associated genetic alterations can affect proteins participating in nearly all levels of transcriptional control,
including trans-factors (TFs, signaling proteins, cofactors, chromatin regulators, and chromosome structuring proteins) and
cis-elements (enhancers, promoters, and insulators) (Figure 3).
Many excellent reviews have described an ever-expanding catalog of these alterations (Bywater et al., 2013; Garraway and
Lander, 2013; Kandoth et al., 2013; Lawrence et al., 2014; Stratton et al., 2009; Sur and Taipale, 2016; Vogelstein et al., 2013;
Watson et al., 2013). Here, we discuss a subset of these alterations—specifically, those that lead to the most profound
changes in the gene expression program, thereby driving the
malignant cell state. We focus on these programs because
632 Cell 168, February 9, 2017
they may direct the discovery and development of new classes
of cancer therapeutics designed to target vulnerabilities of cancer cells—in particular, their addiction to certain transcriptional
programs.
Trans-factors
The TFs that are deregulated in cancer cells and have the potential to produce profound changes in gene expression programs can be considered to fall into three classes: master
TFs involved in organizing cell identity, proliferation control
TFs that amplify transcriptional output, and signaling TFs
involved in dynamic changes in the control machinery occurring
in response to extracellular signals. The activation of a master
TF that is normally expressed in early development, such as
the pluripotency TF OCT4, or activation of a master TF that is
normally expressed early in a specific lineage, such as TAL1
in T cells, can alter core regulatory circuitry and activate additional genes that are normally expressed in more embryonic
states (Figure 4A) (Sanda et al., 2012). Genes encoding the
MYC and P53 proliferation control TFs, a classic oncogene
and tumor suppressor gene, are among the most frequently
mutated genes in cancer. MYC can have profound effects
because it can function to amplify the entire gene expression
program (Figure 4A) (Lin et al., 2012; Nie et al., 2012), and
P53 is a powerful tumor suppressor because of its ability to arrest progress through the cell cycle or induce apoptosis (Lane
and Levine, 2010). Dysregulation of signaling pathways is a
Figure 4. Common Mechanisms of Dysregulation of Gene Expression Programs in Cancer Involving Trans-factors
(A) Model of a transcriptional regulatory circuit controlling the gene expression program in normal cells is shown on the top. Dysregulation of gene expression
programs in cancer cells can occur through dysregulation of oncogenic master TFs (second circuit from the top), dysregulated signaling (third circuit from the top),
and dysregulation of a transcriptional amplifier (e.g. MYC).
(B) Model of aberrant chromatin modification affecting gene expression programs in cancer cells.
(C) Model of the effect of cohesin mutations and CTCF mutations on enhancer-promoter interactions and insulated neighborhoods in cancer cells.
common feature of cancer cells; a dysregulated signaling TF
can profoundly change the gene expression program through
its binding to enhancers occupied by master TFs (Figure 4A)
(Mullen et al., 2011; Trompouki et al., 2011), and dysregulated
signaling can even stimulate super-enhancer formation (Brown
et al., 2014; Hnisz et al., 2015).
These TFs signal to RNAPII through transcriptional cofactors,
defined here as regulatory components that do not bind directly
Cell 168, February 9, 2017 633
Figure 5. Common Mechanisms of Dysregulation of Gene Expression Programs in
Cancer Involving Cis-factors
(A) Mechanisms leading to the acquisition of super-enhancers to drive oncogenes in cancer cells:
translocation of an existing super-enhancer, focal
amplification of an enhancer element, and nucleation of super-enhancer through somatic insertion
of TF binding sites.
(B) Activation of silent proto-oncogenes by somatic mutations that disrupt insulated neighborhood anchor sites.
to DNA in a sequence-specific manner. Exemplifying this class of
transcriptional signaling proteins are the components of the
Mediator complex, which is recruited to enhancer-promoter regions by TFs in the context of transcription activation (Allen
and Taatjes, 2015; Kagey et al., 2010). Genetic alterations of
Mediator-complex-encoding genes are observed frequently in
prostate cancer and in many uterine myomas (Allen and Taatjes,
2015; Barbieri et al., 2012; Mäkinen et al., 2011). Beyond these
discrete diseases, the genome-wide activities of Mediator in
gene control would be expected to function broadly in all cancer-associated transcription. Interestingly, few cancer-associated alterations are identified in the core RNAPII complex itself
(Clark et al., 2016), suggesting that coordinated transcriptional
signaling upstream of polymerase favors the neoplastic cell state
more than alterations of this core complex.
Efficient transcriptional signaling from enhancers to promoters
is often chromatin dependent, mediated by specialized transcriptional cofactors that physically associate with or biochemically modify the genome to reinforce gene activation or
repression. Chromatin regulators function globally, so their dysregulation can also have profound effects on the gene expression program of cells (Jones et al., 2016). In some tumors, chromatin regulators have become fused to transcriptional cofactors,
producing gene-specific effects, such as those observed in
acute lymphoblastic leukemia cells with MLL-AF4 fusions
(Figure 4B) (Guenther et al., 2008; Krivtsov et al., 2008).
Insulated neighborhoods contribute to proper positive and
negative gene control, so alterations in chromosome-structuring
proteins that establish and maintain insulated neighborhood
boundaries would be expected to have profound effects on
a cell’s overall gene expression program. Cancer genome
sequencing has revealed that somatic mutations occur in
634 Cell 168, February 9, 2017
CTCF and cohesin coding sequences in
various solid tumors and leukemias (Lawrence et al., 2014), and it seems likely that
these mutations contribute to oncogenesis by altering insulated neighborhoods throughout the genome, perhaps
rendering chromatin generally more
permissive to oncogenic transcriptional
signaling (Figure 4C) (Viny et al., 2015).
Cis-elements
Cis-regulatory elements within the
genome that contribute to cancer pathogenesis were first recognized in pioneering studies of cancer-associated chromosomal translocation
(e.g., IgH-MYC in Burkitt’s lymphoma [Taub et al., 1982]).
Following early efforts in cancer genome sequencing, which
concentrated on coding regions of the human genome, recent
focused and genome-wide sequencing efforts have revealed
frequent alteration of cis-elements in both solid and hematological tumors. Two types of cis-elements that play prominent roles
in cancer biology—SEs and insulators—are discussed here.
In normal cells, SEs—which, as we noted above, are large
clusters of enhancers that bind high densities of transcriptional
components—control genes that play prominent roles in specific
cell identities. Tumor cells acquire SEs at oncogenic driver
genes, and they do so through many different mechanisms
(Figure 5A) (Chapuy et al., 2013; Drier et al., 2016; Hnisz et al.,
2013; Kennedy et al., 2015; Lovén et al., 2013; Tomazou et al.,
2015; Wang et al., 2015; Yang et al., 2015). The genetic mechanisms that lead to SE acquisition in cancer include DNA translocation (Affer et al., 2014; Drier et al., 2016; Gröschel et al., 2014;
Northcott et al., 2014; Walker et al., 2014), focal amplification
(Hnisz et al., 2013; Shi et al., 2013; Zhang et al., 2016b), and
nucleation by small insertions/deletions (INDELs) that create
master TF binding sites (Mansour et al., 2014). Additional epigenomic mechanisms also lead to SE formation in cancer, such as
those associated with oncogenic TF overexpression (Hnisz et al.,
2013), the global function of oncogenic TF fusions (e.g., EWSFLI) (Kennedy et al., 2015; Tomazou et al., 2015), and the consequences of upstream oncogenic signaling (e.g., RAS-dependent
signaling to chromatin) (Nabet et al., 2015).
Mutations that alter insulator sequences of oncogene-containing insulated neighborhoods appear to make important contributions to the dysregulation of gene expression observed in some
cancers (Figure 5B) (Flavahan et al., 2016; Hnisz et al., 2016b;
Katainen et al., 2015). Somatic mutations occur frequently and
recurrently in loop anchors of oncogene-containing insulated
neighborhoods in a broad spectrum of cancer cells. Thus, the
CTCF DNA-binding motif in loop anchor regions is among the
most altered human TF-binding sequences in cancer cells (Hnisz
et al., 2016b; Ji et al., 2016; Katainen et al., 2015). DNA hypermethylation occurs in some cancer cells, and tumor-specific DNA
methylation has been implicated in the disruption of CTCF binding, alterations of chromosome structure, and dysregulation of
oncogene expression in a subset of gliomas (Flavahan et al.,
2016). Moreover, chromosomal rearrangements that disrupt
insulated neighborhoods can activate oncogenes without
altering the sequences of the oncogenes themselves (Gröschel
et al., 2014; Hnisz et al., 2016b).
Transcriptional Addiction and Cancer Therapeutics
The concept of oncogene addiction (Weinstein and Joe, 2006)
refers to the behavior of cancer cells that exhibit an absolute
dependence on oncogenes that were initially acquired during
multi-step tumorigenesis and remain critical to the ongoing proliferation and viability of these cells long after they have progressed to a fully neoplastic state. As of late, this concept has
been extended to include other changes acquired during tumor
progression that fostered the early development of a tumor and
continue to be absolutely essential to its continued growth.
Included among these are the dysregulated transcriptional programs operating in certain tumor cells, yielding the concept of
transcriptional addiction.
Various types of transcriptional addiction appear to operate in
specific subsets of cancer. Thus, the majority of human cancers
exhibit genetic amplification or transcriptional dysregulation of
MYC, which is accompanied by an anabolic transcriptional
response driving proliferation and metabolic adaptation (Beroukhim et al., 2010). Research in model systems of MYC addiction
and withdrawal have validated addiction to this master regulatory TF in solid and hematologic malignancies, both with and
without structural changes to the MYC locus itself (Felsher and
Bishop, 1999; Jain et al., 2002; Soucek et al., 2008, 2013). As
for many TFs, direct pharmacologic inhibition of MYC remains
an elusive challenge in drug discovery. We discuss below the
strategy of leveraging mechanistic insights into transcriptional
dysregulation toward a more immediate therapeutic benefit in
cancer.
Direct Inhibition of Oncogenic TFs and Cofactors
Gene-targeted therapy has emerged as a paradigm of cancer
medicine. Where available and where actionable, somatic
alteration of driver oncogenes has provided strong guidance
to the discovery and focused development of cancer therapeutics (Darnell, 2002; Pagliarini et al., 2015; Stuart and
Sellers, 2009). Among the large number of somatically altered
oncogenic TFs, only very few have been successfully approached by coordinated efforts in drug discovery (Bhagwat
and Vakoc, 2015). Most compounds of this class are, at
best, experimental tools for the study of TFs in cell biology,
including our own work to inhibit the NOTCH1 transactivation
complex with constrained alpha helical peptides (Moellering
et al., 2009). Nonetheless, in the longer term, small molecule
agents directed at oncogenic TFs have the promise to confer
significant clinical responses for patients bearing certain types
of genetically defined cancers.
Among the most impactful examples of direct inhibition of an
oncogenic TF is the development of all-trans retinoic acid
(ATRA) as therapy for acute promyelocytic leukemia (APML).
ATRA was initially investigated because it induces differentiation
in cultivated APML cells, an observation made without the guidance of cancer genetics. In 1985, a 5-year-old girl with anthracycline-refractory APML was administered ATRA at Shanghai
Children’s Hospital, achieving a complete remission (CR) and ultimately long-term remission from ATRA with chemotherapy
(Wang and Chen, 2008). A first case series reported by ZhenYi Wang 3 years later would establish ATRA as a highly effective
therapy for APML, conferring CR in 23 of 24 patients as a single
agent (Huang et al., 1988). 11 years after the identification of the
t(15;17) translocation by Rowley (Rowley et al., 1977) and seven
years after the development of ATRA, Chen et al. (1993) reported
that the reciprocal t(15;17) translocation encodes a novel oncogene comprising the gene encoding the retinoic acid receptor
alpha (RARA) fused to a second gene (PML) specifying a novel
Krüppel-like zinc finger protein. The ensuing decade of laboratory research would firmly establish ATRA as targeted therapy
for APML via modulation and destabilization of the encoded
chimeric PML-RARA oncogenic TF (Wang and Chen, 2008).
Notably, innovation in clinical investigation would ultimately
pair ATRA with arsenic trioxide (also associated with PMLRARA degradation) in low- to intermediate-risk APML as effective first-line treatment, equivalent in efficacy to chemotherapy
in this disease (Lo-Coco et al., 2013). APML exemplifies biology
of transcriptional addiction, revealed through the effective development of transcriptional therapy with curative intent.
An emerging example of direct inhibition of an oncogenic transcriptional cofactor is bromodomain inhibition in carcinomas
harboring fusions of BET bromodomain coactivators. Among
the most aggressive subtypes of lung and head and neck cancer
are tumors expressing the chimeric, oncogenic cofactors BRD4NUT or BRD3-NUT (so-called NUT midline carcinoma or NMC).
NMC is a poorly differentiated, chemoresistant, and aggressive
malignancy that lacks effective, FDA-approved therapy. In
2010, we reported the first effective inhibitors of human bromodomains targeting the BET family (BRD2, BRD3, and BRD4),
exemplified by the chemical probe JQ1 (Filippakopoulos et al.,
2010). Using preclinical models of NMC, JQ1 was shown to
displace BRD4 from chromatin, resulting in potent and irreversible squamous differentiation. In murine models harboring primary human NMC xenografts, JQ1 prompted a robust pro-differentiation and anti-proliferative response associated with
durable responses by PET-CT imaging. Based on this rationale,
drug-like derivatives of JQ1 have been transitioned to clinical
investigation by our group and others. Early reports of index trials
confirm unambiguous antitumor activity in advanced disease
(Stathis et al., 2016).
Targeting Tissue Identity and Homeostasis
Oncogenic events occur in the context of cell identity, which is
established and maintained by TF-defined core regulatory circuits and signals from the tissue environment (Lee and Young,
2013). Targeting transcriptional identity has thus emerged as
an important therapeutic strategy in cancer. Initial efforts have
Cell 168, February 9, 2017 635
Table 1. Examples of Transcriptional Dependencies that Occur as a Consequence of Transcriptional Dysregulation
Cancer type
Component of transcriptional
control perturbed by genetic change
Genetic change
Transcriptional dependency
T cell leukemia
Master transcription factor
TAL1 overexpression
CDK7
Multiple myeloma
Enhancer/Super-enhancer
IgH-MYC translocation
BRD4
Glioma
Insulated neighborhood anchor
IDH1 mutation
PDGFRA
Mixed lineage leukemia
Chromatin regulator
MLL-AF9 fusion
DOT1L, BRD4
T cell leukemia
Signaling factor
NOTCH1 mutation
BRD4
attempted to deplete the tissue type in which the tumor arose
or to interfere with the signaling pathways that contribute to
cell identity by connecting to cell-type-specific core regulatory
circuits.
Tissue depletion may be accomplished surgically, as with
excision of neoplastic and surrounding unaffected tissue in
mastectomy or radical prostatectomy. Medical approaches
also exist, as with anti-CD20 monoclonal antibody therapy or
CD19-directed chimeric antigen receptor T cell (CART) therapy
to eradicate B cell acute lymphoblastic leukemia (B-ALL). A
consequence of B cell depletion therapy is the temporary or permanent depletion of systemic, normal B cells as a considerable
side effect.
In contrast to tissue depletion, direct targeting of cell identity
undermines the requisite interaction of oncogenic and cellular
circuitry. The best-studied example of this strategy led to the
development of nuclear hormone receptor antagonists. Nuclear
hormone receptors facilitate differentiation and growth of
discrete tissues during development. In post-pubertal adults,
the estrogen receptor (ER) and androgen receptor (AR) TFs
maintain tissue homeostasis and sex hormone responsive function of breast and prostate tissues, respectively, among other tissues. In low-grade breast and prostate cancer expressing ER
and AR, respectively, these TFs are not oncogenes, per se,
and are genetically unaltered in the majority of patients. Indeed,
these tumors are each characterized by diverse genetic alterations to oncogenes and tumor suppressors, the majority of
which lack direct-acting agents (Kandoth et al., 2013; Lawrence
et al., 2014; Vogelstein et al., 2013). The impactful and important
contribution of anti-estrogen and anti-androgen therapy in these
diseases, then, defines the opportunity of disrupting transcriptional pathways of cellular identity alone and in combination
with oncogene-targeted therapy as available. Further, the broad
use of glucocorticoids in lymphoid malignancies exemplifies the
same conceptual strategy (Inaba and Pui, 2010) and establishes
the feasibility of agonizing a core transcriptional pathway for
therapeutic benefit.
Context-Specific Transcriptional Dependencies
As discussed earlier, eukaryotic transcription is a dynamic
network with multiprotein complexes collaborating as nodes
of activating, repressing, remodeling, and insulating function.
More than a thousand human proteins contribute to gene control at the level of nuclear chromatin, spatially distributed to tens
of thousands of sites in the human genome (ENCODE Project
Consortium et al., 2012; Levine et al., 2014; Roadmap Epigenomics Consortium et al., 2015). Despite this complexity, specific oncogenic impulses can engender exceptional reliance
636 Cell 168, February 9, 2017
on discrete protein complexes and even individual factors.
The identification, validation, and prosecution of these targets
can yield important mechanistic insights and therapeutic opportunities (Table 1).
Indeed, our research on BET bromodomains began with a
mechanistic hypothesis—specifically, that BRD4 might mediate
transcriptional addiction to MYC. Because MYC had been
shown to play a key role in the transcriptional elongation operating at many genes, we hypothesized that BRD4 acted by mediating chromatin-dependent transcription elongation signaling to
RNAPII. BRD4 possesses twin acetyl-lysine recognition domains
(bromodomains), suggesting that BRD4 might localize to hyperacetylated regions of euchromatin occupied in cancer by MYC
and associated lysine acetyltransferases. BRD4 binds the pTEFb
elongation factor, establishing the possibility that BRD4 may
function to facilitate elongation downstream of MYC function in
cancer (Bisgrove et al., 2007; Rahl et al., 2010). Indeed, in
models of MYC-addicted hematologic cancers, BRD4 localizes
with MYC throughout the active genome and contributes to the
MYC-mediated proximal promoter pause release that enables
elevated transcriptional elongation (Chapuy et al., 2013; Lovén
et al., 2013). These mechanistic studies and accompanying
translational research establish BRD4 inhibition as a therapeutic
strategy to inhibit MYC-dependent transcriptional signaling in
multiple myeloma, diffuse large B cell lymphoma, and mixed lineage leukemia (MLL) (Chapuy et al., 2013; Dawson et al., 2011;
Delmore et al., 2011; Zuber et al., 2011).
Of relevance to the present subject of targeting transcriptional
addiction in cancer, these studies established that pharmacologic inhibition of a TF or cofactor that is presumed to act widely
on countless genes throughout the genome can nevertheless
exert highly selective effects on gene control. Thus, the mechanistic investigation of this phenomenon identified regions of
disproportionately high levels of BRD4 occupancy in cis-regulatory regions of many genes that were associated with massive Mediator enrichment, i.e., super-enhancers (Lovén et al.,
2013). Moreover, BRD4 function has emerged as a transcriptional addiction in MYCN-amplified neuroblastomas (Puissant
et al., 2013), as a positive regulator of anti-apoptotic gene
expression in AMLs (Dawson et al., 2011), and as a mediator of
resistance to Notch pathway inhibition in T-ALLs (Yashiro-Ohtani
et al., 2014).
Several features of TF gene regulation are thought to
contribute to the exceptional sensitivity of large SEs to BRD4 inhibition in tumor cells. TFs and their mRNAs generally have
short half-lives, so the genes that encode key TFs may evolve
SEs in order to maintain a high transcriptional output of these
short-lived proteins (Figure 6A). Many TFs bind and auto-regulate the genes that encode them so that disruption of TF levels
may have an especially pronounced effect on auto-regulated
SE control of these genes (Figure 6A). The cooperative features
of enhancer components may make SEs especially vulnerable to
inhibition of enhancer factors (Figure 6B). Genes encoding tumor
cell master TFs acquire especially large SEs, with exceptionally
high densities of enhancer factors, and this may further
contribute to the exceptional vulnerability of SE-driven oncogenic TFs seen with inhibition of BRD4 (Bhagwat et al., 2016;
Chapuy et al., 2013; Lovén et al., 2013; Shi and Vakoc, 2014).
SE-driven transcription may also be especially sensitive to disruptions in cooperative interactions that contribute to the 3D
chromatin architecture at SEs (Dowen et al., 2014; Ji et al.,
2016; Kieffer-Kwon et al., 2013).
The transcriptional CDKs have likewise emerged as compelling targets for transcriptional therapy in cancer. CDK7 is a key
component of the general transcription initiation apparatus,
which governs RNAPII activity by phosphorylating residues on
its C-terminal domain. ChIP-seq data indicate that CDK7
densely occupies the SEs that drive expression of oncogenes
in a wide variety of cancers, including T-ALL (Kwiatkowski
et al., 2014), NSCLC (Christensen et al., 2014), neuroblastoma
(Chipumuro et al., 2014), and triple-negative breast cancer
(Wang et al., 2015). Small molecule inhibition of CDK7 by THZ1
leads to rapid loss of transcripts for TFs that contribute to oncogenesis and whose expression is driven by SEs in these cells,
consistent with the idea that these SE-driven TF genes are especially vulnerable to inhibition of CDK7. The preferential loss of
SE-driven TF gene expression, along with expression of genes
involved in the DNA damage response, also occurs when tumor
cells are treated with inhibitors of CDK12 and CDK13 (Zhang
et al., 2016a). Interestingly, inhibition of the CDK8 and CDK19,
components of Mediator that contribute to transcriptional
repression, leads to hyperactivation of SE-driven genes in
AML, and this is as deleterious to the leukemia cells as BRD4 inhibition, which has the opposite effect on these SE-driven genes
(Pelish et al., 2015). These and other results show that the leukemia cells are addicted to a specific level of SE-associated gene
transcription (Pelish et al., 2015).
The tumor-promoting effect of fusion oncogenes can produce
actionable transcriptional dependencies. For example, many
AML and MLL leukemias contain oncogenic fusions between
MLL and AF genes; the AF part of the fusion product binds the
DOT1L histone methyltransferase, which plays important roles
in the control of transcriptional elongation by dimethylating histone H3 on lysine 79 (H3K79) (Cai et al., 2015; Deshpande
et al., 2013). The MLL-AF fusions thus recruit DOT1L to MLLtarget genes, including the TF genes HOXA9 and MEIS1, and
the hypermethylation of histones leads to aberrant expression
of those genes, which in turn drive leukemogenesis (Figure 6C)
(Chen and Armstrong, 2015). These leukemias are especially
vulnerable to genetic and chemical perturbation of DOT1L, and
DOT1L inhibitors are now in clinical trials for these malignancies
(Daigle et al., 2013).
Evidence that tumor cells can develop additional dependencies
on specific transcriptional and chromatin regulators, rendering
them susceptible to transcriptional drugs, has emerged from
studies of drug resistance. Treatment of tumor cells with various
anti-cancer agents can select for populations of drug-tolerant
cells that have become dependent on specific chromatin regulators such as histone demethylases (KDM5A) and deacetylases
(HDACs) (Figure 6D) (Sharma et al., 2010). The drug-tolerant
tumor cells can, in turn, be ablated with histone deacetylase inhibitors, establishing a paradigm of combination therapy using inhibitors of chromatin regulators against drug resistance (Sharma
et al., 2010).
Targeting Chromosomal Neighborhoods and Associated
Addictions
As described above, insulated neighborhoods, whose ends are
established by CTCF anchor sites, are frequently perturbed in
cancer genomes by somatic mutations or aberrant DNA methylation; these perturbations have been implicated in the activation
of cellular proto-oncogenes (Flavahan et al., 2016; Hnisz et al.,
2016b; Katainen et al., 2015). Cancer cells containing aberrantly
methylated neighborhood anchors can develop transcriptional
dependencies that are amenable to small molecule therapeutics
(Flavahan et al., 2016). The recent development of genetic and
epigenetic editing technologies that can manipulate or repair
chromosomal neighborhoods suggests novel therapeutic strategies for such modified cancer genomes (Amabile et al., 2016;
Doudna and Charpentier, 2014; Liu et al., 2016b).
As an example, the expression of active genes can be reduced
by disrupting the boundaries of its insulated neighborhood,
which can, in turn, cause enhancers within a neighborhood to
loop to other gene targets outside the neighborhood (Dowen
et al., 2014). Targeted methylation of a neighborhood anchor
site with a dCas9-DNA-methyltransferase-3 fusion protein has
been shown to disrupt the neighborhood through the eviction
of CTCF (Liu et al., 2016b). Thus, targeted methylation of an
oncogene enhancer or the anchor sites of oncogene-containing
neighborhoods might lead to selective downregulation of oncogene expression (Figure 6E).
In glioma cells that harbor mutations in the IDH1 gene, hypermethylation of insulated neighborhood anchor sites occurs,
which leads to loss of CTCF binding at those anchors and activation of oncogenes that were previously silent in the intact
neighborhoods (Flavahan et al., 2016). This suggests that restoration of the perturbed neighborhood boundaries might lead to
silencing of the oncogene. Targeted DNA 5-hydroxy-methylation
with a dCas9-TET fusion protein has recently been demonstrated (Amabile et al., 2016; Liu et al., 2016b), and this strategy
could be used to restore an insulated neighborhood whose anchor site has been disrupted by aberrant DNA methylation
(Figure 6E).
The identification of oncogenes activated by perturbations of
insulated neighborhood CTCF anchor sites by aberrant methylation has revealed transcriptional dependencies of these tumor
cells. For example, the IDH1 mutant glioma cells described
above overexpress the PDGFRA gene due to hypermethylation
of the CTCF anchor site of the insulated neighborhood containing PDGFRA, which disrupts a boundary of the neighborhood
and allows enhancers located elsewhere to activate PDGFRA.
The growth of these cells is inhibited by small molecules that
block PDGFRA, while gliomas containing a wild-type IDH1
gene and an intact PDGFR neighborhood are unaffected by
Cell 168, February 9, 2017 637
Figure 6. Drugging Transcriptional Dependencies
(A) Model of the features of super-enhancer-associated oncogene control that contribute to transcriptional dependencies.
(B) Model of the mechanistic basis of higher transcriptional activity and vulnerability of super-enhancers: cooperative interactions between the coactivators
recruited to these sites.
(C) Therapeutic strategy to inhibit transcriptional dependencies due to aberrant recruitment of DOT1L to driver genes in MLL-AF fusion leukemias.
(D) Therapeutic strategy to attack small populations of drug-tolerant tumor cells with histone deacetylase inhibitors.
(E) Therapeutic strategy to downregulate oncogenes through manipulation and repair of insulated neighborhoods.
(F) Targeted degradation of transcriptional regulators using degronimids that recruit an E3-ubiquitin ligase for proteasome-mediated degradation.
638 Cell 168, February 9, 2017
such inhibition (Flavahan et al., 2016). In this case, the dependence of the cancer cells in these tumors on PDGFRA was not
identified by cancer genome sequencing (which revealed the
IDH1 mutation), but by hypothesis-driven study of the effect of
IDH1-associated hypermethylation on transcriptional control in
these tumor cells.
Inhibition of DNA methyltransferases (DNMTs) by 5-azanucleoside drugs, such as 5-azacitidine, have broad impacts on
gene expression. Though toxic in high doses, these compounds
have shown therapeutic benefits at optimized low doses and are
approved for the treatment of myelodysplastic syndrome (MDS)
and acute myeloid leukemia (AML) (Jones et al., 2016). These
drugs were notable for their ability to alter the identities of
cultured mammalian cells (Taylor and Jones, 1979) and are
thought to activate tumor suppressor genes that have been
silenced by DNA methylation and to reverse the tumor-promoting effects of cancer-specific genetic alterations (Jones et al.,
2016). With our new understanding of the roles of insulated
neighborhoods and DNA methylation in gene control, it is
possible that a major effect of these drugs is to alter neighborhood structures and thus the gene expression programs to
which the tumor cells are addicted.
Future Challenges
The fields of cancer biology and transcriptional biology are
rapidly maturing and converging, establishing a compelling
role for transcriptional dysregulation in cancer. We think it is
imperative to commit to more aggressive study of the mechanisms that produce transcriptional addiction and learn how to
exploit these for new therapies. Despite tremendous progress,
there exist significant challenges ahead before these nascent insights can be broadly leveraged for patient benefit.
Where feasible, targeting oncogenic TFs in cancer (e.g., PMLRARA) has demonstrated profound clinical benefit. However,
therapeutics capable of disrupting oncogenic TFs are lacking.
We need a new science of transcriptional therapeutics and a
generational commitment to the pursuit of the high-hanging fruit.
Oncogenic MYC remains a holy grail of cancer therapy, but as of
yet, there are no compounds that directly target this TF. One solution may be new platforms of discovery chemistry capable of
identifying and optimizing TF-directed compounds, delivery solutions to bring biomolecules nimbly across cell membranes,
and, quite plausibly, entirely new classes of agents. As an
example, the general chemical strategy for drug-induced degradation of targeted proteins provides one path to this goal
(Figure 6F) (Winter et al., 2015).
Where known, targeting tissue-specifying master TFs in cancer (ER/PR, AR) has profound clinical benefit. However, our
knowledge of tissue-specifying TFs and dependencies remains
limited. Here, there is a need for more epigenome-based insights
into cancer core regulatory circuitry, elaborated from primary human tumors. Some important insights into specific TF dependencies have recently emerged using such approaches (e.g.,
OCA-B in DLBCL [Chapuy et al., 2013]; LMX1A in a medulloblastoma subtype [Lin et al., 2016]).
The development of clinically useful inhibitors of various components of the transcriptional apparatus will require further
mechanistic understanding of the role of transcriptional addic-
tion in cancer pathogenesis. As an example, CDK9 inhibition is
highly active in CLL; at present, however, the drugs that have
been developed exhibit only a narrow therapeutic index, and
no biomarkers of pharmacologic efficacy exist. Clarity on the
mechanisms that create such dependencies may provide solutions such as therapeutic synergies.
A ubiquitous challenge of targeted cancer therapy is the emergence of tumor cells resistant to the therapeutic agent. While
emerging evidence suggests that transcriptional inhibitors can
suppress the emergence of drug-resister cells when combined
with other therapeutic agents (Sharma et al., 2010; Yashiro-Ohtani et al., 2014), recent studies indicate that tumor cells can
develop resistance against the transcriptional inhibitors (Fong
et al., 2015; Rathert et al., 2015; Shu et al., 2016). Future investigations will need to evaluate the effects of combination therapies that include transcriptional inhibitors.
While genome structural alterations have long been understood to have roles in cancer, the science of genome structure
and function is still in its infancy. More complete mechanistic understanding of insulators, CTCF, and regional roles of chromatinassociated complexes is needed in cancer. Careful dissection of
genome structure and its regulators in experimental and translational model systems will surely reveal new targets.
Improved understanding of dysregulated transcription in the
context of the cancer cell heterogeneity present in individual tumors will be important for the development of effective therapies;
however, precise measures of this heterogeneity at the level of
the epigenome are currently limited by the fact that the generation of genome-wide maps of DNA-associated proteins often requires samples of > 106 cells. As a consequence, current analyses obscure the heterogeneity created by intermingled clonal
subpopulations within individual tumors that may differ from
one another in important ways. The need to understand transcriptional control in, for example, tumor-initiating cells, AKTlow dormant cells, and metastatic carcinoma cells generated
by epithelial-mesenchymal transitions will require single-cell integrated epigenomic analyses. Recent insights into intratumoral
heterogeneity in primary glioblastoma, obtained using single-cell
RNA sequencing (RNA-seq), highlight this need (Patel et al.,
2014).
Finally, we need to study the dynamics of transcriptional control in the most relevant model system: the cancer patient. We do
not presently have the capability to study the dynamic effects of
treatment in order to elucidate mechanisms of response and
resistance. Indeed, future technologies that reveal epigenetic responses to targeted therapy are likely to provide novel, critically
important mechanistic insights that will greatly benefit patients.
ACKNOWLEDGMENTS
We are grateful to Brian Abraham, Brad Bernstein, Daniel Day, Matthew
Fearer, Isaac Klein, Nick Kwiatkowski, Charles Lin, Marc Mansour, Peter
Rahl, Jussi Taipale, William Tansey, Marc Timmers, Chris Vakoc, and Robert
Weinberg for critical comments. Supported by an NIH Grant HG002668
(R.A.Y.), an Erwin Schrödinger Fellowship (J3490) from the Austrian Science
Fund (D.H.), and a Margaret and Herman Sokol Postdoctoral Award (D.H.).
R.A.Y. and J.E.B are a founders and shareholders of Syros Pharmaceuticals.
R.A.Y. is a founder and shareholder of Marauder Therapeutics. J.E.B. is a
shareholder and executive of Novartis Pharmaceuticals, and a shareholder
Cell 168, February 9, 2017 639
of Acetylon Pharmaceuticals. Syros, Novartis, and Acetylon are discovering
and developing therapeutics directed at transcriptional pathways in cancer.
REFERENCES
Affer, M., Chesi, M., Chen, W.D., Keats, J.J., Demchenko, Y.N., Tamizhmani,
K., Garbitt, V.M., Riggs, D.L., Brents, L.A., Roschke, A.V., et al. (2014). Promiscuous MYC locus rearrangements hijack enhancers but mostly super-enhancers to dysregulate MYC expression in multiple myeloma. Leukemia 28,
1725–1735.
and the retinoic acid receptor-alpha locus due to a variant t(11;17) translocation associated with acute promyelocytic leukaemia. EMBO J. 12, 1161–1167.
Chipumuro, E., Marco, E., Christensen, C.L., Kwiatkowski, N., Zhang, T., Hatheway, C.M., Abraham, B.J., Sharma, B., Yeung, C., Altabef, A., et al. (2014).
CDK7 inhibition suppresses super-enhancer-linked oncogenic transcription in
MYCN-driven cancer. Cell 159, 1126–1139.
Christensen, C.L., Kwiatkowski, N., Abraham, B.J., Carretero, J., Al-Shahrour,
F., Zhang, T., Chipumuro, E., Herter-Sprie, G.S., Akbay, E.A., Altabef, A., et al.
(2014). Targeting transcriptional addictions in small cell lung cancer with a covalent CDK7 inhibitor. Cancer Cell 26, 909–922.
Alarcón, C., Zaromytidou, A.I., Xi, Q., Gao, S., Yu, J., Fujisawa, S., Barlas, A.,
Miller, A.N., Manova-Todorova, K., Macias, M.J., et al. (2009). Nuclear CDKs
drive Smad transcriptional activation and turnover in BMP and TGF-beta pathways. Cell 139, 757–769.
Clark, V.E., Harmancı, A.S., Bai, H., Youngblood, M.W., Lee, T.I., Baranoski,
J.F., Ercan-Sencicek, A.G., Abraham, B.J., Weintraub, A.S., Hnisz, D., et al.
(2016). Recurrent somatic mutations in POLR2A define a distinct subset of meningiomas. Nat. Genet. 48, 1253–1259.
Allen, B.L., and Taatjes, D.J. (2015). The Mediator complex: a central integrator
of transcription. Nat. Rev. Mol. Cell Biol. 16, 155–166.
Daigle, S.R., Olhava, E.J., Therkelsen, C.A., Basavapathruni, A., Jin, L., Boriack-Sjodin, P.A., Allain, C.J., Klaus, C.R., Raimondi, A., Scott, M.P., et al.
(2013). Potent inhibition of DOT1L as treatment of MLL-fusion leukemia. Blood
122, 1017–1025.
Amabile, A., Migliara, A., Capasso, P., Biffi, M., Cittaro, D., Naldini, L., and
Lombardo, A. (2016). Inheritable Silencing of Endogenous Genes by Hitand-Run Targeted Epigenetic Editing. Cell 167, 219–232.
Barbieri, C.E., Baca, S.C., Lawrence, M.S., Demichelis, F., Blattner, M., Theurillat, J.P., White, T.A., Stojanov, P., Van Allen, E., Stransky, N., et al. (2012).
Exome sequencing identifies recurrent SPOP, FOXA1 and MED12 mutations
in prostate cancer. Nat. Genet. 44, 685–689.
Baubec, T., Ivánek, R., Lienert, F., and Schübeler, D. (2013). Methylationdependent and -independent genomic targeting principles of the MBD protein
family. Cell 153, 480–492.
Beroukhim, R., Mermel, C.H., Porter, D., Wei, G., Raychaudhuri, S., Donovan,
J., Barretina, J., Boehm, J.S., Dobson, J., Urashima, M., et al. (2010). The landscape of somatic copy-number alteration across human cancers. Nature 463,
899–905.
Bhagwat, A.S., Roe, J.S., Mok, B.Y., Hohmann, A.F., Shi, J., and Vakoc, C.R.
(2016). BET Bromodomain Inhibition Releases the Mediator Complex from
Select cis-Regulatory Elements. Cell Rep. 15, 519–530.
Bhagwat, A.S., and Vakoc, C.R. (2015). Targeting Transcription Factors in
Cancer. Trends Cancer 1, 53–65.
Bisgrove, D.A., Mahmoudi, T., Henklein, P., and Verdin, E. (2007). Conserved
P-TEFb-interacting domain of BRD4 inhibits HIV transcription. Proc. Natl.
Acad. Sci. USA 104, 13690–13695.
Brown, J.D., Lin, C.Y., Duan, Q., Griffin, G., Federation, A.J., Paranal, R.M.,
Bair, S., Newton, G., Lichtman, A.H., Kung, A.L., et al. (2014). NF-kB directs
dynamic super enhancer formation in inflammation and atherogenesis. Mol.
Cell 56, 219–231.
Buganim, Y., Faddah, D.A., and Jaenisch, R. (2013). Mechanisms and models
of somatic cell reprogramming. Nat. Rev. Genet. 14, 427–439.
Bulger, M., and Groudine, M. (2011). Functional and mechanistic diversity of
distal transcription enhancers. Cell 144, 327–339.
Bywater, M.J., Pearson, R.B., McArthur, G.A., and Hannan, R.D. (2013). Dysregulation of the basal RNA polymerase transcription apparatus in cancer. Nat.
Rev. Cancer 13, 299–314.
Cai, S.F., Chen, C.W., and Armstrong, S.A. (2015). Drugging Chromatin in
Cancer: Recent Advances and Novel Approaches. Mol. Cell 60, 561–570.
Campos, E.I., and Reinberg, D. (2009). Histones: annotating chromatin. Annu.
Rev. Genet. 43, 559–599.
Chapuy, B., McKeown, M.R., Lin, C.Y., Monti, S., Roemer, M.G., Qi, J., Rahl,
P.B., Sun, H.H., Yeda, K.T., Doench, J.G., et al. (2013). Discovery and characterization of super-enhancer-associated dependencies in diffuse large B cell
lymphoma. Cancer Cell 24, 777–790.
Chen, C.W., and Armstrong, S.A. (2015). Targeting DOT1L and HOX gene
expression in MLL-rearranged leukemia and beyond. Exp. Hematol. 43,
673–684.
Chen, Z., Brand, N.J., Chen, A., Chen, S.J., Tong, J.H., Wang, Z.Y., Waxman,
S., and Zelent, A. (1993). Fusion between a novel Krüppel-like zinc finger gene
640 Cell 168, February 9, 2017
Darnell, J.E., Jr. (2002). Transcription factors as targets for cancer therapy.
Nat. Rev. Cancer 2, 740–749.
Dawson, M.A., Prinjha, R.K., Dittmann, A., Giotopoulos, G., Bantscheff, M.,
Chan, W.I., Robson, S.C., Chung, C.W., Hopf, C., Savitski, M.M., et al.
(2011). Inhibition of BET recruitment to chromatin as an effective treatment
for MLL-fusion leukaemia. Nature 478, 529–533.
Delmore, J.E., Issa, G.C., Lemieux, M.E., Rahl, P.B., Shi, J., Jacobs, H.M.,
Kastritis, E., Gilpatrick, T., Paranal, R.M., Qi, J., et al. (2011). BET bromodomain inhibition as a therapeutic strategy to target c-Myc. Cell 146, 904–917.
Deshpande, A.J., Chen, L., Fazio, M., Sinha, A.U., Bernt, K.M., Banka, D., Dias,
S., Chang, J., Olhava, E.J., Daigle, S.R., et al. (2013). Leukemic transformation
by the MLL-AF6 fusion oncogene requires the H3K79 methyltransferase Dot1l.
Blood 121, 2533–2541.
Doudna, J.A., and Charpentier, E. (2014). Genome editing. The new frontier of
genome engineering with CRISPR-Cas9. Science 346, 1258096.
Dowen, J.M., Fan, Z.P., Hnisz, D., Ren, G., Abraham, B.J., Zhang, L.N., Weintraub, A.S., Schuijers, J., Lee, T.I., Zhao, K., and Young, R.A. (2014). Control of
cell identity genes occurs in insulated neighborhoods in mammalian chromosomes. Cell 159, 374–387.
Drier, Y., Cotton, M.J., Williamson, K.E., Gillespie, S.M., Ryan, R.J., Kluk, M.J.,
Carey, C.D., Rodig, S.J., Sholl, L.M., Afrogheh, A.H., et al. (2016). An oncogenic MYB feedback loop drives alternate cell fates in adenoid cystic carcinoma. Nat. Genet. 48, 265–272.
Eick, D., and Geyer, M. (2013). The RNA polymerase II carboxy-terminal
domain (CTD) code. Chem. Rev. 113, 8456–8490.
ENCODE Project Consortium (2012). An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74.
Felsher, D.W., and Bishop, J.M. (1999). Reversible tumorigenesis by MYC in
hematopoietic lineages. Mol. Cell 4, 199–207.
Filippakopoulos, P., Qi, J., Picaud, S., Shen, Y., Smith, W.B., Fedorov, O.,
Morse, E.M., Keates, T., Hickman, T.T., Felletar, I., et al. (2010). Selective inhibition of BET bromodomains. Nature 468, 1067–1073.
Flavahan, W.A., Drier, Y., Liau, B.B., Gillespie, S.M., Venteicher, A.S., Stemmer-Rachamimov, A.O., Suvà, M.L., and Bernstein, B.E. (2016). Insulator
dysfunction and oncogene activation in IDH mutant gliomas. Nature 529,
110–114.
Fong, C.Y., Gilan, O., Lam, E.Y., Rubin, A.F., Ftouni, S., Tyler, D., Stanley, K.,
Sinha, D., Yeh, P., Morison, J., et al. (2015). BET inhibitor resistance emerges
from leukaemia stem cells. Nature 525, 538–542.
Garraway, L.A., and Lander, E.S. (2013). Lessons from the cancer genome.
Cell 153, 17–37.
Geng, F., Wenzel, S., and Tansey, W.P. (2012). Ubiquitin and proteasomes in
transcription. Annu. Rev. Biochem. 81, 177–201.
Ghirlando, R., and Felsenfeld, G. (2016). CTCF: making the right connections.
Genes Dev. 30, 881–891.
Gibcus, J.H., and Dekker, J. (2013). The hierarchy of the 3D genome. Mol. Cell
49, 773–782.
Gorkin, D.U., Leung, D., and Ren, B. (2014). The 3D genome in transcriptional
regulation and pluripotency. Cell Stem Cell 14, 762–775.
Graf, T., and Enver, T. (2009). Forcing cells to change lineages. Nature 462,
587–594.
Gröschel, S., Sanders, M.A., Hoogenboezem, R., de Wit, E., Bouwman, B.A.,
Erpelinck, C., van der Velden, V.H., Havermans, M., Avellino, R., van Lom, K.,
et al. (2014). A single oncogenic enhancer rearrangement causes concomitant
EVI1 and GATA2 deregulation in leukemia. Cell 157, 369–381.
Guenther, M.G., Lawton, L.N., Rozovskaia, T., Frampton, G.M., Levine, S.S.,
Volkert, T.L., Croce, C.M., Nakamura, T., Canaani, E., and Young, R.A.
(2008). Aberrant chromatin at genes encoding stem cell regulators in human
mixed-lineage leukemia. Genes Dev. 22, 3403–3408.
Hah, N., Benner, C., Chong, L.W., Yu, R.T., Downes, M., and Evans, R.M.
(2015). Inflammation-sensitive super enhancers form domains of coordinately
regulated enhancer RNAs. Proc. Natl. Acad. Sci. USA 112, E297–E302.
Hay, D., Hughes, J.R., Babbs, C., Davies, J.O., Graham, B.J., Hanssen, L.L.,
Kassouf, M.T., Oudelaar, A.M., Sharpe, J.A., Suciu, M.C., et al. (2016). Genetic
dissection of the a-globin super-enhancer in vivo. Nat. Genet. 48, 895–903.
Hnisz, D., Abraham, B.J., Lee, T.I., Lau, A., Saint-André, V., Sigova, A.A., Hoke,
H.A., and Young, R.A. (2013). Super-enhancers in the control of cell identity
and disease. Cell 155, 934–947.
Hnisz, D., Schuijers, J., Lin, C.Y., Weintraub, A.S., Abraham, B.J., Lee, T.I.,
Bradner, J.E., and Young, R.A. (2015). Convergence of developmental and
oncogenic signaling pathways at transcriptional super-enhancers. Mol. Cell
58, 362–370.
Hnisz, D., Day, D.S., and Young, R.A. (2016a). Insulated Neighborhoods:
Structural and Functional Units of Mammalian Gene Control. Cell 167, 1188–
1200.
Hnisz, D., Weintraub, A.S., Day, D.S., Valton, A.L., Bak, R.O., Li, C.H., Goldmann, J., Lajoie, B.R., Fan, Z.P., Sigova, A.A., et al. (2016b). Activation of
proto-oncogenes by disruption of chromosome neighborhoods. Science
351, 1454–1458.
Huang, M.E., Ye, Y.C., Chen, S.R., Chai, J.R., Lu, J.X., Zhoa, L., Gu, L.J., and
Wang, Z.Y. (1988). Use of all-trans retinoic acid in the treatment of acute promyelocytic leukemia. Blood 72, 567–572.
Inaba, H., and Pui, C.H. (2010). Glucocorticoid use in acute lymphoblastic
leukaemia. Lancet Oncol. 11, 1096–1106.
Jain, M., Arvanitis, C., Chu, K., Dewey, W., Leonhardt, E., Trinh, M., Sundberg,
C.D., Bishop, J.M., and Felsher, D.W. (2002). Sustained loss of a neoplastic
phenotype by brief inactivation of MYC. Science 297, 102–104.
Ji, X., Dadon, D.B., Powell, B.E., Fan, Z.P., Borges-Rivera, D., Shachar, S.,
Weintraub, A.S., Hnisz, D., Pegoraro, G., Lee, T.I., et al. (2016). 3D Chromosome Regulatory Landscape of Human Pluripotent Cells. Cell Stem Cell 18,
262–275.
Jiang, T., Raviram, R., Snetkova, V., Rocha, P.P., Proudhon, C., Badri, S., Bonneau, R., Skok, J.A., and Kluger, Y. (2016). Identification of multi-loci hubs from
4C-seq demonstrates the functional importance of simultaneous interactions.
Nucleic Acids Res. 44, 8714–8725.
Jones, P.A., Issa, J.P., and Baylin, S. (2016). Targeting the cancer epigenome
for therapy. Nat. Rev. Genet. 17, 630–641.
Kagey, M.H., Newman, J.J., Bilodeau, S., Zhan, Y., Orlando, D.A., van Berkum, N.L., Ebmeier, C.C., Goossens, J., Rahl, P.B., Levine, S.S., et al.
(2010). Mediator and cohesin connect gene expression and chromatin architecture. Nature 467, 430–435.
Kandoth, C., McLellan, M.D., Vandin, F., Ye, K., Niu, B., Lu, C., Xie, M., Zhang,
Q., McMichael, J.F., Wyczalkowski, M.A., et al. (2013). Mutational landscape
and significance across 12 major cancer types. Nature 502, 333–339.
Katainen, R., Dave, K., Pitkänen, E., Palin, K., Kivioja, T., Välimäki, N., Gylfe,
A.E., Ristolainen, H., Hänninen, U.A., Cajuso, T., et al. (2015). CTCF/cohesin-binding sites are frequently mutated in cancer. Nat. Genet. 47, 818–821.
Kennedy, A.L., Vallurupalli, M., Chen, L., Crompton, B., Cowley, G., Vazquez,
F., Weir, B.A., Tsherniak, A., Parasuraman, S., Kim, S., et al. (2015). Functional,
chemical genomic, and super-enhancer screening identify sensitivity to cyclin
D1/CDK4 pathway inhibition in Ewing sarcoma. Oncotarget 6, 30178–30193.
Kieffer-Kwon, K.R., Tang, Z., Mathe, E., Qian, J., Sung, M.H., Li, G., Resch, W.,
Baek, S., Pruett, N., Grøntved, L., et al. (2013). Interactome maps of mouse
gene regulatory domains reveal basic principles of transcriptional regulation.
Cell 155, 1507–1520.
Kouzarides, T. (2007). Chromatin modifications and their function. Cell 128,
693–705.
Krivtsov, A.V., Feng, Z., Lemieux, M.E., Faber, J., Vempati, S., Sinha, A.U., Xia,
X., Jesneck, J., Bracken, A.P., Silverman, L.B., et al. (2008). H3K79 methylation profiles define murine and human MLL-AF4 leukemias. Cancer Cell 14,
355–368.
Roadmap Epigenomics Consortium, Kundaje, A., Meuleman, W., Ernst, J., Bilenky, M., Yen, A., Heravi-Moussavi, A., Kheradpour, P., Zhang, Z., Wang, J.,
Ziller, M.J., et al. (2015). Integrative analysis of 111 reference human epigenomes. Nature 518, 317–330.
Kwiatkowski, N., Zhang, T., Rahl, P.B., Abraham, B.J., Reddy, J., Ficarro, S.B.,
Dastur, A., Amzallag, A., Ramaswamy, S., Tesar, B., et al. (2014). Targeting
transcription regulation in cancer with a covalent CDK7 inhibitor. Nature 511,
616–620.
Lai, F., Orom, U.A., Cesaroni, M., Beringer, M., Taatjes, D.J., Blobel, G.A., and
Shiekhattar, R. (2013). Activating RNAs associate with Mediator to enhance
chromatin architecture and transcription. Nature 494, 497–501.
Lane, D., and Levine, A. (2010). p53 Research: the past thirty years and the
next thirty years. Cold Spring Harb. Perspect. Biol. 2, a000893.
Lawrence, M.S., Stojanov, P., Mermel, C.H., Robinson, J.T., Garraway, L.A.,
Golub, T.R., Meyerson, M., Gabriel, S.B., Lander, E.S., and Getz, G. (2014).
Discovery and saturation analysis of cancer genes across 21 tumour types.
Nature 505, 495–501.
Lee, T.I., and Young, R.A. (2013). Transcriptional regulation and its misregulation in disease. Cell 152, 1237–1251.
Levine, M., Cattoglio, C., and Tjian, R. (2014). Looping back to leap forward:
transcription enters a new era. Cell 157, 13–25.
Li, W., Notani, D., and Rosenfeld, M.G. (2016). Enhancers as non-coding RNA
transcription units: recent insights and future perspectives. Nat. Rev. Genet.
17, 207–223.
Lin, C.Y., Lovén, J., Rahl, P.B., Paranal, R.M., Burge, C.B., Bradner, J.E., Lee,
T.I., and Young, R.A. (2012). Transcriptional amplification in tumor cells with
elevated c-Myc. Cell 151, 56–67.
Lin, C.Y., Erkek, S., Tong, Y., Yin, L., Federation, A.J., Zapatka, M., Haldipur,
P., Kawauchi, D., Risch, T., Warnatz, H.J., et al. (2016). Active medulloblastoma enhancers reveal subgroup-specific cellular origins. Nature 530, 57–62.
Jonkers, I., and Lis, J.T. (2015). Getting up to speed with transcription elongation by RNA polymerase II. Nat. Rev. Mol. Cell Biol. 16, 167–177.
Liu, M.Y., DeNizio, J.E., Schutsky, E.K., and Kohli, R.M. (2016a). The expanding scope and impact of epigenetic cytosine modifications. Curr. Opin. Chem.
Biol. 33, 67–73.
Joo, J.Y., Schaukowitch, K., Farbiak, L., Kilaru, G., and Kim, T.K. (2016). Stimulus-specific combinatorial functionality of neuronal c-fos enhancers. Nat.
Neurosci. 19, 75–83.
Liu, X.S., Wu, H., Ji, X., Stelzer, Y., Wu, X., Czauderna, S., Shu, J., Dadon, D.,
Young, R.A., and Jaenisch, R. (2016b). Editing DNA Methylation in the
Mammalian Genome. Cell 167, 233–247.
Kadoch, C., and Crabtree, G.R. (2015). Mammalian SWI/SNF chromatin remodeling complexes and cancer: Mechanistic insights gained from human genomics. Sci. Adv. 1, e1500447.
Lo-Coco, F., Avvisati, G., Vignetti, M., Thiede, C., Orlando, S.M., Iacobelli, S.,
Ferrara, F., Fazi, P., Cicconi, L., Di Bona, E., et al.; Gruppo Italiano Malattie
Ematologiche dell’Adulto; German-Austrian Acute Myeloid Leukemia Study
Cell 168, February 9, 2017 641
Group; Study Alliance Leukemia (2013). Retinoic acid and arsenic trioxide for
acute promyelocytic leukemia. N. Engl. J. Med. 369, 111–121.
Long, H.K., Prescott, S.L., and Wysocka, J. (2016). Ever-Changing Landscapes: Transcriptional Enhancers in Development and Evolution. Cell 167,
1170–1187.
Lovén, J., Hoke, H.A., Lin, C.Y., Lau, A., Orlando, D.A., Vakoc, C.R., Bradner,
J.E., Lee, T.I., and Young, R.A. (2013). Selective inhibition of tumor oncogenes
by disruption of super-enhancers. Cell 153, 320–334.
Mäkinen, N., Mehine, M., Tolvanen, J., Kaasinen, E., Li, Y., Lehtonen, H.J.,
Gentile, M., Yan, J., Enge, M., Taipale, M., et al. (2011). MED12, the mediator
complex subunit 12 gene, is mutated at high frequency in uterine leiomyomas.
Science 334, 252–255.
Malik, S., and Roeder, R.G. (2010). The metazoan Mediator co-activator complex as an integrative hub for transcriptional regulation. Nat. Rev. Genet. 11,
761–772.
Mansour, M.R., Abraham, B.J., Anders, L., Berezovskaya, A., Gutierrez, A.,
Durbin, A.D., Etchin, J., Lawton, L., Sallan, S.E., Silverman, L.B., et al.
(2014). Oncogene regulation. An oncogenic super-enhancer formed through
somatic mutation of a noncoding intergenic element. Science 346, 1373–1377.
Mellén, M., Ayata, P., Dewell, S., Kriaucionis, S., and Heintz, N. (2012). MeCP2
binds to 5hmC enriched within active genes and accessible chromatin in the
nervous system. Cell 151, 1417–1430.
Moellering, R.E., Cornejo, M., Davis, T.N., Del Bianco, C., Aster, J.C., Blacklow, S.C., Kung, A.L., Gilliland, D.G., Verdine, G.L., and Bradner, J.E. (2009).
Direct inhibition of the NOTCH transcription factor complex. Nature 462,
182–188.
Morris, S.A., and Daley, G.Q. (2013). A blueprint for engineering cell fate: current technologies to reprogram cell identity. Cell Res. 23, 33–48.
Mullen, A.C., Orlando, D.A., Newman, J.J., Lovén, J., Kumar, R.M., Bilodeau,
S., Reddy, J., Guenther, M.G., DeKoter, R.P., and Young, R.A. (2011). Master
transcription factors determine cell-type-specific responses to TGF-b
signaling. Cell 147, 565–576.
Nabet, B., Ó Broin, P., Reyes, J.M., Shieh, K., Lin, C.Y., Will, C.M., Popovic, R.,
Ezponda, T., Bradner, J.E., Golden, A.A., and Licht, J.D. (2015). Deregulation
of the Ras-Erk Signaling Axis Modulates the Enhancer Landscape. Cell Rep.
12, 1300–1313.
Nie, Z., Hu, G., Wei, G., Cui, K., Yamane, A., Resch, W., Wang, R., Green, D.R.,
Tessarollo, L., Casellas, R., et al. (2012). c-Myc is a universal amplifier of expressed genes in lymphocytes and embryonic stem cells. Cell 151, 68–79.
Northcott, P.A., Lee, C., Zichner, T., Stütz, A.M., Erkek, S., Kawauchi, D., Shih,
D.J., Hovestadt, V., Zapatka, M., Sturm, D., et al. (2014). Enhancer hijacking
activates GFI1 family oncogenes in medulloblastoma. Nature 511, 428–434.
Ong, C.T., and Corces, V.G. (2011). Enhancer function: new insights into the
regulation of tissue-specific gene expression. Nat. Rev. Genet. 12, 283–293.
Pagliarini, R., Shao, W., and Sellers, W.R. (2015). Oncogene addiction: pathways of therapeutic response, resistance, and road maps toward a cure.
EMBO Rep. 16, 280–296.
Patel, A.P., Tirosh, I., Trombetta, J.J., Shalek, A.K., Gillespie, S.M., Wakimoto,
H., Cahill, D.P., Nahed, B.V., Curry, W.T., Martuza, R.L., et al. (2014). Singlecell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma.
Science 344, 1396–1401.
Rahl, P.B., Lin, C.Y., Seila, A.C., Flynn, R.A., McCuine, S., Burge, C.B., Sharp,
P.A., and Young, R.A. (2010). c-Myc regulates transcriptional pause release.
Cell 141, 432–445.
Rathert, P., Roth, M., Neumann, T., Muerdter, F., Roe, J.S., Muhar, M., Deswal,
S., Cerny-Reiterer, S., Peter, B., Jude, J., et al. (2015). Transcriptional plasticity promotes primary and acquired resistance to BET inhibition. Nature
525, 543–547.
Rowley, J.D., Golomb, H.M., and Dougherty, C. (1977). 15/17 translocation, a
consistent chromosomal change in acute promyelocytic leukaemia. Lancet 1,
549–550.
Sancho-Martinez, I., Baek, S.H., and Izpisua Belmonte, J.C. (2012). Lineage
conversion methodologies meet the reprogramming toolbox. Nat. Cell Biol.
14, 892–899.
Sanda, T., Lawton, L.N., Barrasa, M.I., Fan, Z.P., Kohlhammer, H., Gutierrez,
A., Ma, W., Tatarek, J., Ahn, Y., Kelliher, M.A., et al. (2012). Core transcriptional
regulatory circuit controlled by the TAL1 complex in human T cell acute
lymphoblastic leukemia. Cancer Cell 22, 209–221.
Sharma, S.V., Lee, D.Y., Li, B., Quinlan, M.P., Takahashi, F., Maheswaran, S.,
McDermott, U., Azizian, N., Zou, L., Fischbach, M.A., et al. (2010). A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations.
Cell 141, 69–80.
Shi, J., and Vakoc, C.R. (2014). The mechanisms behind the therapeutic activity of BET bromodomain inhibition. Mol. Cell 54, 728–736.
Shi, J., Whyte, W.A., Zepeda-Mendoza, C.J., Milazzo, J.P., Shen, C., Roe,
J.S., Minder, J.L., Mercan, F., Wang, E., Eckersley-Maslin, M.A., et al.
(2013). Role of SWI/SNF in acute leukemia maintenance and enhancer-mediated Myc regulation. Genes Dev. 27, 2648–2662.
Shin, H.Y., Willi, M., Yoo, K.H., Zeng, X., Wang, C., Metser, G., and Hennighausen, L. (2016). Hierarchy within the mammary STAT5-driven Wap superenhancer. Nat. Genet. 48, 904–911.
Shu, S., Lin, C.Y., He, H.H., Witwicki, R.M., Tabassum, D.P., Roberts, J.M.,
Janiszewska, M., Huh, S.J., Liang, Y., Ryan, J., et al. (2016). Response and
resistance to BET bromodomain inhibitors in triple-negative breast cancer.
Nature 529, 413–417.
Soshnev, A.A., Josefowicz, S.Z., and Allis, C.D. (2016). Greater Than the Sum
of Parts: Complexity of the Dynamic Epigenome. Mol. Cell 62, 681–694.
Soucek, L., Whitfield, J., Martins, C.P., Finch, A.J., Murphy, D.J., Sodir, N.M.,
Karnezis, A.N., Swigart, L.B., Nasi, S., and Evan, G.I. (2008). Modelling Myc inhibition as a cancer therapy. Nature 455, 679–683.
Soucek, L., Whitfield, J.R., Sodir, N.M., Massó-Vallés, D., Serrano, E., Karnezis, A.N., Swigart, L.B., and Evan, G.I. (2013). Inhibition of Myc family proteins
eradicates KRas-driven lung cancer in mice. Genes Dev. 27, 504–513.
Spitz, F., and Furlong, E.E. (2012). Transcription factors: from enhancer binding to developmental control. Nat. Rev. Genet. 13, 613–626.
Stathis, A., Zucca, E., Bekradda, M., Gomez-Roca, C., Delord, J.P., de La
Motte Rouge, T., Uro-Coste, E., de Braud, F., Pelosi, G., and French, C.A.
(2016). Clinical Response of Carcinomas Harboring the BRD4-NUT Oncoprotein to the Targeted Bromodomain Inhibitor OTX015/MK-8628. Cancer Discov. 6, 492–500.
Stratton, M.R., Campbell, P.J., and Futreal, P.A. (2009). The cancer genome.
Nature 458, 719–724.
Pelish, H.E., Liau, B.B., Nitulescu, I.I., Tangpeerachaikul, A., Poss, Z.C., Da
Silva, D.H., Caruso, B.T., Arefolov, A., Fadeyi, O., Christie, A.L., et al. (2015).
Mediator kinase inhibition further activates super-enhancer-associated genes
in AML. Nature 526, 273–276.
Stuart, D., and Sellers, W.R. (2009). Linking somatic genetic alterations in cancer to therapeutics. Curr. Opin. Cell Biol. 21, 304–310.
Phillips-Cremins, J.E., and Corces, V.G. (2013). Chromatin insulators: linking
genome organization to cellular function. Mol. Cell 50, 461–474.
Sur, I., and Taipale, J. (2016). The role of enhancers in cancer. Nat. Rev. Cancer
16, 483–493.
Piunti, A., and Shilatifard, A. (2016). Epigenetic balance of gene expression by
Polycomb and COMPASS families. Science 352, aad9780.
Taub, R., Kirsch, I., Morton, C., Lenoir, G., Swan, D., Tronick, S., Aaronson, S.,
and Leder, P. (1982). Translocation of the c-myc gene into the immunoglobulin
heavy chain locus in human Burkitt lymphoma and murine plasmacytoma
cells. Proc. Natl. Acad. Sci. USA 79, 7837–7841.
Puissant, A., Frumm, S.M., Alexe, G., Bassil, C.F., Qi, J., Chanthery, Y.H., Nekritz, E.A., Zeid, R., Gustafson, W.C., Greninger, P., et al. (2013). Targeting
MYCN in neuroblastoma by BET bromodomain inhibition. Cancer Discov. 3,
308–323.
642 Cell 168, February 9, 2017
Taylor, S.M., and Jones, P.A. (1979). Multiple new phenotypes induced in
10T1/2 and 3T3 cells treated with 5-azacytidine. Cell 17, 771–779.
Tessarz, P., and Kouzarides, T. (2014). Histone core modifications regulating
nucleosome structure and dynamics. Nat. Rev. Mol. Cell Biol. 15, 703–708.
Watson, I.R., Takahashi, K., Futreal, P.A., and Chin, L. (2013). Emerging patterns of somatic mutations in cancer. Nat. Rev. Genet. 14, 703–718.
Thomas, D., and Tyers, M. (2000). Transcriptional regulation: Kamikaze activators. Curr. Biol. 10, R341–R343.
Weinstein, I.B., and Joe, A.K. (2006). Mechanisms of disease: Oncogene
addiction–a rationale for molecular targeting in cancer therapy. Nat. Clin.
Pract. Oncol. 3, 448–457.
Tomazou, E.M., Sheffield, N.C., Schmidl, C., Schuster, M., Schönegger, A.,
Datlinger, P., Kubicek, S., Bock, C., and Kovar, H. (2015). Epigenome mapping
reveals distinct modes of gene regulation and widespread enhancer reprogramming by the oncogenic fusion protein EWS-FLI1. Cell Rep. 10, 1082–1095.
Trompouki, E., Bowman, T.V., Lawton, L.N., Fan, Z.P., Wu, D.C., DiBiase, A.,
Martin, C.S., Cech, J.N., Sessa, A.K., Leblanc, J.L., et al. (2011). Lineage regulators direct BMP and Wnt pathways to cell-specific programs during differentiation and regeneration. Cell 147, 577–589.
Vaquerizas, J.M., Kummerfeld, S.K., Teichmann, S.A., and Luscombe, N.M.
(2009). A census of human transcription factors: function, expression and evolution. Nat. Rev. Genet. 10, 252–263.
Vierbuchen, T., and Wernig, M. (2012). Molecular roadblocks for cellular reprogramming. Mol. Cell 47, 827–838.
Viny, A.D., Ott, C.J., Spitzer, B., Rivas, M., Meydan, C., Papalexi, E., Yelin, D.,
Shank, K., Reyes, J., Chiu, A., et al. (2015). Dose-dependent role of the cohesin
complex in normal and malignant hematopoiesis. J. Exp. Med. 212, 1819–
1832.
Vogelstein, B., Papadopoulos, N., Velculescu, V.E., Zhou, S., Diaz, L.A., Jr.,
and Kinzler, K.W. (2013). Cancer genome landscapes. Science 339, 1546–
1558.
Walker, B.A., Wardell, C.P., Brioli, A., Boyle, E., Kaiser, M.F., Begum, D.B., Dahir, N.B., Johnson, D.C., Ross, F.M., Davies, F.E., and Morgan, G.J. (2014).
Translocations at 8q24 juxtapose MYC with genes that harbor superenhancers
resulting in overexpression and poor prognosis in myeloma patients. Blood
Cancer J. 4, e191.
Whyte, W.A., Orlando, D.A., Hnisz, D., Abraham, B.J., Lin, C.Y., Kagey, M.H.,
Rahl, P.B., Lee, T.I., and Young, R.A. (2013). Master transcription factors and
mediator establish super-enhancers at key cell identity genes. Cell 153, 307–
319.
Winter, G.E., Buckley, D.L., Paulk, J., Roberts, J.M., Souza, A., Dhe-Paganon,
S., and Bradner, J.E. (2015). DRUG DEVELOPMENT. Phthalimide conjugation
as a strategy for in vivo target protein degradation. Science 348, 1376–1381.
Yamanaka, S. (2012). Induced pluripotent stem cells: past, present, and future.
Cell Stem Cell 10, 678–684.
Yang, H., Schramek, D., Adam, R.C., Keyes, B.E., Wang, P., Zheng, D., and
Fuchs, E. (2015). ETS family transcriptional regulators drive chromatin dynamics and malignancy in squamous cell carcinomas. eLife 4, e10870.
Yashiro-Ohtani, Y., Wang, H., Zang, C., Arnett, K.L., Bailis, W., Ho, Y., Knoechel, B., Lanauze, C., Louis, L., Forsyth, K.S., et al. (2014). Long-range
enhancer activity determines Myc sensitivity to Notch inhibitors in T cell leukemia. Proc. Natl. Acad. Sci. USA 111, E4946–E4953.
Zhang, T., Kwiatkowski, N., Olson, C.M., Dixon-Clarke, S.E., Abraham, B.J.,
Greifenberg, A.K., Ficarro, S.B., Elkins, J.M., Liang, Y., Hannett, N.M., et al.
(2016a). Covalent targeting of remote cysteine residues to develop CDK12
and CDK13 inhibitors. Nat. Chem. Biol. 12, 876–884.
Zhang, X., Choi, P.S., Francis, J.M., Imielinski, M., Watanabe, H., Cherniack,
A.D., and Meyerson, M. (2016b). Identification of focally amplified lineage-specific super-enhancers in human epithelial cancers. Nat. Genet. 48, 176–182.
Wang, Z.Y., and Chen, Z. (2008). Acute promyelocytic leukemia: from highly
fatal to highly curable. Blood 111, 2505–2515.
Ziller, M.J., Gu, H., Müller, F., Donaghey, J., Tsai, L.T., Kohlbacher, O., De Jager, P.L., Rosen, E.D., Bennett, D.A., Bernstein, B.E., et al. (2013). Charting
a dynamic DNA methylation landscape of the human genome. Nature 500,
477–481.
Wang, Y., Zhang, T., Kwiatkowski, N., Abraham, B.J., Lee, T.I., Xie, S., Yuzugullu, H., Von, T., Li, H., Lin, Z., et al. (2015). CDK7-dependent transcriptional
addiction in triple-negative breast cancer. Cell 163, 174–186.
Zuber, J., Shi, J., Wang, E., Rappaport, A.R., Herrmann, H., Sison, E.A., Magoon, D., Qi, J., Blatt, K., Wunderlich, M., et al. (2011). RNAi screen identifies
Brd4 as a therapeutic target in acute myeloid leukaemia. Nature 478, 524–528.
Cell 168, February 9, 2017 643
Leading Edge
Review
Endogenous DNA Damage as a Source of Genomic
Instability in Cancer
Anthony Tubbs1 and André Nussenzweig1,*
1Laboratory of Genome Integrity, NIH, Bethesda, MD 20892, USA
*Correspondence: andre_nussenzweig@nih.gov
http://dx.doi.org/10.1016/j.cell.2017.01.002
Genome instability, defined as higher than normal rates of mutation, is a double-edged sword. As a
source of genetic diversity and natural selection, mutations are beneficial for evolution. On the other
hand, genomic instability can have catastrophic consequences for age-related diseases such as
cancer. Mutations arise either from inactivation of DNA repair pathways or in a repair-competent
background due to genotoxic stress from celluar processes such as transcription and replication
that overwhelm high-fidelity DNA repair. Here, we review recent studies that shed light on endogenous sources of mutation and epigenomic features that promote genomic instability during cancer
evolution.
Introduction
DNA is the template for the basic processes of replication and
transcription, making the maintenance of genetic stability critical
for viability. Even before the discovery of the double helix in 1953,
it was known that exogenous agents that we are exposed to in
our everyday lives, such as X-rays, ultraviolet (UV) light, and
various chemicals, can cause genetic changes that can promote
cancer (Friedberg, 2008). It took an additional 10 years after its
stable structure was elucidated to recognize that DNA is also
subject to constant assault from endogenous sources during
normal metabolism (Lindahl, 1993; Lindahl and Nyberg, 1972).
Although both exogenous and endogenous lesions have the potential to modify the basic building blocks of genetic information,
the relative contributions of intrinsic and extrinsic factors to
organ-specific cancer incidence remains unclear (Tomasetti
and Vogelstein, 2015; Wu et al., 2016).
It has been estimated that each human cell is subject to
approximately 70,000 lesions per day (Figure 1) (Lindahl and
Barnes, 2000). The majority of lesions (75%) are single-strand
DNA (ssDNA) breaks, which can arise from oxidative damage
during metabolism or base hydrolysis. ssDNA breaks can also
be converted to DNA double-strand breaks (DSBs), which
although much less frequent, are more dangerous. Given this
extraordinary daily barrage of endogenous and exogenous DNA
damage, it became evident that cells must have acquired enzymes during evolution that repaired DNA anomalies and thereby
restored genome integrity (Friedberg, 2008; Lindahl and Barnes,
2000). In a personal perspective 21 years after the discovery of
the double helix, Francis Crick wrote, ‘‘We totally missed the
possible role of enzymes in repair although . I later came to
realize that DNA is so precious that probably many distinct mechanisms would exist. Nowadays, one could hardly discuss mutation without considering repair at the same time’’ (Crick, 1974).
Thanks to the pioneering work of Tomas Lindahl, Paul Modrich,
and Aziz Sancar (who shared the 2015 Nobel Prize in Chemistry),
among others, various DNA repair pathways were identified that
644 Cell 168, February 9, 2017 Published by Elsevier Inc.
protect cells from different lesions to which they are subjected
(Figure 2) (Kunkel, 2015). Tomas Lindahl discovered the pathway
that repairs modified bases (base excision repair [BER]), and
Paul Modrich discovered a distinct pathway that detects and removes bases that are mis-incorporated during DNA replication
(mismatch repair [MMR]) (Figure 2D), whereas the mechanism
for removal of bulky adducts in DNA (nucleotide excision repair
[NER]) (Figure 2C), was proposed by Aziz Sancar. Each of these
DNA repair pathways excise a damaged region and insert new
bases to fill the gap (Figures 2A–2D).
The mechanisms for DSB repair are distinct and consist of two
sub-pathways called homologous recombination (HR) and nonhomologous end-joining (NHEJ). During HR, the DSB is repaired
by exchanges of equivalent regions of DNA between homologous or sister chromosomes, whereas NHEJ religates the ends
without the use of a template. NHEJ frequently leaves insertions
or deletions at the breakpoint and therefore tends to be error
prone whereas HR-mediated repair is of higher fidelity. Each of
the DNA repair pathways needs to be coordinated with a series
of signaling responses that arrest cell division or trigger cell
death in case the lesions are irreparable (Haber, 2015).
Although DNA repair pathways have been extensively studied,
the precise identity and sources of DNA damage that shape the
mutational landscape of the cancer genome remain unclear. The
Cancer Genome Project and the Cancer Genome Atlas have intersected with the field of DNA repair in the last few years as genetic mutations responsible for cancer have been identified.
These large scale sequencing and bioinformatics approaches
have revealed a remarkable diversity of somatic mutations, as
well as specific signatures of DNA damage and errors in DNA
repair in various cancers. These errors may occur if specific
DNA repair and/or checkpoint pathways are inactivated, if DNA
editing enzymes are deregulated, or if the damage load overwhelms DNA restorative capacity. Although exogenous agents
such as carcinogens were once thought to be the major source
of mutation, technological advances in the detection of DNA
Figure 1. Estimated Frequencies of DNA
Lesions and Mutations Associated with
Dysfunctional DNA Repair
damage have revealed diverse and abundant types of endogenous DNA damage. Like noxious foreign chemicals, baseline
DNA damage by endogenous processes may also overwhelm
functional DNA repair machinery to generate mutations. Here,
we review how endogenous sources of DNA damage and chromatin organization contribute to mutational processes that have
been recorded in cancer genomes.
Inactivation of DNA Repair Pathways in Cancer
The classic mutator hypothesis (Loeb et al., 1974; Nowell, 1976)
postulates that loss of DNA repair genes leads to genome instability, which in turn increases mutation rates at other genomic
sites, leading to cellular transformation. Consistent with this hypothesis, an elevated mutation load associated with defective
DNA repair underlies various hereditary cancers. For example,
hereditary non-polyposis colon cancer is caused by defective
MMR, and a large fraction of hereditary breast and ovarian cancers are accounted for by mutations in genes (BRCA1 and
BRCA2) that control DSB repair by HR (Figure 2E). Until recently,
it has been difficult to detect germline or somatic DNA repair deficiencies in most types of cancer. However, next-generation
sequencing has revealed more widespread deficits in DNA repair
pathways across different cancers. For example, it has been
found that men with advanced high-risk prostate cancer are
more than five times as likely to harbor either heritable or somatic
mutations in HR-mediated repair genes (most commonly BRCA2
and ATM) and MMR genes (MLH1 and MLH2) than patients with
low-risk tumors are (Pritchard et al., 2016). Whole-genome sequences of breast cancers have also revealed deficiencies in
both HR and MMR (Morganella et al., 2016; Nik-Zainal et al.,
2016). Although these discoveries will undoubtedly provide
new targets for therapy, it remains puzzling why malfunction of
DNA repair genes is associated with organ-specific cancers
despite their ubiquitous expression and function in all cell and
tissue types. Variations in cancer risk could result from tissuespecific vulnerabilities to distinct intrinsic or extrinsic factors
leading to distinct mutational processes.
Base Substitution and Rearrangement Signatures in
Cancer
The classification of different mutation and rearrangement types
into distinct signatures has uncovered pathways associated with
defective DNA maintenance (Alexandrov
and Stratton, 2014). Although different
mutational processes inevitably operate
during the course of malignancy, there
can be specific mutation types that dominate the repertoire for a particular cancer
type. For example, a high proportion of
thymine to guanine (T>G) base mutations
found at the immunoglobulin (Ig) genes is
the dominant feature of chronic lymphocytic leukemia, which reflects the activity of the error-prone
DNA polymerase h during somatic hypermutation (Figure 3C)
(Alexandrov et al., 2013). In melanoma, the majority of mutations
are cytosine to thymine (C>T) at adjacent pyrimidine residues,
which reflect the generation of thymine dimers after exposure
to UV light. By classifying mutations according to the substitution
and the sequence context immediately surrounding the mutated
base and chromosomal rearrangements types (ie. deletions,
duplication, inversions, and translocations), it is possible to
deconvolve the spectrum of mutations into combinations of
individual mutation types (Morganella et al., 2016; Nik-Zainal
et al., 2016). In this way, tumors can be characterized by a combination of distinct signatures with different relative strengths
that reflect the underlying mutational processes that occurred
during tumorigenesis.
These analyses have revealed distinct signatures associated
with deficiencies in HR, MMR, and NER in various cancers (Helleday et al., 2014). Interestingly, a mutational signature attributed
to impaired BER activity has not been found so far, perhaps
because loss of BER would compromise fitness. In some instances, the mutation(s) underlying the deregulated DNA repair
pathway has not been determined, suggesting that new caretaker genes could be uncovered. It is also conceivable that in
some cases, signatures might not reflect constitutive loss of
DNA repair activity per se, but rather a strong intermittent process that temporarily overwhelmed the DNA repair capacity or
provoked error-prone repair. Just as transient UV exposure will
produce a signature similar to that expected in NER-deficient
cancers, cellular processes such as oncogene activation, hormone signaling, and inflammation may occur sporadically in
bursts, which could overwhelm an intact DNA repair pathway.
In this context, recent single-cell analyses of triple negative
breast cancer suggest that the majority of copy number aberrations are acquired in moments of crisis, followed by clonal
expansions that produce the tumor mass (Gao et al., 2016).
Examples include chromothripsis, in which a massive rearrangement was acquired in a single event, and chromoplexy, which
involves coordinate and simultaneous rearrangements of multiple chromosomes (Zhang et al., 2013). Recent work also
challenges the classic theory that pancreatic cancer evolves
gradually through random acquisition of driving mutations,
instead favoring a model where multiple driving mutations occur
Cell 168, February 9, 2017 645
(legend on next page)
646 Cell 168, February 9, 2017
simultaneously through genome rearrangements (Notta et al.,
2016). These bursts of genomic instability may be facilitated by
replication stress or transcriptional stress (described below) in
an otherwise DNA-repair-proficient background.
Stochastic Errors in DNA Replication or Spontaneous
Deamination as a Source of Mutation
Among the estimated 70,000 lesions per day that a normal cell is
forced to cope with, each has the potential to be converted into a
mutation. The most frequent single-nucleotide substitutions
found in normal tissues and human tumors are transversion mutations that change a purine nucleotide to another purine (A>G)
or a pyrimidine nucleotide to another pyrimidine (C>T)(Alexandrov and Stratton, 2014). Likely sources of these mutations are
misincorporation by DNA polymerases or spontaneous deamination of cytosine to uracil.
Each time a cell divides, 6 3 109 nucleotides are replicated by
DNA polymerases. DNA polymerases are also responsible for
filling in single-strand gaps of DNA when mutated bases are
excised during NER, MMR, and BER. Replication is predicted to
generate mutations at a low but constant rate, and the variation
in cancer risk across different tissues might be explained by the
total number of cell divisions in normal stem cells derived from
that tissue (Tomasetti and Vogelstein, 2015). Recently, mutation
rates in human stem cells have been determined to accumulate
at a steady rate of approximately 40 novel mutations per year in
various tissue types (Blokzijl et al., 2016). Some of these mutations
could be generated during translesion synthesis (TLS), which allows cells to efficiently complete DNA replication across sites of
DNA damage (Goodman and Woodgate, 2013). DNA translesion
polymerases are error prone and are thought to promote tolerance of DNA damage, which can contribute to cancer evolution.
The most recurrent set of mutations that accumulate in stem
cells with high division rate represents the aging signature (Alexandrov and Stratton, 2014), which has also been found to be the
most common base substitution signature in cancer. This signature, which correlates with patient age, is made up of predominantly C>T transitions at CpG dinucleotide motifs, reflecting
the inherent mutability of methylated cytosine (Lindahl, 1993).
Indeed, C>T substitutions at CpGs make up 25% of somatic mutations in TP53 codons, implying that mutations associated with
this signature might drive tumorigenesis (Olivier et al., 2010).
The deaminated product of methylated cytosine is thymine,
which leads to transition mutations if the T:G mismatch is not
recognized and repaired prior to DNA replication (Figure 2A). In
addition to incorporating the mutation, DNA replication also increases the probability of spontaneous deamination through
transient exposure of ssDNA and during post-replicative conversion of cytosine to 5-methylcytosine (Lindahl, 1993). Although
70% to 80% of CpG cytosines are methylated in mammalian
cells, the CpG dinucleotide occurs with a frequency of less
than one quarter of that which would be expected by random
chance. It has been proposed that CpG deficiency is due to
the increased mutation rate of 5-methylcytosine (Scarano
et al., 1967), which likely stems from inefficient excision of T:G
mismatches. 5-methylcytosine is demethylated by TET family
DNA demethylases through a 5-hydroxymethylcytosine intermediate. Recent work shows that sites of 5-hydroxymethylcytosine
are associated with C>G transversions in cancer genomes, and
increased C>G correlates with the expression of TET enzymes
(Supek et al., 2014). Thus, modified cytosine might also be inherently mutagenic.
Aging poses the greatest risk for cancer. The free radical theory of aging posits that reactive oxygen species produced by
normal cellular metabolism are responsible for the damage to
biomolecules resulting in organismal aging (Harman, 1956).
One of the most common lesions arising from reactive oxygen
is 8-oxoguanine (8-oxoG), which can result in mispairing with
adenine resulting in G>T substitutions if 8-oxoG is not removed
prior to DNA replication (Figure 2B). Even though 8-oxoG has
long been believed to be a primary driver of mutagenesis in
both nuclear and mitochondrial DNA, a mutational signature
attributed to oxidative DNA lesions has not been found in primary
cancers or in mitochondrial DNA during aging (Helleday et al.,
2014; Kennedy et al., 2013). The lack of G>T substitutions associated with age might indicate that removal of 8-oxoG is
extremely efficient (Banerjee et al., 2006) (Figure 2B). In contrast,
T:G mismatches that arise during spontaneous deamination of
5-methylcytosine might either occur more frequently or might
not be recognized as efficiently by the MMR machinery or the
thymine DNA glycosylase (TDG), which recognizes T:G mismatches (Figure 2A).
DNA Editing as a Source of Mutation
ssDNA is a substrate for the cytosine deaminase family of enzymes, which includes APOBECs and activation-induced
cytidine deaminase (AID) (see below) (Swanton et al., 2015).
APOBECs normally function in protecting against retroviruses
Figure 2. DNA Repair Pathways and Mutations Associated with Dysfunctional Repair
(A) Spontaneous deamination of 5-methylcytosine (C*) generates an aging signature, characterized by CpG>TpG transition mutations. 5-methylcytosine
deamination converts C>T, which is recognized by thymine DNA glycosylase (TDG). Apurinic/apyrimidinic endodeoxyribonuclease (APE) excises the abasic
nucleotide, which is replaced by polymerase beta (POLb).
(B) Reactive oxygen species (ROS) oxidizes guanine to form 8-oxoG. 8-oxoG can pair with adenine and may generate G>T mutations if 8-oxoG:C pairing is not
detected by OGG1 before replication or if the 8-oxoG:A pairing is not recognized by MUTYH after the first round of replication.
(C) Exposure to UV light generates cyclobutane pyrimidine dimers, which are recognized by nucleotide excision repair (NER) factors and excised by XPG and XPF
flap endonucleases. Polymerase delta (POLd) synthesizes DNA to restore C:G pairing. Failure to complete NER results in C>T transition mutations at CC
dinucleotides.
(D) DNA polymerase errors induce mismatch repair (MMR) through MSH2/MSH6. Exonuclease I (EXO1) removes a short patch of DNA that is resynthesized by
POLd. Failure to complete MMR results in base substitutions.
(E) DNA DSBs are repaired using non-homologous end joining (NHEJ) or homologous recombination (HR). Normal NHEJ may result in small insertions or
deletions, as does an inherently error-prone form of end-joining (alternative end joining), which relies on short homologous sequences (microhomologies) for base
pairing. Normal HR is an error-free repair mechanism, but abortive HR may yield chromosomal rearrangements, such as tandem duplications or large deletions,
which are common in HR-deficient cancers. BRCA1 and CtIP promote DNA end resection during HR, whereas BRCA2 promotes strand invasion to facilitate
normal HR.
Cell 168, February 9, 2017 647
(legend on next page)
648 Cell 168, February 9, 2017
and other mobile elements (Swanton et al., 2015). APOBEC3B is
upregulated in approximately half of all cancers and promotes
chemotherapy resistance in estrogen-receptor-positive breast
cancer (Alexandrov et al., 2013; Burns et al., 2013; Law et al.,
2016). The APOBEC mutational burden ranks second only to
the aging signature in cancer (Alexandrov et al., 2013; Burns
et al., 2013). APOBEC-induced mutations occur either as
strand-coordinated clusters (kataegis) or as scattered mutations
(Alexandrov et al., 2013). APOBEC-associated mutational signatures occur at a higher rate on the lagging strand (Haradhvala
et al., 2016; Morganella et al., 2016), which during DNA synthesis
endures longer exposure as ssDNA relative to the leading strand
(Figure 3A). In addition, collapsed replication forks are repaired
by HR, which generates a long 30 ssDNA overhang during end
resection that, in theory, can be a substrate for APOBEC
(Figure 3A). Interestingly, localized APOBEC hypermutations
show marked colocalization and appear to be coupled to
genomic rearrangements (Alexandrov et al., 2013), perhaps
because of high levels of ssDNA formed as a result of replication
stress and transcription stress (see Gene Activation Can Destabilize the Genome) (Kanu et al., 2016) (Figures 3A and 3B).
The Ig genes are targets of high density AID-mediated cytidine
deamination during somatic hypermutation and class-switch
recombination in B cells (reviewed in (Di Noia and Neuberger,
2007). AID has a preference for deaminating cytosine residues
flanked by a 50 purine, making its signature distinguishable
from APOBECs, which show a variety of sequence preferences
(Swanton et al., 2015). In either case, the U:G mismatches, if
replicated, will generate C>T transition mutations (Figure 3C).
Alternatively, these mismatches can be processed into DSBs,
leading to class-switch recombination (Figure 3D). In addition
to the Ig heavy- and light-chain loci, AID initiates recurrent
DSBs in non-Ig loci, predominantly within B cell super enhancers
and regulatory clusters (Meng et al., 2014; Qian et al., 2014).
Some of these off-target sites are proto-oncogenes that, when
translocated to potent Ig enhancers, lead to their constitutive
expression. Despite the risk of B cell genomic instability associated with AID off-target deamination activity and DSB formation
(Figure 3D), the majority of the AID-generated uracils in non-Ig
loci are corrected by high-fidelity DNA repair and hence are
not detectable as mutations (Liu et al., 2008). Thus, AID-mediated cytosine deamination seems to be differentially repaired
depending on genomic context. The mechanisms that allow
high-fidelity BER and/or MMR machineries to perform errorprone repair at some loci whereas others are spared from muta-
tion is not known but might be linked to B cell transformation (Liu
et al., 2008). Numerous alternative possibilities have been suggested for differential repair, including distinct chromatin structures that might recruit error-prone versus high-fidelity repair
factors, variation in repair capacity during the cell cycle, and differences in the replication timing of the targeted genes (Liu and
Schatz, 2009).
Influence of Chromatin Organization on Regional
Mutation Rates
There is a great variation in the mutation load in cancer, ranging
from frequencies of less than one mutation per Mb in some childhood cancers to greater than 250 mutations per Mb in ultramutated biallellic MMR-deficient brain tumors (Shlien et al., 2015).
Most cancers have been found to harbor elevated rates of
base substitutions in heterochromatic late replicating regions
and reduced rates in early replicating regions (Schuster-Böckler
and Lehner, 2012). High mutation densities correlate with repressive histone modifications (e.g., H3K9me3 and H4K20me3) and
anti-correlate with marks of open chromatin such as H3K4me3,
GC content, and genes with higher expression (Schuster-Böckler and Lehner, 2012) (Figure 4A). This association is upheld
across diverse tissue and signature types and is also conserved
in S. cerevisiae, in which nucleotide mis-incorporations are
dominant during late phases of the cell cycle (Waters and
Walker, 2006). This suggests that replication timing and the chromatin landscape can influence lesion formation and DNA repair
fidelity.
It has been suggested that base substitutions in cancer are enriched in genomic regions that undergo late replication because
of differential DNA repair rather than differences in mutation rates
(Supek and Lehner, 2015; Zheng et al., 2014). One possibility is
that early S phase templates have more time to recognize and
repair mutations prior to mitosis. In addition, the DNA MMR machinery, which is coupled to DNA replication, appears to be more
effective in euchromatic early replicating regions (Supek and
Lehner, 2015; Zheng et al., 2014) (Figure 4B). NER also reduces
the rate of mutation in genic regions (Zheng et al., 2014), and
differences in repair rates between transcribed and non-transcribed DNA have been linked to high-fidelity-transcriptioncoupled NER (Haradhvala et al., 2016; Morganella et al., 2016;
Nik-Zainal et al., 2012) (Figure 4B). Consistent with differences
in repair across the genome, when either the NER or MMR
pathway is lost, mutations are more evenly distributed (Supek
and Lehner, 2015; Zheng et al., 2014).
Figure 3. Causes and Consequences of Enzyme-Induced Cytosine Deamination
(A) ssDNA exposed during replication fork progression is a substrate for cytosine deamination. Active replication forks expose ssDNA on the lagging strand.
Stalled replication forks expose ssDNA on both the leading and lagging strand, which activates ATR to stop replication and stabilize the fork through controlled
fork reversal. Endonuclease cleavage causes replication fork collapse into one-ended DNA DSBs, which must be resected before repair by HR.
(B) ssDNA exposed during transcription is a substrate for cytosine deamination. RNA polymerase (RNAP) exposes ssDNA while synthesizing nascent RNA (red) at
the transcription bubble and behind the RNAP as a result of negative supercoiling (left). Transcription stress induces the formation of R-loops (middle), which
exposes long stretches of ssDNA on the non-template strand. R-loops may be resolved in DSBs through cleavage by TC-NER components XPF and XPG (right).
(C) Mechanism of enzyme-induced somatic hypermutation (SHM). Cytosine deamination generates a U:G mismatch, which is recognized by uracil glycosylase
UNG to generate an abasic site opposite the G. Apurinic/apyrimidinic endodeoxyribonuclease (APE) excises the abasic nucleotide which is repaired by POLb. If
the abasic site is replicated over prior to repair, this can give rise to transition or transversion mutations (N). If the U:G mismatch is replicated prior to repair, C>T
substitution occurs. Mutations at neighboring A:T base pairs can occur if the U:G pair is recognized by the MMR pathway and repaired by error prone polymerase
eta (Polh).
(D) Mechanisms of enzyme-induced rearrangements and class-switch recombination (CSR). If cytosine deamination occurs nearby on opposite strands, repair
through BER or MMR will form DNA DSBs, leading to chromosomal rearrangements or CSR.
Cell 168, February 9, 2017 649
Figure 4. Replication Timing and Protein-DNA Interactions Influence Regional Mutation Rates
(A) Early replicating regions of the genome are associated with high levels of H3K4me3, transcription, increased chromatin accessibility, and low levels of
H3K9me3.
(B) In cancer genomes, accessible early replicating regions have fewer base-pair substitutions. MMR is more likely to recognize mismatches in early replication
regions. NER is also more efficient in transcribed, accessible chromatin.
(C) DNA accessibility determines efficiency of NER (adapted from Sabarinathan et al., 2016). Repair of UV-induced lesions is efficient within DNase I hypersensitive sites (DHSs), but is impaired by nucleosome occupancy and transcription factor binding. Conversely, UV-induced mutations are more frequent where
DNA-binding proteins interfere with NER machinery.
(D) Whereas DSBs are more efficiently repaired in accessible chromatin, chromosome rearrangements in cancer genomes accumulate at early replicating
regions.
Even within accessible chromatin, there is variation in mutation density (Figure 4C). For example, in several cancer types,
including melanoma, ovarian, lung, astrocytoma, esophageal,
and prostate cancers, there is increased mutation at the peaks
of DNase I hypersensitive sites relative to accessible flanking regions, even though both regions are nucleosome free (Perera
et al., 2016; Sabarinathan et al., 2016) (Figure 4C). In melanoma,
such regulatory regions have mutation rates four times higher
than flanking sequences. Evidence suggests that peaks of
accessible chromatin within regulatory DNA (i.e., promoters)
are bound by proteins such as the transcription initiation machineries that exclude NER and other repair proteins (Perera et al.,
2016; Sabarinathan et al., 2016) (Figure 4C). Genome-wide
analysis of excision repair after UV damage in primary skin fibroblasts has demonstrated a reduction in repair within transcription factor binding sites at DNase I hypersensitive sites, whereas
NER-deficient cells do not show any increased promoter mutation density (Perera et al., 2016; Sabarinathan et al., 2016). Interestingly, recent sequencing of individual neurons from the
human prefrontal cortex revealed that each neuron has a distinct
genome with more than 1,000 mutations per cell (Lodato et al.,
2015). Mutation hotspots in neurons occurred predominantly at
transcriptionally-active loci and at DNase I hypersensitive sites
(Lodato et al., 2015). This suggests the possibility that even in
neurons, mutation hotspots can result from the exclusion of
repair proteins by DNA-bound factors. Impaired DNA damage
repair is also a common feature of neurodegenerative diseases
(McKinnon, 2013). In neuronal progenitors, however, recurrent
650 Cell 168, February 9, 2017
DNA DSB clusters occur in gene bodies of late-replicating
genes, possibly contributing to functional diversity during
neuronal development (Wei et al., 2016).
Influence of Chromatin Organization on Large-Scale
Chromosomal Rearrangements
Chromosomal rearrangements (deletions, tandem duplications,
inversions, and translocations) are initiated by DNA DSBs. Overall, rearrangements are 10- to 1,000-fold less frequent than somatic mutations but also show regional variation across a cancer
genome. However, in stark contrast to substitution mutations,
which are predominant in heterochromatin, rearrangements
are enriched in early replicating regions (Morganella et al.,
2016; Sima and Gilbert, 2014), including those that may arise
through a single catastrophic event (Baca et al., 2013; Zhang
et al., 2013).
Unlike the case for base substitutions, it appears that
increased DNA damage rather than deficient DNA repair might
underlie the predominance of rearrangements in early replicating
regions (Figure 4D). While different chromatin environments may
selectively utilize NHEJ or HR pathways (Chiolo et al., 2011;
Tsouroula et al., 2016), it is generally believed that the heterochromatin superstructure confers a barrier to DSB repair.
Consistent with this, euchromatic DSBs are more rapidly repaired than heterochromatic DSBs (Goodarzi et al., 2008)
(Figure 4D). Thus, early replicating regions of the genome might
be expected to harbor fewer rearrangements, unless there were
an overabundance of DSBs in these regions relative to late
replicating regions. One hallmark feature of early replicating
euchromatin is the association with gene-rich, transcriptionally
active, and DNase I accessible regions (Figure 4A). Nearly three
quarters of expressed genes are estimated to be early replicating
(Sima and Gilbert, 2014). As discussed below, there is evidence
that high levels of transcription and the increased probability of
conflicts with the replication machinery pose a substantial threat
to genome integrity (Aguilera, 2002).
Oncogenic Stress Promotes DNA Damage and Genomic
Rearrangements in Early Replicating Regions
Oncogene activation is a major driving force in cancer development (Halazonetis et al., 2008). Oncogenes stimulate new replication origin firing, which depletes nucleotide pools, decreases
replication fork speed, and results in stalled replication forks,
which are prone to collapse into DSBs (Halazonetis et al.,
2008). Replication stress leads to exposure of tracts of ssDNA
that are substrates for endonuclease cleavage, which generates
DSBs (Cortez, 2015), or for APOBEC3-deaminase-mediated
mutagenesis (Kanu et al., 2016). By deregulating the timing of
replication initiation and progression, oncogenes might also
disrupt the spatio-temporal segregation that normally prevents
conflicts between transcription and DNA replication (Halazonetis
et al., 2008). This might be exacerbated by oncogenes such as
c-myc, which amplifies gene expression at all active genes
(Nie et al., 2012), thereby posing a further impediment to replication. Recent studies in bacteria indicate that replication-transcription collisions can cause replication fork stalling and DNA
breaks leading to two types of mutation signatures: duplications
and deletions within the transcriptional unit and promoter-localized base substitutions (Sankar et al., 2016).
Oncogene-induced DNA damage is not random, but preferentially occurs at fragile sites (Sarni and Kerem, 2016). In contrast
to late replicating common fragile sites (CFSs), early replicating
fragile sites (ERFSs) appear near mammalian origins of replication within highly expressed gene clusters enriched for repetitive
elements and CpG dinucleotides (Barlow et al., 2013). These regions may be particularly susceptible to replication-transcription
collisions, creating the potential for genome rearrangements.
Indeed, ERFSs identified in primary B cells account for a large
fraction of recurrent amplifications and/or deletions in human
diffuse large B cell lymphomas (Barlow et al., 2013).
In addition to oncogenes, recent cancer genome sequencing
studies have identified the histone methyltransferases KMT2C
and KMT2D, among the most broadly mutated genes in human
cancer (Lawrence et al., 2014). KMT2C and KMT2D belong to
the family of mammalian mixed-lineage leukemia (MLL) genes
that encode histone methyltransferases, responsible for methylation at H3K4 at a subset of active gene promoters and enhancers (Lee et al., 2013). Despite their mutation in many
cancers, the roles of KMT2C and KMT2D in genome stability
have only recently been uncovered. KMT2C- and KMT2Ddependent H3K4 methylation was found to be induced at replication forks upon replication stress (Ray Chaudhuri et al.,
2016), and KMT2D physically interacts with the helicase
RECQL5 to prevent collisions between transcription and replication machineries (Kantidakis et al., 2016; Saponaro et al., 2014).
Like RECQL5-deficient cells, KMT2D-deficient cells exhibited
transcription stress associated with RNAPII pausing, stalling,
and backtracking, leading to chromosomal aberrations primarily
within ERFSs (Kantidakis et al., 2016). Thus, cancer-driving
oncogenes and mutations in KMT2C and KMT2D could potentiate DNA breaks and chromosomal rearrangements preferentially in early replicating regions.
Gene Activation Can Destabilize the Genome
Like oncogenes, hormone stimulation of gene expression may
induce genome instability (Haffner et al., 2010; Ju et al., 2006;
Lin et al., 2009; Williamson and Lees-Miller, 2011). In estrogenreceptor-positive breast cancer cells, estrogen stimulation
boosts expression of a broad cohort of genes, promotes cell
cycle entry, and increases DNA damage (Ju et al., 2006; Williamson and Lees-Miller, 2011). Robust transcription at estrogenresponsive genes results in co-transcriptional structures known
as R-loops, or RNA:DNA hybrids associated with unannealed
ssDNA (Stork et al., 2016) (Figure 3B). Although RNA processing
factors normally resolve RNA:DNA hybrids, it has been suggested that hormones or other stimuli that promote high levels
of transcription might overload the system, resulting in the accumulation of R-loops (Stork et al., 2016). The displaced ssDNA
non-transcribed strand can then act as a substrate for cytosine
deamination by APOBEC, which through the activation of BER
pathways can lead to DSB formation (Periyasamy et al., 2015)
(Figure 3B). Alternatively, unprocessed R-loops can activate
the transcription-coupled NER (TC-NER) endonucleolytic machinery, which can excise the RNA:DNA hybrid, generating a
DSB upon encounter with a replication fork (Cortez, 2015; Stork
et al., 2016) (Figure 3B). Interestingly, a correlation was observed
between genomic rearrangements in breast tumors and estrogen-responsive genes, suggesting a tissue-specific vulnerability
to a endogenous but potentially genotoxic agent (Stork et al.,
2016). In summary, ssDNA associated with R-loops can be mutagenetic and recombinogenic.
Independent of collisions with replication forks, transcription
itself might contribute to genome instability. Over 20 years
ago, it was demonstrated that an increased transcription level
stimulates spontaneous mutagenesis in yeast (Datta and
Jinks-Robertson, 1995). Recent genome-wide sequencing approaches in mammalian cells have revealed that transcriptionally-active regions also harbor more translocations and are
more fragile than less-active regions (Barlow et al., 2013; Chiarle
et al., 2011; Klein et al., 2011). One potential source of transcriptional stress is the superhelical tension produced as RNA polymerase traverses DNA. For example, negative supercoiling
(underwinding) can lead to stretches of ssDNA behind the
advancing polymerase, with potential for enzyme-induced cytosine deamination and mutation (Figure 5A). Torsional stress is
normally relieved by topoisomerases, enzymes that catalyze
transient cleavage of the DNA backbone, followed by one rotation of the DNA strand before resealing (Pommier et al., 2016).
The broken DNA intermediates are thought to be sequestered
from the DNA damage surveillance machinery. However, at
some unknown frequency, topoisomerases fail to complete the
reaction, which can result in trapped cleavage complexes and
persistent DNA damage on one or both DNA strands
(Figure 5B). Notably, several DNA alterations, including oxidative
Cell 168, February 9, 2017 651
Figure 5. Canonical and Non-canonical Topoisomerase Activity and Gene Activation
(A) Upon gene activation, canonical topoisomerase activity promotes gene expression. The enzymes TOP1 and TOP2 cleave one or both strands to alleviate
positive supercoiling ahead of and negative supercoiling behind RNA polymerase (Pol II). ssDNA behind Pol II can be a substrate for cytosine deamination.
(B) Non-canonical, abortive TOP2 reactions result in protein-linked DNA breaks that are resolved the NHEJ repair pathway. Abortive TOP2 activity also promotes
transcription by an unknown mechanism. Persistent DNA damage may also result in chromosomal alterations.
lesions, base modifications, mismatches, and ssDNA breaks
enhance the frequency of persistent topoisomerase cleavage
complexes (Pommier et al., 2016). These can eventually give
rise to DSBs and short deletions (Lippert et al., 2011; Takahashi
et al., 2011), implicating trapped cleavage complexes as a
source of mutation.
Recently, non-canonical roles for topoisomerases have also
been observed, in which prolonged topoisomerase-mediated
DNA damage is necessary to promote transcription (Figure 5B).
In cells activated by diverse stimuli, including heat shock, serum
induction, insulin, androgen, estrogen, and various neuronal
stimuli, the transcriptional program appears to require persistent
topoisomerase-mediated DNA damage that is recognized by
the DNA damage transducers ATM, DNA-PKcs, and PARP1
(Bunch et al., 2015; Haffner et al., 2010; Ju et al., 2006;
Lin et al., 2009; Madabhushi et al., 2015; Perillo et al., 2008;
Wong et al., 2009). For example, upon activation of postmitotic
neurons with NMDA, H2AX is rapidly phosphorylated in transcribed regions of early-response genes (Madabhushi et al.,
2015). It is postulated that DNA damage signaling might be
necessary for the release of paused RNA polymerase at promoters or for conformational changes that facilitate promoterenhancer interactions (Calderwood, 2016).
Although persistent DNA damage induced by non-canonical
topoisomerase activity promotes transcription, it can also lead
to genome instability. For example, androgen stimulation in
prostate cells leads to TOPO2b-dependent androgen receptor
target gene expression, as well as genomic recombination
events at activated loci, leading to prostate cancer rearrange652 Cell 168, February 9, 2017
ments (Haffner et al., 2010; Lin et al., 2009). The probability of
aberrant rearrangements is likely to be exacerbated by deficiencies in HR, which are found in as many as 25% of advanced
prostate cancers (Pritchard et al., 2016). It has been speculated
that the inflammatory mileau might stabilize TOP2-DNA cleavage complexes, perhaps due to increases in reactive oxygen
species surrounding the prostate lesion (De Marzo et al., 2007;
Pommier et al., 2016). Inflammation also induces the expression
of hundreds of pathogen-associated genes, which require
topoisomerase I for nucleosome remodeling and transcriptional
elongation (Rialdi et al., 2016). If such inflammatory transcriptional responses similarly result in persistent topoisomerase
lesions, they might contribute to mutation and/or chromosomal
rearrangements.
Conclusions and Future Perspectives
The mutational landscape of cancer is generated through a combination of environmental and endogenous stresses that cause
base substitutions, insertions, deletions, and chromosomal rearrangements. A major lesson from recent work is that even in
repair-sufficient cells, endogenous and oncogenic stress can
occasionally overwhelm normal genome maintenance pathways
(Figure 6). One major outcome of these stresses is exposure of
ssDNA, which is subject to nuclease activity, spontaneous hydrolysis, and enzyme-induced deamination that often results in
base substitutions or processing into DSBs. Hormonal signaling
and inflammatory responses may act similar to oncogenes by
driving cells to undergo high levels of transcriptional activity or
replication stress that leads to genome instability.
of the various sources, risk factors, and mechanisms that
contribute to genome instability.
Our comprehensive understanding of genome maintenance
pathways is a direct result of decades of work investigating
DNA repair mechanisms. This is the basis for current clinical cancer treatments predicated on chemotherapy and ionizing radiation that damage DNA and in turn overwhelm the cellular DNA
repair capacity in highly proliferative tumor cells. As effective
as they have been, treatments that produce DNA damage on a
global scale frequently result in side effects, secondary tumors,
and drug resistance. Precision therapies based on synthetic
lethality exploit cancer-specific mutations by targeting deficient
DNA repair and thereby reduce collateral damage to normal cells
(Lord et al., 2015). Tumors with defective DNA repair might also
respond well to immune checkpoint inhibitors by generating mutations that produce neoantigens that enhance T cell reactivity
(Schumacher and Schreiber, 2015). Additionally, by manipulating highly intricate DNA repair pathways, it has become
possible to generate site-specific nucleotide substitutions of
DNA through CRISPR/Cas9 base editing (Komor et al., 2016;
Nishida et al., 2016; Paquet et al., 2016), which might provide
a complementary approach to correcting known germline mutations that predispose individuals to cancer. In conclusion, a
deeper understanding of how DNA lesions are generated, processed, and repaired will continue to provide insights and new
opportunities for cancer prevention and treatment.
ACKNOWLEDGMENTS
Figure 6. Replication and Transcription Stress Contribute to the
Mutational Landscape of Cancer
An emerging model for how mitogenic signaling promotes genome instability
at early replicating fragile sites (ERFSs). Oncogenes, hormones, and inflammation can promote unscheduled activation of replication and transcription
programs. This can result in replication stress and transcription stress through
mechanisms that are largely unknown. The generation of ssDNA and DSBs at
ERFSs can lead to genomic instability.
Despite the fact that each cell in the human body is exposed
to thousands of DNA lesions per day, most are efficiently repaired. Unfortunately, it is not possible with current technologies
to accurately quantify the steady-state levels of several types of
DNA lesions. For example, the most reliable method to detect
ssDNA breaks (i.e., the alkali comet assay) is sensitive to only
a few thousand lesions per cell (or one ssDNA break every
million bp). Similarly, it is difficult to estimate the frequency
and location of replication-associated DSBs and proteinblocked lesions arising from abortive topoisomerase activity.
Although the phosphorylation of the histone variant H2AX, or
g-H2AX, has been extensively used as a surrogate marker of
DSBs, it also marks single-stranded regions exposed during
replication stress and spreads over hundreds of kilobases.
However, recent DNA break capture and next-generationsequencing-based methods have begun to quantitate and precisely locate genomic co-ordinates of low frequency DNA
lesions, translocations, and other infrequent complex rearrangements (Canela et al., 2016; Frock et al., 2015; Lensing et al.,
2016; Zhang et al., 2015). These emerging technological and
computational advances should enable a better understanding
We are especially grateful to Joshua Waterfall, Ferenc Livak, Avinash Bhandoola, and Sam John for comments on the manuscript and to Jiri Lukas, Keith
Caldecott, and Yossi Shiloh for discussions. This work was supported by the
Intramural Research Program of the NIH, the National Cancer Institute, and the
Center for Cancer Research. A.N. was also supported by the US Department
of Defense (BCRP DOD Idea Expansion Award BC133858 and BCRP Breakthrough Award BC151331), the Ellison Foundation Award for Aging Research,
and Alex’s Lemonade Stand Foundation Reach Award. A.T. has been supported by a fellowship from the American Cancer Society (PF-16-03701-DMC).
REFERENCES
Aguilera, A. (2002). The connection between transcription and genomic instability. EMBO J. 21, 195–201.
Alexandrov, L.B., and Stratton, M.R. (2014). Mutational signatures: the patterns of somatic mutations hidden in cancer genomes. Curr. Opin. Genet.
Dev. 24, 52–60.
Alexandrov, L.B., Nik-Zainal, S., Wedge, D.C., Aparicio, S.A., Behjati, S., Biankin, A.V., Bignell, G.R., Bolli, N., Borg, A., Børresen-Dale, A.L., et al.; Australian
Pancreatic Cancer Genome Initiative; ICGC Breast Cancer Consortium; ICGC
MMML-Seq Consortium; ICGC PedBrain (2013). Signatures of mutational processes in human cancer. Nature 500, 415–421.
Baca, S.C., Prandi, D., Lawrence, M.S., Mosquera, J.M., Romanel, A., Drier,
Y., Park, K., Kitabayashi, N., MacDonald, T.Y., Ghandi, M., et al. (2013). Punctuated evolution of prostate cancer genomes. Cell 153, 666–677.
Banerjee, A., Santos, W.L., and Verdine, G.L. (2006). Structure of a DNA glycosylase searching for lesions. Science 311, 1153–1157.
Barlow, J.H., Faryabi, R.B., Callén, E., Wong, N., Malhowski, A., Chen, H.T.,
Gutierrez-Cruz, G., Sun, H.W., McKinnon, P., Wright, G., et al. (2013). Identification of early replicating fragile sites that contribute to genome instability. Cell
152, 620–632.
Cell 168, February 9, 2017 653
Blokzijl, F., de Ligt, J., Jager, M., Sasselli, V., Roerink, S., Sasaki, N., Huch, M.,
Boymans, S., Kuijk, E., Prins, P., et al. (2016). Tissue-specific mutation accumulation in human adult stem cells during life. Nature 538, 260–264.
Ju, B.G., Lunyak, V.V., Perissi, V., Garcia-Bassets, I., Rose, D.W., Glass, C.K.,
and Rosenfeld, M.G. (2006). A topoisomerase IIbeta-mediated dsDNA break
required for regulated transcription. Science 312, 1798–1802.
Bunch, H., Lawney, B.P., Lin, Y.F., Asaithamby, A., Murshid, A., Wang, Y.E.,
Chen, B.P., and Calderwood, S.K. (2015). Transcriptional elongation requires
DNA break-induced signalling. Nat. Commun. 6, 10191.
Kantidakis, T., Saponaro, M., Mitter, R., Horswell, S., Kranz, A., Boeing, S.,
Aygün, O., Kelly, G.P., Matthews, N., Stewart, A., et al. (2016). Mutation of cancer driver MLL2 results in transcription stress and genome instability. Genes
Dev. 30, 408–420.
Burns, M.B., Lackey, L., Carpenter, M.A., Rathore, A., Land, A.M., Leonard, B.,
Refsland, E.W., Kotandeniya, D., Tretyakova, N., Nikas, J.B., et al. (2013).
APOBEC3B is an enzymatic source of mutation in breast cancer. Nature
494, 366–370.
Calderwood, S.K. (2016). A critical role for topoisomerase IIb and DNA double
strand breaks in transcription. Transcription 7, 75–83.
Canela, A., Sridharan, S., Sciascia, N., Tubbs, A., Meltzer, P., Sleckman, B.P.,
and Nussenzweig, A. (2016). DNA Breaks and End Resection Measured
Genome-wide by End Sequencing. Mol. Cell 63, 898–911.
Chiarle, R., Zhang, Y., Frock, R.L., Lewis, S.M., Molinie, B., Ho, Y.J., Myers,
D.R., Choi, V.W., Compagno, M., Malkin, D.J., et al. (2011). Genome-wide
translocation sequencing reveals mechanisms of chromosome breaks and rearrangements in B cells. Cell 147, 107–119.
Chiolo, I., Minoda, A., Colmenares, S.U., Polyzos, A., Costes, S.V., and Karpen,
G.H. (2011). Double-strand breaks in heterochromatin move outside of a dynamic HP1a domain to complete recombinational repair. Cell 144, 732–744.
Cortez, D. (2015). Preventing replication fork collapse to maintain genome
integrity. DNA Repair (Amst.) 32, 149–157.
Crick, F. (1974). The double helix: a personal view. Nature 248, 766–769.
Datta, A., and Jinks-Robertson, S. (1995). Association of increased spontaneous mutation rates with high levels of transcription in yeast. Science 268,
1616–1619.
De Marzo, A.M., Platz, E.A., Sutcliffe, S., Xu, J., Grönberg, H., Drake, C.G., Nakai, Y., Isaacs, W.B., and Nelson, W.G. (2007). Inflammation in prostate carcinogenesis. Nat. Rev. Cancer 7, 256–269.
Kanu, N., Cerone, M.A., Goh, G., Zalmas, L.P., Bartkova, J., Dietzen, M.,
McGranahan, N., Rogers, R., Law, E.K., Gromova, I., et al. (2016). DNA replication stress mediates APOBEC3 family mutagenesis in breast cancer.
Genome Biol. 17, 185.
Kennedy, S.R., Salk, J.J., Schmitt, M.W., and Loeb, L.A. (2013). Ultra-sensitive
sequencing reveals an age-related increase in somatic mitochondrial mutations that are inconsistent with oxidative damage. PLoS Genet. 9, e1003794.
Klein, I.A., Resch, W., Jankovic, M., Oliveira, T., Yamane, A., Nakahashi, H., Di
Virgilio, M., Bothmer, A., Nussenzweig, A., Robbiani, D.F., et al. (2011). Translocation-capture sequencing reveals the extent and nature of chromosomal
rearrangements in B lymphocytes. Cell 147, 95–106.
Komor, A.C., Kim, Y.B., Packer, M.S., Zuris, J.A., and Liu, D.R. (2016). Programmable editing of a target base in genomic DNA without double-stranded
DNA cleavage. Nature 533, 420–424.
Kunkel, T.A. (2015). Celebrating DNA’s Repair Crew. Cell 163, 1301–1303.
Law, E.K., Sieuwerts, A.M., LaPara, K., Leonard, B., Starrett, G.J., Molan,
A.M., Temiz, N.A., Vogel, R.I., Meijer-van Gelder, M.E., Sweep, F.C., et al.
(2016). The DNA cytosine deaminase APOBEC3B promotes tamoxifen resistance in ER-positive breast cancer. Sci. Adv. 2, e1601737.
Lawrence, M.S., Stojanov, P., Mermel, C.H., Robinson, J.T., Garraway, L.A.,
Golub, T.R., Meyerson, M., Gabriel, S.B., Lander, E.S., and Getz, G. (2014).
Discovery and saturation analysis of cancer genes across 21 tumour types.
Nature 505, 495–501.
Friedberg, E.C. (2008). A brief history of the DNA repair field. Cell Res. 18, 3–7.
Lee, J.E., Wang, C., Xu, S., Cho, Y.W., Wang, L., Feng, X., Baldridge, A.,
Sartorelli, V., Zhuang, L., Peng, W., and Ge, K. (2013). H3K4 mono- and dimethyltransferase MLL4 is required for enhancer activation during cell differentiation. eLife 2, e01503.
Frock, R.L., Hu, J., Meyers, R.M., Ho, Y.J., Kii, E., and Alt, F.W. (2015).
Genome-wide detection of DNA double-stranded breaks induced by engineered nucleases. Nat. Biotechnol. 33, 179–186.
Lensing, S.V., Marsico, G., Hänsel-Hertsch, R., Lam, E.Y., Tannahill, D., and
Balasubramanian, S. (2016). DSBCapture: in situ capture and sequencing of
DNA breaks. Nat. Methods 13, 855–857.
Gao, R., Davis, A., McDonald, T.O., Sei, E., Shi, X., Wang, Y., Tsai, P.C., Casasent, A., Waters, J., Zhang, H., et al. (2016). Punctuated copy number evolution
and clonal stasis in triple-negative breast cancer. Nat. Genet. 48, 1119–1130.
Lin, C., Yang, L., Tanasa, B., Hutt, K., Ju, B.G., Ohgi, K., Zhang, J., Rose, D.W.,
Fu, X.D., Glass, C.K., and Rosenfeld, M.G. (2009). Nuclear receptor-induced
chromosomal proximity and DNA breaks underlie specific translocations in
cancer. Cell 139, 1069–1083.
Di Noia, J.M., and Neuberger, M.S. (2007). Molecular mechanisms of antibody
somatic hypermutation. Annu. Rev. Biochem. 76, 1–22.
Goodarzi, A.A., Noon, A.T., Deckbar, D., Ziv, Y., Shiloh, Y., Löbrich, M., and
Jeggo, P.A. (2008). ATM signaling facilitates repair of DNA double-strand
breaks associated with heterochromatin. Mol. Cell 31, 167–177.
Goodman, M.F., and Woodgate, R. (2013). Translesion DNA polymerases.
Cold Spring Harb. Perspect. Biol. 5, a010363.
Haber, J.E. (2015). Deciphering the DNA Damage Response. Cell 162,
1183–1185.
Haffner, M.C., Aryee, M.J., Toubaji, A., Esopi, D.M., Albadine, R., Gurel, B.,
Isaacs, W.B., Bova, G.S., Liu, W., Xu, J., et al. (2010). Androgen-induced
TOP2B-mediated double-strand breaks and prostate cancer gene rearrangements. Nat. Genet. 42, 668–675.
Halazonetis, T.D., Gorgoulis, V.G., and Bartek, J. (2008). An oncogeneinduced DNA damage model for cancer development. Science 319,
1352–1355.
Haradhvala, N.J., Polak, P., Stojanov, P., Covington, K.R., Shinbrot, E., Hess,
J.M., Rheinbay, E., Kim, J., Maruvka, Y.E., Braunstein, L.Z., et al. (2016). Mutational Strand Asymmetries in Cancer Genomes Reveal Mechanisms of DNA
Damage and Repair. Cell 164, 538–549.
Lindahl, T. (1993). Instability and decay of the primary structure of DNA. Nature
362, 709–715.
Lindahl, T., and Barnes, D.E. (2000). Repair of endogenous DNA damage. Cold
Spring Harb. Symp. Quant. Biol. 65, 127–133.
Lindahl, T., and Nyberg, B. (1972). Rate of depurination of native deoxyribonucleic acid. Biochemistry 11, 3610–3618.
Lippert, M.J., Kim, N., Cho, J.E., Larson, R.P., Schoenly, N.E., O’Shea, S.H.,
and Jinks-Robertson, S. (2011). Role for topoisomerase 1 in transcriptionassociated mutagenesis in yeast. Proc. Natl. Acad. Sci. USA 108, 698–703.
Liu, M., and Schatz, D.G. (2009). Balancing AID and DNA repair during somatic
hypermutation. Trends Immunol. 30, 173–181.
Liu, M., Duke, J.L., Richter, D.J., Vinuesa, C.G., Goodnow, C.C., Kleinstein,
S.H., and Schatz, D.G. (2008). Two levels of protection for the B cell genome
during somatic hypermutation. Nature 451, 841–845.
Harman, D. (1956). Aging: a theory based on free radical and radiation chemistry. J. Gerontol. 11, 298–300.
Lodato, M.A., Woodworth, M.B., Lee, S., Evrony, G.D., Mehta, B.K., Karger, A.,
Lee, S., Chittenden, T.W., D’Gama, A.M., Cai, X., et al. (2015). Somatic mutation in single human neurons tracks developmental and transcriptional history.
Science 350, 94–98.
Helleday, T., Eshtad, S., and Nik-Zainal, S. (2014). Mechanisms underlying
mutational signatures in human cancers. Nat. Rev. Genet. 15, 585–598.
Loeb, L.A., Springgate, C.F., and Battula, N. (1974). Errors in DNA replication
as a basis of malignant changes. Cancer Res. 34, 2311–2321.
654 Cell 168, February 9, 2017
Lord, C.J., Tutt, A.N., and Ashworth, A. (2015). Synthetic lethality and cancer
therapy: lessons learned from the development of PARP inhibitors. Annu.
Rev. Med. 66, 455–470.
Madabhushi, R., Gao, F., Pfenning, A.R., Pan, L., Yamakawa, S., Seo, J.,
Rueda, R., Phan, T.X., Yamakawa, H., Pao, P.C., et al. (2015). Activity-Induced
DNA Breaks Govern the Expression of Neuronal Early-Response Genes. Cell
161, 1592–1605.
Repair Gene Mutations in Men with Metastatic Prostate Cancer. N. Engl. J.
Med. 375, 443–453.
Qian, J., Wang, Q., Dose, M., Pruett, N., Kieffer-Kwon, K.R., Resch, W., Liang,
G., Tang, Z., Mathé, E., Benner, C., et al. (2014). B cell super-enhancers and
regulatory clusters recruit AID tumorigenic activity. Cell 159, 1524–1537.
McKinnon, P.J. (2013). Maintaining genome stability in the nervous system.
Nat. Neurosci. 16, 1523–1529.
Ray Chaudhuri, A., Callen, E., Ding, X., Gogola, E., Duarte, A.A., Lee, J.E.,
Wong, N., Lafarga, V., Calvo, J.A., Panzarino, N.J., et al. (2016). Replication
fork stability confers chemoresistance in BRCA-deficient cells. Nature 535,
382–387.
Meng, F.L., Du, Z., Federation, A., Hu, J., Wang, Q., Kieffer-Kwon, K.R.,
Meyers, R.M., Amor, C., Wasserman, C.R., Neuberg, D., et al. (2014). Convergent transcription at intragenic super-enhancers targets AID-initiated genomic
instability. Cell 159, 1538–1548.
Rialdi, A., Campisi, L., Zhao, N., Lagda, A.C., Pietzsch, C., Ho, J.S., MartinezGil, L., Fenouil, R., Chen, X., Edwards, M., et al. (2016). Topoisomerase 1 inhibition suppresses inflammatory genes and protects from death by inflammation. Science 352, aad7993.
Morganella, S., Alexandrov, L.B., Glodzik, D., Zou, X., Davies, H., Staaf, J.,
Sieuwerts, A.M., Brinkman, A.B., Martin, S., Ramakrishna, M., et al. (2016).
The topography of mutational processes in breast cancer genomes. Nat. Commun. 7, 11383.
Sabarinathan, R., Mularoni, L., Deu-Pons, J., Gonzalez-Perez, A., and LópezBigas, N. (2016). Nucleotide excision repair is impaired by binding of transcription factors to DNA. Nature 532, 264–267.
Nie, Z., Hu, G., Wei, G., Cui, K., Yamane, A., Resch, W., Wang, R., Green, D.R.,
Tessarollo, L., Casellas, R., et al. (2012). c-Myc is a universal amplifier of expressed genes in lymphocytes and embryonic stem cells. Cell 151, 68–79.
Nik-Zainal, S., Alexandrov, L.B., Wedge, D.C., Van Loo, P., Greenman, C.D.,
Raine, K., Jones, D., Hinton, J., Marshall, J., Stebbings, L.A., et al.; Breast Cancer Working Group of the International Cancer Genome Consortium (2012).
Mutational processes molding the genomes of 21 breast cancers. Cell 149,
979–993.
Nik-Zainal, S., Davies, H., Staaf, J., Ramakrishna, M., Glodzik, D., Zou, X.,
Martincorena, I., Alexandrov, L.B., Martin, S., Wedge, D.C., et al. (2016). Landscape of somatic mutations in 560 breast cancer whole-genome sequences.
Nature 534, 47–54.
Sankar, T.S., Wastuwidyaningtyas, B.D., Dong, Y., Lewis, S.A., and Wang,
J.D. (2016). The nature of mutations induced by replication–transcription collisions. Nature 535, 178–181.
Saponaro, M., Kantidakis, T., Mitter, R., Kelly, G.P., Heron, M., Williams, H.,
Söding, J., Stewart, A., and Svejstrup, J.Q. (2014). RECQL5 controls transcript
elongation and suppresses genome instability associated with transcription
stress. Cell 157, 1037–1049.
Sarni, D., and Kerem, B. (2016). The complex nature of fragile site plasticity and
its importance in cancer. Curr. Opin. Cell Biol. 40, 131–136.
Scarano, E., Iaccarino, M., Grippo, P., and Parisi, E. (1967). The heterogeneity
of thymine methyl group origin in DNA pyrimidine isostichs of developing sea
urchin embryos. Proc. Natl. Acad. Sci. USA 57, 1394–1400.
Schumacher, T.N., and Schreiber, R.D. (2015). Neoantigens in cancer immunotherapy. Science 348, 69–74.
Nishida, K., Arazoe, T., Yachie, N., Banno, S., Kakimoto, M., Tabata, M., Mochizuki, M., Miyabe, A., Araki, M., Hara, K.Y., et al. (2016). Targeted nucleotide
editing using hybrid prokaryotic and vertebrate adaptive immune systems.
Science 353.
Schuster-Böckler, B., and Lehner, B. (2012). Chromatin organization is a major
influence on regional mutation rates in human cancer cells. Nature 488,
504–507.
Notta, F., Chan-Seng-Yue, M., Lemire, M., Li, Y., Wilson, G.W., Connor, A.A.,
Denroche, R.E., Liang, S.B., Brown, A.M., Kim, J.C., et al. (2016). A renewed
model of pancreatic cancer evolution based on genomic rearrangement patterns. Nature 538, 378–382.
Shlien, A., Campbell, B.B., de Borja, R., Alexandrov, L.B., Merico, D., Wedge,
D., Van Loo, P., Tarpey, P.S., Coupland, P., Behjati, S., et al.; Biallelic
Mismatch Repair Deficiency Consortium (2015). Combined hereditary and somatic mutations of replication error repair genes result in rapid onset of ultrahypermutated cancers. Nat. Genet. 47, 257–262.
Nowell, P.C. (1976). The clonal evolution of tumor cell populations. Science
194, 23–28.
Olivier, M., Hollstein, M., and Hainaut, P. (2010). TP53 mutations in human cancers: origins, consequences, and clinical use. Cold Spring Harb. Perspect.
Biol. 2, a001008.
Paquet, D., Kwart, D., Chen, A., Sproul, A., Jacob, S., Teo, S., Olsen, K.M.,
Gregg, A., Noggle, S., and Tessier-Lavigne, M. (2016). Efficient introduction
of specific homozygous and heterozygous mutations using CRISPR/Cas9.
Nature 533, 125–129.
Perera, D., Poulos, R.C., Shah, A., Beck, D., Pimanda, J.E., and Wong, J.W.
(2016). Differential DNA repair underlies mutation hotspots at active promoters
in cancer genomes. Nature 532, 259–263.
Perillo, B., Ombra, M.N., Bertoni, A., Cuozzo, C., Sacchetti, S., Sasso, A.,
Chiariotti, L., Malorni, A., Abbondanza, C., and Avvedimento, E.V. (2008).
DNA oxidation as triggered by H3K9me2 demethylation drives estrogeninduced gene expression. Science 319, 202–206.
Periyasamy, M., Patel, H., Lai, C.F., Nguyen, V.T., Nevedomskaya, E., Harrod,
A., Russell, R., Remenyi, J., Ochocka, A.M., Thomas, R.S., et al. (2015).
APOBEC3B-Mediated Cytidine Deamination Is Required for Estrogen Receptor Action in Breast Cancer. Cell Rep. 13, 108–121.
Pommier, Y., Sun, Y., Huang, S.N., and Nitiss, J.L. (2016). Roles of eukaryotic
topoisomerases in transcription, replication and genomic stability. Nat. Rev.
Mol. Cell Biol. 17, 703–721.
Pritchard, C.C., Mateo, J., Walsh, M.F., De Sarkar, N., Abida, W., Beltran, H.,
Garofalo, A., Gulati, R., Carreira, S., Eeles, R., et al. (2016). Inherited DNA-
Sima, J., and Gilbert, D.M. (2014). Complex correlations: replication timing and
mutational landscapes during cancer and genome evolution. Curr. Opin.
Genet. Dev. 25, 93–100.
Stork, C.T., Bocek, M., Crossley, M.P., Sollier, J., Sanz, L.A., Chédin, F., Swigut, T., and Cimprich, K.A. (2016). Co-transcriptional R-loops are the main
cause of estrogen-induced DNA damage. eLife 5.
Supek, F., and Lehner, B. (2015). Differential DNA mismatch repair underlies
mutation rate variation across the human genome. Nature 521, 81–84.
Supek, F., Lehner, B., Hajkova, P., and Warnecke, T. (2014). Hydroxymethylated cytosines are associated with elevated C to G transversion rates. PLoS
Genet. 10, e1004585.
Swanton, C., McGranahan, N., Starrett, G.J., and Harris, R.S. (2015). APOBEC
Enzymes: Mutagenic Fuel for Cancer Evolution and Heterogeneity. Cancer
Discov. 5, 704–712.
Takahashi, T., Burguiere-Slezak, G., Van der Kemp, P.A., and Boiteux, S.
(2011). Topoisomerase 1 provokes the formation of short deletions in repeated
sequences upon high transcription in Saccharomyces cerevisiae. Proc. Natl.
Acad. Sci. USA 108, 692–697.
Tomasetti, C., and Vogelstein, B. (2015). Cancer etiology. Variation in cancer
risk among tissues can be explained by the number of stem cell divisions. Science 347, 78–81.
Tsouroula, K., Furst, A., Rogier, M., Heyer, V., Maglott-Roth, A., Ferrand, A.,
Reina-San-Martin, B., and Soutoglou, E. (2016). Temporal and Spatial Uncoupling of DNA Double Strand Break Repair Pathways within Mammalian Heterochromatin. Mol. Cell 63, 293–305.
Cell 168, February 9, 2017 655
Waters, L.S., and Walker, G.C. (2006). The critical mutagenic translesion DNA
polymerase Rev1 is highly expressed during G(2)/M phase rather than
S phase. Proc. Natl. Acad. Sci. USA 103, 8971–8976.
Wei, P.C., Chang, A.N., Kao, J., Du, Z., Meyers, R.M., Alt, F.W., and Schwer, B.
(2016). Long Neural Genes Harbor Recurrent DNA Break Clusters in Neural
Stem/Progenitor Cells. Cell 164, 644–655.
Williamson, L.M., and Lees-Miller, S.P. (2011). Estrogen receptor a-mediated
transcription induces cell cycle-dependent DNA double-strand breaks. Carcinogenesis 32, 279–285.
Wong, R.H., Chang, I., Hudak, C.S., Hyun, S., Kwan, H.Y., and Sul, H.S. (2009).
A role of DNA-PK for the metabolic gene regulation in response to insulin. Cell
136, 1056–1072.
656 Cell 168, February 9, 2017
Wu, S., Powers, S., Zhu, W., and Hannun, Y.A. (2016). Substantial contribution
of extrinsic risk factors to cancer development. Nature 529, 43–47.
Zhang, C.Z., Leibowitz, M.L., and Pellman, D. (2013). Chromothripsis and
beyond: rapid genome evolution from complex chromosomal rearrangements.
Genes Dev. 27, 2513–2530.
Zhang, C.Z., Spektor, A., Cornils, H., Francis, J.M., Jackson, E.K., Liu, S.,
Meyerson, M., and Pellman, D. (2015). Chromothripsis from DNA damage in
micronuclei. Nature 522, 179–184.
Zheng, C.L., Wang, N.J., Chung, J., Moslehi, H., Sanborn, J.Z., Hur, J.S., Collisson, E.A., Vemula, S.S., Naujokas, A., Chiotti, K.E., et al. (2014). Transcription restores DNA repair to heterochromatin, determining regional mutation
rates in cancer genomes. Cell Rep. 9, 1228–1234.
Leading Edge
Review
Understanding the Intersections
between Metabolism and Cancer Biology
Matthew G. Vander Heiden1,2,* and Ralph J. DeBerardinis3,*
1The Koch Institute for Integrative Cancer Research and Department of Biology, Massachusetts Institute of Technology, Cambridge,
MA 02139, USA
2Dana–Farber Cancer Institute, Boston, MA 02115, USA
3Children’s Medical Center Research Institute, Department of Pediatrics, and Eugene McDermott Center for Human Growth and
Development, The University of Texas Southwestern Medical Center, Dallas, TX 75390
*Correspondence: mvh@mit.edu (M.G.V.H.), ralph.deberardinis@utsouthwestern.edu (R.J.D.)
http://dx.doi.org/10.1016/j.cell.2016.12.039
Transformed cells adapt metabolism to support tumor initiation and progression. Specific metabolic activities can participate directly in the process of transformation or support the biological
processes that enable tumor growth. Exploiting cancer metabolism for clinical benefit requires
defining the pathways that are limiting for cancer progression and understanding the context specificity of metabolic preferences and liabilities in malignant cells. Progress toward answering these
questions is providing new insight into cancer biology and can guide the more effective targeting of
metabolism to help patients.
Introduction
Changes in cell metabolism can contribute to transformation and
tumor progression. Metabolic phenotypes can also be exploited
to image tumors, provide prognostic information, and treat
cancer. Thus, understanding cancer metabolism has implications for understanding basic cancer pathophysiology and for
clinical oncology.
Cell-Autonomous Reprogramming of Cancer
Metabolism
Many recent cancer metabolism studies have focused on the
cell-autonomous effects of cancer-promoting mutations. This
has uncovered new principles in metabolic regulation and in
crosstalk between signaling and metabolic networks. No
pathway has received more attention than aerobic glycolysis
(the Warburg effect) (Vander Heiden et al., 2009). This phenomenon involves the propensity for proliferating cells, including
cancer cells, to take up glucose and secrete the carbon as
lactate even when oxygen is present. Principles governing glycolytic regulation in cancer cells have been extensively reviewed
(Cairns and Mak, 2016).
Warburg interpreted tumor lactate secretion as an indication
that oxidative metabolism (i.e., respiration) was damaged. However, numerous studies, including Warburg’s original work, fail
to demonstrate defective respiration as a general feature of
malignant cells (Koppenol et al., 2011). Instead, respiration and
other mitochondrial activities are required for tumor growth
(DeBerardinis and Chandel, 2016). Furthermore, in non-transformed cells the Warburg effect is a reversible phenomenon
tethered to proliferation, indicating that it reflects proliferationassociated changes in metabolism rather than a unique feature
of malignancy.
Proliferating cells tend to express glucose transporters and
glycolytic enzymes out of proportion to the machinery required
to oxidize pyruvate (Curi et al., 1988), consistent with preferential conversion of glucose to lactate without loss of respiration. This distinction is important because intermediates
generated by the tricarboxylic acid (TCA) cycle are precursors
for lipids, amino acids, and nucleotides. These precursors complement precursor metabolites from glycolysis and other pathways and are necessary to support proliferation (DeBerardinis
et al., 2007).
Fuels besides glucose also contribute to core metabolic
functions of cancer cells: energy formation, biomass assimilation, and redox control. Glutamine is a prominent example
(Altman et al., 2016); however, recent work has revealed that
a diversity of nutrients and pathways support these functions. The expanding metabolic repertoire of cancer cells
has been reviewed extensively, with acetate and other fatty
acids, lactate, branched chain amino acids, serine, and
glycine representing some of the nutrients that are needed to
fuel different cancers (DeBerardinis and Chandel, 2016; White,
2013).
The growing list of cancer fuels contrasts with prevailing views
from just a few years ago, which suggested that oncogenic
signaling imposed a rigid dependence on specific nutrients (Vander Heiden et al., 2009). The current picture is more logical when
one considers that cancer cells must compete for fuels in a
crowded, nutrient-limited tissue environment. The ability to use
a panoply of fuels is advantageous, with some cancer cells
even relying on autophagic degradation or scavenging of macromolecules (Commisso et al., 2013; Guo et al., 2016; Palm
et al., 2015).
This Review builds on the increasingly sophisticated understanding of cancer metabolism derived from work over the last
decade. Our goal is to convey emerging paradigms and questions in order to guide future research so that altered metabolism
can be exploited to improve patient care.
Cell 168, February 9, 2017 ª 2017 Elsevier Inc. 657
Figure 1. Classification of Reprogrammed Metabolic Activities
Some enzyme mutations result in perturbed metabolic activities or metabolite
levels that contribute to cancer initiation (transforming activities). Mutations in
oncogenes and tumor-suppressor genes transform cells, in part, by activating
proliferative signaling and/or by inducing broad changes in gene expression
that favor cell proliferation, both of which induce metabolic reprogramming.
Some metabolic alterations enable cancer progression, whereas others are
neutral and not required for cancer cell proliferation or survival. Enzyme mutations resulting in transforming activities may also induce metabolic network
changes that are enabling or neutral for tumor progression.
Functional Classification of Reprogrammed Metabolic
Activities in Cancer Cells
Not all reprogrammed metabolic activities contribute equally to
cancer. With many metabolic activities under oncogenic control,
categorizing them based on whether they are transforming,
enabling, or neutral can clarify the role of each activity in cancer
biology and predict how it might be exploited in basic research
and clinical oncology (Figure 1).
Transforming Activities
These activities directly contribute to cell transformation, and
blocking them might prevent tumorigenesis in susceptible patients or antagonize disease progression. At present, only a
handful of metabolic activities can be considered as transforming based on genetic evidence. These include metabolic
perturbations caused by enzyme mutations that either arise
somatically in a recurrent fashion and/or are inherited in the
germlines of patients with heritable cancer predisposition syndromes. Three examples of these types of mutations have
been intensely studied: mutations in the genes encoding isocitrate dehydrogenases-1 and -2 (IDH1, IDH2); mutations in
components of the succinate dehydrogenase (SDH) complex;
and mutations in fumarate hydratase (FH) (Losman and Kaelin, 2013).
Somatic mutations in IDH1 and IDH2 occur in several tumor
types (Losman and Kaelin, 2013). These monoallelic mutations
generate an enzyme with neomorphic ability to convert a-ketoglutarate (aKG) to (D)-2-hydroxyglutarate (D-2HG) (Dang et al.,
658 Cell 168, February 9, 2017
2009). D-2HG accumulates to high levels in IDH mutant tumors
and interferes with the function of several aKG-dependent dioxygenases, including the prolyl hydroxylases that target HIF-a
subunits for degradation and epigenetic enzymes that regulate
the methylation status of histones and DNA (Losman and Kaelin,
2013). This interferes with expression of genes required for
differentiation.
SDH and FH catalyze sequential reactions in the TCA cycle
(Figure 2). Both enzymes act as tumor suppressors in familial
cancer syndromes where patients inherit one loss-of-function
allele and lose expression of the other allele in tumors (Baysal
et al., 2000; Tomlinson et al., 2002). Consequently, tumors
accumulate high levels of succinate and/or fumarate. Like
D-2HG, these dicarboxylic acids interfere with dioxygenase
function (Sciacovelli et al., 2016; Selak et al., 2005; Xiao
et al., 2012).
In mouse models, mutations in these metabolic enzymes do
not transform cells on their own but cooperate with other mutations to promote neoplasia. Loss of FH in the kidney leads to cyst
formation, and hyperproliferation results when p21 is also
deleted (Zheng et al., 2015). Idh2 mutations synergize with oncogenic Kras to produce intrahepatic cholangiocarcinoma (Saha
et al., 2014). IDH mutations increase the number of hematopoietic progenitor cells (Sasaki et al., 2012) and cooperate with
increased HoxA9 and Meis1a expression or with mutant Flt3 or
Nras to cause AML (Chen et al., 2013; Kats et al., 2014).
An important concept emerged from these studies. In all
cases, an accumulated metabolite inhibits enzymes using aKG
as a co-substrate to promote tumorigenesis (Kaelin and
McKnight, 2013). Thus, a unifying feature of D-2HG, succinate,
and fumarate as ‘‘oncometabolites’’ is causing non-metabolic
effects that promote transformation in receptive contexts. Interestingly, modestly elevated levels of the L-enantiomer of 2HG
might also contribute to malignancy (Shim et al., 2014). Given
the importance of dioxygenases in regulating epigenetics, fluctuations in metabolite levels may also play functional roles in
normal development. Consistent with this idea, extreme abnormalities in metabolite levels observed in monogenic metabolic
diseases, such as deficiency in the dehydrogenases converting
L- and D-2HG to aKG, are associated with developmental abnormalities (Erez and DeBerardinis, 2015).
Enabling Activities
These activities are altered in cancer cells but are not involved in
transformation. They carry out conventional metabolic tasks
such as supporting energetics, generating macromolecules,
and maintaining redox state and are required for tumor progression (Figure 1). In many cases they are effectors of oncogenes
and tumor-suppressor genes. For example, oncogene expression is sufficient to enhance glucose uptake in fibroblasts (Flier
et al., 1987). Transcriptional targets of c-MYC include metabolite
transporters and enzymes required for glucose metabolism, glutaminolysis, and biosynthesis, and inhibiting these activities suppresses growth of c-MYC-driven tumors (Altman et al., 2016).
Oncogenic KRAS also regulates nutrient acquisition, macromolecular synthesis, and redox homeostasis, and inhibiting these
pathways suppresses oncogenic KRAS-driven tumor growth
(White, 2013). Activation of mTORC1, an essential component
of growth factor signaling networks, enables protein, lipid, and
Figure 2. Nutrient Availability and the Metabolic Network Both Influence Metabolic
Phenotypes
Both nutrient availability and metabolic network
configuration affect how cells use metabolism to
produce ATP, generate macromolecules, and
regulate redox state. Key reactions in central
carbon metabolism are shown, including how the
TCA cycle and the electron transport chain are
involved in purine and pyrimidine synthesis. Some
of the reactions catalyzed by enzymes and metabolites discussed in this Review are also shown
for reference. PKM2, pyruvate kinase M2; LDH,
lactate dehydrogenase; PDH, pyruvate dehydrogenase; PDHK, pyruvate dehydrogenase kinase;
PC, pyruvate carboxylase; IDH, isocitrate dehydrogenase; SDH, succinate dehydrogenase; FH,
fumarate hydratase; GLS, glutaminase; DHODH,
dihydroorotate dehydrogenase; Ox PPP, oxidative
pentose phosphate pathway; Non-Ox PPP, nonoxidative pentose phosphate pathway; aKG,
a-ketoglutarate; OAA, oxaloacetate; ROS, reactive
oxygen species.
nucleic acid synthesis through a variety of mechanisms (Howell
et al., 2013). Genetic suppression of metabolic activities may
also enable tumor growth. For example, somatic deletion of
genes involved in arginine synthesis allows aspartate to be
diverted toward nucleotide synthesis, thereby supporting cell
proliferation (Rabinovich et al., 2015). Most existing and investigational metabolic therapies are directed against enabling activities (Table 1).
Neutral Activities
Many metabolic features of cancer cells are dispensable for tumor growth (Figure 1). In a given context, these activities are
predicted to be poor therapeutic targets. Fluctuating nutrient
access may cause activities to be required in some contexts
and dispensable in others. Thus, confidently classifying an activity as neutral is challenging and requires definitive proof that loss
of the activity does not impair tumor progression.
Pyruvate kinase M2 (PKM2) provides an example of an
enzyme dispensable for tumor growth. PKM2 is an isoform of
the glycolytic enzyme pyruvate kinase
that is expressed in most cancers
(Figure 2) and is regulated by oncogenic
signaling (Israelsen and Vander Heiden,
2015). PKM2 expression is not required
for growth of breast tumors, liver tumors,
leukemia, or xenograft tumors from
various human cancer cell lines in mice
(Dayton et al., 2016; Israelsen and Vander
Heiden, 2015). However, studies of PKM2
still provide insight into cancer metabolic
regulation. PKM2 loss promotes cancer
in some mouse models, and characterizing these models has argued that how
glucose is metabolized can influence tumor progression (Dayton et al., 2016; Israelsen et al., 2013). In fact, activating
PKM2 can slow growth of some tumors,
but the ability of tumors to lose pyruvate
kinase expression argues against PKM2 being a good therapeutic target for most cancers (Israelsen and Vander Heiden, 2015).
What Products of Metabolism are Limiting for
Proliferation?
One approach to identify enabling activities is to determine
which aspects of metabolism are limiting for cancer cell proliferation. Targeting activities that supply limiting materials for proliferation is therapeutically attractive, especially if the pathways
used are less important in normal proliferative tissues. Although
several metabolic products have been proposed as critical outputs of cancer metabolism, which are rate limiting for proliferation remains controversial.
Is ATP Limiting for Proliferating Cells?
ATP plays a critical role in all cells to enable otherwise unfavorable processes. Aerobic glycolysis has been historically interpreted as a shift from oxidative phosphorylation toward
fermentation to generate ATP (Gatenby and Gillies, 2004; Vander
Cell 168, February 9, 2017 659
Table 1. Drugs Targeting Metabolic Activities to Treat Malignancies or Suppress Immune Cell Proliferation
DRUG
TARGET(S)
EXAMPLE INDICATIONS
Methotrexate
dihydrofolate reductase (DHFR)
lymphoma
breast cancer
choriocarcinoma
osteosarcoma
Pemetrexed
DHFR
glycinamide ribonucleotide formyltransferase (GARFT)
thymidylate synthase
lung cancer
mesothelioma
5-Fluorouracil
thymidylate synthase
colon cancer
pancreatic cancer
breast cancer
Gemcitabine
Cytarabine
Fludarabine
ribonucleotide reductase
DNA synthesis
pancreatic, lung cancer
leukemia
lymphoma
Hydroxyurea
ribonucleotide reductase
acute myeloid leukemia (AML)
Leflunomide
DHODH
rheumatoid arthritis
psoriatic arthritis
Azathioprine
DNA synthesis
immunosuppression for organ transplant
asparagine depletion
acute lymphoblastic leukemia (ALL)
AG120, AG221, AG881, IDH305,
BAY1436032, FT-2102
IDH1 and/or IDH2
clinical trials for AML, MDS, and solid tumors
with mutant IDH1 or IDH2
Epacadostat
indoleamine 2,3-dioxygenase-1 (IDO1)
clinical trials for chronic lymphocytic leukemia
(CLL), ovarian cancer, melanoma, other cancers
CB-839
glutaminase
clinical trials for AML, ALL, solid tumors
TVB-2640
Fatty-acid synthase
clinical trials for solid tumors
PEG-BCT-100
(ADI-PEG20)
AEB-1102
arginine depletion
clinical trials for hepatocellular carcinoma and
other hematological and solid tumors including
those with ASS1 deficiency
Nucleotide Metabolism
Amino Acid Metabolism
L-Asparginase
Select Drugs in Trials
The cell target(s) for several drugs as well as some approved indications are shown. Some drugs targeting metabolism in clinical trials are also shown.
Heiden et al., 2009). This route of ATP production yields less ATP
per mole of glucose but allows fast ATP generation even in low
oxygen (Koppenol et al., 2011). Energy is derived from ATP hydrolysis to ADP (or AMP) and thus depends on the ATP/ADP
(or ATP/AMP) ratio. Therefore, cells must continuously oxidize
nutrients to regenerate ATP to maintain homeostasis. Although
many proliferative processes also require ATP hydrolysis, the
additional ATP needed for proliferation is small relative to the requirements for homeostasis (Vander Heiden et al., 2009). Thus,
although producing ATP is critical for survival, it may not be
limiting for proliferation.
Little evidence supports ATP limitation as a reason cells use
aerobic glycolysis. First, many proliferating cells use aerobic
glycolysis regardless of nutrient and oxygen availability (Vander
Heiden et al., 2009). Second, cancer cells retain the capacity
to increase fermentation when respiration is inhibited and also
can increase respiration when mitochondria are uncoupled
(Andrzejewski et al., 2014; Birsoy et al., 2014). The existence of
‘‘spare respiratory capacity’’ argues that the ATP/ADP ratio is
660 Cell 168, February 9, 2017
sufficiently high to limit electron transport, and substrates are
available in excess of the demand for ATP synthesis. Finally,
increasing ATP consumption can promote proliferation (Fang
et al., 2010), providing further evidence that the ability to
generate ATP is greater than what is needed by proliferating cells.
Is NADPH Limiting for Proliferating Cells?
Cells store energy as reduced carbon in carbohydrates and
lipids, and the anabolic reactions used to synthesize these materials requires electrons from NADPH (Figure 3), leading to the hypothesis that NADPH may be limiting for proliferation (Vander
Heiden et al., 2009). However, the fact that some cancers,
such as clear cell renal cell carcinoma, accumulate lipids (Qiu
et al., 2015) argues that these cancers have sufficient NADPH
to produce lipids in excess of those needed for cell membranes.
Most macromolecular biosynthesis occurs in the cytosol where
several enzymes can produce NADPH, including malic enzyme 1,
IDH1, and folate metabolism; however, the oxidative pentose
phosphate pathway (oxPPP) is a major source of cytosolic
Figure 3. Role of Redox Cofactors in Energy
Generation and Biosynthesis
Cells rely on nutrient oxidation to generate ATP
(energy), either through glycolysis or via NADH
generation to fuel oxidative phosphorylation.
Generating biomass can involve either nutrient
reduction or nutrient oxidation. Continued nutrient
oxidation requires cycling of NADH back to NAD+,
which necessitates transfer of electrons to an
electron acceptor such as oxygen.
NAPDH (Fan et al., 2014; Lewis et al., 2014) (Figure 2). Despite
this, most ribose is produced via the non-oxidative pentose
phosphate pathway, which does not produce NAPDH (Boros
et al., 2000; Ying et al., 2012). A major regulator of oxPPP flux
is the NADP+/NADPH ratio, and exposing cells to oxidative
stress increases oxPPP flux (Kuehne et al., 2015), arguing that
cells have excess capacity to produce NADPH even when
ROS levels increase. Thus, although NADPH is important to
help cancer cells cope with ROS, generating enough cytosolic
NADPH for competing metabolic pathways may not be limiting
for proliferation. This assertion is supported by both classic literature (Reitzer et al., 1980) and the fact that reduced oxPPP flux
resulting from G6PD deficiency has no impact on human cancer
risk (Cocco, 1987).
Products of the TCA Cycle Can Be Limiting for
Proliferation
Clues to which aspects of metabolism are most limiting for proliferation come from studies where proliferation arrest from
metabolic disruption is rescued by re-supplying metabolites.
A classic example is glutamine metabolism for TCA cycle anaplerosis (DeBerardinis et al., 2007; Yuneva et al., 2007). Most
cultured cells require glutamine, and providing cells with aKG
or other TCA cycle intermediates can rescue proliferation in
low-glutamine conditions (Yuneva et al., 2007). Activating pyruvate carboxylase to allow anaplerosis from glucose can also
enable proliferation in low-glutamine and contexts where glutamine is not used (Figure 2) (Cheng et al., 2011; Davidson et al.,
2016b; Sellers et al., 2015).
The TCA cycle carbon requirement suggests that a product of
this cycle can be limiting for proliferation. TCA cycle intermediates are precursors for several biosynthetic intermediates, but
production of amino acids accounts for the dominant disposition
of TCA cycle carbon in biomass (Hosios
et al., 2016). Aspartate, asparagine, proline, and glutamate can all be produced
from TCA cycle intermediates, and cells
synthesize these amino acids from glutamine even when they are present in the
environment (Hosios et al., 2016). Indeed,
synthesis of both aspartate (Birsoy et al.,
2015; Sullivan et al., 2015) and proline
(Loayza-Puch et al., 2016) can be limiting
under some conditions.
Nucleotide Synthesis Can Be
Limiting for Proliferation
Supplying nucleotide bases can rescue
proliferation of glutamine-deprived cells
and cells with high pyruvate kinase activity, suggesting that
nucleotide base production is important downstream of major
nutrient pathways (Gaglio et al., 2009; Lunt et al., 2015).
Acquiring deoxyribonucleotide bases is limiting in other cancer
contexts (Aird et al., 2013), and nucleotide synthesis is the
target of several chemotherapeutics (Table 1), arguing that
acquiring nucleotides might be a metabolic bottleneck for
diverse cancers. Why might nucleotide base synthesis be
limiting on a biochemical level? One important factor is that
unlike lipids and proteins, nucleotides may not be provided in
sufficient quantities or proportions by the tumor microenvironment, and synthesis of these complex molecules requires the
integrated metabolism of non-essential amino acids, ribose,
and one-carbon donors.
The DHODH step of pyrimidine biosynthesis is directly
coupled to the mitochondrial electron transport chain and oxygen consumption (Figure 2) and is important for both AML and
melanoma progression (Sykes et al., 2016; White et al., 2011).
Whereas oxygen is abundant in standard tissue culture, oxygen
levels in even the best perfused tissues are 9% or less (Bertout
et al., 2008), a value closer to the oxygen levels used to study
hypoxia responses in culture. Oxygen delivery can be more
limiting than glucose for normal tissue physiology, as evidenced
by lactic acid build-up in muscle during exercise, by increased
hematocrit enhancing the performance of elite athletes, and by
the fact that angiogenesis is regulated by oxygen levels rather
than by nutrients even though tissue delivery of both relies on
the vasculature. Oxygen can also be limiting for tumor growth.
Individuals living at high altitude have lower cancer incidence
(Burtscher, 2013), and increasing oxygen delivery to tissues
with erythropoietin promotes cancer progression (Szenajch
et al., 2010).
Cell 168, February 9, 2017 661
Figure 4. Intrinsic and Extrinsic Influences on Cancer Cell Metabolic
Reprogramming
Reprogrammed metabolic pathways, including pathways involved in bioenergetics, anabolism, and redox homeostasis, are common features of tumor
tissue. The metabolic phenotype of cancer cells is the cumulative result of a
variety of processes both intrinsic and extrinsic to the malignant cell. An integrated understanding of the relative contributions of all processes, in the
context of intact tumors, will be necessary to exploit metabolic reprogramming
in the clinic. Both intrinsically and extrinsically regulated pathways provide
opportunities for clinical translation, including new targets for therapy and for
non-invasive imaging to detect and monitor cancer.
Cancer cells require oxidative metabolism to form tumors
(Davidson et al., 2016b; Weinberg et al., 2010). Oxygen consumption is coupled to ATP synthesis, but maintaining an
adequate ATP/ADP ratio may not be why respiration is limiting
for proliferation. Respiration also regenerates NAD+ for oxidation
reactions (Figure 3), which are required for aspartate synthesis
(Birsoy et al., 2015; Sullivan et al., 2015), providing another link
between oxygen consumption and nucleotide synthesis. Nucleotide base carbon is more oxidized than the carbon in most nutrients. Aspartate is a key oxidized precursor for both purines
and pyrimidines (Figure 2), blood aspartate levels are low, and
most cells are unable to take up aspartate from the environment
(Birsoy et al., 2015). Thus, stoichiometric reduction of oxygen or
another molecule is required for aspartate and nucleotide base
synthesis (Sullivan et al., 2015). The ability to produce aspartate
can be limiting for some tumors (Gui et al., 2016), and respiration
also supports production of folate species for purine synthesis
(Bao et al., 2016; Meiser et al., 2016). Taken together, these
studies argue that access to oxygen or alternative electron acceptors limits nucleotide synthesis for some cancer cells.
Consequences of Electron Acceptor Limitation in
Cancer
The notion that chemical disposal of excess electrons to synthesize nucleotides is limiting for proliferation is attractive because it
explains several cancer metabolism phenotypes. First, it provides an explanation for the propensity of cancer cells to produce lactate. NAD+ regeneration from NADH requires electron
disposal (Figure 3). Conversion of pyruvate to lactate is driven
by the NAD+/NADH ratio, and decreasing this ratio in proliferating cells because of increased nucleotide production may be
sufficient to drive increased lactate production. All proliferating
cells must replicate DNA, so the accompanying reduction in
662 Cell 168, February 9, 2017
NAD+/NADH ratio could also explain the use of fermentation
by many proliferating cells across organisms and conditions.
Cancer cells have a very low NAD+/NADH ratio (Hung et al.,
2011), and changes in NAD+/NADH ratio correlate better with
tumor growth rate than changes in ATP/ADP ratio (Gui et al.,
2016). Regenerating NAD+ via orthogonal pathways can also increase cell proliferation when respiration is impaired (Birsoy
et al., 2015; Sullivan et al., 2015; Titov et al., 2016), and disposal
of excess electrons can limit proliferation in prokaryotic systems
(Dietrich et al., 2013).
A low NAD+/NADH ratio also could explain increased ROS in
cancer. ROS levels might increase because cells are deficient
in electron donors to detoxify ROS. However, electron acceptor
deficiency leading to inefficient mitochondrial electron transport
chain function will also lead to increased ROS (Figure 2). The
latter explanation is consistent with angiogenesis and oxygen
delivery being a limitation for tumor growth (Gatenby and Gillies,
2004).
What Determines How Different Tumors Use
Metabolism?
Pathways downstream of oncogenes and tumor suppressors
regulate cancer cell metabolism. Genomic alterations can also
result in copy-number gains and losses of genes encoding metabolic enzymes, and this may induce vulnerabilities (Li et al., 2014;
Locasale et al., 2011; Possemato et al., 2011). However, the
extent to which metabolic preferences are hard-wired by the tumor genotype is less clear because many non-genetic factors
also influence tumor metabolism (Figure 2). As in all tissues, tumor metabolism is dictated by a variety of intrinsic and extrinsic
factors (Figure 4). We need to understand how these factors are
integrated to create metabolic dependencies.
The Environment Can Affect Cancer Cell Metabolism
Cancer cells in culture have a different metabolic phenotype than
tumors. Whereas many cancer cell lines quantitatively convert
glucose to lactate, glucose oxidation is prevalent in tumors
(Davidson et al., 2016b; Hensley et al., 2016; Maher et al.,
2012; Marin-Valencia et al., 2012). Cultured lung cancer cells
use glutamine to supply TCA cycle carbon, whereas lung tumors
in mice prefer glucose as a TCA cycle fuel (Davidson et al.,
2016b; Hensley et al., 2016; Sellers et al., 2015). These differences translate into altered vulnerabilities, as lung cancer cell
lines require glutaminase for proliferation whereas tumors
derived from these same cells do not; the converse is true for enzymes involved in glucose oxidation (Davidson et al., 2016b).
The environment can also affect the efficacy of drugs targeting
metabolism. Metformin and other biguanides are mitochondrial
complex I inhibitors that slow tumor growth by preventing complex I-mediated NAD+ regeneration (Wheaton et al., 2014). Thus,
alternative NAD+ regeneration pathways decrease complex I
dependence and promote metformin resistance (Gui et al.,
2016). Lipid depletion potentiates the effect of acetyl-coA
carboxylase inhibitors in culture, but the same drugs impair
lung tumor growth in vivo despite the presumed availability of
fatty acids (Svensson et al., 2016). Tumors vary in the fraction
of actively proliferating cells, which influences metabolic dependency (Coloff et al., 2016). Both genetic and drug screens have
identified context-specific metabolic liabilities, implying that
the tumor environment can influence sensitivity to metabolic inhibitors (Possemato et al., 2011; Wenzel et al., 2014).
Cell Lineage Can Also Affect Cancer Metabolism
Most cytotoxic chemotherapies, including those that target
nucleotide metabolism, are only effective against select cancers.
A genetic predictor of response is lacking for most of these
drugs, with sensitive cancer types defined instead by the cancer
tissue of origin. This suggests that some metabolic dependencies might be defined by tumor lineage. Indeed, acute
lymphoblastic leukemia (ALL) cells are auxotrophic for asparagine and are sensitive to asparagine depletion by L-asparaginase. DHODH activity is a lineage-specific requirement of
myeloid cells that impacts differentiation (Sykes et al., 2016).
MYC induces distinct metabolic phenotypes in mouse liver and
lung tumors, with liver tumors displaying enhanced glutamine
catabolism and lung tumors retaining the ability to synthesize
glutamine from glucose (Yuneva et al., 2012). Mouse lung and
pancreatic tumors initiated by the same driver mutations
also exhibit differences in amino acid metabolism with lung tumors relying on uptake of free amino acids from the blood
(Mayers et al., 2016), whereas pancreatic tumors obtain
amino acids from extracellular protein (Davidson et al., 2016a).
These differences in amino acid metabolism translate into a differential dependency of lung and pancreatic tumors on
branched-chain amino acid nitrogen for nucleotide synthesis
(Mayers et al., 2016).
An explanation for how lineage influences tumor metabolism is
suggested by metabolic gene-expression studies. The metabolic network of tumors is more similar to the normal tissue
from where the cancer arose than it is to the metabolic network
of tumors arising from another tissue (Hu et al., 2013), and normal
tissues wire metabolism differently to support their unique functions. Mechanistically, tissue identity is established through
epigenetic regulation of gene expression, and despite altered
gene expression in cancer, many aspects of tissue identity are
maintained (Gaude and Frezza, 2016). Metabolism can also influence epigenetic state, but the fact that tumors retain metabolic
features from their parental normal tissue argues that cancers
adapt existing tissue metabolism to support abnormal proliferation rather than converging on a universal proliferation program.
This distinction is critical because it suggests that drugs targeting proliferative metabolism may not be effective against all
tumors.
Tumor lineage also modifies the penetrance of transforming
mutations. The restricted tumor spectrum in patients with familial
cancer syndromes caused by SDH and FH mutations indicates
an exquisite context dependence for these events to cause
malignancy. Why specific mutations cause lineage-restricted
cancers is not understood, but one might speculate that some
metabolic alterations are only tolerated in the context of a specific metabolic network.
Tissue-specific differences in nutrient availability impose
further constraints on metabolism. How cancer cells generate
aspartate in different contexts illustrates how such constraints
can affect production of a key metabolite for nucleotide synthesis. Cells in culture where oxygen is in excess produce aspartate
from glutamine via the TCA cycle, a pathway that involves multiple oxidation steps (Birsoy et al., 2015; DeBerardinis et al., 2007;
Sullivan et al., 2015) (Figure 2). However, in lung tumors where
oxygen is less abundant, aspartate is produced from glucose
via a series of reactions with fewer oxidation steps (Davidson
et al., 2016b; Hensley et al., 2016; Sellers et al., 2015). Pancreatic
tumors are particularly oxygen-limited and use yet another strategy, scavenging amino acids from extracellular protein through
KRAS-dependent macropinocytosis and circumventing any
need to synthesize aspartate from other nutrients (Commisso
et al., 2013; Davidson et al., 2016a; Kamphorst et al., 2015).
Interactions with Benign Cells Can Affect Cancer Cell
Metabolism
Tumors consist of a complex milieu of malignant and non-malignant cell types with distinct metabolic preferences. Specific nutrients might be available to cancer cells because they are
produced in a given tissue, and differential consumption of nutrients by cancer and non-cancer cells might further affect metabolite levels. In the brain, astrocytes metabolize glucose and
secrete lactate to be used by neurons as a source of energy (Bittar et al., 1996). A similar symbiotic relationship may exist within
tumors, where some cells ferment glucose to lactate, which is
then used as a respiratory substrate for other cells (Sonveaux
et al., 2008). Alanine produced by pancreatic stellate cells can
promote pancreatic cancer cell proliferation (Sousa et al.,
2016), and bone marrow stromal cells provide cysteine to promote survival of chronic lymphocytic leukemia cells (Zhang
et al., 2012). Similar relationships likely exist in all cancers, but
dissecting how cell populations share nutrients within tumors remains a challenge.
Roles for Metabolic Reprogramming during Cancer
Progression
Most cancer-related deaths occur after therapy resistance has
emerged and/or after tumor cells have spread from the primary
site. Whether specific metabolic activities acquired during tumor
progression promote therapy resistance or metastasis is being
investigated.
Role of Metabolism in Metastasis
Epithelial-mesenchymal transition (EMT) is a transdifferentiation
program associated with metastasis and chemotherapeutic
resistance (Ye and Weinberg, 2015). Enforcing an EMT-like
phenotype in lung cancer cells reduces lipogenesis and increases respiration (Jiang et al., 2015). Silencing fatty-acid
synthase in these cells induces EMT and enhances metastatic
seeding in mice. Similarly, in a breast cancer model cellular invasion, migration, and metastasis correlate with enhanced respiration and mitochondrial biogenesis driven by the transcriptional
co-activator PGC1a (LeBleu et al., 2014). Silencing PGC1a
reduces metastasis without altering tumor growth at the primary
site. In this model, PGC1a is required for cancer cell invasion but
not proliferation, perhaps explaining its role in metastasis and
dispensability for primary tumor growth.
Several studies have examined metabolic transitions accompanying detachment from extracellular matrix, an early step in
metastasis. In breast epithelial cells, loss of matrix attachment
results in oxidative stress and cell death (Schafer et al., 2009).
Treating detached cells with antioxidants or enhancing NAPDH
production by the oxPPP promotes survival. In lung cancer cells,
detachment suppresses pathways required to maximize cell
Cell 168, February 9, 2017 663
growth and induces shuttling of cytosolic-reducing equivalents
into the mitochondria to combat ROS (Jiang et al., 2016). Specifically, reductive carboxylation of cytosolic aKG is induced to produce citrate, which then enters the mitochondria to produce
NADPH. This pathway is dispensable in monolayer culture but
required in the anchorage-independent state.
Additional bottlenecks can prevent metastasis after cells enter
the circulation. Circulating melanoma cells and metastases have
elevated ROS compared to primary tumors (Piskounova et al.,
2015). Highly metastatic tumors undergo reversible metabolic
changes allowing them to increase mitochondrial NADPH and
combat ROS stress. Inhibiting this response reduces metastasis,
whereas systemic treatment with antioxidants enhances metastasis. Exacerbating oxidative stress potentiates some chemotherapeutic responses (Harris et al., 2015; Raj et al., 2011) and
enhances response to radiation therapy (Robbins and Zhao,
2004). Altogether, these studies underscore the importance of
managing ROS during tumor progression and suggest that this
might be exploited therapeutically.
Metabolism and Therapy Resistance
Therapy-resistant tumors have altered metabolic phenotypes
relative to treatment-naive tumors, with enhanced reliance on
mitochondrial metabolism in the resistant cancers. Genetic inactivation of oncogenic KRASG12D in pancreatic cancer leads to tumor regression; however, a small number of cancer cells survive
in a dormant state and regenerate tumors if KRASG12D is reactivated (Viale et al., 2014). Compared to cells with constitutive
KRASG12D expression, the surviving cells have enhanced respiration, reduced glycolysis, and an impaired ability to increase
glycolysis when respiration is inhibited. This renders surviving
cells susceptible to respiration inhibitors, and treating mice
with an ATP synthase inhibitor during tumor regression delays
relapse after KRASG12D reactivation.
Mitochondria also contribute to melanoma therapy resistance.
Higher PGC1a expression portends poor outcome in human
melanoma (Vazquez et al., 2013). In melanoma cells, PGC1a
expression is regulated by the melanocyte-lineage transcription
factor MITF and is required to maintain high levels of respiration,
low ROS, and resistance to the oxidizing agent piperlongumine.
Silencing PGC1a synergizes with piperlongumine to reduce melanoma xenograft growth. In human melanoma with mutant
BRAF, gene-expression patterns associated with mitochondrial
biogenesis predict reduced survival, and tumors that relapse
after MAPK inhibitor treatment have increased mitochondrial
biogenesis gene expression (Zhang et al., 2016). Inhibiting mitochondrial biogenesis using a mitochondrial-targeted Hsp90 inhibitor decreases MAPK inhibitor resistance in xenografts.
In autochthonous models of breast and lung cancer, antiangiogenic kinase inhibitors enhance AMPK signaling and shift
metabolism toward a more oxidative phenotype. Treated cells
are more sensitive to inhibitors of oxidative metabolism, and
the same drugs synergize with anti-angiogenics to suppress tumor growth (Navarro et al., 2016). Mitochondrial inhibition with
metformin can kill chemotherapy-resistant breast tumor stem
cells (Janzer et al., 2014). Taken together, these studies demonstrate the importance of mitochondrial function in enabling therapeutic resistance and suggest that mitochondrial inhibitors may
help curtail cancer progression.
664 Cell 168, February 9, 2017
Can Cancer Metabolism Be Exploited to Improve
Therapy?
To target metabolism for therapy, limiting metabolic processes
must be identified and understood sufficiently to target the process safely and select responsive patients. Using the classifications described above, transforming and/or enabling activities
must be identified with an adequate therapeutic index. Clinical
experience with cytotoxic chemotherapy highlights the challenges that will likely confront new metabolic therapies. Many
chemotherapies inhibit nucleotide metabolism (Table 1). These
drugs are highly effective in some tumors, but efficacy is not
solely determined by proliferative index. Tumor cells acquire mutations in cell-cycle checkpoint genes that contribute to therapeutic window, but the differential sensitivity of different cancers
to these drugs is incompletely understood. Evidence suggests
that tumors use diverse mechanisms to maintain nucleotide
pools, and this may contribute to efficacy. However, a better understanding of differential sensitivity to metabolism-targeted
chemotherapy may help in using these drugs more effectively
and in suggesting new targets.
Recent Challenges Associated with Targeting Cancer
Metabolism
Whether studies of cell-autonomous metabolic reprogramming
will identify the best therapeutic targets is unproven. Success
targeting glycolysis has been limited. Exposing patients to high
levels of 2-deoxyglucose generated tumor responses but
caused hypoglycemia symptoms, and more tolerable doses in
modern trials have been disappointing (Vander Heiden, 2011).
Attempts to target the Warburg effect directly have also been
met with limited success. Dependence on the enzyme lactate
dehydrogenase (LDH) was demonstrated both genetically and
pharmacologically (Fantin et al., 2006; Le et al., 2010; Shim
et al., 1997), leading to development of LDHA inhibitors (Boudreau et al., 2016), but none progressed to clinical trials, suggesting either unacceptable toxicity, insufficient drug exposure,
or a lack of LDHA dependence in human tumors. Lactate transport inhibition is currently being tested as an alternative
approach. Activating pyruvate dehydrogenase (PDH) by targeting its inhibitory kinase (PDHK) has been proposed as a therapeutic strategy. The PDHK inhibitor dichloroacetate (DCA) elicits
metabolic effects in human tumors (Michelakis et al., 2010), but
minimal efficacy data have been reported. Recent data indicate
that PDH is active and required by some tumors (Davidson et al.,
2016b; Hensley et al., 2016), tempering hope that further activating the enzyme will be therapeutically useful.
One success in targeting cancer metabolism is the reversal of
2HG-mediated reprogramming in IDH mutant tumors (Losman
and Kaelin, 2013). Drugs targeting mutant IDH can elicit responses in patients with hematological malignancies but have
been less effective in glioma models (Tateishi et al., 2015). The
mechanism of 2HG-mediated transformation may differ in
different cancers; however, liabilities caused by high 2HG might
be exploited as an alternative approach to treat these cancers.
Identifying Tumors Susceptible to Metabolic Therapy
Transforming mutations in metabolic enzymes present clinical
opportunities where patients can be selected based on tumor
genetics. However, most metabolic changes are not driven by
metabolic enzyme mutations, and even in the case of IDH
mutations, a durable state of mutant enzyme dependence may
not always be present (Tateishi et al., 2015). Thus, how to select
patients for drugs targeting metabolic enzymes remains a critical
therapeutic question.
Based on successes developing protein kinase inhibitors, one
approach has been to screen cell-line panels to identify genetically susceptible cancers and then to validate these hypotheses
in animal models. However, differential nutrient use in culture
and in tumors suggests that the use of cell lines to identify sensitive cancers could identify activities that are neutral in vivo.
Conversely, enabling pathways in vivo may be less important
in culture. A requirement for glutamine metabolism has been
observed in many cancer types (Altman et al., 2016), but both tissue of origin and genetics contribute to this phenotype (Yuneva
et al., 2012). The auxotrophy of some cancers for individual
amino acids is also influenced by a combination of genetic and
non-genetic factors. Auxotrophy for arginine or asparagine can
be explained in some cases by decreased expression of synthesis enzymes; however, treatment of ALL with asparaginase is
effective even when asparagine synthase is expressed, suggesting that additional factors underlie the need for exogenous
asparagine (Krall et al., 2016; Stams et al., 2003). Increased
biguanide sensitivity has been attributed to various mutations
(Buzzai et al., 2007; Shackelford et al., 2013), but tissue environment also affects response to these drugs (Gui et al., 2016).
Selecting patients for chemotherapy based on tumor type
argues that cell lineage can be an effective tool to stratify patients and argues that in addition to genetics, tumor type and
location should be considered when developing metabolism-targeted drugs.
Genetically Defined Metabolic Targets
Sensitivity to some enabling metabolic targets is determined by
genetics, including passenger deletion of enzymes that prevent
metabolic compensation in tumors. For example, most glioblastoma cells express two isoforms of the glycolytic enzyme
enolase encoded by the genes ENO1 and ENO2. ENO1 resides
on chromosome 1p36 at a tumor-suppressor locus that is homozygously deleted in 1%–5% of glioblastomas (Muller et al.,
2012). This reduces metabolic redundancy and creates a therapeutic opportunity to target ENO2. The gene encoding the
enzyme methylthioadenosine phosphorylase (MTAP) can be
deleted with the p16/CDKN2A tumor-suppressor gene. The resulting accumulation of MTA reduces PRMT5 methyltransferase
activity, creating another therapeutic opportunity (Kryukov et al.,
2016; Marjon et al., 2016; Mavrakis et al., 2016). Gene-expression changes that create auxotrophies for specific amino acids
can also be targeted, with L-asparaginase providing an example
of success. Drugs that deplete arginine are being tested in patients with ASS1-deficient tumors (Phillips et al., 2013). Limiting
serine in the diet can be efficacious in some p53 mutant cancers
(Maddocks et al., 2013).
Non-Cell-Autonomous Targeting of Tumor Metabolism
Understanding the interactions between cancer cell and immune
cell metabolism will be critical for combining metabolism-targeted therapies with immunotherapies. Nutrient availability and
metabolism affect T cell and macrophage differentiation, raising
the possibility that targeting tumor cell metabolism may either
promote or prevent anti-cancer immune responses.
Cancer cells compete with T cells for glucose in tumors, and
restricting T cell glucose metabolism causes lymphocyte
exhaustion (Chang et al., 2015; Ho et al., 2015). Thus, therapies
that decrease glucose use by cancer cells could make glucose
available for T cells and enhance immune-effector functions to
further limit tumor growth. However, if the same therapies inhibit
T cell glucose metabolism, they may limit anti-tumor immunity.
Other nutrient levels can affect both cancer and immune cells.
High arginine levels promote enhanced T cell survival and antitumor activity (Geiger et al., 2016), and therapies that deplete tumor arginine promote immune-suppressor cell accumulation
(Fletcher et al., 2015). Lactate can blunt T and NK cell function
(Brand et al., 2016). Kynurenine, a breakdown product of tryptophan, can modulate both the innate and adaptive immune
system and has been implicated in cancer-associated immunosuppression (Platten et al., 2015), and indoleamine-2,3-dioxygenase (IDO) inhibitors that decrease tryptophan metabolism can
help break immune tolerance and potentiate chemotherapy
(Muller et al., 2005). Increased MTA levels in MTAP null tumors
can also suppress T cell activation (Henrich et al., 2016) and
suggest that some metabolic interventions might enhance anticancer immune responses.
The relationship between tumors and other organ systems is
not limited to the immune system. Endothelial cells also undergo
factor-induced metabolic reprogramming, and metabolic therapies can limit endothelial cell proliferation and angiogenesis (De
Bock et al., 2013). Cancer also causes alterations in whole-body
metabolism that may influence tumor-nutrient availability. Modulating the amino acid composition of the diet can slow cancer
growth (Maddocks et al., 2013), and investigations into how
diet affects tumor growth remain an underexplored area.
Clinical Utility of Neutral Metabolic Activities
By definition, neutral metabolic activities are not therapeutic targets, but some may still present useful opportunities in clinical
oncology, particularly as predictive biomarkers or for tumor imaging. PKM2 expression in tumors is the basis for a positron
emission tomography (PET) imaging agent. N,N-diarylsulfonamide (DASA) compounds bind to PKM2, and an 11C-labeled
analog of DASA-23 is taken up by PKM2-expressing orthotopic
gliomas in mice, providing imaging contrast with the normal brain
(Witney et al., 2015). Other investigational PET agents exploit the
ability of some tumors to take up and retain glutamine (Venneti
et al., 2015). Conversion of 13C-labeled pyruvate to lactate can
be imaged in patients by hyperpolarized MRI (Nelson et al.,
2013). Some targetable activities are also the basis of new imaging techniques, with the detection of 2HG by magnetic resonance spectroscopy a prominent recent example (Andronesi
et al., 2012; Choi et al., 2012). Finally, the heterogeneity of metabolic phenotypes among human cancer patients may provide
biomarkers that correlate with disease progression (Hakimi
et al., 2016).
Future Perspectives
Analysis of cell-intrinsic metabolic preferences imposed by the
oncogenotype has been informative to uncover new paradigms
in proliferating cell metabolic regulation. However, the next
phase of cancer metabolism research will need to address
increasingly complex questions about how intrinsic and extrinsic
Cell 168, February 9, 2017 665
influences integrate to create exploitable metabolic phenotypes
in cancer. This will require consideration of the metabolic
preferences hard-wired into cancer cells by tissue of origin, interactions between benign and malignant cells within the microenvironment, and influences of the diet and microbiome on the
host. Cell culture systems must be improved to better reflect
the metabolic limitations in tumors, and these studies will be propelled by improvements in quantitative assessment of metabolic
fluxes in different contexts. Ultimately this will enable matching
of the right therapies with the right patients to exploit altered
metabolism and improve cancer outcomes.
ACKNOWLEDGMENTS
We regret that due to space limitations we were unable to cite many excellent
studies that shaped our modern understanding of cancer metabolism. We
thank members of the R.J.D. and M.G.V.H. laboratories for insight and extensive critique of the manuscript and Brooke Bevis for help with figures. M.G.V.H.
is supported by the NCI (CA168653, CA201276), the Ludwig Center at MIT,
SU2C, the Lustgarten Foundation, and the Howard Hughes Medical Institute
(Faculty Scholars Award). R.J.D. is supported by grants from the NCI
(CA157996), Cancer Prevention and Research Institute of Texas (RP160089),
V Foundation (Translational Research Award), Robert A. Welch Foundation
(I-1733), and the Howard Hughes Medical Institute (Faculty Scholars Award).
M.G.V.H. is on the scientific advisory board of Agios Pharmaceuticals and
Aeglea Biotherapeutics, and R.J.D. is on the scientific advisory board of Agios
Pharmaceuticals.
REFERENCES
Aird, K.M., Zhang, G., Li, H., Tu, Z., Bitler, B.G., Garipov, A., Wu, H., Wei, Z.,
Wagner, S.N., Herlyn, M., and Zhang, R. (2013). Suppression of nucleotide
metabolism underlies the establishment and maintenance of oncogeneinduced senescence. Cell Rep. 3, 1252–1265.
Altman, B.J., Stine, Z.E., and Dang, C.V. (2016). From Krebs to clinic: glutamine metabolism to cancer therapy. Nat. Rev. Cancer 16, 619–634.
Andronesi, O.C., Kim, G.S., Gerstner, E., Batchelor, T., Tzika, A.A., Fantin,
V.R., Vander Heiden, M.G., and Sorensen, A.G. (2012). Detection of 2-hydroxyglutarate in IDH-mutated glioma patients by in vivo spectral-editing
and 2D correlation magnetic resonance spectroscopy. Sci. Transl. Med. 4,
116ra4.
Andrzejewski, S., Gravel, S.P., Pollak, M., and St-Pierre, J. (2014). Metformin
directly acts on mitochondria to alter cellular bioenergetics. Cancer Metab.
2, 12.
Bao, X.R., Ong, S.E., Goldberger, O., Peng, J., Sharma, R., Thompson, D.A.,
Vafai, S.B., Cox, A.G., Marutani, E., Ichinose, F., et al. (2016). Mitochondrial
dysfunction remodels one-carbon metabolism in human cells. eLife 5, 5.
Baysal, B.E., Ferrell, R.E., Willett-Brozick, J.E., Lawrence, E.C., Myssiorek, D.,
Bosch, A., van der Mey, A., Taschner, P.E., Rubinstein, W.S., Myers, E.N., et al.
(2000). Mutations in SDHD, a mitochondrial complex II gene, in hereditary paraganglioma. Science 287, 848–851.
Bertout, J.A., Patel, S.A., and Simon, M.C. (2008). The impact of O2 availability
on human cancer. Nat. Rev. Cancer 8, 967–975.
Birsoy, K., Possemato, R., Lorbeer, F.K., Bayraktar, E.C., Thiru, P., Yucel, B.,
Wang, T., Chen, W.W., Clish, C.B., and Sabatini, D.M. (2014). Metabolic determinants of cancer cell sensitivity to glucose limitation and biguanides. Nature
508, 108–112.
Birsoy, K., Wang, T., Chen, W.W., Freinkman, E., Abu-Remaileh, M., and Sabatini, D.M. (2015). An essential role of the mitochondrial electron transport
chain in cell proliferation is to enable aspartate synthesis. Cell 162, 540–551.
Bittar, P.G., Charnay, Y., Pellerin, L., Bouras, C., and Magistretti, P.J. (1996).
Selective distribution of lactate dehydrogenase isoenzymes in neurons and astrocytes of human brain. J. Cereb. Blood Flow Metab. 16, 1079–1089.
666 Cell 168, February 9, 2017
Boros, L.G., Torday, J.S., Lim, S., Bassilian, S., Cascante, M., and Lee, W.N.
(2000). Transforming growth factor beta2 promotes glucose carbon incorporation into nucleic acid ribose through the nonoxidative pentose cycle in lung
epithelial carcinoma cells. Cancer Res. 60, 1183–1185.
Boudreau, A., Purkey, H.E., Hitz, A., Robarge, K., Peterson, D., Labadie, S.,
Kwong, M., Hong, R., Gao, M., Del Nagro, C., et al. (2016). Metabolic plasticity
underpins innate and acquired resistance to LDHA inhibition. Nat. Chem. Biol.
12, 779–786.
Brand, A., Singer, K., Koehl, G.E., Kolitzus, M., Schoenhammer, G., Thiel, A.,
Matos, C., Bruss, C., Klobuch, S., Peter, K., et al. (2016). LDHA-associated lactic acid production blunts tumor immunosurveillance by T and NK cells. Cell
Metab. 24, 657–671.
Burtscher, M. (2013). Effects of living at higher altitudes on mortality: a narrative review. Aging Dis. 5, 274–280.
Buzzai, M., Jones, R.G., Amaravadi, R.K., Lum, J.J., DeBerardinis, R.J., Zhao,
F., Viollet, B., and Thompson, C.B. (2007). Systemic treatment with the antidiabetic drug metformin selectively impairs p53-deficient tumor cell growth.
Cancer Res. 67, 6745–6752.
Cairns, R.A., and Mak, T.W. (2016). The current state of cancer metabolism.
Nat. Rev. Cancer 16, 613–614.
Chang, C.H., Qiu, J., O’Sullivan, D., Buck, M.D., Noguchi, T., Curtis, J.D.,
Chen, Q., Gindin, M., Gubin, M.M., van der Windt, G.J., et al. (2015). Metabolic
competition in the tumor microenvironment is a driver of cancer progression.
Cell 162, 1229–1241.
Chen, C., Liu, Y., Lu, C., Cross, J.R., Morris, J.P., 4th, Shroff, A.S., Ward, P.S.,
Bradner, J.E., Thompson, C., and Lowe, S.W. (2013). Cancer-associated IDH2
mutants drive an acute myeloid leukemia that is susceptible to Brd4 inhibition.
Genes Dev. 27, 1974–1985.
Cheng, T., Sudderth, J., Yang, C., Mullen, A.R., Jin, E.S., Matés, J.M., and
DeBerardinis, R.J. (2011). Pyruvate carboxylase is required for glutamine-independent growth of tumor cells. Proc. Natl. Acad. Sci. USA 108, 8674–8679.
Choi, C., Ganji, S.K., DeBerardinis, R.J., Hatanpaa, K.J., Rakheja, D., Kovacs,
Z., Yang, X.L., Mashimo, T., Raisanen, J.M., Marin-Valencia, I., et al. (2012).
2-hydroxyglutarate detection by magnetic resonance spectroscopy in IDHmutated patients with gliomas. Nat. Med. 18, 624–629.
Cocco, P. (1987). Does G6PD deficiency protect against cancer? A critical review. J. Epidemiol. Community Health 41, 89–93.
Coloff, J.L., Murphy, J.P., Braun, C.R., Harris, I.S., Shelton, L.M., Kami, K.,
Gygi, S.P., Selfors, L.M., and Brugge, J.S. (2016). Differential glutamate metabolism in proliferating and quiescent mammary epithelial cells. Cell Metab. 23,
867–880.
Commisso, C., Davidson, S.M., Soydaner-Azeloglu, R.G., Parker, S.J., Kamphorst, J.J., Hackett, S., Grabocka, E., Nofal, M., Drebin, J.A., Thompson,
C.B., et al. (2013). Macropinocytosis of protein is an amino acid supply route
in Ras-transformed cells. Nature 497, 633–637.
Curi, R., Newsholme, P., and Newsholme, E.A. (1988). Metabolism of pyruvate
by isolated rat mesenteric lymphocytes, lymphocyte mitochondria and isolated mouse macrophages. Biochem. J. 250, 383–388.
Dang, L., White, D.W., Gross, S., Bennett, B.D., Bittinger, M.A., Driggers, E.M.,
Fantin, V.R., Jang, H.G., Jin, S., Keenan, M.C., et al. (2009). Cancer-associated
IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739–744.
Davidson, S.M., Jonas, O., Keibler, M.A., Hou, H.W., Luengo, A., Mayers, J.R.,
Wyckoff, J., Del Rosario, A.M., Whitman, M., Chin, C.R., et al. (2016a). Direct
evidence for cancer-cell-autonomous extracellular protein catabolism in
pancreatic tumors. Nat. Med. Published online December 26, 2016. http://
dx.doi.org/10.1038/nm.4256.
Davidson, S.M., Papagiannakopoulos, T., Olenchock, B.A., Heyman, J.E., Keibler, M.A., Luengo, A., Bauer, M.R., Jha, A.K., O’Brien, J.P., Pierce, K.A., et al.
(2016b). Environment impacts the metabolic dependencies of Ras-driven nonsmall cell lung cancer. Cell Metab. 23, 517–528.
Dayton, T.L., Gocheva, V., Miller, K.M., Israelsen, W.J., Bhutkar, A., Clish,
C.B., Davidson, S.M., Luengo, A., Bronson, R.T., Jacks, T., and Vander
Heiden, M.G. (2016). Germline loss of PKM2 promotes metabolic distress and
hepatocellular carcinoma. Genes Dev. 30, 1020–1033.
De Bock, K., Georgiadou, M., Schoors, S., Kuchnio, A., Wong, B.W., Cantelmo, A.R., Quaegebeur, A., Ghesquière, B., Cauwenberghs, S., Eelen, G.,
et al. (2013). Role of PFKFB3-driven glycolysis in vessel sprouting. Cell 154,
651–663.
DeBerardinis, R.J., and Chandel, N.S. (2016). Fundamentals of cancer metabolism. Sci. Adv. 2, e1600200.
DeBerardinis, R.J., Mancuso, A., Daikhin, E., Nissim, I., Yudkoff, M., Wehrli, S.,
and Thompson, C.B. (2007). Beyond aerobic glycolysis: transformed cells can
engage in glutamine metabolism that exceeds the requirement for protein and
nucleotide synthesis. Proc. Natl. Acad. Sci. USA 104, 19345–19350.
Henrich, F.C., Singer, K., Poller, K., Bernhardt, L., Strobl, C.D., Limm, K., Ritter,
A.P., Gottfried, E., Völkl, S., Jacobs, B., et al. (2016). Suppressive effects of tumor cell-derived 50 -deoxy-50 -methylthioadenosine on human T cells. OncoImmunology 5, e1184802.
Hensley, C.T., Faubert, B., Yuan, Q., Lev-Cohain, N., Jin, E., Kim, J., Jiang, L.,
Ko, B., Skelton, R., Loudat, L., et al. (2016). Metabolic heterogeneity in human
lung tumors. Cell 164, 681–694.
Ho, P.C., Bihuniak, J.D., Macintyre, A.N., Staron, M., Liu, X., Amezquita, R.,
Tsui, Y.C., Cui, G., Micevic, G., Perales, J.C., et al. (2015). Phosphoenolpyruvate is a metabolic checkpoint of anti-tumor T cell responses. Cell 162,
1217–1228.
Dietrich, L.E., Okegbe, C., Price-Whelan, A., Sakhtah, H., Hunter, R.C., and
Newman, D.K. (2013). Bacterial community morphogenesis is intimately linked
to the intracellular redox state. J. Bacteriol. 195, 1371–1380.
Hosios, A.M., Hecht, V.C., Danai, L.V., Johnson, M.O., Rathmell, J.C., Steinhauser, M.L., Manalis, S.R., and Vander Heiden, M.G. (2016). Amino acids
rather than glucose account for the majority of cell mass in proliferating
mammalian cells. Dev. Cell 36, 540–549.
Erez, A., and DeBerardinis, R.J. (2015). Metabolic dysregulation in monogenic
disorders and cancer - finding method in madness. Nat. Rev. Cancer 15,
440–448.
Howell, J.J., Ricoult, S.J., Ben-Sahra, I., and Manning, B.D. (2013). A growing
role for mTOR in promoting anabolic metabolism. Biochem. Soc. Trans. 41,
906–912.
Fan, J., Ye, J., Kamphorst, J.J., Shlomi, T., Thompson, C.B., and Rabinowitz,
J.D. (2014). Quantitative flux analysis reveals folate-dependent NADPH production. Nature 510, 298–302.
Hu, J., Locasale, J.W., Bielas, J.H., O’Sullivan, J., Sheahan, K., Cantley, L.C.,
Vander Heiden, M.G., and Vitkup, D. (2013). Heterogeneity of tumor-induced
gene expression changes in the human metabolic network. Nat. Biotechnol.
31, 522–529.
Fang, M., Shen, Z., Huang, S., Zhao, L., Chen, S., Mak, T.W., and Wang, X.
(2010). The ER UDPase ENTPD5 promotes protein N-glycosylation, the Warburg effect, and proliferation in the PTEN pathway. Cell 143, 711–724.
Fantin, V.R., St-Pierre, J., and Leder, P. (2006). Attenuation of LDH-A expression uncovers a link between glycolysis, mitochondrial physiology, and tumor
maintenance. Cancer Cell 9, 425–434.
Fletcher, M., Ramirez, M.E., Sierra, R.A., Raber, P., Thevenot, P., Al-Khami,
A.A., Sanchez-Pino, D., Hernandez, C., Wyczechowska, D.D., Ochoa, A.C.,
and Rodriguez, P.C. (2015). l-Arginine depletion blunts antitumor T-cell responses by inducing myeloid-derived suppressor cells. Cancer Res. 75,
275–283.
Flier, J.S., Mueckler, M.M., Usher, P., and Lodish, H.F. (1987). Elevated levels
of glucose transport and transporter messenger RNA are induced by ras or src
oncogenes. Science 235, 1492–1495.
Gaglio, D., Soldati, C., Vanoni, M., Alberghina, L., and Chiaradonna, F. (2009).
Glutamine deprivation induces abortive s-phase rescued by deoxyribonucleotides in k-ras transformed fibroblasts. PLoS ONE 4, e4715.
Gatenby, R.A., and Gillies, R.J. (2004). Why do cancers have high aerobic
glycolysis? Nat. Rev. Cancer 4, 891–899.
Gaude, E., and Frezza, C. (2016). Tissue-specific and convergent metabolic
transformation of cancer correlates with metastatic potential and patient survival. Nat. Commun. 7, 13041.
Geiger, R., Rieckmann, J.C., Wolf, T., Basso, C., Feng, Y., Fuhrer, T., Kogadeeva, M., Picotti, P., Meissner, F., Mann, M., et al. (2016). L-Arginine modulates T cell metabolism and enhances survival and anti-tumor activity. Cell 167,
829–842.e13.
Gui, D.Y., Sullivan, L.B., Luengo, A., Hosios, A.M., Bush, L.N., Gitego, N., Davidson, S.M., Freinkman, E., Thomas, C.J., and Vander Heiden, M.G. (2016).
Environment dictates dependence on mitochondrial complex I for NAD+ and
aspartate production and determines cancer cell sensitivity to metformin.
Cell Metab. 24, 716–727.
Hung, Y.P., Albeck, J.G., Tantama, M., and Yellen, G. (2011). Imaging cytosolic
NADH-NAD(+) redox state with a genetically encoded fluorescent biosensor.
Cell Metab. 14, 545–554.
Israelsen, W.J., Dayton, T.L., Davidson, S.M., Fiske, B.P., Hosios, A.M., Bellinger, G., Li, J., Yu, Y., Sasaki, M., Horner, J.W., et al. (2013). PKM2 isoform-specific deletion reveals a differential requirement for pyruvate kinase in tumor
cells. Cell 155, 397–409.
Israelsen, W.J., and Vander Heiden, M.G. (2015). Pyruvate kinase: Function,
regulation and role in cancer. Semin. Cell Dev. Biol. 43, 43–51.
Janzer, A., German, N.J., Gonzalez-Herrera, K.N., Asara, J.M., Haigis, M.C.,
and Struhl, K. (2014). Metformin and phenformin deplete tricarboxylic acid
cycle and glycolytic intermediates during cell transformation and NTPs in cancer stem cells. Proc. Natl. Acad. Sci. USA 111, 10574–10579.
Jiang, L., Xiao, L., Sugiura, H., Huang, X., Ali, A., Kuro-o, M., Deberardinis,
R.J., and Boothman, D.A. (2015). Metabolic reprogramming during TGFb1induced epithelial-to-mesenchymal transition. Oncogene 34, 3908–3916.
Jiang, L., Shestov, A.A., Swain, P., Yang, C., Parker, S.J., Wang, Q.A., Terada,
L.S., Adams, N.D., McCabe, M.T., Pietrak, B., et al. (2016). Reductive carboxylation supports redox homeostasis during anchorage-independent growth.
Nature 532, 255–258.
Kaelin, W.G., Jr., and McKnight, S.L. (2013). Influence of metabolism on epigenetics and disease. Cell 153, 56–69.
Kamphorst, J.J., Nofal, M., Commisso, C., Hackett, S.R., Lu, W., Grabocka, E.,
Vander Heiden, M.G., Miller, G., Drebin, J.A., Bar-Sagi, D., et al. (2015). Human
pancreatic cancer tumors are nutrient poor and tumor cells actively scavenge
extracellular protein. Cancer Res. 75, 544–553.
Kats, L.M., Reschke, M., Taulli, R., Pozdnyakova, O., Burgess, K., Bhargava,
P., Straley, K., Karnik, R., Meissner, A., Small, D., et al. (2014). Proto-oncogenic role of mutant IDH2 in leukemia initiation and maintenance. Cell Stem
Cell 14, 329–341.
Guo, J.Y., Teng, X., Laddha, S.V., Ma, S., Van Nostrand, S.C., Yang, Y., Khor,
S., Chan, C.S., Rabinowitz, J.D., and White, E. (2016). Autophagy provides
metabolic substrates to maintain energy charge and nucleotide pools in
Ras-driven lung cancer cells. Genes Dev. 30, 1704–1717.
Koppenol, W.H., Bounds, P.L., and Dang, C.V. (2011). Otto Warburg’s contributions to current concepts of cancer metabolism. Nat. Rev. Cancer 11,
325–337.
Hakimi, A.A., Reznik, E., Lee, C.H., Creighton, C.J., Brannon, A.R., Luna, A.,
Aksoy, B.A., Liu, E.M., Shen, R., Lee, W., et al. (2016). An integrated metabolic
atlas of clear cell renal cell carcinoma. Cancer Cell 29, 104–116.
Krall, A.S., Xu, S., Graeber, T.G., Braas, D., and Christofk, H.R. (2016). Asparagine promotes cancer cell proliferation through use as an amino acid exchange factor. Nat. Commun. 7, 11457.
Harris, I.S., Treloar, A.E., Inoue, S., Sasaki, M., Gorrini, C., Lee, K.C., Yung,
K.Y., Brenner, D., Knobbe-Thomsen, C.B., Cox, M.A., et al. (2015). Glutathione
and thioredoxin antioxidant pathways synergize to drive cancer initiation and
progression. Cancer Cell 27, 211–222.
Kryukov, G.V., Wilson, F.H., Ruth, J.R., Paulk, J., Tsherniak, A., Marlow, S.E.,
Vazquez, F., Weir, B.A., Fitzgerald, M.E., Tanaka, M., et al. (2016). MTAP deletion confers enhanced dependency on the PRMT5 arginine methyltransferase
in cancer cells. Science 351, 1214–1218.
Cell 168, February 9, 2017 667
Kuehne, A., Emmert, H., Soehle, J., Winnefeld, M., Fischer, F., Wenck, H., Gallinat, S., Terstegen, L., Lucius, R., Hildebrand, J., and Zamboni, N. (2015).
Acute activation of oxidative pentose phosphate pathway as first-line
response to oxidative stress in human skin cells. Mol. Cell 59, 359–371.
Muller, A.J., DuHadaway, J.B., Donover, P.S., Sutanto-Ward, E., and Prendergast, G.C. (2005). Inhibition of indoleamine 2,3-dioxygenase, an immunoregulatory target of the cancer suppression gene Bin1, potentiates cancer
chemotherapy. Nat. Med. 11, 312–319.
Le, A., Cooper, C.R., Gouw, A.M., Dinavahi, R., Maitra, A., Deck, L.M., Royer,
R.E., Vander Jagt, D.L., Semenza, G.L., and Dang, C.V. (2010). Inhibition of
lactate dehydrogenase A induces oxidative stress and inhibits tumor progression. Proc. Natl. Acad. Sci. USA 107, 2037–2042.
Muller, F.L., Colla, S., Aquilanti, E., Manzo, V.E., Genovese, G., Lee, J., Eisenson, D., Narurkar, R., Deng, P., Nezi, L., et al. (2012). Passenger deletions
generate therapeutic vulnerabilities in cancer. Nature 488, 337–342.
LeBleu, V.S., O’Connell, J.T., Gonzalez Herrera, K.N., Wikman, H., Pantel, K.,
Haigis, M.C., de Carvalho, F.M., Damascena, A., Domingos Chinen, L.T., Rocha, R.M., et al. (2014). PGC-1a mediates mitochondrial biogenesis and oxidative phosphorylation in cancer cells to promote metastasis. Nat. Cell Biol. 16,
992–1003, 1–15.
Lewis, C.A., Parker, S.J., Fiske, B.P., McCloskey, D., Gui, D.Y., Green, C.R.,
Vokes, N.I., Feist, A.M., Vander Heiden, M.G., and Metallo, C.M. (2014).
Tracing compartmentalized NADPH metabolism in the cytosol and mitochondria of mammalian cells. Mol. Cell 55, 253–263.
Li, B., Qiu, B., Lee, D.S., Walton, Z.E., Ochocki, J.D., Mathew, L.K., Mancuso,
A., Gade, T.P., Keith, B., Nissim, I., and Simon, M.C. (2014). Fructose-1,6-bisphosphatase opposes renal carcinoma progression. Nature 513, 251–255.
Loayza-Puch, F., Rooijers, K., Buil, L.C., Zijlstra, J., Oude Vrielink, J.F., Lopes,
R., Ugalde, A.P., van Breugel, P., Hofland, I., Wesseling, J., et al. (2016).
Tumour-specific proline vulnerability uncovered by differential ribosome
codon reading. Nature 530, 490–494.
Locasale, J.W., Grassian, A.R., Melman, T., Lyssiotis, C.A., Mattaini, K.R.,
Bass, A.J., Heffron, G., Metallo, C.M., Muranen, T., Sharfi, H., et al. (2011).
Phosphoglycerate dehydrogenase diverts glycolytic flux and contributes to
oncogenesis. Nat. Genet. 43, 869–874.
Losman, J.A., and Kaelin, W.G., Jr. (2013). What a difference a hydroxyl makes:
mutant IDH, (R)-2-hydroxyglutarate, and cancer. Genes Dev. 27, 836–852.
Lunt, S.Y., Muralidhar, V., Hosios, A.M., Israelsen, W.J., Gui, D.Y., Newhouse,
L., Ogrodzinski, M., Hecht, V., Xu, K., Acevedo, P.N., et al. (2015). Pyruvate kinase isoform expression alters nucleotide synthesis to impact cell proliferation. Mol. Cell 57, 95–107.
Maddocks, O.D., Berkers, C.R., Mason, S.M., Zheng, L., Blyth, K., Gottlieb, E.,
and Vousden, K.H. (2013). Serine starvation induces stress and p53-dependent metabolic remodelling in cancer cells. Nature 493, 542–546.
Maher, E.A., Marin-Valencia, I., Bachoo, R.M., Mashimo, T., Raisanen, J., Hatanpaa, K.J., Jindal, A., Jeffrey, F.M., Choi, C., Madden, C., et al. (2012). Metabolism of [U-13 C]glucose in human brain tumors in vivo. NMR Biomed. 25,
1234–1244.
Marin-Valencia, I., Yang, C., Mashimo, T., Cho, S., Baek, H., Yang, X.L., Rajagopalan, K.N., Maddie, M., Vemireddy, V., Zhao, Z., et al. (2012). Analysis of tumor metabolism reveals mitochondrial glucose oxidation in genetically diverse
human glioblastomas in the mouse brain in vivo. Cell Metab. 15, 827–837.
Marjon, K., Cameron, M.J., Quang, P., Clasquin, M.F., Mandley, E., Kunii, K.,
McVay, M., Choe, S., Kernytsky, A., Gross, S., et al. (2016). MTAP deletions in
cancer create vulnerability to targeting of the MAT2A/PRMT5/RIOK1 axis. Cell
Rep. 15, 574–587.
Mavrakis, K.J., McDonald, E.R., 3rd, Schlabach, M.R., Billy, E., Hoffman, G.R.,
deWeck, A., Ruddy, D.A., Venkatesan, K., Yu, J., McAllister, G., et al. (2016).
Disordered methionine metabolism in MTAP/CDKN2A-deleted cancers leads
to dependence on PRMT5. Science 351, 1208–1213.
Mayers, J.R., Torrence, M.E., Danai, L.V., Papagiannakopoulos, T., Davidson,
S.M., Bauer, M.R., Lau, A.N., Ji, B.W., Dixit, P.D., Hosios, A.M., et al. (2016).
Tissue of origin dictates branched-chain amino acid metabolism in mutant
Kras-driven cancers. Science 353, 1161–1165.
Meiser, J., Tumanov, S., Maddocks, O., Labuschagne, C.F., Athineos, D., Van
Den Broek, N., Mackay, G.M., Gottlieb, E., Blyth, K., Vousden, K., et al. (2016).
Serine one-carbon catabolism with formate overflow. Sci. Adv. 2, e1601273.
Michelakis, E.D., Sutendra, G., Dromparis, P., Webster, L., Haromy, A., Niven,
E., Maguire, C., Gammer, T.L., Mackey, J.R., Fulton, D., et al. (2010). Metabolic
modulation of glioblastoma with dichloroacetate. Sci. Transl. Med. 2, 31ra34.
668 Cell 168, February 9, 2017
Navarro, P., Bueno, M.J., Zagorac, I., Mondejar, T., Sanchez, J., Mourón, S.,
Muñoz, J., Gómez-López, G., Jimenez-Renard, V., Mulero, F., et al. (2016).
Targeting tumor mitochondrial metabolism overcomes resistance to antiangiogenics. Cell Rep. 15, 2705–2718.
Nelson, S.J., Kurhanewicz, J., Vigneron, D.B., Larson, P.E., Harzstark, A.L.,
Ferrone, M., van Criekinge, M., Chang, J.W., Bok, R., Park, I., et al. (2013).
Metabolic imaging of patients with prostate cancer using hyperpolarized
[1-13C]pyruvate. Sci. Transl. Med. 5, 198ra108.
Palm, W., Park, Y., Wright, K., Pavlova, N.N., Tuveson, D.A., and Thompson,
C.B. (2015). The utilization of extracellular proteins as nutrients is suppressed
by mTORC1. Cell 162, 259–270.
Phillips, M.M., Sheaff, M.T., and Szlosarek, P.W. (2013). Targeting argininedependent cancers with arginine-degrading enzymes: opportunities and challenges. Cancer Res. Treat. 45, 251–262.
Piskounova, E., Agathocleous, M., Murphy, M.M., Hu, Z., Huddlestun, S.E.,
Zhao, Z., Leitch, A.M., Johnson, T.M., DeBerardinis, R.J., and Morrison, S.J.
(2015). Oxidative stress inhibits distant metastasis by human melanoma cells.
Nature 527, 186–191.
Platten, M., von Knebel Doeberitz, N., Oezen, I., Wick, W., and Ochs, K. (2015).
Cancer immunotherapy by targeting IDO1/TDO and their downstream effectors. Front. Immunol. 5, 673.
Possemato, R., Marks, K.M., Shaul, Y.D., Pacold, M.E., Kim, D., Birsoy, K., Sethumadhavan, S., Woo, H.K., Jang, H.G., Jha, A.K., et al. (2011). Functional genomics reveal that the serine synthesis pathway is essential in breast cancer.
Nature 476, 346–350.
Qiu, B., Ackerman, D., Sanchez, D.J., Li, B., Ochocki, J.D., Grazioli, A., Bobrovnikova-Marjon, E., Diehl, J.A., Keith, B., and Simon, M.C. (2015). HIF2adependent lipid storage promotes endoplasmic reticulum homeostasis in
clear-cell renal cell carcinoma. Cancer Discov. 5, 652–667.
Rabinovich, S., Adler, L., Yizhak, K., Sarver, A., Silberman, A., Agron, S., Stettner,
N., Sun, Q., Brandis, A., Helbling, D., et al. (2015). Diversion of aspartate in ASS1deficient tumours fosters de novo pyrimidine synthesis. Nature 527, 379–383.
Raj, L., Ide, T., Gurkar, A.U., Foley, M., Schenone, M., Li, X., Tolliday, N.J., Golub, T.R., Carr, S.A., Shamji, A.F., et al. (2011). Selective killing of cancer cells by
a small molecule targeting the stress response to ROS. Nature 475, 231–234.
Reitzer, L.J., Wice, B.M., and Kennell, D. (1980). The pentose cycle. Control
and essential function in HeLa cell nucleic acid synthesis. J. Biol. Chem.
255, 5616–5626.
Robbins, M.E., and Zhao, W. (2004). Chronic oxidative stress and radiationinduced late normal tissue injury: a review. Int. J. Radiat. Biol. 80, 251–259.
Saha, S.K., Parachoniak, C.A., Ghanta, K.S., Fitamant, J., Ross, K.N., Najem,
M.S., Gurumurthy, S., Akbay, E.A., Sia, D., Cornella, H., et al. (2014). Mutant
IDH inhibits HNF-4a to block hepatocyte differentiation and promote biliary
cancer. Nature 513, 110–114.
Sasaki, M., Knobbe, C.B., Munger, J.C., Lind, E.F., Brenner, D., Brüstle, A.,
Harris, I.S., Holmes, R., Wakeham, A., Haight, J., et al. (2012). IDH1(R132H)
mutation increases murine haematopoietic progenitors and alters epigenetics.
Nature 488, 656–659.
Schafer, Z.T., Grassian, A.R., Song, L., Jiang, Z., Gerhart-Hines, Z., Irie, H.Y.,
Gao, S., Puigserver, P., and Brugge, J.S. (2009). Antioxidant and oncogene
rescue of metabolic defects caused by loss of matrix attachment. Nature
461, 109–113.
Sciacovelli, M., Gonçalves, E., Johnson, T.I., Zecchini, V.R., da Costa, A.S.,
Gaude, E., Drubbel, A.V., Theobald, S.J., Abbo, S.R., Tran, M.G., et al.
(2016). Fumarate is an epigenetic modifier that elicits epithelial-to-mesenchymal transition. Nature 537, 544–547.
Selak, M.A., Armour, S.M., MacKenzie, E.D., Boulahbel, H., Watson, D.G.,
Mansfield, K.D., Pan, Y., Simon, M.C., Thompson, C.B., and Gottlieb, E.
(2005). Succinate links TCA cycle dysfunction to oncogenesis by inhibiting
HIF-alpha prolyl hydroxylase. Cancer Cell 7, 77–85.
Sellers, K., Fox, M.P., Bousamra, M., 2nd, Slone, S.P., Higashi, R.M., Miller,
D.M., Wang, Y., Yan, J., Yuneva, M.O., Deshpande, R., et al. (2015). Pyruvate
carboxylase is critical for non-small-cell lung cancer proliferation. J. Clin.
Invest. 125, 687–698.
Shackelford, D.B., Abt, E., Gerken, L., Vasquez, D.S., Seki, A., Leblanc, M.,
Wei, L., Fishbein, M.C., Czernin, J., Mischel, P.S., and Shaw, R.J. (2013).
LKB1 inactivation dictates therapeutic response of non-small cell lung cancer
to the metabolism drug phenformin. Cancer Cell 23, 143–158.
Shim, H., Dolde, C., Lewis, B.C., Wu, C.S., Dang, G., Jungmann, R.A., Dalla-Favera, R., and Dang, C.V. (1997). c-Myc transactivation of LDH-A: implications
for tumor metabolism and growth. Proc. Natl. Acad. Sci. USA 94, 6658–6663.
Shim, E.H., Livi, C.B., Rakheja, D., Tan, J., Benson, D., Parekh, V., Kho, E.Y.,
Ghosh, A.P., Kirkman, R., Velu, S., et al. (2014). L-2-Hydroxyglutarate: an
epigenetic modifier and putative oncometabolite in renal cancer. Cancer Discov. 4, 1290–1298.
Sonveaux, P., Végran, F., Schroeder, T., Wergin, M.C., Verrax, J., Rabbani,
Z.N., De Saedeleer, C.J., Kennedy, K.M., Diepart, C., Jordan, B.F., et al.
(2008). Targeting lactate-fueled respiration selectively kills hypoxic tumor cells
in mice. J. Clin. Invest. 118, 3930–3942.
Sousa, C.M., Biancur, D.E., Wang, X., Halbrook, C.J., Sherman, M.H., Zhang,
L., Kremer, D., Hwang, R.F., Witkiewicz, A.K., Ying, H., et al. (2016). Pancreatic
stellate cells support tumour metabolism through autophagic alanine secretion. Nature 536, 479–483.
Vazquez, F., Lim, J.H., Chim, H., Bhalla, K., Girnun, G., Pierce, K., Clish, C.B.,
Granter, S.R., Widlund, H.R., Spiegelman, B.M., and Puigserver, P. (2013).
PGC1a expression defines a subset of human melanoma tumors with
increased mitochondrial capacity and resistance to oxidative stress. Cancer
Cell 23, 287–301.
Venneti, S., Dunphy, M.P., Zhang, H., Pitter, K.L., Zanzonico, P., Campos, C.,
Carlin, S.D., La Rocca, G., Lyashchenko, S., Ploessl, K., et al. (2015). Glutamine-based PET imaging facilitates enhanced metabolic evaluation of gliomas
in vivo. Sci. Transl. Med. 7, 274ra17.
Viale, A., Pettazzoni, P., Lyssiotis, C.A., Ying, H., Sánchez, N., Marchesini, M.,
Carugo, A., Green, T., Seth, S., Giuliani, V., et al. (2014). Oncogene ablationresistant pancreatic cancer cells depend on mitochondrial function. Nature
514, 628–632.
Weinberg, F., Hamanaka, R., Wheaton, W.W., Weinberg, S., Joseph, J., Lopez, M., Kalyanaraman, B., Mutlu, G.M., Budinger, G.R., and Chandel, N.S.
(2010). Mitochondrial metabolism and ROS generation are essential for
Kras-mediated tumorigenicity. Proc. Natl. Acad. Sci. USA 107, 8788–8793.
Wenzel, C., Riefke, B., Gründemann, S., Krebs, A., Christian, S., Prinz, F., Osterland, M., Golfier, S., Räse, S., Ansari, N., et al. (2014). 3D high-content
screening for the identification of compounds that target cells in dormant tumor spheroid regions. Exp. Cell Res. 323, 131–143.
Wheaton, W.W., Weinberg, S.E., Hamanaka, R.B., Soberanes, S., Sullivan,
L.B., Anso, E., Glasauer, A., Dufour, E., Mutlu, G.M., Budigner, G.S., and Chandel, N.S. (2014). Metformin inhibits mitochondrial complex I of cancer cells to
reduce tumorigenesis. eLife 3, e02242.
White, E. (2013). Exploiting the bad eating habits of Ras-driven cancers. Genes
Dev. 27, 2065–2071.
Stams, W.A., den Boer, M.L., Beverloo, H.B., Meijerink, J.P., Stigter, R.L., van
Wering, E.R., Janka-Schaub, G.E., Slater, R., and Pieters, R. (2003). Sensitivity
to L-asparaginase is not associated with expression levels of asparagine synthetase in t(12;21)+ pediatric ALL. Blood 101, 2743–2747.
White, R.M., Cech, J., Ratanasirintrawoot, S., Lin, C.Y., Rahl, P.B., Burke, C.J.,
Langdon, E., Tomlinson, M.L., Mosher, J., Kaufman, C., et al. (2011). DHODH
modulates transcriptional elongation in the neural crest and melanoma. Nature
471, 518–522.
Sullivan, L.B., Gui, D.Y., Hosios, A.M., Bush, L.N., Freinkman, E., and Vander
Heiden, M.G. (2015). Supporting aspartate biosynthesis is an essential function of respiration in proliferating cells. Cell 162, 552–563.
Witney, T.H., James, M.L., Shen, B., Chang, E., Pohling, C., Arksey, N.,
Hoehne, A., Shuhendler, A., Park, J.H., Bodapati, D., et al. (2015). PET imaging
of tumor glycolysis downstream of hexokinase through noninvasive measurement of pyruvate kinase M2. Sci. Transl. Med. 7, 310ra169.
Svensson, R.U., Parker, S.J., Eichner, L.J., Kolar, M.J., Wallace, M., Brun,
S.N., Lombardo, P.S., Van Nostrand, J.L., Hutchins, A., Vera, L., et al.
(2016). Inhibition of acetyl-CoA carboxylase suppresses fatty acid synthesis
and tumor growth of non-small-cell lung cancer in preclinical models. Nat.
Med. 22, 1108–1119.
Sykes, D.B., Kfoury, Y.S., Mercier, F.E., Wawer, M.J., Law, J.M., Haynes,
M.K., Lewis, T.A., Schajnovitz, A., Jain, E., Lee, D., et al. (2016). Inhibition of
dihydroorotate dehydrogenase overcomes differentiation blockade in acute
myeloid leukemia. Cell 167, 171–186.e15.
Szenajch, J., Wcislo, G., Jeong, J.Y., Szczylik, C., and Feldman, L. (2010). The
role of erythropoietin and its receptor in growth, survival and therapeutic
response of human tumor cells From clinic to bench - a critical review. Biochim. Biophys. Acta 1806, 82–95.
Tateishi, K., Wakimoto, H., Iafrate, A.J., Tanaka, S., Loebel, F., Lelic, N., Wiederschain, D., Bedel, O., Deng, G., Zhang, B., et al. (2015). Extreme vulnerability of IDH1 mutant cancers to NAD+ depletion. Cancer Cell 28, 773–784.
Titov, D.V., Cracan, V., Goodman, R.P., Peng, J., Grabarek, Z., and Mootha,
V.K. (2016). Complementation of mitochondrial electron transport chain by
manipulation of the NAD+/NADH ratio. Science 352, 231–235.
Tomlinson, I.P., Alam, N.A., Rowan, A.J., Barclay, E., Jaeger, E.E., Kelsell, D.,
Leigh, I., Gorman, P., Lamlum, H., Rahman, S., et al.; Multiple Leiomyoma
Consortium (2002). Germline mutations in FH predispose to dominantly inherited uterine fibroids, skin leiomyomata and papillary renal cell cancer.
Nat. Genet. 30, 406–410.
Vander Heiden, M.G. (2011). Targeting cancer metabolism: a therapeutic window opens. Nat. Rev. Drug Discov. 10, 671–684.
Vander Heiden, M.G., Cantley, L.C., and Thompson, C.B. (2009). Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033.
Xiao, M., Yang, H., Xu, W., Ma, S., Lin, H., Zhu, H., Liu, L., Liu, Y., Yang, C., Xu,
Y., et al. (2012). Inhibition of a-KG-dependent histone and DNA demethylases
by fumarate and succinate that are accumulated in mutations of FH and SDH
tumor suppressors. Genes Dev. 26, 1326–1338.
Ye, X., and Weinberg, R.A. (2015). Epithelial-mesenchymal plasticity: a central
regulator of cancer progression. Trends Cell Biol. 25, 675–686.
Ying, H., Kimmelman, A.C., Lyssiotis, C.A., Hua, S., Chu, G.C., Fletcher-Sananikone, E., Locasale, J.W., Son, J., Zhang, H., Coloff, J.L., et al. (2012). Oncogenic Kras maintains pancreatic tumors through regulation of anabolic glucose
metabolism. Cell 149, 656–670.
Yuneva, M., Zamboni, N., Oefner, P., Sachidanandam, R., and Lazebnik, Y.
(2007). Deficiency in glutamine but not glucose induces MYC-dependent
apoptosis in human cells. J. Cell Biol. 178, 93–105.
Yuneva, M.O., Fan, T.W., Allen, T.D., Higashi, R.M., Ferraris, D.V., Tsukamoto,
T., Matés, J.M., Alonso, F.J., Wang, C., Seo, Y., et al. (2012). The metabolic
profile of tumors depends on both the responsible genetic lesion and tissue
type. Cell Metab. 15, 157–170.
Zhang, G., Frederick, D.T., Wu, L., Wei, Z., Krepler, C., Srinivasan, S., Chae, Y.C.,
Xu, X., Choi, H., Dimwamwa, E., et al. (2016). Targeting mitochondrial biogenesis
to overcome drug resistance to MAPK inhibitors. J. Clin. Invest. 126, 1834–1856.
Zhang, W., Trachootham, D., Liu, J., Chen, G., Pelicano, H., Garcia-Prieto, C.,
Lu, W., Burger, J.A., Croce, C.M., Plunkett, W., et al. (2012). Stromal control of
cystine metabolism promotes cancer cell survival in chronic lymphocytic
leukaemia. Nat. Cell Biol. 14, 276–286.
Zheng, L., Cardaci, S., Jerby, L., MacKenzie, E.D., Sciacovelli, M., Johnson,
T.I., Gaude, E., King, A., Leach, J.D., Edrada-Ebel, R., et al. (2015). Fumarate
induces redox-dependent senescence by modifying glutathione metabolism.
Nat. Commun. 6, 6001.
Cell 168, February 9, 2017 669
Leading Edge
Review
Emerging Biological Principles of Metastasis
Arthur W. Lambert,1 Diwakar R. Pattabiraman,1 and Robert A. Weinberg1,2,*
1Whitehead
Institute for Biomedical Research, 9 Cambridge Center, Cambridge, MA 02142, USA
of Biology, Massachusetts Institute of Technology and the MIT Ludwig Center for Molecular Oncology, Cambridge,
MA 02142, USA
*Correspondence: weinberg@wi.mit.edu
http://dx.doi.org/10.1016/j.cell.2016.11.037
2Department
Metastases account for the great majority of cancer-associated deaths, yet this complex process
remains the least understood aspect of cancer biology. As the body of research concerning metastasis continues to grow at a rapid rate, the biological programs that underlie the dissemination and
metastatic outgrowth of cancer cells are beginning to come into view. In this review we summarize
the cellular and molecular mechanisms involved in metastasis, with a focus on carcinomas where
the most is known, and we highlight the general principles of metastasis that have begun to emerge.
Introduction
The diversity of cancers that arise in humans exceeds 200
distinct disease entities—reflecting differences in the normal
cells of origin, acquired somatic mutations, variably altered transcriptional networks, and influences of local tissue microenvironments. Attempts have been made to distill this complexity into a
unifying set of organizing principles termed cancer hallmarks
(Hanahan and Weinberg, 2000, 2011). In spite of significant
advances in the study, diagnosis, and treatment of cancer, the
vast majority of patients with advanced metastatic disease
confront a terminal illness that is, with rare exception, incurable
by current therapeutic regimens. Stated differently, the overwhelming majority of cancer-associated deaths (about 90%)
are caused by metastatic disease rather than primary tumors.
The dissemination of cancer cells from primary tumors and
their subsequent seeding of new tumor colonies in distant tissues involves a multi-step process known as the invasionmetastasis cascade (Fidler, 2003; Gupta and Massagué, 2006;
Talmadge and Fidler, 2010). This sequence of events involves
the local invasion of primary tumor cells into surrounding tissues;
intravasation of these cells into the circulatory system and survival during hematogenous transit; arrest and extravasation
through vascular walls into the parenchyma of distant tissues;
formation of micrometastatic colonies in this parenchyma; and
the subsequent proliferation of microscopic colonies into overt,
clinically detectable metastatic lesions, this last process being
termed colonization.
In contrast to the large body of findings that have revealed the
detailed pathogenetic mechanisms leading to primary tumor formation, the biological underpinnings of metastatic disease
remain poorly understood. Furthermore, relatively few principles
have emerged that would unify our understanding of how diverse
types of metastases arise and how similar or different each may
be relative to the behavior of its corresponding primary tumor.
Nonetheless, over the past 15 years significant progress has
been made in elucidating various aspects of the metastatic
program, particularly for carcinomas, which in aggregate account for 80% of cancer cases and thus the majority of cancer
deaths.
670 Cell 168, February 9, 2017 ª 2016 Elsevier Inc.
Here we summarize important advances that have revealed
some of the mechanisms that underlie the dissemination and
metastatic outgrowth of carcinoma cells. Drawing from this
increasingly large and complex body of work, we suggest that
a few key biological principles have begun to emerge for certain
aspects of the metastatic cascade, while for other steps of the
cascade a unifying conceptual framework remains more elusive.
Dissemination of Carcinoma Cells
The process of dissemination subsumes the initial steps of the
invasion-metastasis cascade that enable malignant tumor cells
to acquire traits that equip them with the ability to leave the
primary site and travel to distant tissues (Figure 1A). As with
almost all of the discussions in this review, we describe these
processes in the context of the intensively studied carcinomas.
One centrally important process enabling these steps is the
cell-biological program termed the epithelial-mesenchymal
transition (EMT), a developmental program that is normally employed during embryogenesis (and in adults for the healing of
epithelial tissues) and is hijacked by carcinoma cells, endowing
them with multiple malignant traits associated with the loss of
epithelial properties and the acquisition of certain mesenchymal
features in their stead (Thiery, 2002).
The Epithelial-Mesenchymal Transition
The EMT program confers on epithelial cells, both normal and
neoplastic, properties that are critical to invasion and metastatic
dissemination, notably increased motility, invasiveness, and the
ability to degrade components of the extracellular matrix (ECM)
(Kalluri and Weinberg, 2009; Nieto et al., 2016; Thiery, 2002). In
fact, the EMT is really a group of cell-biological programs that
share features in common but differ in certain critical details,
depending on the tissue site, the degree of malignancy, and
the contextual signals experienced by individual neoplastic cells.
These complex programs are orchestrated and coordinated by a
series of master EMT-inducing transcription factors (EMT-TFs),
notably Snail, Slug, Twist, and Zeb1, which have been explored
in great experimental detail (De Craene and Berx, 2013;
Lamouille et al., 2014). Yet other TFs capable of inducing components of the EMT program have also been described (e.g., Zeb2,
Figure 1. Dissemination of Carcinoma Cells
(A) Carcinoma cell dissemination occurs via two mechanisms: single-cell dissemination through an EMT (gray arrow) or the collective dissemination of tumor
clusters (black arrow). Recent evidence suggests that the leader cells of tumor clusters also undergo certain phenotypic changes associated with the EMT.
(B) The epithelial state can be portrayed as the default state of residence; as cells undergo an EMT they enter into a succession of multiple epigenetic states,
depicted here as free energy wells, with each state moving toward a more mesenchymal phenotype representing a higher energy state.
(C) However, the barriers between states, depicted here again as free energy wells, may be relatively low, resulting in substantial spontaneous interconversion
between them, this being manifested as phenotypic plasticity.
Foxc2, Prrx1, among others), but their roles in cancer pathogenesis remain less well documented. Although traditional models
of tumorigenesis posit that metastasis is a late event during the
course of multi-step tumor progression, some studies have
shown that acquisition of EMT-associated traits and the process
of dissemination can actually occur relatively early, being evident
even in certain preneoplastic lesions (Hüsemann et al., 2008;
Rhim et al., 2012; Harper et al., 2016).
Of additional relevance is the fact that several types of
carcinoma cells have been found to acquire tumor-initiating
capability after induction of EMT programs. These include breast
(Mani et al., 2008; Morel et al., 2008), colorectal (Brabletz
et al., 2005; Fan et al., 2012; Pang et al., 2010), ovarian
(Long et al., 2015), pancreatic (Rasheed et al., 2010), prostate
(Kong et al., 2010), and renal (Zhou et al., 2016), among other
types of carcinomas. Tumor-initiating ability, usually depicted
as the defining trait of cancer stem cells (CSCs), is generally
gauged by implantation of populations of neoplastic cells in
appropriate mouse hosts. Such tests indicate that CSCs are
almost always present as relatively small subpopulations of
neoplastic cells residing within individual tumors among larger
populations of cancer cells that lack tumor-initiating powers.
Residence of a disseminating carcinoma cell in the CSC state
would seem to be critical for progression through the invasionmetastasis cascade, since disseminated tumor cells must
presumably be endowed with tumor-initiating ability in order to
function as the founders of new metastatic colonies. Moreover,
acquisition of more mesenchymal traits, as driven by an EMT
program, has been found to elevate the resistance of carcinoma
cells to various types of cytotoxic treatments, including both radio- and chemotherapies (Gupta et al., 2009; Kurrey et al., 2009),
providing one explanation of the often-observed phenomenon
that CSCs tend to be more therapy resistant than their nonCSC counterparts (Singh and Settleman, 2010).
While the EMT program might be depicted as operating much
like a binary switch, in which cancer cells reside either in an
epithelial or a mesenchymal state, the truth is more complex,
in that EMT programs activated in carcinoma cells usually drive
the acquisition of certain mesenchymal traits while permitting the
retention of some epithelial traits, leaving carcinoma cells with
mixed epithelial/mesenchymal phenotypes (Figures 1B and 1C).
EMT programs seem almost invariably to be triggered in
carcinoma cells by heterotypic signals that these cells receive
from the nearby tumor-associated stroma. Thus, during the
course of tumor progression, the stroma—which is composed
of a variety of fibroblasts, myofibroblasts, endothelial, myeloid,
and lymphoid cells recruited from host tissues—increasingly
takes on the appearance of a stroma that typically forms during
the healing of wounded epithelial tissues. Such a ‘‘reactive’’
stroma releases various signals, including TGF-bs, Wnts, and
certain interleukins that impinge on nearby carcinoma cells,
inducing the latter to activate their previously silent EMT
Cell 168, February 9, 2017 671
programs. This activation is generally reversible, and indeed
carcinoma cells that have activated EMT programs may revert
via a mesenchymal-epithelial transition (MET) to the phenotypic
state in which their ancestors resided prior to induction of the
EMT program.
While the EMT program appears to be critical to invasion and
dissemination of most and possibly all carcinoma types (see
below), to date there have been no rules formulated to predict
expression of its various components in different tissue contexts. Among the unresolved fundamental issues are: (1) the
nature of the heterotypic signals that converge on carcinoma
cells and collaborate to activate previously silent EMT programs
in these cells; (2) the extent to which these programs are activated at various stages of carcinoma progression; (3) the extent
to which the differentiation programs of normal cells of origin
influence the expression of various components of the EMT program; (4) the respective roles of the various EMT-TFs cited above
in collaborating with one another in choreographing various
types of EMT programs; (5) the influence of somatic mutations
sustained during primary tumor formation on the activation and
expression of EMT programs; and (6) the roles of intracellular
and extracellular signaling pathways in sustaining the expression
of already-activated EMT programs.
Invasion by Collective Migration
Although the EMT is widely embraced as an important mode of
carcinoma cell dissemination, its precise roles in primary tumor
behavior remain unresolved. For example, invasion by primary
tumor cells generally involves the collective migration of large,
cohesive cohorts of cells into adjacent tissues rather than
the dispersal of individual carcinoma cells (Figure 1A; Friedl
et al., 2012). The organization of these cohorts appears to conflict with the behavior of cells that have passed through an
EMT and have lost cohesive cell-cell interactions, notably those
mediated by adherens junctions. Thus, these cohorts provoke
the question of whether EMT programs are indeed central to
eventual carcinoma cell dissemination, as implied above, or
instead represent only one of several alternative cell-biological
programs that enable dissemination to occur.
Collective migration involving groups of cells, which is
commonly seen at the borders of invasive carcinomas, is best
documented in the case of carcinomas of the breast and lungs
(Friedl et al., 2012); similar invasive cohorts undoubtedly participate in invasion by other types of carcinoma cells as well (Chung
et al., 2016; Veracini et al., 2015). Cells residing within these
invasive cell phalanxes continue to express key epithelial
markers such as E-cadherin, which helps to sustain the cohesion
between the individual epithelial cells within these cohorts.
Moreover, the polyclonal nature of metastatic colonies of certain
breast cancers raises the possibility that they arose from genetically heterogeneous clusters of disseminated cells, rather than
arising clonally from single disseminated cells (Cheung et al.,
2016). This raises the question of whether collective migration
represents an alternative to EMT and whether the two cell-biological programs are essentially mutually exclusive.
In fact, detailed histopathological analyses of invasive cohorts
often suggest that the EMT does indeed participate in collective
migration (Ye et al., 2015). Thus, these cohorts are themselves
internally complex, with invading cells at the leading edges
672 Cell 168, February 9, 2017
paving the way for large populations of followers to which they
remain attached via cell-cell junctions (Cheung et al., 2013). In
some cases, careful examination has revealed that certain
mesenchymal traits are exhibited by the leading cells at the
invasive fronts during collective migration (Revenu and Gilmour,
2009; Westcott et al., 2015; Ye et al., 2015). Such invading
leaders are likely to release various proteases that degrade the
extracellular matrix that would otherwise impede the forward
progress of the cohort as a whole. Moreover, such leader cells
may also possess the EMT-associated motility to enable forward
motion of the cohort as a whole. Together, the cells at invasive
edges may therefore pave the way for the followers that constitute the bulk of the cell phalanxes.
Unresolved is a key experimental test of this model: can collective invasion occur if activation of EMT programs is totally
blocked? Yet other studies report the presence of cancer-associated fibroblasts, rather than carcinoma cells that have undergone an EMT, as leader cells at the invasive edges of carcinomas
(Gaggioli et al., 2007). Thus, additional experimental evidence is
required to address and clarify more precisely the events occurring at the invasive edges of carcinomas and the nature of the
normal and neoplastic cell types involved.
An Essential Role of the EMT Program in Metastasis
Two studies have recently undertaken to refute the essential role
of the EMT program in the process of metastasis (Fischer et al.,
2015; Zheng et al., 2015). In both instances, the proofs that EMTs
did not occur while metastasis proceeded were not supported
by the evidence presented, leaving open the continuing question
of whether EMT is indeed critical to the metastatic ability of all
types of carcinoma cells. Moreover, the reports of these findings
coincide with a time when the definition of the EMT is undergoing
re-evaluation, as suggested above. Thus, EMT programs are
increasingly viewed as generating cells residing in a spectrum
of multiple intermediate states lying between epithelial and
mesenchymal poles (Figures 1B and 1C; Bednarz-Knoll et al.,
2012; Grosse-Wilde et al., 2015; Li and Kang, 2016; Nieto
et al., 2016). It is therefore likely that in some cases, metastasizing carcinomas may exhibit overt mesenchymal properties
that aid in metastatic spread (Bonnomet et al., 2012; Trimboli
et al., 2008), whereas in other cases they may not require
the same suite of EMT-associated traits (Celià-Terrassa et al.,
2012).
In fact, a large number of reports highlight the existence
of the ‘‘partial EMT’’ state and its propensity to enhance
tumor progression and metastasis (Bednarz-Knoll et al., 2012;
Grosse-Wilde et al., 2015; Hong et al., 2015; Jordan et al.,
2011; Lundgren et al., 2009; Sampson et al., 2014; Schliekelman
et al., 2015). In contrast, induction of a fully mesenchymal state,
as achieved experimentally through the actions of introduced,
highly expressed EMT-TFs and resulting completion of an entire
EMT program, yields cells that have lost tumor-initiating ability
and thus the power to found metastatic colonies (Ocaña et al.,
2012; Tsai et al., 2012). Stated differently, the phenotypic plasticity associated with carcinoma cells inhabiting the middle of
the epithelial-mesenchymal spectrum appears to be critical to
the founding of metastatic colonies and their subsequent robust
outgrowth. Unaddressed by this discussion is the behavior of
ovarian carcinomas, whose spread through the peritoneal space
operates through principles very different from those characteristic of most solid tumors.
Circulating Tumor Cells
Individual invasive carcinoma cells and invasive cohorts arising
from primary tumors may, sooner or later, invade into the vasculature either of adjacent normal tissues or the neovasculature
that has been assembled within the tumors themselves. The resulting intravasation provides access to an avenue for circulating
tumor cells (CTCs) to travel to distant sites, where they may seed
new metastatic colonies (Kang and Pantel, 2013). Such travelers
may move as individual cells or as multi-cellular clumps that can
persist in the circulation until they encounter the small-bore
microvessels of distant tissues (which often possess luminal diameters as small as 8 mm). The consequent physical trapping
would seem to ensure that the vast majority of intravasated
CTCs dwell in the general circulation for only seconds or minutes
after their initial entry into the vasculature. Although most CTCs
may be rapidly cleared, it has been recently reported that even
clusters of CTCs are capable of maneuvering through capillary-sized vessels, doing so as a single-cell chain still held
together through adhesive interactions (Au et al., 2016). CTC
clusters introduced experimentally into the venous circulation
are far more efficient than individual carcinoma cells in seeding
metastatic colonies, ostensibly because, relative to single
CTCs, they are more resistant to apoptosis and may have an
advantage in physically lodging in the lumia of vessels (Aceto
et al., 2014). In addition, these clusters might be shielded from
various types of attacks, such as those launched by natural killer
(NK) cells, and may benefit from certain poorly understood
advantages in post-extravasation proliferation that could
contribute to their increased metastatic efficiency.
Nonetheless, single CTCs have been extensively studied in
recent years because of technical improvements in their isolation
from the blood of cancer patients (Aceto et al., 2015). Implicit in
these surveys is the notion that these cells represent intermediaries between primary tumors and eventually formed metastatic
colonies. However, in light of the considerations discussed
above, it remains unclear which types of CTCs (single versus
clusters) are actually responsible for the lion’s share of metastasis formation. Indeed, the probability of a single CTC successfully founding a metastatic colony is vanishingly small (Baccelli
et al., 2013). Independent of these considerations is the notion
that single and clustered CTCs released by primary tumors could
often be produced in a certain ratio, in which case the solitary
CTCs may serve as surrogate markers of the cell clusters that
may indeed be responsible for the formation of the great majority
of metastatic colonies.
Of additional relevance here is the fact the CTCs, traveling
either as individual cells or as clusters, often exhibit combinations
of epithelial and mesenchymal traits, reinforcing the role of the
EMT program in the process of intravasation and cancer cell
dissemination (Yu et al., 2013). Moreover, in longitudinal studies
of individual patients, the fraction of mesenchymal CTCs has
been found to increase progressively with acquired treatment
resistance and disease progression. One concern here derives
from the fact that CTC enrichment methods that rely on the
display by CTCs of cell-surface epithelial markers may well
miss capturing a sizeable, clinically relevant portion of the
CTCs that are responsible for seeding distant metastases but
have shed the bulk of their epithelial cell-surface markers as a
consequence of extensive progression through EMT programs.
All of these provisos do not detract from certain alreadyproven uses of CTC technology. Single CTCs may indeed be
useful for certain types of diagnoses, since the presence of
CTCs has been repeatedly found in commonly occurring carcinomas, including those of the breast, prostate, lung, and colon
(Aceto et al., 2015). In particular, the longitudinal monitoring of
CTC concentrations through ‘‘liquid biopsies’’ may provide highly useful information about the responses of a patient’s tumor to
various types of therapies. Another clearly useful application is
the measurement of CTCs in patients whose primary tumors
have been removed in order to determine whether residual,
occult metastatic deposits persist and continue to empty carcinoma cells into the circulation.
In addition, the isolation, ex vivo expansion, and analysis of
viable CTCs can be used to profile genetic mutations and drug
sensitivities of the cells residing within primary tumors (Yu
et al., 2014). This may allow the prediction of patient responses
to various types of therapy, especially when the lesions being
treated are not readily biopsied, for example those in the brain.
Indeed, one already published report demonstrates that CTCs
isolated from prostate cancer patients can be harbingers of
eventually acquired drug resistance, such as those carrying molecular changes that can confer resistance to androgen receptor
antagonists (Miyamoto et al., 2015). Ideally, early detection and
characterization of CTCs prior to the appearance of clinically
detectable metastatic growths could be used to initiate or switch
treatment before the eruption of life-threatening metastases. At
present, however, this seems to be impractical, given the fact
that even actively growing, aggressive tumors tend to release
relatively low numbers of detectable CTCs into the circulation.
Interactions in Transit: Fates of Intravasated Carcinoma
Cells
In fact, carcinoma cells that have successfully invaded stromal
environments surrounding primary tumors can intravasate either
into blood or lymphatic vessels. The dissemination of cancer
cells to draining lymph nodes represents an important clinical
parameter that is incorporated into the histopathological staging
of the disease and thus is associated with particular prognoses
(de Boer et al., 2010). While carcinoma cells may promote the
growth of lymphatic vessels through the process of lymphangiogenesis (Karaman and Detmar, 2014)—a process that is correlated with disease progression (Skobe et al., 2001)—there is
scant evidence for the notion that the draining lymph nodes
represent temporary staging areas that enable significant
numbers of cancer cells to pause before proceeding further
into the bloodstream and thereafter to distant sites in the body.
Hence, these small metastatic deposits probably represent
dead ends for cancer cells and primarily function as surrogate
markers that reveal the extent of parallel, concomitant dissemination from the primary tumor into the general circulation. For
this reason, the discussion below is focused on the hematogenous transport of carcinoma cells, as this is likely the main route
that metastatic cancer cells traverse prior to entering and colonizing distant tissues.
Cell 168, February 9, 2017 673
Figure 2. Interactions in Transit
Carcinoma cells escaping from primary tumors can intravasate into the circulation, either as single circulating tumor cells (CTCs) or as multicellular CTC clusters.
The bloodstream represents a hostile environment for CTCs, exposing them to rapid clearance by natural killer (NK) cells or fragmentation due to the physical
stresses encountered in transit through the circulation. Carcinoma cells gain physical and immune protection through the actions of platelets, which coat CTCs
shortly after intravasation. Neutrophils can provide protection from NK cell attacks as well, while also contributing to the physical entrapment and extravasation of
CTCs. Once lodged in a capillary, activated platelets and carcinoma cells secrete a number of bioactive factors that can act on monocytes, endothelial cells, and
the carcinoma cells themselves. The collective effects of these interactions promote the transendothelial migration (TEM) of carcinoma cells, which can be aided
by metastasis-associated macrophages (MAMs) in the target parenchyma. In lieu of TEM, arrested carcinoma cells may also proliferate intraluminally (not shown)
or induce necroptosis in endothelial cells.
The safe passage of intravasated cancer cells to distant
anatomical sites is hardly guaranteed. Although the transit time
of a cancer cell through the bloodstream may amount to only a
few minutes, CTCs encounter multiple obstacles en route to
the parenchyma of distant tissues. Foremost here are the physical challenges associated with life in circulation, which include
loss of attachment to a substrate, hydrodynamic flow, and shear
stress (Headley et al., 2016). In addition, carcinoma cells in the
circulation are vulnerable to an immune attack, notably by NK
cells that target them for rapid elimination. However, certain
interactions between circulating carcinoma cells and other cell
types in the circulation can actually facilitate their passage
to and extravasation at distant sites, notably those involving
platelets, neutrophils, monocytes/macrophages, and endothelial cells (Figure 2).
Interactions with Platelets
Once in the circulation, CTCs rapidly associate with platelets, an
interaction that is triggered by tissue factor displayed on the
surface of the carcinoma cells (Labelle and Hynes, 2012). Depending on the rate of CTC introduction into the circulation,
this can lead to imbalances in the normal homeostatic controls
on coagulation, which can result in certain clotting symptoms
that are seen in patients with cancer, specifically microthrombi,
674 Cell 168, February 9, 2017
disseminated intravascular coagulation, and even large pulmonary emboli (Gay and Felding-Habermann, 2011).
At the same time, platelets facilitate tumor metastasis. Indeed,
the contribution of platelets to the metastatic process has been
appreciated since the 1960s, when studies revealed that experimental induction of thrombocytopenia can exert an anti-metastatic effect (Gasic et al., 1968), while a high platelet count has
for years been known to be associated with a poor clinical
prognosis across diverse types of carcinomas (Gay and Felding-Habermann, 2011). Platelets contain a plethora of bioactive
molecules that can potentially impact cancer progression and
work in more recent years has revealed a number of mechanisms by which platelets can alter the fate of carcinoma cells
in transit (Franco et al., 2015; Gay and Felding-Habermann,
2011).
Of relevance here is the fact that platelets can protect CTCs
from elimination by cellular arms of the immune system. More
specifically, adhered platelets can prevent tumor cell recognition
and lysis by NK cells (Kopp et al., 2009; Nieswandt et al., 1999;
Palumbo et al., 2005). This effect can be mediated by soluble
factors derived from platelets, including TGF-b and PDGF that
inhibit NK cell activity (Labelle and Hynes, 2012), and, quite
possibly, by physically shielding cancer cells from NK cells
through the formation of protective cloaks around CTCs and the
deposition of fibrinogen on the cancer cells (Palumbo et al.,
2005, 2007). Such protection specifically against NK cellmediated attack may represent the most important benefit
conferred on intravascular carcinoma cells by platelets, since
the pro-metastatic effects of the thrombocytes are no longer
apparent in mice depleted of NK cells (Palumbo et al., 2005).
In addition to protecting circulating tumor cells from external
insults, platelets can also alter intracellular signaling pathways
within carcinoma cells that ultimately affect the ability of the
latter to establish metastatic growths. Notably, TGF-b secreted
by degranulating platelets can act in coordination with contactdependent signals that activate the NF-kB pathway in carcinoma cells, thereby inducing or sustaining the expression of
EMT programs in the CTCs (Labelle et al., 2011). This direct
signaling between platelets and carcinoma cells can presumably substitute for the absence of stroma-derived signals that
previously led, in the context of the primary tumor, to the induction of an EMT. In the absence of such heterotypic interactions,
CTCs may revert via a MET to the epithelial state of their
ancestors in the primary tumor, thereby losing the invasive
traits and tumor-initiating ability that would seem to be critical
for subsequent extravasation and the founding of metastatic
colonies.
Once activated by cancer cells, platelets can signal to nearby
endothelial cells as well. Tumor cells elicit ATP secretion from
activated platelets, which can proceed to render the vasculature
more permeable by acting on P2Y2 receptors expressed by
endothelial cells (Schumacher et al., 2013). Moreover, physical
interactions between platelets and endothelial cells, for example
those mediated by selectins, have been proposed to be important for the adhesion of platelet-cancer cell clusters to the walls
of the vasculature (Köhler et al., 2010). It remains unclear, however, whether such adhesive interactions are actually critical to
the intraluminal arrest and eventual entrance by the neoplastic
cells into the parenchyma of various tissues.
Interactions with Neutrophils
Neutrophils can exist in distinct and dynamically changing
phenotypic states that can be shaped by the primary tumor as
well as other host cells (Coffelt et al., 2016; Fridlender et al.,
2009; Sagiv et al., 2015). We focus here on their actions in circulation, where evidence is beginning to clarify their role during this
phase of the metastatic cascade. In certain instances neutrophils have been found to inhibit metastasis. For example, primary tumors can educate neutrophils via CCL2 secretion, giving
rise to tumor-entrained neutrophils (Granot et al., 2011). These
cells appear to accumulate in the circulation and the lungs of tumor-bearing mice even prior to metastatic progression and have
been found to prevent carcinoma cells from seeding the lungs.
Neutrophils mobilized by G-CSF treatment lack this power
(Granot et al., 2011), highlighting the fact that neutrophils can
be primed to adopt different functional states.
In large part, however, the molecular and cellular physiology of
neutrophils appears to dictate that their predominant role is one
that favors metastatic seeding. For example, neutrophil extracellular traps (NETs), which are formed from released DNA molecules, are designed to entangle pathogens during a response
to infection but can also be deployed by neutrophils to capture
tumor cells in the circulation (Cools-Lartigue et al., 2013). Such
entangled CTCs may be more apt to survive intraluminally,
adhere to endothelial cells, and extravasate. Neutrophils can
directly interact with tumor cells trapped in the vasculature, prolonging their retention in the lung after intravenous injection (Huh
et al., 2010). In a similar manner, neutrophils can facilitate
adhesive interactions within liver sinusoids, thereby serving
as physical platforms on which CTCs can dock prior to extravasation (Spicer et al., 2012). Additionally, neutrophils enhance
the extravasation of tumor cells after arrest, mainly through the
secretion of various matrix metalloproteinases (MMPs) (Spiegel
et al., 2016).
Neutrophils have also been shown to exert immunosuppressive functions. Often mobilized through systemic signaling by a
primary tumor, neutrophils can inhibit both cytotoxic CD8+
T cell responses (Coffelt et al., 2015) and the intraluminal clearance of carcinoma cells by NK cells (Spiegel et al., 2016). Such
protection from attack by arms of the innate and adaptive immune system offers a clear advantage to tumor cells in transit.
Finally, some of the effects mediated by neutrophils may occur
in response to the aggregation of platelets and tumor cells noted
previously. Thus, the release of platelet-derived chemokines can
recruit neutrophils, which can then, as described here, enhance
the seeding and metastatic outgrowth of carcinoma cells in
circulation (Labelle et al., 2014).
Extravasation
Many of the intravascular interactions described above influence
the ability of CTCs to extravasate and thereby enter into the
parenchyma of distant tissues. Extravasation requires carcinoma cells to traverse the endothelial wall through a process
that is termed transendothelial migration (TEM) (Reymond
et al., 2013). Earlier we cited the ability of ATP released by
activated platelets to render the capillary walls more permeable;
in more detail, this is achieved by causing endothelial cells to
retract from one another. In addition, breast carcinoma cells
primed by TGF-b in the primary tumor acquire the ability to
produce angiopoietin-like 4 (ANGPTL4), which enhances the
permeability of the lung vasculature, promotes TEM of carcinoma cells, and leads to an increased capacity for metastatic
outgrowth (Padua et al., 2008). Several other proteins produced
by carcinoma cells have been reported to function as disruptors
of vascular integrity, including VEGF, MMPs, and ADAM12;
these secreted molecules seem to enhance both intravasation
as well as extravasation (Gupta et al., 2007; Reymond et al.,
2013), indicating that certain traits that were advantageous previously in the course of primary tumor invasion may also prove
useful at later steps in the invasion-metastasis cascade.
The recruitment of monocytes has also been demonstrated to
play a functional role in tumor cell extravasation. In particular, the
recruitment of CCR2+ inflammatory monocytes in response to
CCL2 secretion by carcinoma or host cells can facilitate extravasation and subsequent metastatic growth in the lung parenchyma (Qian et al., 2011; Wolf et al., 2012). These inflammatory
monocytes may differentiate into metastasis-associated macrophages, which similarly enhance the seeding, survival, and
growth of carcinoma cells in the lung through the release of
VEGF (Qian et al., 2009, 2011). CCL2 can also act directly on
endothelial cells to enhance vascular permeability (Wolf et al.,
Cell 168, February 9, 2017 675
2012). Although inhibition of the CCL2-CCR2 axis would seem to
represent an ideal anti-metastatic therapy, the termination of
anti-CCL2 therapy actually leads to an enhanced monocyte infiltration of tumors and lungs with a corresponding acceleration of
disease progression (Bonapace et al., 2014), underscoring the
dynamic and unpredictable nature of targeting such microenvironmental interactions.
Most experimental models of metastasis have, for various reasons, focused on the lungs as destination sites of disseminated
tumor cells. However, the requirements for successful extravasation and the relevant interactions that facilitate this process
are likely to be quite different in various tissue sites. For instance,
the fenestrated sinusoids of the bone marrow and liver are more
likely to permit the passive entry of CTCs, obviating many of the
complex interactions and mechanisms enumerated above. In
the case of the brain, the dissemination of carcinoma cells would
seem to require passage through the blood-brain barrier, which
may in fact necessitate the actions of a tissue-specific program
for extravasation that is very different from those enabling metastatic seeding elsewhere in the body. Indeed, breast cancer
cells selected for preferential metastasis to the brain express
at high levels a number of genes that are known to facilitate passage through the blood-brain barrier (Bos et al., 2009; Sevenich
et al., 2014).
In certain cases, TEM migration may not be required at all, as
arrested carcinoma cells have been found to proliferate in the
lumina of blood vessels, leading to the growth of large intraluminal tumor colonies that eventually rupture nearby endothelial
walls, enabling direct access to the tissue parenchyma (Al-Mehdi
et al., 2000). Finally, a novel mechanism has recently been
described, in which tumor cells can extravasate and generate
lung metastases via induction of programmed necrosis (necroptosis) in endothelial cells (Strilic et al., 2016).
Metastatic Colonization
The growth of an overt metastatic colony represents the final
and most deadly phase in the malignant progression of a tumor.
Still, the vast majority of carcinoma cells in circulation seem ill
prepared for growth in a distant organ environment; some experimental evidence has yielded estimates of the efficiency of
metastasis after intravenous injection of tumor cells as low as
0.01% (Chambers et al., 2002). Even carcinoma cells that have
managed to extravasate seem almost invariably destined to
either be eliminated from the tissue parenchyma or to enter
into a state of dormancy (Luzzi et al., 1998), in which they persist
in an indolent state as single disseminated tumor cells (DTCs)
or as small micrometastatic clusters—sometimes for weeks,
months, even years.
Having traveled far from the primary tumor, DTCs find themselves in a new tissue microenvironment that is devoid of the
familiar stromal cells, growth factors, and ECM constituents
that previously sustained the lives of their predecessors in the
primary site. Hence, their inability to continue proliferating and
the resulting entrance into a prolonged growth-arrested state
may often be attributable to a microenvironment to which these
cells are poorly adapted when they first arrive after extravasation. When portrayed in this way, metastatic dormancy reflects
a failure of DTCs to adapt to and colonize a given tissue. Impor676 Cell 168, February 9, 2017
tantly, a dormant state can also be actively imposed by certain
anti-proliferative signals encountered by recently arrived cells
in the parenchyma of foreign tissues. We first consider the programs operative in dormant DTCs before turning to those that
enable colonization.
Dormancy Programs
The latent, clinically inapparent phase of metastasis might well
be the result of factors beyond those cited here that render
carcinoma cells unable to proliferate, such as an inability to
induce angiogenesis or active suppression by the immune system (Aguirre-Ghiso, 2007). These two particular mechanisms
are thought to permit a low level of proliferation that is counterbalanced by ongoing elimination, resulting in no net increase in
the sizes of micrometastatic clusters.
From a clinical perspective, patients successfully treated for
their primary tumors but potentially harboring such dormant
cancer cells are considered to have asymptomatic minimal residual disease (MRD) (Figure 3A). For certain carcinomas, such
as those of the breast, prostate, and kidney, this period of
dormancy may last for many years, even decades after ostensibly successful courses of initial therapy. And while it is difficult
to formally prove that a metastatic colony directly developed
from a preexisting dormant DTC, the presence of DTCs in the
bone marrow is clearly correlated with an increased risk of eventual clinical recurrence (Braun et al., 2005). This reveals why an
understanding of the biologic bases of dormancy is of utmost
clinical importance, if only because the period of dormancy represents a critical time window during which therapeutic interventions directed at DTCs—either targeting them for elimination or
restraining their proliferation—may well succeed in preventing
the eventual eruption of life-threatening metastatic disease.
Dormancy programs (Figure 3B) can be initiated from either an
active response to signals encountered in the new tissue microenvironment or from an absence of contextual cues that carcinoma cells previously depended on while residing in their sites
of origin within primary tumors (Giancotti, 2013; Sosa et al.,
2014). As an example, DTCs that respond to survival signals
present in the microenvironment can avoid destruction and
persist for extended periods within a tissue parenchyma. In
one well-studied case, breast cancer cells that have lodged in
the bone marrow and possess high SRC activity and expression
of CXCR4 are able to activate pro-survival pathways in response
to bone-derived CXCL12 (Zhang et al., 2009). DTCs capable of
sensing and responding to these survival cues are able to counteract TRAIL-induced apoptosis, a conserved tissue defense
mechanism that can work in the opposite direction to eliminate
DTCs. The survival of DTCs may also be related to their ability
to withstand anoikis, for example through the expression of the
tyrosine kinase receptor TrkB (Douma et al., 2004) or through
non-canonical Wnt signaling mediated by WNT2 (Yu et al., 2012).
Even if DTCs benefit from such survival signals in their new tissue environment, in the absence of additional mitogenic cues,
including interactions with the extracellular matrix (ECM), these
cells may languish in a dormant state. Thus, dormancy has
been reported to ensue when disseminated carcinoma cells fail
to engage integrin b1 and the downstream activation of focal
adhesion kinase (FAK) (Aguirre Ghiso et al., 1999; Barkan
et al., 2008; Shibue and Weinberg, 2009). The ability of DTCs
Figure 3. Dormancy Programs and Niches
(A) Carcinoma cells that have disseminated prior to the surgical removal of the primary tumor may persist in distant tissue environments as dormant disseminated
tumor cells (DTCs). Patients harboring such reservoirs of occult carcinoma cells are considered to have minimal residual disease and are at increased risk of
eventual metastatic recurrence. Although DTCs are most frequently examined in the bone, the delayed outgrowth of metastases in other organs suggests that
they, too, can harbor dormant DTCs.
(B) Dormant DTCs rely on unique biochemical signaling pathways that sustain their survival and impose programs of quiescence. Signals from the microenvironment, such as CXCL12, can activate SRC and AKT to promote DTC survival. Reduced integrin-mediated mitogenic signaling, coupled with the actions of
certain dormancy-inducing cytokines, enacts a quiescent program in DTCs that is associated with an ERKlow/p38high signaling state.
(C) DTCs may reside in dormant niches such as the hematopoietic stem cell niche (not shown) or the perivascular niche illustrated here. Thrombospondin-1
(TSP1), present in the basement membrane surrounding mature blood vessels, promotes dormancy. Dormant cells can evade detection by NK cells through the
repression of NK cell-activating ligands and are likely subject to surveillance by the adaptive immune system, which may keep cancer cells in a dormant state
through the actions of IFNg.
to productively interact with the matrix, at least in the context of
the lung, appears to be contingent upon the formation of filopodium-like protrusions (FLPs) that are coated with integrin b1 (Shibue et al., 2012, 2013). DTCs that are unable to sense or respond
to such adhesive signals fail to activate proliferative programs
that are primarily driven by FAK, SRC, and ERK signaling (Barkan
et al., 2010; Shibue et al., 2012). Accordingly, combined inhibition of both the SRC and ERK pathways blocks the escape of
DTCs from dormancy and thus prevents their subsequent
success in metastatic colonization (El Touny et al., 2014).
Several dormancy-inducing signals found in the microenvironment of certain target tissues have been identified as well.
For instance, TGF-b2, present in high concentrations in the bone
marrow and acting through stimulation of TGF-b-RI and TGFb-RIII displayed by DTCs, can impose a state of dormancy upon
head-and-neck squamous carcinoma cells (Bragado et al.,
2013). Members of the related BMP ligand family have also been
linked to metastatic dormancy. BMP7, which can be produced
by bone stromal cells, can induce dormancy in prostate cancer
cells (Kobayashi et al., 2011). In the lung, too, a number of alternative BMP ligands are expressed, including BMP4, and these have
been implicated as factors that maintain a state of dormancy in
disseminated mammary carcinoma cells (Gao et al., 2012).
Many of these dormancy-inducing cytokines lead to activation
of the p38 MAPK pathway; coupled with the absence of mitogenic
signals, this has the net effect of promoting an ERKlow/p38high
state in DTCs, which leads in turn to arrest in the G0/G1 phases
of the cell cycle and associated quiescence (Sosa et al., 2011).
The Dormant Niche
Dormant DTCs may reside in specialized niches (Figure 3C)
that support their survival, restrain their proliferation, and quite
possibly provide resistance to therapeutic agents (Ghajar,
2015). Of particular interest here is the idea that dormant DTCs
can co-opt a niche that is otherwise reserved for tissue-resident
stem cell populations. A compelling demonstration of this phenomenon is provided by the case of prostate cancer cells that
metastasize to the bone, where these carcinoma cells have
been found to compete with hematopoietic stem cells (HSCs)
for occupancy of sites in the endosteal niche; this occurs via
the CXCL12-CXCR4 signaling axis that is normally reserved for
the physiologic regulation of HSCs (Shiozawa et al., 2011). The
fact that DTCs can specifically target a stem-cell niche suggests
that they may be poised to respond to the quiescent and survival
signals present within the HSC microenvironment.
In multiple organs—including the lung, bone, and brain—DTCs
have been found to reside in the microenvironment surrounding
the vasculature, a region known as the perivascular niche (Ghajar, 2015). Whether this represents their active retention in this
niche or simply indicates an inability to move farther from the
vasculature after initial extravasation is unclear. An alternative
mechanism is suggested by the finding that factors present
in the perivascular niche have been demonstrated to actively
promote dormancy. Thus, thrombospondin-1, produced from
mature endothelial cells and deposited in the microvascular
basement membrane, is able to confine DTCs to residence in a
quiescent state (Ghajar et al., 2013). Moreover, in a study using
Cell 168, February 9, 2017 677
real-time imaging to examine the process of brain metastasis,
the rare solitary DTCs that achieved long-term dormancy were
invariably localized to the perivascular region (Kienast et al.,
2010), suggesting a critical role for this niche in sustaining
dormant DTCs in the brain as well.
DTCs must protect themselves from immune attack when
dwelling as isolated single cells lodged far from the confines
of the immunosuppressive primary tumor microenvironment.
Breast and lung carcinoma cells selected for their ability to
persist in a latent state after seeding of distant organ sites succeed in evading clearance by NK cells through the repression
of various NK cell-activating ligands, a program that appears
to be tightly coupled with entrance into a quiescent state (Malladi
et al., 2016). Indeed, these latency-competent cells have been
observed to grow out when injected into mice that lack NK cells,
indicating that the innate immune system is an important component of the dormant niche that effectively forces many cancer
cells into a quiescent state. A quite different process is suggested by the observation that antigen-presenting dendritic cells
can protect against metastasis (Headley et al., 2016), implying a
role of the adaptive immune system in controlling the growth of
metastatic deposits. Both CD4+ and CD8+ T cells have been
implicated in the control of dormant primary tumor cells through
the secretion of IFNg (Koebel et al., 2007; Müller-Hermelink et al.,
2008) and there is evidence that CD8+ T cells can hold disseminated uveal melanoma cells in a dormant state (Eyles et al.,
2010). However, at present very little is known about such immune-mediated dormancy mechanisms in the context of DTCs
originating from carcinomas.
Cancer Stem Cell Programs and the Initiation of
Metastatic Colonization
As mentioned above, activation of the EMT program, which is
capable of driving the physical dissemination of carcinoma cells
to distant anatomical sites, can also confer upon these cells
important stem cell traits (Mani et al., 2008; Morel et al., 2008)
that would appear to be highly relevant to metastatic colonization. Thus, an apparent prerequisite to the successful formation
of a metastatic colony is the property of tumor initiation as
embodied in CSCs. At least in principle, it is only those DTCs
that reside in the CSC state that are qualified to serve as the
founders of metastatic colonies.
Accumulating evidence, mostly from animal models, largely
supports this notion. In the MMTV-PyMT mammary tumor
model, a rare population of CSCs has been shown to be responsible for the initiation of metastatic growths in the lung and,
accordingly, the ability of these tumors to metastasize is dependent on the maintenance of this stem cell population through
enhanced Wnt signaling (Malanchi et al., 2011). In human breast
cancer cells, the activation of key stem cells pathways, such as
Wnt and Notch signaling, is also important for supporting their
colonization in xenograft mouse models (Oskarsson et al.,
2011). And mouse models of lung adenocarcinoma have
revealed that metastatic progression is associated with a dedifferentiation program, mediated by loss of Nkx2-1 expression,
which resembles programs operating in stem-like states (Li
et al., 2015; Winslow et al., 2011). Thus, it appears that the metastatic potential of a carcinoma is closely related to its ability to
dispatch populations of CSCs that can re-initiate tumor growth
678 Cell 168, February 9, 2017
after arrival at distant sites (Oskarsson et al., 2014). This notion
implies that cell state is a critical determinant of successful
metastasis, more specifically residence in the epigenetic state
associated with CSCs.
As discussed extensively above, an alternative to metastatic
outgrowth proceeding immediately after dissemination is the
entrance of DTCs into an indolent state in which they may
persist for extended periods of time before their progeny eventually erupt into readily detectable macroscopic metastases.
Such persistence may be favored by the acquisition of stem
cell characteristics. Thus, DTCs detected in the bone marrow
of breast cancer patients exhibit features of CSCs (Balic
et al., 2006). Consistent with this, cells that remain in a latent
state in distant tissues also show CSC attributes, including
expression of the SOX2 and SOX9 transcription factors (Malladi et al., 2016). In addition, single-cell expression analyses
have been applied to DTCs isolated from the organs of patient-derived xenograft (PDX) models of breast cancer; some
organs harbored low-burden metastatic disease due to the
presence of small numbers of ostensibly dormant carcinoma
cells (Lawson et al., 2015). These cells exhibited a distinctive
gene expression profile, relative to carcinoma cells from
advanced metastatic lesions, that was characterized by the
expression of EMT, stem cell, and survival/dormancy genes.
Most intriguingly, when neoplastic cells isolated from such
low-burden tissues were implanted into new recipient animals,
they retained their tumorigenic potential and could readily
generate more differentiated carcinomas (Lawson et al.,
2015). These studies provide further evidence in support of
the notion that stem-like cancer cells often serve as the founders of metastatic colonies, even when such colonies appear
only after great delay.
This scheme implicating the EMT and stem-cell programs as
critical prerequisites to the successful founding of metastatic
colonies must be reconciled with the commonly observed fact
that carcinoma metastases tend to recapitulate key histopathological traits of their corresponding primary tumors. Among other
traits, this usually includes significant epithelial features (Brabletz, 2012). On its surface, this notion this would seem incompatible with the proposition that EMT plays a central role in
launching carcinoma metastases through its ability to impart
mesenchymal and stem cell attributes to the disseminating cells.
In fact, this paradox is resolved by numerous studies, some cited
here in passing, that have found that the disseminated progeny
of carcinoma cells appear to undergo the reverse of the EMT
program at some point after dissemination, i.e., they pass
through a MET. This reversion to an epithelial state should
restore many of the cellular traits that were lost during the prior
passage through an EMT (Brabletz, 2012) and enable reconstruction of hierarchical cell organizations similar to those present in the initial primary tumors. Indeed, such reversals by
many cells within an early metastatic growth to a more epithelial
state may actually be essential for metastatic colonization (Del
Pozo Martin et al., 2015; Korpal et al., 2011; Ocaña et al.,
2012; Tsai et al., 2012). Of note, it remains unclear precisely
why highly mesenchymal CSCs cannot generate robustly
growing metastatic colonies in the absence of the epithelial
progeny generated by such METs.
Mechanisms of Colonization
Metastatic colonization appears, at least as presently understood,
to depend critically on two preconditions of the disseminated carcinomas cells: they must possess tumor-initiating ability, as
argued above, and they must in some fashion contrive adaptive
programs that enable them to thrive in the microenvironment
present in the parenchyma of distant tissues. The ‘‘seed and
soil’’ hypothesis, put forth by Paget in the late 19th century, suggested a complementary notion—essentially, that certain types
of carcinoma cells are more able to generate metastases in certain
foreign tissue microenvironments than are others (Fidler, 2003).
Unspoken by Paget was the notion that even in such favored metastatic sites, DTCs must still undergo some form of phenotypic
adaptation in order to proliferate robustly in those sites. Thus,
the proclivity of prostate and breast carcinomas to metastasize
to the bone would seem to imply some preexisting ability of the
corresponding DTCs to more readily assemble adaptive programs
suited to that tissue, whereas other less-favored tissue sites
might require more elaborate, less readily assembled adaptive
programs.
To be sure, in certain cases, the organ-specific tropism of metastatic cells is influenced by the design of the circulatory system.
Colorectal carcinoma (CRC) metastasis to the liver is strongly
favored simply because the portal vein draining the gut empties
directly into the liver (Gupta and Massagué, 2006). Hence, even if
disseminated CRC cells were intrinsically poorly adaptable for
liver colonization, the sheer numbers of these cells that are trapped in the liver after passage through the portal vein may, on its
own, pre-ordain metastases eventually arising at this site.
Importantly, the layout of the circulatory system explains only
a small proportion of the organ-specific metastases commonly
observed in the oncology clinic. Often cited in this context is
the proclivity of breast and prostate cancer cells, as mentioned
above, to colonize the bone marrow, usually termed osteotropic
metastasis. We highlight below specific examples that illustrate
the nature of the adaptive programs that seem critical to successful metastatic outgrowth.
To begin, we note that some of these programs may act generally by conferring a survival advantage in a number of distinct
target organs. For instance, cancer cells have been shown to
experience higher levels of oxidative stress both in the circulation
and in the parenchyma of a distant tissue (Piskounova et al.,
2015). As a consequence, metabolic adaptations, including the
synthesis of antioxidants, may promote the survival and eventual
metastatic outgrowth in diverse sites. Adhesive interactions that
substitute for those encountered in the primary tumor, such as
homotypic cell-cell interactions in the case of disseminating
CTC clusters (Aceto et al., 2014) or FLP-ECM interactions in
the case of single DTCs (Shibue et al., 2012), may be capable
of activating crucial survival pathways in a manner that could
be independent of specific target organs and would thus qualify
as more general adaptations promoting colonization.
These general adaptive programs may be nothing more than
preludes to the challenging tasks of contriving more narrowly
applicable, tissue-specific adaptations. Indeed, a diverse array
of organ-specific metastatic programs that mediate colonization
of the bone, lung, liver, and brain have been reported and studied
in mechanistic detail (Nguyen et al., 2009; Obenauf and Mas-
sague, 2015; Sethi and Kang, 2011). In the brain, for example,
cancer cells encounter reactive astrocytes that produce plasminogen activator, leading to the production of plasmin that
induces carcinoma cell death (Valiente et al., 2014). The ability
of carcinoma cells to survive in this hostile environment is
dependent upon the expression of serpins, which are typically
produced by neurons and protect against plasminogen activator-mediated cell death. In the lung, VCAM-1-expressing
carcinoma cells are able to activate their own AKT signaling
by physically engaging with integrin a4 on macrophages that
are particularly abundant in the pulmonary microenvironment
(Chen et al., 2011). The survival of carcinoma cells in the liver
has been linked to an ability to utilize creatine and ATP present
in the extracellular microenvironment to generate and import
phosphocreatine, which may confer a significant survival advantage on DTCs subject to metabolic stress (Loo et al., 2015). The
diversity of these survival mechanisms is a clear reflection of the
varied cellular and molecular determinants of successful colonization that operate within different target organs.
More generally, the mechanisms that permit and/or promote
the proliferation of various types of cancer cells in diverse distant
tissue microenvironments remain obscure. Arguably, the bestunderstood example to date involves the metastatic colonization
of the bone, which has been documented in the case of the
osteolytic metastases formed by breast cancers (Nguyen et al.,
2009; Obenauf and Massague, 2015; Weilbaecher et al.,
2011). Breast carcinoma cells produce a number of molecules,
including parathyroid hormone-related protein (PTHrP), IL-11,
and MMPs, that favor RANKL stimulation of osteoclast activity,
which in turn liberates growth factors from the bone matrix that
reciprocally promote tumor cell proliferation and the secretion
of even more factors that enhance osteoclast activity. The resulting self-reinforcing positive-feedback loop has been termed
the ‘‘vicious cycle’’ of osteolytic metastasis (Mundy, 2002). In
contrast, prostate carcinoma cells tend to spawn predominantly
osteoblastic metastases that occur as a result of induced osteoblast differentiation (Weilbaecher et al., 2011). Presumably, the
appearance of macroscopic metastases in other target organs
is similarly dependent on the ability of carcinoma cells to subvert
normal cell types residing within these organs, but the details of
these heterotypic interactions largely remain to be defined. In
one recent example, breast carcinoma cells that colonize the
brain have been found to benefit from communication with astrocytes through the assembly of gap junctions established between cancer cells and astrocytes (Chen et al., 2016).
The growth of a metastatic colony may also ensue when
dormant DTCs are awakened from their indolent state. The
awakening of previously dormant micrometastases may depend
on the successful assembly of functional adaptive programs,
which may be achieved only rarely per cell generation, explaining
the extraordinary low efficiency of metastasis formation. For
example, we note that dormant micrometastases in the
bone that somehow gain expression of VCAM-1 can transition
to an active colonization phase through the recruitment of osteoclast progenitor cells expressing integrin a4b1, a receptor for
VCAM-1, which enables bone resorption and initiation of the vicious cycle described above (Lu et al., 2011). Carcinoma cells in
the lung are able to escape dormancy through the production of
Cell 168, February 9, 2017 679
Coco, a secreted inhibitor of BMP signaling that promotes colonization (Gao et al., 2012). Unspoken here are the mechanisms
by which such adaptive programs are actually acquired. Thus,
it seems likely that continuous, low-level proliferation of the cells
within individual micrometastatic deposits—this occurring over
extended periods of time—is essential to the ability of DTCs to
stumble through trial and error on highly effective gene expression programs and adaptive behaviors that enable them to thrive
in the tissue microenvironment in which they happen to have
landed.
Programs that confer multi-organ colonization potential may
exist as well. Interestingly, the few examples of these programs that have been described center on interactions between DTCs and the ECM. For example, carcinoma cells
selected in vivo for their ability to re-initiate tumor growth in
subsequent xenotransplantation injections are also highly
competent in establishing metastatic growths in multiple
different organs (Ross et al., 2015). In this case, the capacity
for multi-organ colonization has been traced to the production
of the matrix protein laminin-a4 (LAMA4), which seems to be
critical for the initial proliferation of DTCs. Similarly, the
collagen receptor DDR1, in collaboration with the TM4SF1
adaptor protein, has recently been identified as a signaling
axis that regulates CSCs and thereby enables the outgrowth
of otherwise-dormant carcinoma cells in multiple organ sites
(Gao et al., 2016). The activation of such programs could
account for the apparently synchronous appearance of metastases in various organs—metastatic showers—that are occasionally observed in patients.
The Metastatic Microenvironment
The above discussions fail to address in any detail the nature of
the resident cells within various types of normal tissues that
sprout metastatic colonies. At least in the case of carcinomas,
these residents are essentially the various types of more mesenchymal cells that constitute the tissue-associated stroma
together with the ECM laid down by these cells. To begin, in
the same way that primary tumors are highly dependent on their
recruited stromal microenvironment, metastatic growths seem
equally reliant on stromal support (Hanahan and Coussens,
2012; Quail and Joyce, 2013; Wan et al., 2013). Indeed, the transition of carcinoma cells from a dormant state to one of robust
outgrowth may be provoked by changes in their local environment. For example, the apparent dormancy-inducing actions
of the perivascular niche noted above seem to be reversed during neo-vascularization as sprouting endothelial tip cells secrete
TGF-b1 and periostin (POSTN), which can break dormancy and
promote tumor cell proliferation (Ghajar et al., 2013). Consistent
with this idea, the outgrowth of dormant DTCs in the brain also
seems to be dependent on angiogenesis (Kienast et al., 2010).
Another recent report describes the outgrowth of previously
latent DTCs in the lungs being provoked by inflammation (as
mediated by pro-inflammatory cells) induced in this tissue (De
Cock et al., 2016).
Other findings suggest that metastatic colonization requires,
or at least can be aided by, a supportive ECM. This idea is
bolstered by the identification of specific ECM components,
such as tenascin C (TNC) (Oskarsson et al., 2011) and POSTN
(Malanchi et al., 2011), that drive colonization of the lung by
680 Cell 168, February 9, 2017
breast carcinoma cells. Tumor cells may themselves produce
these ECM components or, alternatively, they may evoke their
secretion by resident stromal fibroblasts. In addition, separate
but complementary lines of evidence have reported a connection between fibrosis and metastasis (Barkan et al., 2010; Cox
and Erler, 2014), suggesting that the local fibroblast and ECM
composition can influence the ability of carcinoma cells to colonize an organ. ECM stiffness (Levental et al., 2009; Mouw et al.,
2014), which can be modulated by the collagen-crosslinking
enzyme lysl oxidase (LOX), may also be important for the creation of pro-metastatic microenvironment (Erler et al., 2006,
2009). Indeed, the well-described contribution of hypoxia to
metastasis may be substantially related to the production of
LOX downstream of the transcription driven by hypoxia-inducible factor (Rankin and Giaccia, 2016).
Metastatic colonization is also likely to be impacted by cells of
both the innate and adaptive immune system (Kitamura et al.,
2015; Quail and Joyce, 2013). Thus, both NK cells and CD8+
T cells have been implicated in the suppression of metastasis
(Bidwell et al., 2012; Malladi et al., 2016). Conversely, the oxygen-rich environment in the lung acts to restrain T cell responses
and induces tolerance against innocuous antigens, but in the
context of cancer this actually provides a more hospitable environment for metastatic colonization (Clever et al., 2016). Myeloid
cells have also been identified as important contributors to the
formation of a favorable metastatic microenvironment (Kitamura
et al., 2015), where a unique population of metastasis-associated macrophages may be responsible for not only provoking
but also sustaining metastatic growth, perhaps by stimulation
of angiogenesis (Qian et al., 2009). Finally, acute inflammatory
responses have been found to trigger the outgrowth of carcinoma cells, an effect that may be primarily driven by neutrophils
(De Cock et al., 2016).
The establishment of a supportive metastatic environment
may occur prior to the arrival of any carcinoma cells, through
the formation of what has been termed a pre-metastatic niche.
This niche formation may involve the actions of VEGFR+ bone
marrow progenitors (Kaplan et al., 2005), myeloid-derived suppressor cells (MDSCs) (Psaila and Lyden, 2009), or neutrophils
(Wculek and Malanchi, 2015). Some have also reported that
tumor-derived exosomes—small tumor-derived vesicles that
contain DNA, mRNAs, microRNAs, and protein—can re-shape
the pre-metastatic environment in preparation for the arrival of
carcinoma cells (Costa-Silva et al., 2015; Peinado et al., 2012).
Thus, the formation of a pre-metastatic niche may represent
one consequence of far-ranging systemic effects induced by
primary tumors. More generally, the presence of a primary tumor
can lead to the production of numerous systemic signaling
factors that, by acting on distant tissues, can elicit responses
that may thereafter affect primary tumor growth, pre-metastatic
niches, and the outgrowth of previously latent micrometastases
(McAllister and Weinberg, 2014).
Genetic and Epigenetic Drivers of Colonization
The classic description of multi-step tumorigenesis implies that
the successive accumulation of genetic and/or epigenetic alterations drives primary tumor progression (Fearon and Vogelstein,
1990). A logical extension of this concept would suggest that the
outgrowth of a metastatic colony depends on the acquisition of
yet another somatic mutation or set of mutations that empower
cancer cells to disseminate and thereafter proliferate in a distant
organ. However, more than 25 years after the pioneering work on
multi-step progression of colorectal carcinoma (Fearon and
Vogelstein, 1990), no genetic mutations have been identified
that are characteristically associated with progression to metastatic disease. Indeed, even large-scale genomic sequencing efforts have yet to uncover recurrent genetic mutations that can
adequately explain the eruption of metastatic growths (Garraway
and Lander, 2013; Vogelstein et al., 2013). This suggests that the
development of metastasis is not contingent upon the accumulation of somatic driver mutations beyond those selected for
during primary tumor formation.
In particular, these findings have focused attention on non-genetic mechanisms enabling colonization. According to one idea,
colonization may depend on the amplification in metastatic cells
of oncogenic signaling pathways that were previously activated
in the cells of primary tumors (Vanharanta and Massagué, 2013),
for example, through the enrichment of existing clones with
elevated signaling through the MAP kinase pathway (Campbell
et al., 2010; Jacob et al., 2015). Metastatic carcinoma cells
may also need to evade the actions of metastasis suppressor
genes, which have been proposed to specifically block the
later stages of the invasion-metastasis cascade (Steeg, 2003).
Another alternative mechanism may involve defined epigenetic
alterations that drive colonization, such as aberrant DNA methylation patterns (Ozturk et al., 2016).
In addition to the actions of individual genes, recent data
suggest that metastatic carcinoma cells often exhibit global
changes in the structure of their chromatin. Thus, in a mouse
model of small cell lung cancer pathogenesis, carcinoma cells
competent for metastasis displayed a distinct open chromatin
configuration at distal regulatory regions, which were established and bound by the transcription factor Nfib; this change
in chromatin structure facilitated, in turn, a shift toward expression of a pro-metastatic neuronal gene expression program
(Denny et al., 2016). Such altered epigenetic states may ease
the adaptation of DTCs to foreign microenvironments. These advances notwithstanding, the difficulties involved in the procurement and analysis of metastatic samples have led to a continued
dearth of information concerning the genetic and epigenetic
landscapes found within the neoplastic cells that form human
metastases.
To summarize, in spite of the findings described above, metastatic colonization continues to represent the most puzzling
phase of malignant progression and the most challenging to
model experimentally. The physical dissemination of tumor cells
from the primary tumor into the parenchyma of distant tissues
can, at least in the context of many carcinomas, be largely
understood through the actions of a single cell-biological
program—the EMT. This contrasts starkly with the extraordinary
complexity of the last step of the invasion-metastasis cascade—
colonization. This complexity, highlighted by the apparently
myriad heterotypic interactions between populations of disseminated carcinoma cells and constituents of their newfound
homes in distant tissues, has complicated attempts at deriving
broadly applicable mechanistic principles underlying colonization. Nevertheless, we suggest that the weight of current evi-
dence points to three main prerequisites that must be met in order for metastatic colonization to succeed (Figure 4): (1) the
capacity to seed and maintain a population of tumor-initiating
cancer stem cells; (2) the ability to contrive adaptive, often organ-specific, colonization programs; and (3) the development
of a supportive microenvironmental niche.
Metastatic Evolution
The process of multi-step tumor progression and the subsequent seeding of metastases appears, at least superficially, to
operate as a linear path beginning in the primary tumor and
ending in macroscopic metastatic colonies. In truth, however,
each of the intervening steps is confounded by multiple factors,
many discussed above. Similarly, the processes that occur subsequent to the establishment of metastatic colonies and the
mechanisms by which they evolve have been a subject of
research and discussion over the past few decades. The notion
that tumor progression operates according to the Darwinian
model of evolutionary growth has become widely accepted
and influential in our thinking about metastatic progression
(Cairns, 1975; Nowell, 1976). Recent genomic studies have often
revealed close genetic relationships between primary tumors
and metastases in a variety of cancer types, implying that, at
least in certain cases, the cells forming a metastatic colony
derive from a dominant clonal subpopulation within the primary
tumor that managed to complete all of the steps required both
for primary tumor formation and the subsequent multi-step invasion-metastasis cascade (Naxerova and Jain, 2015). Implicit in
this depiction once again is the notion that the genetic alterations
required for completion of the invasion-metastasis cascade are
already present in the genomes of disseminating tumor cells
and that completion of this cascade depends only on non-genetic changes, specifically epigenetically organized programs
that complement the previously acquired genetic mutations.
Unanswered by such a scheme is the nature of the genetic and
epigenetic alterations that render neoplastic cells especially fit to
thrive within the context of the primary tumors and how such
alterations affect the proclivity of primary tumor cells to disseminate. Thus, it may be that phenotypic changes (of genetic and
epigenetic origin) that are selectively advantageous within the
context of primary tumor formation may, through happenstance,
also make primary carcinoma cells more capable of disseminating. If so, the resulting metastases may arise as incidental
side products of primary tumor progression. Alternatively,
many of the traits selected during primary tumor formation may
prove irrelevant to the success of metastasis formation.
Such logic forces consideration of the genetic and nongenetic factors operating within primary tumors that favor the
process of metastatic dissemination. To date, little attention
has been placed on these factors. As a specific mechanistic
example: what combination of epigenetic programs and somatic
mutations render a primary carcinoma cell especially responsive
to EMT-inducing heterotypic signals, enabling it to advance to a
state of high-grade malignancy? Among important non-genetic
factors may be the nature of the normal cells of origin and the differentiation programs that they bequeath to their neoplastic
progeny (Latil et al., 2016). At present, we possess relatively little
information on the fidelity with which preexisting differentiation
Cell 168, February 9, 2017 681
Figure 4. Prerequisites for Metastatic Colonization
The ability of carcinoma cells to outgrow as lethal metastases appears to be dependent on three essential conditions.
(A) The capacity to seed and maintain a population of cancer stem cells, which are competent to re-initiate tumor growth, appears to be an initial prerequisite for
metastatic growth. Dormant DTCs also exhibit key cancer stem cell attributes that probably contribute to their prolonged persistence in a quiescent state and
their ability to eventually spawn a metastatic colony.
(B) Although cancer stem cells are endowed with the potential to re-initiate tumor growth, the proliferative expansion to an overt metastatic colony is dependent
on the ability to contrive organ-specific colonization programs that allow these cells to thrive in a foreign tissue microenvironment. An array of organ-specific
metastatic programs has been described in the literature but there is also evidence for the existence of colonization programs that confer multi-organ metastatic
potential.
(C) During many stages of metastatic growth, cancer cells depend on interactions with their microenvironmental niche and cross talk with various stromal cells,
including endothelial cells, fibroblasts, and cells of the innate and adaptive immune system. The ECM is also an important component of the niche and can be
modified in ways that support metastatic colonization. In some cases the formation of a metastatic niche may actually precede the arrival of cancer cells, in what is
referred to as a pre-metastatic niche. Selected niche interactions discussed in the text are depicted here.
programs operating in cells of origin are transmitted in a cell-heritable fashion to the distant descendants of the founders of
neoplastic cell clones. Such programs could well represent the
dominant determinants of metastatic dissemination and may
explain why certain subtypes of human cancers disseminate
characteristically with predictable frequency to specific sites of
metastatic colony formation (Gupta et al., 2005; Ince et al.,
2007; Lim et al., 2009; Molyneux et al., 2010; Proia et al.,
2011). Unanswered by all of this is another question of great interest: is metastatic ability a trait that is selected for during multistep primary tumor evolution, or is it nothing more than an unselected, incidental consequence of primary tumor progression?
Dynamics of Tumor Progression and Metastasis
The development of metastasis has traditionally been considered as a relatively late event in multi-step tumor progression.
More recent reports, however, suggest that dissemination can
often occur early during the process of neoplastic transformation, perhaps even before departing cells are fully transformed
(Hüsemann et al., 2008; Podsypanina et al., 2008; Rhim et al.,
2012). At least in certain cases, this has been attributed to the
presence of pre-neoplastic cells residing within inflammatory
microenvironments that are able, via heterotypic signaling, to
activate EMT programs, resulting in expression of invasive
phenotypes (Rhim et al., 2012). Embedded in this thinking is
the notion that EMTs operate both in fully normal epithelial cells
682 Cell 168, February 9, 2017
and in neoplastic epithelial cells, suggesting that EMTs may also
function in all of the intermediate cell states that define the multistep progression of primary tumors.
Additionally, the kinetics of metastasis formation in certain
mouse models of breast cancer are in line with the idea that
dissemination, and hence metastasis, are early events during tumor progression (Weng et al., 2012).The mechanistic details of
this early dissemination program have recently been described
in murine models of HER2+ breast cancer, where in the early
stages of primary tumor formation a migratory and stem-like program predominates, before the well-established proliferative
pathways take hold during the later stages of tumor growth (Hosseini et al., 2016; Harper et al., 2016). Both studies suggest the
possibility that such early disseminated cells may subsequently
generate overt metastases. However, in other cases the actual
formation of distant metastases appears to be a late event, taking place many years or decades after initial neoplastic transformation (Yachida et al., 2010). Although physical dissemination
itself could be an early event, it may have little bearing on the
remaining steps of the cascade that result in the generation of
macrometastatic foci. Stated differently, it is unclear whether
early-disseminated carcinoma cells are ever able to evolve at
distant anatomical sites to states of high-grade malignancy
and spawn metastatic colonies, this situation representing the
‘‘parallel progression’’ model of metastasis formation.
Two general models of metastatic dissemination have been
proposed: the parallel progression model and the linear progression model. According to the latter, clones capable of spawning
metastases arise at the later stages of tumorigenesis with a small
degree of genetic divergence between those cells in the primary
tumor that actually spawned a metastasis and the cells in the
metastasis itself (Turajlic and Swanton, 2016). However, such
genetic divergence may, in real life, be very difficult to gauge,
given the clonal diversity that may have arisen within a primary
tumor (Gerlinger et al., 2012) and the fact that various genetically
distinct clonal subpopulations may be represented within the primary tumor in dramatically different sizes. Given the possibility
that a minor subpopulation within a primary tumor can serve as
the source of a metastasis (Haffner et al., 2013), how can one
know with any certainty that a sampling of the genomes of primary tumor cells has been able to detect and gauge the genome
of this minority population responsible for metastasis and its somatic mutations?
Yet another confounding factor when assessing the linear
progression of metastasis is the difference in time between
resection and sampling of the primary tumor and that of the
metastasis. In fact, a majority of studies have carried out comparisons between primary and secondary (metastatic) tissues
that were resected synchronously, while others have compared
metastases sampled up to 17 years after resection of the corresponding primary tumors; both have found genetic similarities
between the two tissues (Campbell et al., 2010; Ding et al.,
2010; Haffner et al., 2013; Liu et al., 2009). These studies favor
the linear progression model rather than the parallel progression
model, which posits that metastasis occurs as an early event
during tumorigenesis, after which the primary tumor and disseminated colonies evolve independently at sites far removed from
one another (Klein, 2009).
The parallel progression model, for its part, is encumbered
with its own complications. It assumes that the cells disseminating early from the primary tumor are able to proliferate sufficiently to allow for the acquisition of additional mutations that
would render them fully transformed and thus capable of forming
significant tumor masses. Given that metastatic colonization is a
highly inefficient process and given the complexity of essential
adaptive programs, it seems unlikely that disseminated preneoplastic cells will actually continuously proliferate after their arrival
in distant tissue microenvironments; in the absence of ongoing
proliferation, it seems implausible that such cells can acquire,
via stochastically occurring mutations, the complex repertoire
of mutant alleles that are needed, in aggregate, for continuous
growth and clonal expansion. Resolving between these models
of metastatic progression may be further complicated by the
fact that metastases have been reported to result from polyclonal populations ostensibly derived from CTC clusters
(Cheung et al., 2016) and by the observation that metastatic
clones may be transferred between different metastatic lesions
in the same patient (Gundem et al., 2015).
Treatment and Resistance
Metastatic cancer most often represents a terminal illness and
patients eventually succumb to the disease or from complications that result from their course of treatment, indicating the
current dearth of effective therapies (Steeg, 2016). Moreover, it
remains unclear precisely whether the cells within metastases are intrinsically more resistant to therapy or whether they
respond to therapies at rates comparable to the cells in their
corresponding primary tumors. Comparable rates of responsiveness of metastases would certainly be compatible with the
known genetic similarities between primary tumors and their
derived metastases. Any heightened resistance might be explained by the fact that metastases derive from especially
aggressive subpopulations of cells that resided within primary
tumors or, alternatively, from further evolution to higher grades
of malignancy after dissemination to distant sites.
Treatment of Primary Tumors and Metastatic Growths
Current therapeutic strategies for eliminating metastases are
essentially the same as those directed at the corresponding
primary tumors, the exception being surgery, which is infrequently employed to remove metastatic deposits. While cells
that have succeeded in colonizing distant tissue microenvironments have often and perhaps always evolved adaptive
programs that enable their robust proliferation at these secondary sites, it remains unclear whether this additional evolution,
much of it achieved through epigenetic reprogramming, confers
elevated therapeutic resistance. The alternative is that successful colonization of distant sites depends on the acquisition of
adaptive traits that ultimately have no direct effect on therapeutic
resistance.
The fact that metastatic lesions represent the progeny of minority subpopulations of the neoplastic cells present in a primary
tumor (Ding et al., 2010; Yachida et al., 2010; Yates et al., 2015)
suggests that metastatic colonies could be quite different from
the primary tumor in terms of their clonal architecture and
biology. And while numerous studies have examined the genetic
and phenotypic diversity of the neoplastic cells that compose
primary tumors (Marusyk et al., 2012), the level of genetic and
epigenetic heterogeneity and phenotypic plasticity that operates
in metastatic growths is still in question.
A formidable obstacle to treating the minimal residual disease
(MRD) that may remain after initial chemo- or radiotherapy derives from the fact that dormant carcinoma cells appear to
perpetuate this disease and form the precursors of eventual
metastatic relapses. Unfortunately, almost all currently deployed cytotoxic therapies preferentially kill proliferating cells
rather than those that have exited the active cell cycle,
rendering dormant cells intrinsically more resistant to almost
all currently available therapies (Ghajar, 2015; Goss and Chambers, 2010). This stark contrast in the behavior of these dormant
DTCs and the actively cycling cells of the primary tumor may ultimately prove to be far more critical in determining susceptibility to therapeutic elimination than any genetic or epigenetic differences distinguishing MRD from the corresponding primary
tumors. Further complicating the development of novel agents
directed at dormant metastatic deposits is the fact that, for
various types of cancer, true efficacy can be judged only after
extremely long follow-up periods when the much-feared relapses may appear (Steeg, 2016). Still, preventative adjuvant
therapies directed at dormant DTCs and their ability to spawn
clinical relapses arguably offer the best opportunity to prevent
these outcomes.
Cell 168, February 9, 2017 683
Therapeutic Resistance
The mechanisms of therapeutic resistance acquired by metastatic growths may closely parallel those operating within corresponding tumors. In the context of targeted adjuvant therapy,
drug-resistant clones may emerge in primary tumors and metastatic lesions, as has been observed, for example, in ER+ breast
cancer patients receiving hormone therapy (Alluri et al., 2014)
and patients with EGFR mutant non-small-cell lung cancer
treated with targeted kinase inhibitors (Gazdar, 2009). Of special
interest to the present discussion are resistance mechanisms
that are particular to the sites of dissemination. One possibility
is that the metastatic microenvironment favors the induction of
biological programs that confer drug resistance. For example,
it has been reported that CXCL1/2, which actively supports the
establishment of metastases through the recruitment of myeloid
cells, can also mediate resistance to chemotherapy (Acharyya
et al., 2012), providing support for the idea that certain traits
involved in metastatic dissemination may also contribute to therapeutic resistance.
More generally, the effect of chemotherapy on either primary
or metastatic growths may elicit the secretion of various paracrine mediators from the surrounding stromal cells that can promote resistance, including CXCL1/2 (Acharyya et al., 2012), IL-6
and Timp1 (Gilbert and Hemann, 2010), WNT16B (Sun et al.,
2012), and HGF (Straussman et al., 2012). Resistance to targeted
kinase inhibitors can also be conferred by a host of secreted
factors that are produced by carcinoma cells after exposure to
a drug (Lee et al., 2014; Obenauf et al., 2015). Although such
effects may indeed promote the emergence of drug-resistant
cell clones within primary tumors, they may operate even more
strongly in sites of metastasis.
According to an alternative view, the cells forming metastases
are intrinsically no more or less resistant to therapies than their
counterparts in primary tumors. Hence, if drug-resistant ancestral metastatic clones were present in the original neoplasm,
then such cells would render this tumor as well as its derived metastases equally resistant to therapy. Following such thinking, a
major benefit of surgically eliminating primary tumors derives
from reducing the sheer number and diversity of neoplastic cells,
thereby increasing the chance that any therapy-resistant variant
clones are removed from the body of a patient. In the context of
metastatic disease this is, it seems, often not possible.
Conclusion: Principles and Outlook
As the preceding discussions have indicated, significant
progress has been made over the past decade in elucidating
the cellular and molecular programs that drive cancer metastasis. Although our understanding of metastasis remains quite
incomplete, we see a number of common biological principles
beginning to emerge. Thus, we suggest that one can take stock
of the information that is currently at hand and conclude that:
1. Metastasis occurs mainly through a sequential, multi-step
process that can be conceptualized as the invasionmetastasis cascade.
2. In the case of carcinomas, the EMT program enables primary tumor cells to accomplish most if not all of the steps
684 Cell 168, February 9, 2017
involved in the physical dissemination of tumor cells to a
distant site.
3. The fate of disseminating carcinoma cells is strongly influenced by interactions that they experience during transit
through the circulatory system.
4. Disseminated carcinoma cells must escape clearance by
the arms of the immune system and subvert the cellular
programs that impose a state of dormancy.
5. The process of active metastatic colonization is contingent upon the dissemination of cancer stem cells that
can re-initiate tumor growth; the ability of their progeny
to assemble adaptive, organ-specific colonization programs; and the establishment of a microenvironment
conducive to metastasis.
The processes that enable the physical translocation of cancer
cells from primary tumors to the parenchyma of distant tissues
are within sight and relatively small in number; in contrast, the
adaptive programs allowing cancer cells arising from diverse primary tumors to thrive in various tissue microenvironments may
be large in number and not readily reducible to a common set
of underlying mechanistic principles.
While these principles articulate general concepts, a number
of key mechanistic details related to these ideas remain to be established. For example, we are beginning to appreciate that the
EMT program is capable of generating a wide spectrum of carcinoma cells with various complements of mesenchymal traits, but
there is little information on the functional role of these different
phenotypic states in the metastatic process. Yet other critical
questions about metastasis fall outside the bounds of the points
outlined above. For one, it is not yet clear what specific factors
determine the efficiency of clinical metastatic disease and why
some patients present with metastatic cancer, while in other patients many years may lapse before the disease advances to this
stage. The literature holds some provocative hints that could
account for this variability (Figure 5), such as different cells of
origin whose differentiation programs strongly predispose to
an aggressive malignancy or to the dissemination of CTC clusters that may more readily establish a metastatic colony. Additionally, the fact that many patients experience metastatic
spread to multiple organs suggests the existence of more universal, multi-organ metastatic programs, but the extent to which
such programs operate is unclear and their biological details
have just begun to be described. Finally, the clinical and biological impact of various immunotherapies, particularly checkpoint
inhibitors (Sharma and Allison, 2015), on metastases is certain
to be a continued area of active research, even offering the
hope of seeking out and eliminating metastatic deposits.
Perhaps most pressing is a better understanding of the biological similarities and differences between primary tumors and
their metastatic descendants, especially in regard to the extent
of heterogeneity, plasticity, and resistance that they exhibit. We
believe that an accurate comparison of the principles that
govern primary tumor growth with those that govern the
dissemination and outgrowth of metastases will be essential
in order to enable the development of new approaches and
therapies that are specifically designed to prevent or treat metastatic disease.
Figure 5. Dynamics of Metastatic Evolution
The progression and evolution of metastatic disease is highly variable, manifesting in ways that must affect the kinetics of metastatic colonization. Five
hypothetical alternatives are presented here.
(A) The dissemination of CTC clusters to distant sites may generate overt metastases with a relatively short latency, since such clusters are highly efficient at
spawning metastatic growths. Their efficiency in forming metastases may derive from advantages during transit in the circulation or because they benefit from
homotypic cell-cell interactions in a foreign tissue environment.
(B) Solitary disseminated carcinoma cells that are adept at recruiting and establishing a supportive metastatic niche, or that are able to generate a microenvironmental niche themselves, may be better able to survive and initiate programs of proliferation.
(C) While the dissemination of tumor-initiating cancer stem cells may be a prerequisite for metastasis, the generation and evolution of progeny that are well
adapted to the local microenvironment could take many months or years.
(D) At later stages of metastatic progression, other dynamics come into play, such as the exchange of metastatic cell clones between different metastatic lesions
in the same patient. The biological and clinical impact of such transfer, however, remains to be firmly established.
(E) Tumor cells may disseminate during the early stages of tumorigenesis and even from pre-malignant lesions, but it remains unclear how such cells are able to
evolve, in parallel with the primary tumor, the full complement of genetic mutations and malignant traits required for successful metastatic colonization.
ACKNOWLEDGMENTS
We would like to thank all members of the R.A.W. laboratory for fruitful discussions and especially Tsukasa Shibue for critical review of the manuscript.
We would also like to thank Meredith Leffler for preparation of the figures.
A.W.L. is supported by an American Cancer Society – New England Division – Ellison Foundation Postdoctoral Fellowship (PF-15-131-01-CSM).
D.R.P. was supported by a C.J. Martin Overseas Biomedical Fellowship
from the National Health and Medical Research Council of Australia
(NHMRC APP1071853) and is currently supported by a K99/R00 Pathway
to Independence Award (NIH/NCI 1K99CA201574-01A1). Work in the
R.A.W. laboratory is supported by grants from the NIH (R01-CA078461),
the Breast Cancer Research Foundation, the Advanced Medical Research
Foundation, and the Ludwig Center for Molecular Oncology. R.A.W. is an
American Cancer Society Research Professor and a Daniel K. Ludwig Cancer Research Professor.
Cell 168, February 9, 2017 685
REFERENCES
Aceto, N., Bardia, A., Miyamoto, D.T., Donaldson, M.C., Wittner, B.S.,
Spencer, J.A., Yu, M., Pely, A., Engstrom, A., Zhu, H., et al. (2014). Circulating
tumor cell clusters are oligoclonal precursors of breast cancer metastasis. Cell
158, 1110–1122.
Aceto, N., Toner, M., Maheswaran, S., and Haber, D.A. (2015). En Route to
Metastasis: Circulating Tumor Cell Clusters and Epithelial-to-Mesenchymal
Transition. Trends Cancer 1, 44–52.
Acharyya, S., Oskarsson, T., Vanharanta, S., Malladi, S., Kim, J., Morris, P.G.,
Manova-Todorova, K., Leversha, M., Hogg, N., Seshan, V.E., et al. (2012). A
CXCL1 paracrine network links cancer chemoresistance and metastasis.
Cell 150, 165–178.
Aguirre-Ghiso, J.A. (2007). Models, mechanisms and clinical evidence for
cancer dormancy. Nat. Rev. Cancer 7, 834–846.
Aguirre Ghiso, J.A., Kovalski, K., and Ossowski, L. (1999). Tumor dormancy
induced by downregulation of urokinase receptor in human carcinoma involves integrin and MAPK signaling. J. Cell Biol. 147, 89–104.
Al-Mehdi, A.B., Tozawa, K., Fisher, A.B., Shientag, L., Lee, A., and Muschel,
R.J. (2000). Intravascular origin of metastasis from the proliferation of endothelium-attached tumor cells: a new model for metastasis. Nat. Med. 6, 100–102.
Alluri, P.G., Speers, C., and Chinnaiyan, A.M. (2014). Estrogen receptor mutations and their role in breast cancer progression. Breast Cancer Res. 16, 494.
Au, S.H., Storey, B.D., Moore, J.C., Tang, Q., Chen, Y.L., Javaid, S., Sarioglu,
A.F., Sullivan, R., Madden, M.W., O’Keefe, R., et al. (2016). Clusters of circulating tumor cells traverse capillary-sized vessels. Proc. Natl. Acad. Sci. USA
113, 4947–4952.
Baccelli, I., Schneeweiss, A., Riethdorf, S., Stenzinger, A., Schillert, A., Vogel,
V., Klein, C., Saini, M., Bäuerle, T., Wallwiener, M., et al. (2013). Identification of
a population of blood circulating tumor cells from breast cancer patients that
initiates metastasis in a xenograft assay. Nat. Biotechnol. 31, 539–544.
Balic, M., Lin, H., Young, L., Hawes, D., Giuliano, A., McNamara, G., Datar,
R.H., and Cote, R.J. (2006). Most early disseminated cancer cells detected
in bone marrow of breast cancer patients have a putative breast cancer
stem cell phenotype. Clin. Cancer Res. 12, 5615–5621.
Barkan, D., Kleinman, H., Simmons, J.L., Asmussen, H., Kamaraju, A.K., Hoenorhoff, M.J., Liu, Z.Y., Costes, S.V., Cho, E.H., Lockett, S., et al. (2008). Inhibition of metastatic outgrowth from single dormant tumor cells by targeting the
cytoskeleton. Cancer Res. 68, 6241–6250.
Barkan, D., El Touny, L.H., Michalowski, A.M., Smith, J.A., Chu, I., Davis, A.S.,
Webster, J.D., Hoover, S., Simpson, R.M., Gauldie, J., and Green, J.E. (2010).
Metastatic growth from dormant cells induced by a col-I-enriched fibrotic environment. Cancer Res. 70, 5706–5716.
Bednarz-Knoll, N., Alix-Panabières, C., and Pantel, K. (2012). Plasticity of
disseminating cancer cells in patients with epithelial malignancies. Cancer
Metastasis Rev. 31, 673–687.
Bidwell, B.N., Slaney, C.Y., Withana, N.P., Forster, S., Cao, Y., Loi, S.,
Andrews, D., Mikeska, T., Mangan, N.E., Samarajiwa, S.A., et al. (2012).
Silencing of Irf7 pathways in breast cancer cells promotes bone metastasis
through immune escape. Nat. Med. 18, 1224–1231.
Brabletz, T., Jung, A., Spaderna, S., Hlubek, F., and Kirchner, T. (2005).
Opinion: migrating cancer stem cells - an integrated concept of malignant
tumour progression. Nat. Rev. Cancer 5, 744–749.
Bragado, P., Estrada, Y., Parikh, F., Krause, S., Capobianco, C., Farina, H.G.,
Schewe, D.M., and Aguirre-Ghiso, J.A. (2013). TGF-b2 dictates disseminated
tumour cell fate in target organs through TGF-b-RIII and p38a/b signalling. Nat.
Cell Biol. 15, 1351–1361.
Braun, S., Vogl, F.D., Naume, B., Janni, W., Osborne, M.P., Coombes, R.C.,
Schlimok, G., Diel, I.J., Gerber, B., Gebauer, G., et al. (2005). A pooled analysis of bone marrow micrometastasis in breast cancer. N. Engl. J. Med. 353,
793–802.
Cairns, J. (1975). Mutation selection and the natural history of cancer. Nature
255, 197–200.
Campbell, P.J., Yachida, S., Mudie, L.J., Stephens, P.J., Pleasance, E.D.,
Stebbings, L.A., Morsberger, L.A., Latimer, C., McLaren, S., Lin, M.L., et al.
(2010). The patterns and dynamics of genomic instability in metastatic pancreatic cancer. Nature 467, 1109–1113.
Celià-Terrassa, T., Meca-Cortés, O., Mateo, F., Martı́nez de Paz, A., Rubio, N.,
Arnal-Estapé, A., Ell, B.J., Bermudo, R., Dı́az, A., Guerra-Rebollo, M., et al.
(2012). Epithelial-mesenchymal transition can suppress major attributes of
human epithelial tumor-initiating cells. J. Clin. Invest. 122, 1849–1868.
Chambers, A.F., Groom, A.C., and MacDonald, I.C. (2002). Dissemination and
growth of cancer cells in metastatic sites. Nat. Rev. Cancer 2, 563–572.
Chen, Q., Zhang, X.H., and Massagué, J. (2011). Macrophage binding to receptor VCAM-1 transmits survival signals in breast cancer cells that invade
the lungs. Cancer Cell 20, 538–549.
Chen, Q., Boire, A., Jin, X., Valiente, M., Er, E.E., Lopez-Soto, A., Jacob, L.S.,
Patwa, R., Shah, H., Xu, K., et al. (2016). Carcinoma-astrocyte gap junctions
promote brain metastasis by cGAMP transfer. Nature 533, 493–498.
Cheung, K.J., Gabrielson, E., Werb, Z., and Ewald, A.J. (2013). Collective invasion in breast cancer requires a conserved basal epithelial program. Cell 155,
1639–1651.
Cheung, K.J., Padmanaban, V., Silvestri, V., Schipper, K., Cohen, J.D., Fairchild, A.N., Gorin, M.A., Verdone, J.E., Pienta, K.J., Bader, J.S., and Ewald,
A.J. (2016). Polyclonal breast cancer metastases arise from collective dissemination of keratin 14-expressing tumor cell clusters. Proc. Natl. Acad. Sci. USA
113, E854–E863.
Chung, Y.C., Wei, W.C., Hung, C.N., Kuo, J.F., Hsu, C.P., Chang, K.J., and
Chao, W.T. (2016). Rab11 collaborates E-cadherin to promote collective cell
migration and indicates a poor prognosis in colorectal carcinoma. Eur. J.
Clin. Invest. 46, 1002–1011.
Clever, D., Roychoudhuri, R., Constantinides, M.G., Askenase, M.H., Sukumar, M., Klebanoff, C.A., Eil, R.L., Hickman, H.D., Yu, Z., Pan, J.H., et al.
(2016). Oxygen sensing by T cells establishes an immunologically tolerant
metastatic niche. Cell 166, 1117–1131.
Coffelt, S.B., Kersten, K., Doornebal, C.W., Weiden, J., Vrijland, K., Hau, C.S.,
Verstegen, N.J., Ciampricotti, M., Hawinkels, L.J., Jonkers, J., and de Visser,
K.E. (2015). IL-17-producing gd T cells and neutrophils conspire to promote
breast cancer metastasis. Nature 522, 345–348.
Bonapace, L., Coissieux, M.M., Wyckoff, J., Mertz, K.D., Varga, Z., Junt, T.,
and Bentires-Alj, M. (2014). Cessation of CCL2 inhibition accelerates breast
cancer metastasis by promoting angiogenesis. Nature 515, 130–133.
Coffelt, S.B., Wellenstein, M.D., and de Visser, K.E. (2016). Neutrophils in cancer: neutral no more. Nat. Rev. Cancer 16, 431–446.
Bonnomet, A., Syne, L., Brysse, A., Feyereisen, E., Thompson, E.W., Noël, A.,
Foidart, J.M., Birembaut, P., Polette, M., and Gilles, C. (2012). A dynamic
in vivo model of epithelial-to-mesenchymal transitions in circulating tumor
cells and metastases of breast cancer. Oncogene 31, 3741–3753.
Cools-Lartigue, J., Spicer, J., McDonald, B., Gowing, S., Chow, S., Giannias,
B., Bourdeau, F., Kubes, P., and Ferri, L. (2013). Neutrophil extracellular traps
sequester circulating tumor cells and promote metastasis. J. Clin. Invest. 123,
3446–3458.
Bos, P.D., Zhang, X.H., Nadal, C., Shu, W., Gomis, R.R., Nguyen, D.X., Minn,
A.J., van de Vijver, M.J., Gerald, W.L., Foekens, J.A., and Massagué, J. (2009).
Genes that mediate breast cancer metastasis to the brain. Nature 459, 1005–
1009.
Costa-Silva, B., Aiello, N.M., Ocean, A.J., Singh, S., Zhang, H., Thakur, B.K.,
Becker, A., Hoshino, A., Mark, M.T., Molina, H., et al. (2015). Pancreatic cancer
exosomes initiate pre-metastatic niche formation in the liver. Nat. Cell Biol. 17,
816–826.
Brabletz, T. (2012). To differentiate or not–routes towards metastasis. Nat.
Rev. Cancer 12, 425–436.
Cox, T.R., and Erler, J.T. (2014). Molecular pathways: connecting fibrosis and
solid tumor metastasis. Clin. Cancer Res. 20, 3637–3643.
686 Cell 168, February 9, 2017
de Boer, M., van Dijck, J.A., Bult, P., Borm, G.F., and Tjan-Heijnen, V.C. (2010).
Breast cancer prognosis and occult lymph node metastases, isolated tumor
cells, and micrometastases. J. Natl. Cancer Inst. 102, 410–425.
De Cock, J.M., Shibue, T., Dongre, A., Keckesova, Z., Reinhardt, F., and Weinberg, R.A. (2016). Inflammation triggers Zeb1-dependent escape from tumor
latency. Cancer Res. 76, 6778–6784.
Gao, H., Chakraborty, G., Zhang, Z., Akalay, I., Gadiya, M., Gao, Y., Sinha, S.,
Hu, J., Jiang, C., Akram, M., et al. (2016). Multi-organ site metastatic reactivation mediated by non-canonical discoidin domain receptor 1 signaling. Cell
166, 47–62.
Garraway, L.A., and Lander, E.S. (2013). Lessons from the cancer genome.
Cell 153, 17–37.
De Craene, B., and Berx, G. (2013). Regulatory networks defining EMT during
cancer initiation and progression. Nat. Rev. Cancer 13, 97–110.
Gasic, G.J., Gasic, T.B., and Stewart, C.C. (1968). Antimetastatic effects associated with platelet reduction. Proc. Natl. Acad. Sci. USA 61, 46–52.
Del Pozo Martin, Y., Park, D., Ramachandran, A., Ombrato, L., Calvo, F., Chakravarty, P., Spencer-Dene, B., Derzsi, S., Hill, C.S., Sahai, E., and Malanchi, I.
(2015). Mesenchymal cancer cell-stroma crosstalk promotes niche activation,
epithelial reversion, and metastatic colonization. Cell Rep. 13, 2456–2469.
Gay, L.J., and Felding-Habermann, B. (2011). Contribution of platelets to
tumour metastasis. Nat. Rev. Cancer 11, 123–134.
Denny, S.K., Yang, D., Chuang, C.H., Brady, J.J., Lim, J.S., Grüner, B.M.,
Chiou, S.H., Schep, A.N., Baral, J., Hamard, C., et al. (2016). Nfib promotes
metastasis through a widespread increase in chromatin accessibility. Cell
166, 328–342.
Ding, L., Ellis, M.J., Li, S., Larson, D.E., Chen, K., Wallis, J.W., Harris, C.C.,
McLellan, M.D., Fulton, R.S., Fulton, L.L., et al. (2010). Genome remodelling
in a basal-like breast cancer metastasis and xenograft. Nature 464, 999–1005.
Douma, S., Van Laar, T., Zevenhoven, J., Meuwissen, R., Van Garderen, E.,
and Peeper, D.S. (2004). Suppression of anoikis and induction of metastasis
by the neurotrophic receptor TrkB. Nature 430, 1034–1039.
El Touny, L.H., Vieira, A., Mendoza, A., Khanna, C., Hoenerhoff, M.J., and
Green, J.E. (2014). Combined SFK/MEK inhibition prevents metastatic
outgrowth of dormant tumor cells. J. Clin. Invest. 124, 156–168.
Erler, J.T., Bennewith, K.L., Nicolau, M., Dornhöfer, N., Kong, C., Le, Q.T., Chi,
J.T., Jeffrey, S.S., and Giaccia, A.J. (2006). Lysyl oxidase is essential for
hypoxia-induced metastasis. Nature 440, 1222–1226.
Erler, J.T., Bennewith, K.L., Cox, T.R., Lang, G., Bird, D., Koong, A., Le, Q.T.,
and Giaccia, A.J. (2009). Hypoxia-induced lysyl oxidase is a critical mediator of
bone marrow cell recruitment to form the premetastatic niche. Cancer Cell 15,
35–44.
Eyles, J., Puaux, A.-L., Wang, X., Toh, B., Prakash, C., Hong, M., Tan, T.G.,
Zheng, L., Ong, L.C., Jin, Y., et al. (2010). Tumor cells disseminate early, but
immunosurveillance limits metastatic outgrowth, in a mouse model of melanoma. J. Clin. Invest. 120, 2030–2039.
Fan, F., Samuel, S., Evans, K.W., Lu, J., Xia, L., Zhou, Y., Sceusi, E., Tozzi, F.,
Ye, X.C., Mani, S.A., and Ellis, L.M. (2012). Overexpression of snail induces
epithelial-mesenchymal transition and a cancer stem cell-like phenotype in
human colorectal cancer cells. Cancer Med. 1, 5–16.
Fearon, E.R., and Vogelstein, B. (1990). A genetic model for colorectal tumorigenesis. Cell 61, 759–767.
Fidler, I.J. (2003). The pathogenesis of cancer metastasis: the ‘seed and soil’
hypothesis revisited. Nat. Rev. Cancer 3, 453–458.
Fischer, K.R., Durrans, A., Lee, S., Sheng, J., Li, F., Wong, S.T., Choi, H., El
Rayes, T., Ryu, S., Troeger, J., et al. (2015). Epithelial-to-mesenchymal transition is not required for lung metastasis but contributes to chemoresistance.
Nature 527, 472–476.
Franco, A.T., Corken, A., and Ware, J. (2015). Platelets at the interface of
thrombosis, inflammation, and cancer. Blood 126, 582–588.
Fridlender, Z.G., Sun, J., Kim, S., Kapoor, V., Cheng, G., Ling, L., Worthen,
G.S., and Albelda, S.M. (2009). Polarization of tumor-associated neutrophil
phenotype by TGF-beta: ‘‘N1’’ versus ‘‘N2’’ TAN. Cancer Cell 16, 183–194.
Friedl, P., Locker, J., Sahai, E., and Segall, J.E. (2012). Classifying collective
cancer cell invasion. Nat. Cell Biol. 14, 777–783.
Gaggioli, C., Hooper, S., Hidalgo-Carcedo, C., Grosse, R., Marshall, J.F., Harrington, K., and Sahai, E. (2007). Fibroblast-led collective invasion of carcinoma cells with differing roles for RhoGTPases in leading and following cells.
Nat. Cell Biol. 9, 1392–1400.
Gao, H., Chakraborty, G., Lee-Lim, A.P., Mo, Q., Decker, M., Vonica, A.,
Shen, R., Brogi, E., Brivanlou, A.H., and Giancotti, F.G. (2012). The BMP inhibitor Coco reactivates breast cancer cells at lung metastatic sites. Cell 150,
764–779.
Gazdar, A.F. (2009). Activating and resistance mutations of EGFR in nonsmall-cell lung cancer: role in clinical response to EGFR tyrosine kinase inhibitors. Oncogene 28(Suppl 1 ), S24–S31.
Gerlinger, M., Rowan, A.J., Horswell, S., Larkin, J., Endesfelder, D., Gronroos,
E., Martinez, P., Matthews, N., Stewart, A., Tarpey, P., et al. (2012). Intratumor
heterogeneity and branched evolution revealed by multiregion sequencing.
N. Engl. J. Med. 366, 883–892.
Ghajar, C.M. (2015). Metastasis prevention by targeting the dormant niche.
Nat. Rev. Cancer 15, 238–247.
Ghajar, C.M., Peinado, H., Mori, H., Matei, I.R., Evason, K.J., Brazier, H.,
Almeida, D., Koller, A., Hajjar, K.A., Stainier, D.Y., et al. (2013). The perivascular
niche regulates breast tumour dormancy. Nat. Cell Biol. 15, 807–817.
Giancotti, F.G. (2013). Mechanisms governing metastatic dormancy and reactivation. Cell 155, 750–764.
Gilbert, L.A., and Hemann, M.T. (2010). DNA damage-mediated induction of a
chemoresistant niche. Cell 143, 355–366.
Goss, P.E., and Chambers, A.F. (2010). Does tumour dormancy offer a therapeutic target? Nat. Rev. Cancer 10, 871–877.
Granot, Z., Henke, E., Comen, E.A., King, T.A., Norton, L., and Benezra, R.
(2011). Tumor entrained neutrophils inhibit seeding in the premetastatic
lung. Cancer Cell 20, 300–314.
Grosse-Wilde, A., Fouquier d’Hérouël, A., McIntosh, E., Ertaylan, G., Skupin,
A., Kuestner, R.E., del Sol, A., Walters, K.A., and Huang, S. (2015). Stemness
of the hybrid epithelial/mesenchymal state in breast cancer and its association
with poor survival. PLoS ONE 10, e0126522.
Gundem, G., Van Loo, P., Kremeyer, B., Alexandrov, L.B., Tubio, J.M.,
Papaemmanuil, E., Brewer, D.S., Kallio, H.M., Högnäs, G., Annala, M., et al.;
ICGC Prostate UK Group (2015). The evolutionary history of lethal metastatic
prostate cancer. Nature 520, 353–357.
Gupta, G.P., and Massagué, J. (2006). Cancer metastasis: building a framework. Cell 127, 679–695.
Gupta, P.B., Kuperwasser, C., Brunet, J.P., Ramaswamy, S., Kuo, W.L., Gray,
J.W., Naber, S.P., and Weinberg, R.A. (2005). The melanocyte differentiation
program predisposes to metastasis after neoplastic transformation. Nat.
Genet. 37, 1047–1054.
Gupta, G.P., Nguyen, D.X., Chiang, A.C., Bos, P.D., Kim, J.Y., Nadal, C.,
Gomis, R.R., Manova-Todorova, K., and Massagué, J. (2007). Mediators of
vascular remodelling co-opted for sequential steps in lung metastasis. Nature
446, 765–770.
Gupta, P.B., Onder, T.T., Jiang, G., Tao, K., Kuperwasser, C., Weinberg, R.A.,
and Lander, E.S. (2009). Identification of selective inhibitors of cancer stem
cells by high-throughput screening. Cell 138, 645–659.
Haffner, M.C., Mosbruger, T., Esopi, D.M., Fedor, H., Heaphy, C.M., Walker,
D.A., Adejola, N., Gürel, M., Hicks, J., Meeker, A.K., et al. (2013). Tracking
the clonal origin of lethal prostate cancer. J. Clin. Invest. 123, 4918–4922.
Hanahan, D., and Coussens, L.M. (2012). Accessories to the crime: functions
of cells recruited to the tumor microenvironment. Cancer Cell 21, 309–322.
Hanahan, D., and Weinberg, R.A. (2000). The hallmarks of cancer. Cell 100,
57–70.
Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of cancer: the next generation. Cell 144, 646–674.
Cell 168, February 9, 2017 687
Harper, K.L., Sosa, M.S., Entenberg, D., Hosseini, H., Cheung, J.F., Nobre,
R., Avivar-Valderas, A., Nagi, C., Girnius, N., Davis, R.J., et al. (2016). Mechanism of early dissemination and metastasis in Her2+ mammary cancer.
Nature. Published online December 14, 2016. http://dx.doi.org/10.1038/
nature20609.
Headley, M.B., Bins, A., Nip, A., Roberts, E.W., Looney, M.R., Gerard, A., and
Krummel, M.F. (2016). Visualization of immediate immune responses to
pioneer metastatic cells in the lung. Nature 531, 513–517.
Hong, T., Watanabe, K., Ta, C.H., Villarreal-Ponce, A., Nie, Q., and Dai, X.
(2015). An Ovol2-Zeb1 mutual inhibitory circuit governs bidirectional and
multi-step transition between epithelial and mesenchymal states. PLoS Comput. Biol. 11, e1004569.
, M.M., Hoffmann, M., Harper, K.L., Sosa, M.S.,
Hosseini, H., Obradovic
Werner-Klein, M., Nanduri, L.K., Werno, C., Ehrl, C., Maneck, M., et al.
(2016). Early dissemination seeds metastasis in breast cancer. Nature. Published online December 14, 2016. http://dx.doi.org/10.1038/nature20785.
Huh, S.J., Liang, S., Sharma, A., Dong, C., and Robertson, G.P. (2010). Transiently entrapped circulating tumor cells interact with neutrophils to facilitate
lung metastasis development. Cancer Res. 70, 6071–6082.
Hüsemann, Y., Geigl, J.B., Schubert, F., Musiani, P., Meyer, M., Burghart, E.,
Forni, G., Eils, R., Fehm, T., Riethmüller, G., and Klein, C.A. (2008). Systemic
spread is an early step in breast cancer. Cancer Cell 13, 58–68.
Ince, T.A., Richardson, A.L., Bell, G.W., Saitoh, M., Godar, S., Karnoub, A.E.,
Iglehart, J.D., and Weinberg, R.A. (2007). Transformation of different human
breast epithelial cell types leads to distinct tumor phenotypes. Cancer Cell
12, 160–170.
Jacob, L.S., Vanharanta, S., Obenauf, A.C., Pirun, M., Viale, A., Socci, N.D.,
and Massagué, J. (2015). Metastatic competence can emerge with selection
of preexisting oncogenic alleles without a need of new mutations. Cancer
Res. 75, 3713–3719.
Jordan, N.V., Johnson, G.L., and Abell, A.N. (2011). Tracking the intermediate
stages of epithelial-mesenchymal transition in epithelial stem cells and cancer.
Cell Cycle 10, 2865–2873.
Kalluri, R., and Weinberg, R.A. (2009). The basics of epithelial-mesenchymal
transition. J. Clin. Invest. 119, 1420–1428.
Kang, Y., and Pantel, K. (2013). Tumor cell dissemination: emerging biological
insights from animal models and cancer patients. Cancer Cell 23, 573–581.
Kaplan, R.N., Riba, R.D., Zacharoulis, S., Bramley, A.H., Vincent, L., Costa, C.,
MacDonald, D.D., Jin, D.K., Shido, K., Kerns, S.A., et al. (2005). VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic
niche. Nature 438, 820–827.
Karaman, S., and Detmar, M. (2014). Mechanisms of lymphatic metastasis.
J. Clin. Invest. 124, 922–928.
Kienast, Y., von Baumgarten, L., Fuhrmann, M., Klinkert, W.E., Goldbrunner,
R., Herms, J., and Winkler, F. (2010). Real-time imaging reveals the single
steps of brain metastasis formation. Nat. Med. 16, 116–122.
Kitamura, T., Qian, B.Z., and Pollard, J.W. (2015). Immune cell promotion of
metastasis. Nat. Rev. Immunol. 15, 73–86.
Kong, D., Banerjee, S., Ahmad, A., Li, Y., Wang, Z., Sethi, S., and Sarkar, F.H.
(2010). Epithelial to mesenchymal transition is mechanistically linked with stem
cell signatures in prostate cancer cells. PLoS ONE 5, e12445.
Kopp, H.G., Placke, T., and Salih, H.R. (2009). Platelet-derived transforming
growth factor-beta down-regulates NKG2D thereby inhibiting natural killer
cell antitumor reactivity. Cancer Res. 69, 7775–7783.
Korpal, M., Ell, B.J., Buffa, F.M., Ibrahim, T., Blanco, M.A., Celià-Terrassa, T.,
Mercatali, L., Khan, Z., Goodarzi, H., Hua, Y., et al. (2011). Direct targeting of
Sec23a by miR-200s influences cancer cell secretome and promotes metastatic colonization. Nat. Med. 17, 1101–1108.
Kurrey, N.K., Jalgaonkar, S.P., Joglekar, A.V., Ghanate, A.D., Chaskar, P.D.,
Doiphode, R.Y., and Bapat, S.A. (2009). Snail and slug mediate radioresistance and chemoresistance by antagonizing p53-mediated apoptosis and
acquiring a stem-like phenotype in ovarian cancer cells. Stem Cells 27,
2059–2068.
Labelle, M., and Hynes, R.O. (2012). The initial hours of metastasis: the importance of cooperative host-tumor cell interactions during hematogenous
dissemination. Cancer Discov. 2, 1091–1099.
Labelle, M., Begum, S., and Hynes, R.O. (2011). Direct signaling between
platelets and cancer cells induces an epithelial-mesenchymal-like transition
and promotes metastasis. Cancer Cell 20, 576–590.
Labelle, M., Begum, S., and Hynes, R.O. (2014). Platelets guide the formation
of early metastatic niches. Proc. Natl. Acad. Sci. USA 111, E3053–E3061.
Lamouille, S., Xu, J., and Derynck, R. (2014). Molecular mechanisms of epithelial-mesenchymal transition. Nat. Rev. Mol. Cell Biol. 15, 178–196.
Latil, M., Nassar, D., Beck, B., Boumahdi, S., Wang, L., Brisebarre, A.,
Dubois, C., Nkusi, E., Lenglez,, S., Checinska, A., et al. (2016). Cell-TypeSpecific Chromatin States Differentially Prime Squamous Cell Carcinoma
Tumor-Initiating Cells for Epithelial to Mesenchymal Transition. Cell Stem
Cell. Published online November 16, 2016. http://dx.doi.org/10.1016/j.stem.
2016.10.018.
Lawson, D.A., Bhakta, N.R., Kessenbrock, K., Prummel, K.D., Yu, Y., Takai, K.,
Zhou, A., Eyob, H., Balakrishnan, S., Wang, C.Y., et al. (2015). Single-cell
analysis reveals a stem-cell program in human metastatic breast cancer cells.
Nature 526, 131–135.
Lee, H.J., Zhuang, G., Cao, Y., Du, P., Kim, H.J., and Settleman, J. (2014). Drug
resistance via feedback activation of Stat3 in oncogene-addicted cancer cells.
Cancer Cell 26, 207–221.
Levental, K.R., Yu, H., Kass, L., Lakins, J.N., Egeblad, M., Erler, J.T., Fong,
S.F., Csiszar, K., Giaccia, A., Weninger, W., et al. (2009). Matrix crosslinking
forces tumor progression by enhancing integrin signaling. Cell 139, 891–906.
Li, W., and Kang, Y. (2016). Probing the fifty shades of EMT in metastasis.
Trends Cancer 2, 65–67.
Li, C.M., Gocheva, V., Oudin, M.J., Bhutkar, A., Wang, S.Y., Date, S.R., Ng,
S.R., Whittaker, C.A., Bronson, R.T., Snyder, E.L., et al. (2015). Foxa2 and
Cdx2 cooperate with Nkx2-1 to inhibit lung adenocarcinoma metastasis.
Genes Dev. 29, 1850–1862.
Klein, C.A. (2009). Parallel progression of primary tumours and metastases.
Nat. Rev. Cancer 9, 302–312.
Lim, E., Vaillant, F., Wu, D., Forrest, N.C., Pal, B., Hart, A.H., Asselin-Labat,
M.L., Gyorki, D.E., Ward, T., Partanen, A., et al.; kConFab (2009). Aberrant
luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat. Med. 15, 907–913.
Kobayashi, A., Okuda, H., Xing, F., Pandey, P.R., Watabe, M., Hirota, S., Pai,
S.K., Liu, W., Fukuda, K., Chambers, C., et al. (2011). Bone morphogenetic
protein 7 in dormancy and metastasis of prostate cancer stem-like cells in
bone. J. Exp. Med. 208, 2641–2655.
Liu, W., Laitinen, S., Khan, S., Vihinen, M., Kowalski, J., Yu, G., Chen, L.,
Ewing, C.M., Eisenberger, M.A., Carducci, M.A., et al. (2009). Copy number
analysis indicates monoclonal origin of lethal metastatic prostate cancer.
Nat. Med. 15, 559–565.
Koebel, C.M., Vermi, W., Swann, J.B., Zerafa, N., Rodig, S.J., Old, L.J., Smyth,
M.J., and Schreiber, R.D. (2007). Adaptive immunity maintains occult cancer in
an equilibrium state. Nature 450, 903–907.
Long, H., Xiang, T., Qi, W., Huang, J., Chen, J., He, L., Liang, Z., Guo, B., Li, Y.,
Xie, R., and Zhu, B. (2015). CD133+ ovarian cancer stem-like cells promote
non-stem cancer cell metastasis via CCL5 induced epithelial-mesenchymal
transition. Oncotarget 6, 5846–5859.
Köhler, S., Ullrich, S., Richter, U., and Schumacher, U. (2010). E-/P-selectins
and colon carcinoma metastasis: first in vivo evidence for their crucial role in
a clinically relevant model of spontaneous metastasis formation in the lung.
Br. J. Cancer 102, 602–609.
688 Cell 168, February 9, 2017
Loo, J.M., Scherl, A., Nguyen, A., Man, F.Y., Weinberg, E., Zeng, Z., Saltz, L.,
Paty, P.B., and Tavazoie, S.F. (2015). Extracellular metabolic energetics can
promote cancer progression. Cell 160, 393–406.
Lu, X., Mu, E., Wei, Y., Riethdorf, S., Yang, Q., Yuan, M., Yan, J., Hua, Y., Tiede,
B.J., Lu, X., et al. (2011). VCAM-1 promotes osteolytic expansion of indolent
bone micrometastasis of breast cancer by engaging a4b1-positive osteoclast
progenitors. Cancer Cell 20, 701–714.
Lundgren, K., Nordenskjöld, B., and Landberg, G. (2009). Hypoxia, Snail and
incomplete epithelial-mesenchymal transition in breast cancer. Br. J. Cancer
101, 1769–1781.
Luzzi, K.J., MacDonald, I.C., Schmidt, E.E., Kerkvliet, N., Morris, V.L., Chambers, A.F., and Groom, A.C. (1998). Multistep nature of metastatic inefficiency:
dormancy of solitary cells after successful extravasation and limited survival of
early micrometastases. Am. J. Pathol. 153, 865–873.
Malanchi, I., Santamaria-Martı́nez, A., Susanto, E., Peng, H., Lehr, H.A., Delaloye, J.F., and Huelsken, J. (2011). Interactions between cancer stem cells and
their niche govern metastatic colonization. Nature 481, 85–89.
Malladi, S., Macalinao, D.G., Jin, X., He, L., Basnet, H., Zou, Y., de Stanchina,
E., and Massagué, J. (2016). Metastatic latency and immune evasion through
autocrine inhibition of WNT. Cell 165, 45–60.
Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks,
M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelialmesenchymal transition generates cells with properties of stem cells. Cell
133, 704–715.
Marusyk, A., Almendro, V., and Polyak, K. (2012). Intra-tumour heterogeneity:
a looking glass for cancer? Nat. Rev. Cancer 12, 323–334.
McAllister, S.S., and Weinberg, R.A. (2014). The tumour-induced systemic
environment as a critical regulator of cancer progression and metastasis.
Nat. Cell Biol. 16, 717–727.
Miyamoto, D.T., Zheng, Y., Wittner, B.S., Lee, R.J., Zhu, H., Broderick, K.T.,
Desai, R., Fox, D.B., Brannigan, B.W., Trautwein, J., et al. (2015). RNA-Seq
of single prostate CTCs implicates noncanonical Wnt signaling in antiandrogen
resistance. Science 349, 1351–1356.
Molyneux, G., Geyer, F.C., Magnay, F.A., McCarthy, A., Kendrick, H., Natrajan,
R., Mackay, A., Grigoriadis, A., Tutt, A., Ashworth, A., et al. (2010). BRCA1
basal-like breast cancers originate from luminal epithelial progenitors and
not from basal stem cells. Cell Stem Cell 7, 403–417.
Morel, A.P., Lièvre, M., Thomas, C., Hinkal, G., Ansieau, S., and Puisieux, A.
(2008). Generation of breast cancer stem cells through epithelial-mesenchymal transition. PLoS ONE 3, e2888.
Mouw, J.K., Yui, Y., Damiano, L., Bainer, R.O., Lakins, J.N., Acerbi, I., Ou, G.,
Wijekoon, A.C., Levental, K.R., Gilbert, P.M., et al. (2014). Tissue mechanics
modulate microRNA-dependent PTEN expression to regulate malignant
progression. Nat. Med. 20, 360–367.
Müller-Hermelink, N., Braumüller, H., Pichler, B., Wieder, T., Mailhammer, R.,
Schaak, K., Ghoreschi, K., Yazdi, A., Haubner, R., Sander, C.A., et al. (2008).
TNFR1 signaling and IFN-g signaling determine whether T cells induce tumor
dormancy or promote multistage carcinogenesis. Cancer Cell 13, 507–518.
tumour secretomes promote resistance and tumour progression. Nature
520, 368–372.
Ocaña, O.H., Córcoles, R., Fabra, A., Moreno-Bueno, G., Acloque, H., Vega,
S., Barrallo-Gimeno, A., Cano, A., and Nieto, M.A. (2012). Metastatic colonization requires the repression of the epithelial-mesenchymal transition inducer
Prrx1. Cancer Cell 22, 709–724.
Oskarsson, T., Acharyya, S., Zhang, X.H., Vanharanta, S., Tavazoie, S.F.,
Morris, P.G., Downey, R.J., Manova-Todorova, K., Brogi, E., and Massagué,
J. (2011). Breast cancer cells produce tenascin C as a metastatic niche
component to colonize the lungs. Nat. Med. 17, 867–874.
Oskarsson, T., Batlle, E., and Massagué, J. (2014). Metastatic stem cells: sources, niches, and vital pathways. Cell Stem Cell 14, 306–321.
Ozturk, S., Papageorgis, P., Wong, C.K., Lambert, A.W., Abdolmaleky, H.M.,
Thiagalingam, A., Cohen, H.T., and Thiagalingam, S. (2016). SDPR functions
as a metastasis suppressor in breast cancer by promoting apoptosis. Proc.
Natl. Acad. Sci. USA 113, 638–643.
Padua, D., Zhang, X.H., Wang, Q., Nadal, C., Gerald, W.L., Gomis, R.R., and
Massagué, J. (2008). TGFbeta primes breast tumors for lung metastasis seeding through angiopoietin-like 4. Cell 133, 66–77.
Palumbo, J.S., Talmage, K.E., Massari, J.V., La Jeunesse, C.M., Flick, M.J.,
Kombrinck, K.W., Jirousková, M., and Degen, J.L. (2005). Platelets and
fibrin(ogen) increase metastatic potential by impeding natural killer cell-mediated elimination of tumor cells. Blood 105, 178–185.
Palumbo, J.S., Talmage, K.E., Massari, J.V., La Jeunesse, C.M., Flick, M.J.,
Kombrinck, K.W., Hu, Z., Barney, K.A., and Degen, J.L. (2007). Tumor cellassociated tissue factor and circulating hemostatic factors cooperate to
increase metastatic potential through natural killer cell-dependent and-independent mechanisms. Blood 110, 133–141.
Pang, R., Law, W.L., Chu, A.C., Poon, J.T., Lam, C.S., Chow, A.K., Ng, L.,
Cheung, L.W., Lan, X.R., Lan, H.Y., et al. (2010). A subpopulation of CD26+
cancer stem cells with metastatic capacity in human colorectal cancer. Cell
Stem Cell 6, 603–615.
kovic
, M., Lavotshkin, S., Matei, I., Costa-Silva, B., MorenoPeinado, H., Alec
Bueno, G., Hergueta-Redondo, M., Williams, C., Garcı́a-Santos, G., Ghajar,
C., et al. (2012). Melanoma exosomes educate bone marrow progenitor cells
toward a pro-metastatic phenotype through MET. Nat. Med. 18, 883–891.
Piskounova, E., Agathocleous, M., Murphy, M.M., Hu, Z., Huddlestun, S.E.,
Zhao, Z., Leitch, A.M., Johnson, T.M., DeBerardinis, R.J., and Morrison, S.J.
(2015). Oxidative stress inhibits distant metastasis by human melanoma cells.
Nature 527, 186–191.
Podsypanina, K., Du, Y.C., Jechlinger, M., Beverly, L.J., Hambardzumyan, D.,
and Varmus, H. (2008). Seeding and propagation of untransformed mouse
mammary cells in the lung. Science 321, 1841–1844.
Mundy, G.R. (2002). Metastasis to bone: causes, consequences and therapeutic opportunities. Nat. Rev. Cancer 2, 584–593.
Proia, T.A., Keller, P.J., Gupta, P.B., Klebba, I., Jones, A.D., Sedic, M.,
Gilmore, H., Tung, N., Naber, S.P., Schnitt, S., et al. (2011). Genetic predisposition directs breast cancer phenotype by dictating progenitor cell fate. Cell
Stem Cell 8, 149–163.
Naxerova, K., and Jain, R.K. (2015). Using tumour phylogenetics to identify the
roots of metastasis in humans. Nat. Rev. Clin. Oncol. 12, 258–272.
Psaila, B., and Lyden, D. (2009). The metastatic niche: adapting the foreign
soil. Nat. Rev. Cancer 9, 285–293.
Nguyen, D.X., Bos, P.D., and Massagué, J. (2009). Metastasis: from dissemination to organ-specific colonization. Nat. Rev. Cancer 9, 274–284.
Qian, B., Deng, Y., Im, J.H., Muschel, R.J., Zou, Y., Li, J., Lang, R.A., and
Pollard, J.W. (2009). A distinct macrophage population mediates metastatic
breast cancer cell extravasation, establishment and growth. PLoS ONE 4,
e6562.
Nieswandt, B., Hafner, M., Echtenacher, B., and Männel, D.N. (1999). Lysis of
tumor cells by natural killer cells in mice is impeded by platelets. Cancer Res.
59, 1295–1300.
Nieto, M.A., Huang, R.Y., Jackson, R.A., and Thiery, J.P. (2016). EMT: 2016.
Cell 166, 21–45.
Qian, B.Z., Li, J., Zhang, H., Kitamura, T., Zhang, J., Campion, L.R., Kaiser,
E.A., Snyder, L.A., and Pollard, J.W. (2011). CCL2 recruits inflammatory monocytes to facilitate breast-tumour metastasis. Nature 475, 222–225.
Nowell, P.C. (1976). The clonal evolution of tumor cell populations. Science
194, 23–28.
Quail, D.F., and Joyce, J.A. (2013). Microenvironmental regulation of tumor
progression and metastasis. Nat. Med. 19, 1423–1437.
Obenauf, A.C., and Massague, J. (2015). Surviving at a distance: organ specific metastasis. Trends Cancer 1, 76–91.
Rankin, E.B., and Giaccia, A.J. (2016). Hypoxic control of metastasis. Science
352, 175–180.
Obenauf, A.C., Zou, Y., Ji, A.L., Vanharanta, S., Shu, W., Shi, H., Kong, X.,
Bosenberg, M.C., Wiesner, T., Rosen, N., et al. (2015). Therapy-induced
Rasheed, Z.A., Yang, J., Wang, Q., Kowalski, J., Freed, I., Murter, C., Hong,
S.M., Koorstra, J.B., Rajeshkumar, N.V., He, X., et al. (2010). Prognostic
Cell 168, February 9, 2017 689
significance of tumorigenic cells with mesenchymal features in pancreatic
adenocarcinoma. J. Natl. Cancer Inst. 102, 340–351.
Revenu, C., and Gilmour, D. (2009). EMT 2.0: shaping epithelia through collective migration. Curr. Opin. Genet. Dev. 19, 338–342.
Reymond, N., d’Água, B.B., and Ridley, A.J. (2013). Crossing the endothelial
barrier during metastasis. Nat. Rev. Cancer 13, 858–870.
Rhim, A.D., Mirek, E.T., Aiello, N.M., Maitra, A., Bailey, J.M., McAllister, F.,
Reichert, M., Beatty, G.L., Rustgi, A.K., Vonderheide, R.H., et al. (2012).
EMT and dissemination precede pancreatic tumor formation. Cell 148,
349–361.
Ross, J.B., Huh, D., Noble, L.B., and Tavazoie, S.F. (2015). Identification of
molecular determinants of primary and metastatic tumour re-initiation in breast
cancer. Nat. Cell Biol. 17, 651–664.
Sagiv, J.Y., Michaeli, J., Assi, S., Mishalian, I., Kisos, H., Levy, L., Damti, P.,
Lumbroso, D., Polyansky, L., Sionov, R.V., et al. (2015). Phenotypic diversity
and plasticity in circulating neutrophil subpopulations in cancer. Cell Rep.
10, 562–573.
Sampson, V.B., David, J.M., Puig, I., Patil, P.U., de Herreros, A.G., Thomas,
G.V., and Rajasekaran, A.K. (2014). Wilms’ tumor protein induces an epithelial-mesenchymal hybrid differentiation state in clear cell renal cell carcinoma.
PLoS ONE 9, e102041.
Schliekelman, M.J., Taguchi, A., Zhu, J., Dai, X., Rodriguez, J., Celiktas, M.,
Zhang, Q., Chin, A., Wong, C.H., Wang, H., et al. (2015). Molecular portraits
of epithelial, mesenchymal, and hybrid states in lung adenocarcinoma and
their relevance to survival. Cancer Res. 75, 1789–1800.
Schumacher, D., Strilic, B., Sivaraj, K.K., Wettschureck, N., and Offermanns,
S. (2013). Platelet-derived nucleotides promote tumor-cell transendothelial
migration and metastasis via P2Y2 receptor. Cancer Cell 24, 130–137.
Sethi, N., and Kang, Y. (2011). Unravelling the complexity of metastasis - molecular understanding and targeted therapies. Nat. Rev. Cancer 11, 735–748.
Sevenich, L., Bowman, R.L., Mason, S.D., Quail, D.F., Rapaport, F., Elie, B.T.,
Brogi, E., Brastianos, P.K., Hahn, W.C., Holsinger, L.J., et al. (2014). Analysis of
tumour- and stroma-supplied proteolytic networks reveals a brain-metastasispromoting role for cathepsin S. Nat. Cell Biol. 16, 876–888.
Sharma, P., and Allison, J.P. (2015). Immune checkpoint targeting in cancer
therapy: toward combination strategies with curative potential. Cell 161,
205–214.
Shibue, T., and Weinberg, R.A. (2009). Integrin beta1-focal adhesion kinase
signaling directs the proliferation of metastatic cancer cells disseminated in
the lungs. Proc. Natl. Acad. Sci. USA 106, 10290–10295.
Shibue, T., Brooks, M.W., Inan, M.F., Reinhardt, F., and Weinberg, R.A. (2012).
The outgrowth of micrometastases is enabled by the formation of filopodiumlike protrusions. Cancer Discov. 2, 706–721.
Shibue, T., Brooks, M.W., and Weinberg, R.A. (2013). An integrin-linked machinery of cytoskeletal regulation that enables experimental tumor initiation
and metastatic colonization. Cancer Cell 24, 481–498.
Shiozawa, Y., Pedersen, E.A., Havens, A.M., Jung, Y., Mishra, A., Joseph, J.,
Kim, J.K., Patel, L.R., Ying, C., Ziegler, A.M., et al. (2011). Human prostate cancer metastases target the hematopoietic stem cell niche to establish footholds
in mouse bone marrow. J. Clin. Invest. 121, 1298–1312.
Singh, A., and Settleman, J. (2010). EMT, cancer stem cells and drug resistance: an emerging axis of evil in the war on cancer. Oncogene 29, 4741–4751.
Skobe, M., Hawighorst, T., Jackson, D.G., Prevo, R., Janes, L., Velasco, P.,
Riccardi, L., Alitalo, K., Claffey, K., and Detmar, M. (2001). Induction of tumor
lymphangiogenesis by VEGF-C promotes breast cancer metastasis. Nat. Med.
7, 192–198.
Sosa, M.S., Avivar-Valderas, A., Bragado, P., Wen, H.C., and Aguirre-Ghiso,
J.A. (2011). ERK1/2 and p38a/b signaling in tumor cell quiescence: opportunities to control dormant residual disease. Clin. Cancer Res. 17, 5850–5857.
Sosa, M.S., Bragado, P., and Aguirre-Ghiso, J.A. (2014). Mechanisms of
disseminated cancer cell dormancy: an awakening field. Nat. Rev. Cancer
14, 611–622.
690 Cell 168, February 9, 2017
Spicer, J.D., McDonald, B., Cools-Lartigue, J.J., Chow, S.C., Giannias, B.,
Kubes, P., and Ferri, L.E. (2012). Neutrophils promote liver metastasis via
Mac-1-mediated interactions with circulating tumor cells. Cancer Res. 72,
3919–3927.
Spiegel, A., Brooks, M.W., Houshyar, S., Reinhardt, F., Ardolino, M., Fessler,
E., Chen, M.B., Krall, J.A., DeCock, J., Zervantonakis, I.K., et al. (2016). Neutrophils suppress intraluminal NK cell-mediated tumor cell clearance and
enhance extravasation of disseminated carcinoma cells. Cancer Discov. 6,
630–649.
Steeg, P.S. (2003). Metastasis suppressors alter the signal transduction of
cancer cells. Nat. Rev. Cancer 3, 55–63.
Steeg, P.S. (2016). Targeting metastasis. Nat. Rev. Cancer 16, 201–218.
Straussman, R., Morikawa, T., Shee, K., Barzily-Rokni, M., Qian, Z.R., Du, J.,
Davis, A., Mongare, M.M., Gould, J., Frederick, D.T., et al. (2012). Tumour
micro-environment elicits innate resistance to RAF inhibitors through HGF
secretion. Nature 487, 500–504.
Strilic, B., Yang, L., Albarrán-Juárez, J., Wachsmuth, L., Han, K., Müller, U.C.,
Pasparakis, M., and Offermanns, S. (2016). Tumour-cell-induced endothelial
cell necroptosis via death receptor 6 promotes metastasis. Nature 536,
215–218.
Sun, Y., Campisi, J., Higano, C., Beer, T.M., Porter, P., Coleman, I., True, L.,
and Nelson, P.S. (2012). Treatment-induced damage to the tumor microenvironment promotes prostate cancer therapy resistance through WNT16B. Nat.
Med. 18, 1359–1368.
Talmadge, J.E., and Fidler, I.J. (2010). AACR centennial series: the biology of
cancer metastasis: historical perspective. Cancer Res. 70, 5649–5669.
Thiery, J.P. (2002). Epithelial-mesenchymal transitions in tumour progression.
Nat. Rev. Cancer 2, 442–454.
Trimboli, A.J., Fukino, K., de Bruin, A., Wei, G., Shen, L., Tanner, S.M., Creasap, N., Rosol, T.J., Robinson, M.L., Eng, C., et al. (2008). Direct evidence
for epithelial-mesenchymal transitions in breast cancer. Cancer Res. 68,
937–945.
Tsai, J.H., Donaher, J.L., Murphy, D.A., Chau, S., and Yang, J. (2012). Spatiotemporal regulation of epithelial-mesenchymal transition is essential for
squamous cell carcinoma metastasis. Cancer Cell 22, 725–736.
Turajlic, S., and Swanton, C. (2016). Metastasis as an evolutionary process.
Science 352, 169–175.
Valiente, M., Obenauf, A.C., Jin, X., Chen, Q., Zhang, X.H., Lee, D.J., Chaft,
J.E., Kris, M.G., Huse, J.T., Brogi, E., and Massagué, J. (2014). Serpins promote cancer cell survival and vascular co-option in brain metastasis. Cell
156, 1002–1016.
Vanharanta, S., and Massagué, J. (2013). Origins of metastatic traits. Cancer
Cell 24, 410–421.
Veracini, L., Grall, D., Schaub, S., Beghelli-de la Forest Divonne, S., EtienneGrimaldi, M.C., Milano, G., Bozec, A., Babin, E., Sudaka, A., Thariat, J., and
Van Obberghen-Schilling, E. (2015). Elevated Src family kinase activity stabilizes E-cadherin-based junctions and collective movement of head and neck
squamous cell carcinomas. Oncotarget 6, 7570–7583.
Vogelstein, B., Papadopoulos, N., Velculescu, V.E., Zhou, S., Diaz, L.A., Jr.,
and Kinzler, K.W. (2013). Cancer genome landscapes. Science 339, 1546–
1558.
Wan, L., Pantel, K., and Kang, Y. (2013). Tumor metastasis: moving new biological insights into the clinic. Nat. Med. 19, 1450–1464.
Wculek, S.K., and Malanchi, I. (2015). Neutrophils support lung colonization of
metastasis-initiating breast cancer cells. Nature 528, 413–417.
Weilbaecher, K.N., Guise, T.A., and McCauley, L.K. (2011). Cancer to bone: a
fatal attraction. Nat. Rev. Cancer 11, 411–425.
Weng, D., Penzner, J.H., Song, B., Koido, S., Calderwood, S.K., and Gong, J.
(2012). Metastasis is an early event in mouse mammary carcinomas and is
associated with cells bearing stem cell markers. Breast Cancer Res. 14, R18.
Westcott, J.M., Prechtl, A.M., Maine, E.A., Dang, T.T., Esparza, M.A., Sun, H.,
Zhou, Y., Xie, Y., and Pearson, G.W. (2015). An epigenetically distinct breast
cancer cell subpopulation promotes collective invasion. J. Clin. Invest. 125,
1927–1943.
Winslow, M.M., Dayton, T.L., Verhaak, R.G., Kim-Kiselak, C., Snyder, E.L.,
Feldser, D.M., Hubbard, D.D., DuPage, M.J., Whittaker, C.A., Hoersch, S.,
et al. (2011). Suppression of lung adenocarcinoma progression by Nkx2-1.
Nature 473, 101–104.
Wolf, M.J., Hoos, A., Bauer, J., Boettcher, S., Knust, M., Weber, A., Simonavicius, N., Schneider, C., Lang, M., Stürzl, M., et al. (2012). Endothelial CCR2
signaling induced by colon carcinoma cells enables extravasation via the
JAK2-Stat5 and p38MAPK pathway. Cancer Cell 22, 91–105.
Yachida, S., Jones, S., Bozic, I., Antal, T., Leary, R., Fu, B., Kamiyama, M., Hruban, R.H., Eshleman, J.R., Nowak, M.A., et al. (2010). Distant metastasis
occurs late during the genetic evolution of pancreatic cancer. Nature 467,
1114–1117.
Yu, M., Ting, D.T., Stott, S.L., Wittner, B.S., Ozsolak, F., Paul, S., Ciciliano,
J.C., Smas, M.E., Winokur, D., Gilman, A.J., et al. (2012). RNA sequencing
of pancreatic circulating tumour cells implicates WNT signalling in metastasis.
Nature 487, 510–513.
Yu, M., Bardia, A., Wittner, B.S., Stott, S.L., Smas, M.E., Ting, D.T., Isakoff,
S.J., Ciciliano, J.C., Wells, M.N., Shah, A.M., et al. (2013). Circulating breast
tumor cells exhibit dynamic changes in epithelial and mesenchymal composition. Science 339, 580–584.
Yu, M., Bardia, A., Aceto, N., Bersani, F., Madden, M.W., Donaldson, M.C.,
Desai, R., Zhu, H., Comaills, V., Zheng, Z., et al. (2014). Cancer therapy.
Ex vivo culture of circulating breast tumor cells for individualized testing of
drug susceptibility. Science 345, 216–220.
Zhang, X.H., Wang, Q., Gerald, W., Hudis, C.A., Norton, L., Smid, M., Foekens,
J.A., and Massagué, J. (2009). Latent bone metastasis in breast cancer tied to
Src-dependent survival signals. Cancer Cell 16, 67–78.
Yates, L.R., Gerstung, M., Knappskog, S., Desmedt, C., Gundem, G., Van Loo,
P., Aas, T., Alexandrov, L.B., Larsimont, D., Davies, H., et al. (2015). Subclonal
diversification of primary breast cancer revealed by multiregion sequencing.
Nat. Med. 21, 751–759.
Zheng, X., Carstens, J.L., Kim, J., Scheible, M., Kaye, J., Sugimoto, H., Wu,
C.C., LeBleu, V.S., and Kalluri, R. (2015). Epithelial-to-mesenchymal transition
is dispensable for metastasis but induces chemoresistance in pancreatic
cancer. Nature 527, 525–530.
Ye, X., Tam, W.L., Shibue, T., Kaygusuz, Y., Reinhardt, F., Ng Eaton, E., and
Weinberg, R.A. (2015). Distinct EMT programs control normal mammary
stem cells and tumour-initiating cells. Nature 525, 256–260.
Zhou, D., Kannappan, V., Chen, X., Li, J., Leng, X., Zhang, J., and Xuan, S.
(2016). RBP2 induces stem-like cancer cells by promoting EMT and is a prognostic marker for renal cell carcinoma. Exp. Mol. Med. 48, e238.
Cell 168, February 9, 2017 691
Leading Edge
Review
Tumorigenic and Immunosuppressive Effects
of Endoplasmic Reticulum Stress in Cancer
Juan R. Cubillos-Ruiz,1,2,* Sarah E. Bettigole,3 and Laurie H. Glimcher4,5,*
1Department
of Obstetrics and Gynecology, Weill Cornell Medicine, New York, NY 10065, USA
and Edward Meyer Cancer Center, Weill Cornell Medicine, New York, NY 10065, USA
3Quentis Therapeutics, Inc., New York, NY 10016, USA
4Department of Medicine, Harvard Medical School and Brigham and Women’s Hospital, Boston, MA 02215, USA
5Department of Cancer Immunology and Virology, Dana-Farber Cancer Institute, Boston, MA 02215, USA
*Correspondence: jur2016@med.cornell.edu (J.R.C.-R.), laurie_glimcher@dfci.harvard.edu (L.H.G.)
http://dx.doi.org/10.1016/j.cell.2016.12.004
2Sandra
Malignant cells utilize diverse strategies that enable them to thrive under adverse conditions while
simultaneously inhibiting the development of anti-tumor immune responses. Hostile microenvironmental conditions within tumor masses, such as nutrient deprivation, oxygen limitation, high metabolic demand, and oxidative stress, disturb the protein-folding capacity of the endoplasmic reticulum (ER), thereby provoking a cellular state of ‘‘ER stress.’’ Sustained activation of ER stress
sensors endows malignant cells with greater tumorigenic, metastatic, and drug-resistant capacity.
Additionally, recent studies have uncovered that ER stress responses further impede the development of protective anti-cancer immunity by manipulating the function of myeloid cells in the tumor
microenvironment. Here, we discuss the tumorigenic and immunoregulatory effects of ER stress in
cancer, and we explore the concept of targeting ER stress responses to enhance the efficacy of
standard chemotherapies and evolving cancer immunotherapies in the clinic.
Tumor growth persists despite many cell-intrinsic and cellextrinsic stresses, including dysregulated proliferation, oxidative
stress, nutrient and lipid deprivation, hypoxia, and acidic extracellular pH. Tumor progression despite these challenges requires frequent adaptation. The endoplasmic reticulum (ER)
regulates this adaptive capacity by coordinating a wide array
of fundamental cellular processes, including transmembrane
and secretory protein folding, lipid biosynthesis, drug detoxification, and calcium storage and signaling. At steady state, the ER
protein-folding machinery readily handles secretory pathway requirements. However, if misfolded proteins accumulate beyond
a tolerable threshold, ER-resident sensors trigger an unfolded
protein response (UPR) to transcriptionally and translationally
improve ER protein-folding capacity. If these corrective efforts
are insufficient, the cell will undergo apoptosis (Wang and Kaufman, 2014).
Despite these potentially fatal outcomes, robust ER stress responses have been documented in most major types of human
cancer, including breast, pancreatic, lung, skin, prostate, brain,
and even liquid malignancies (Wang and Kaufman, 2014).
Furthermore, ER stress in situ frequently correlates with
advanced-stage disease and chemoresistance. The ability to
tolerate persistent ER stress enhances cancer cell survival,
angiogenesis, metastatic capacity, drug resistance, and immunosuppression. Yet this risky balancing act also endows cancer
cells with selective vulnerabilities that could be harnessed to
therapeutic advantage. In this review, we explore the causes
and consequences of ER stress in malignancy within individual
tumor cells and across their larger microenvironments.
692 Cell 168, February 9, 2017 ª 2016 Elsevier Inc.
Orchestrating an ER Stress Response
Detecting and resolving ER stress requires three major ER-spanning transmembrane proteins, inositol-requiring enzyme
1a (IRE1a) (encoded by ERN1), PKR-like ER kinase (PERK) (encoded by EIF2AK3), and activating transcription factor
6a (ATF6a) (encoded by ATF6). These sensors exhibit a broadly
similar activation mechanism and regulate many unique and
overlapping facets of the ER stress response. Each is bound intraluminally by the chaperone protein BiP (encoded by HSPA5),
which locks them in monomeric, inactive states. If the level of
intraluminal misfolded proteins exceeds the folding capacity of
ER-resident chaperones, glycosylases, and oxido-reductases,
BiP dissociates from IRE1a, PERK, and ATF6a (Bertolotti et al.,
2000; Shen et al., 2002). These sensors subsequently drive
mutually reinforcing signaling pathways to correct the proteinmisfolding stress. If the burden can be reduced quickly, the cells
successfully adapt to the insult, whereas insufficient clearance
results in apoptotic cell death.
IRE1a-XBP1
IRE1a is a highly conserved dual enzyme possessing both kinase
and endoribonuclease activity. After BiP dissociation, IRE1a dimerizes and autophosphorylates, triggering a conformational
shift that allosterically activates its endoribonuclease domain.
This nuclease then catalyzes a unique cytoplasmic mRNAsplicing reaction, specifically cleaving out 26 nucleotides from
the XBP1 mRNA, which is subsequently re-ligated by the tRNA
ligase RCTB (Lu et al., 2014; Yoshida et al., 2001). Re-ligation
causes a reading frameshift and translation of the highly active
transcription factor XBP1, which upregulates multiple foldases,
oxido-reductases, intracellular trafficking components, ERassociated degradation machinery, and glycosylases to correct
ER homeostasis (Shoulders et al., 2013). XBP1 also upregulates
UPR-independent pathways, including pro-inflammatory cytokine production, lipid and hexosamine biosynthesis, and the
hypoxia response (Bettigole and Glimcher, 2015). XBP1 induction favors cell survival, as enforced overexpression rescues
cell viability in vitro and in a transgenic rat model of retinitis pigmentosa (Lin et al., 2007). However, under severe ER stress,
IRE1a can also oligomerize and sequence-specifically degrade
multiple ER-localized mRNAs and microRNAs in a pro-apoptotic
process known as regulated IRE1a-dependent decay (RIDD)
(Hollien et al., 2009; Lerner et al., 2012). Independently of its
endoribonuclease function, phosphorylated IRE1a recruits
TRAF2 to facilitate JNK and NFkB activation upon pharmacological ER stress (Tam et al., 2012; Urano et al., 2000). Similarly,
IRE1a constitutively associates with the transcription factor
STAT3 in mouse primary hepatocytes, and this interaction is
crucial for enhancing STAT3 phosphorylation both in vitro
and in vivo (Liu et al., 2015). IRE1a is thus well positioned to influence several key regulators of tumorigenesis independently
of XBP1.
Interestingly, decreased ER membrane fluidity resulting from
increased ER phospholipid saturation induces IRE1a activation
by forcing transmembrane domains of neighboring IRE1a monomers into contact. Exogenous saturated lipids, such as palmitate, or loss of lipid desaturases like SCD1 perturbs ER membrane composition and can induce this activation mode
(Volmer et al., 2013). Elegant studies using truncated IRE1a variants unable to bind BiP revealed that this lipid-mediated activation mode proceeds in the absence of misfolded proteins.
Interestingly, IRE1a cannot oligomerize when activated by membrane lipid saturation, potentially favoring cell survival outcomes
(Kitai et al., 2013). However, given the immense difficulty in quantifying intracellular protein misfolding, the relative contribution of
membrane rigidity versus unfolded protein accumulation to
IRE1a activation in vivo remains unknown.
PERK
Like IRE1a, PERK homodimerizes and autophosphorylates upon
BiP dissociation or reduced ER membrane fluidity (Volmer et al.,
2013). Activated PERK phosphorylates the translation initiation
factor eIF2a (encoded by EIF2S1), which reduces the influx of
nascent proteins into the ER by restricting 50 cap-dependent
mRNA translation (Harding et al., 1999). Reduced translation
rates facilitate focused refolding efforts by ER-localized chaperones. Paradoxically, global translational inhibition increases selective translation of the transcription factor ATF4, which directly
upregulates the transcription factor C/EBP-homologous protein
(CHOP) (encoded by DDIT3). Subsequently, ATF4 and CHOP
cooperatively induce multiple genes involved in amino acid
biosynthesis, amino acid transport, and the intracellular recycling system autophagy (B’chir et al., 2013; Han et al., 2013).
Accelerated amino acid biosynthesis generates significant
amounts of reactive oxygen species (ROS), which induces
apoptosis if left unabated. However, PERK limits ROS accumulation by phosphorylating and stabilizing NRF2 (Cullinan et al.,
2003), enhancing glutathione synthesis (Rouschop et al., 2013),
and upregulating heme oxygenase-1 (HO-1) (Dey et al., 2015).
PERK also activates NFkB by repressing translation of the
NFkB inhibitor IkBa (Tam et al., 2012).
ATF6a
After BiP dissociation, ATF6a translocates to the Golgi apparatus, where it is cleaved intramembranously by site 1 and site
2 proteases to liberate an active, soluble ATF6a transcription
factor (Shen et al., 2002). Disulfide bonding additionally regulates
this ER-to-Golgi trafficking, and only monomeric, reduced
ATF6a can properly access COPII endosomes (Schindler and
Schekman, 2009). Unlike PERK and IRE1a, reduced ER membrane fluidity does not activate ATF6a, perhaps because dimerization is unfavorable for ATF6a activation. ATF6a fine-tunes the
UPR by upregulating BiP and a subset of XBP1-dependent
chaperones, oxidoreductases, and quality control and degradation machinery (Shoulders et al., 2013). Whereas IRE1a and
PERK conditional or germline knockout mice often exhibit pronounced phenotypes, ATF6a germline knockout mice only yield
clear phenotypes under pharmacological or pathological
stresses (Yamamoto et al., 2010), suggesting that ATF6a finetunes the UPR, which is largely controlled by the more dominant
PERK/IRE1a responses.
Sources of ER Stress in Tumors
Multiple cell-intrinsic and cell-extrinsic mechanisms initiate and
amplify ER stress within the cancer cell and the larger tumor
microenvironment (Figure 1). Spatiotemporal differences in ER
stress burden, driven by genetic, epigenetic, and microenvironmental heterogeneity, likely result in a range of pro-survival and
pro-apoptotic responses. Anticancer interventions, such as
chemotherapy, can also modulate UPR signaling, though the
clinical implications are only beginning to be understood.
Cell-Intrinsic Sources
Cancer initiation and development require both inactivation of
tumor suppressors and/or the acquisition of oncogenic mutations that uncouple proliferation from extracellular, growth factor-mediated regulation. Transformation-associated increases
in protein synthesis often overwhelm ER protein-folding capacity. In particular, highly secretory cancers, such as the B cell
malignancy multiple myeloma, which produces extremely high
levels of immunoglobulins, often undergo persistent ER stress
(Obeng et al., 2006). Specific cellular behaviors can also influence protein secretory rates, as evidenced by PERK activation
during epithelial-mesenchymal transition (EMT) (Feng et al.,
2014). Similarly, oncogenic transformation driven by loss of
the tumor suppressors p53, PTEN, TSC1, or TSC2 dramatically
enhances protein synthesis rates, leading to ER stress (Hart
et al., 2012; Namba et al., 2015; Signer et al., 2014). Enhanced
protein synthesis and concomitant ER stress are also observed
upon overexpression of oncogenic HRAS (G12E), BRAF
(V600E), c-Myc, or Src (Chen et al., 2014; Corazzari et al.,
2015; Denoyelle et al., 2006). Importantly, both the UPR
signaling and leukemia development induced by conditional
PTEN deletion (Signer et al., 2014) or c-Myc overexpression
(Hart et al., 2012) were dramatically reduced or entirely abrogated upon heterozygous deletion of the key translation rate
regulator ribosomal protein RPL24. This strongly implicates
protein synthesis rate as a key driver of ER stress and tumorigenicity in vivo.
Cell 168, February 9, 2017 693
Figure 1. Tumor Microenvironmental Factors and Conditions Perturbing ER Homeostasis
Malignant transformation mediated by oncogene
activation and loss of tumor suppressor function
places intense biosynthetic and bioenergetic demands on available cellular resources, triggering
initial ER stress. Cancer cells may eventually adapt
to these early challenges, yet as tumors expand,
they encounter a host of new environmental
stresses, including oxygen and nutrient deprivation, lactic acidosis, and multiple forms of clinical
intervention. These stimuli can disrupt ER protein
folding by limiting crucial reaction intermediates
(i.e., O2 and N-acetylglucosamine) or by directly
disrupting chaperone function via ROS-mediated
lipid peroxidation and covalent adduct formation.
However, oncogene expression does not always induce ER
stress. In contrast to c-Myc transgene-driven B cell lymphoma
(Hart et al., 2012), high MYC expression insulated a large panel
of human cancer cell lines from ER stress upon exogenous proline depletion (Sahu et al., 2016). Similarly, Ras-transformed,
Mychigh cells exhibited low basal ER stress but activated the
UPR upon Ras inhibition, suggesting that additional layers of
regulation coordinate MYC expression with ER homeostasis
(Yaari-Stark et al., 2010). Furthermore, exogenous desaturated
lipids protected TSC2/ mouse embryonic fibroblasts (MEFs)
from ER stress (Young et al., 2013). Recently transformed cells
may initially undergo ER stress in response to the higher replicative and metabolic demands but can adapt by enhancing
steady-state ER protein-folding capacity (Huber et al., 2013).
However, de novo genetic mutations and other cell-intrinsic
and cell-extrinsic stresses likely contribute to the active UPR
observed in most major cancer types (Wang and Kaufman,
2014). Fundamental differences in the experimental approaches
used, such as overexpression versus endogenous expression,
primary cells versus cell lines, and in vivo versus in vitro models,
may have also contributed to these discrepant findings. Future
work should address how protein-translation rates and related
processes, such as copy number alterations, epigenetic modifications, and microRNA-mediated regulatory mechanisms, influence the prevalence and intensity of ER stress responses in human tumors. Furthermore, identifying genetic defects and cell
biological changes that influence the saturated:unsaturated ER
phospholipid ratio will help distinguish protein misfolding from
lipotoxic sources of ER stress.
Nonsynonymous mutations can also directly destabilize
intrinsic protein folding, triggering the UPR by overwhelming
694 Cell 168, February 9, 2017
ER-resident chaperone capacity. Consistent with this, overexpressing certain
destabilized smoothened (SMO) mutants
induces robust ER stress in Drosophila
in vivo (Marada et al., 2013). Solid tumors
possess dozens of nonsynonymous
mutations, with certain cancers, such as
melanoma and lung cancers, harboring
upward of 200 mutations (Vogelstein
et al., 2013). Identifying the spectrum of
protein-destabilizing mutations that can trigger ER stress will
help clarify the physiological relevance of this mechanism.
Microenvironmental Sources
The tumor microenvironment (TME) predominantly fuels ER
stress via oxygen and nutrient deprivation and acidic waste
accumulation, though hypernutrition can also contribute during
obesity (Nakagawa et al., 2014). Whereas normal cells primarily
rely on oxidative phosphorylation or anaerobic glycolysis to
generate ATP, cancer cells often favor aerobic glycolysis in a
phenomenon known as the Warburg effect. Consequently,
rapidly dividing cancer cells aggressively consume glucose
and release large quantities of lactic acid waste regardless of
local oxygen concentration, which lowers local extracellular
pH. Tumors initially rely on resident tissue microvasculature to
supply key nutrients and oxygen but eventually must generate
their own local neovasculature to sustain growth. Though normal
tissues possess highly ordered and efficient vasculature, tumorgenerated neovasculature is generally leaky and torturous with
slow, inconsistent blood flow. Such intermittent circulation limits
nutrient accessibility, oxygen delivery, and waste drainage,
thereby driving sporadic, acute hypoxia and lactic acidosis (Vaupel et al., 1989).
Each of these extracellular conditions can induce ER stress,
though responsiveness varies depending on cell type. Low oxygen tension activates complex III of the mitochondrial electron
transport chain to increase cytosolic ROS production, required
for stabilizing the key hypoxia response transcription factor
HIF1a (Guzy et al., 2005). ROS can also generate highly reactive
peroxidized lipid byproducts, which form destructive covalent
adducts with various ER chaperones (Cubillos-Ruiz et al.,
2015; Vladykovskaya et al., 2012). Furthermore, both ER
Critically, bortezomib-resistant multiple myelomas partially dedifferentiate to XBP1low, providing clinical evidence that bortezomib efficacy likely relies on UPR inhibition (Leung-Hagesteijn
et al., 2013). Thus, interactions between anticancer drugs and
ER stress signaling can significantly alter disease progression.
Whether additional anticancer drugs induce intratumoral ER
stress in vivo remains to be determined. If they do, intratumoral
ER stress may be a valuable biomarker for determining whether
to use UPR-activating or inhibiting compounds. However, as ER
stress is also highly immunosuppressive in many leukocyte populations (described below), ER-stress-inducing drug dosages
must be carefully optimized to enable selective cancer killing
without compromising anti-tumor immune responses.
Figure 2. Consequences of ER Stress in Cancer Cells
Sublethal UPR activation and signaling via IRE1a, PERK, and ATF6a sustain
multiple cell-intrinsic and cell-extrinsic mechanisms of tumor progression.
ER-stress-mediated activation of central signaling hubs, such as HIF1a,
STAT3, NRF2, and NFkB, facilitates cell survival under harsh microenvironmental conditions and preserves tumor-initiating cell function. Cytokine-driven
angiogenesis delivers much-needed oxygen and nutrients into the tumor bed,
though IRE1a-deficient cells can also utilize vessel co-option. Intrinsic cancer
cell apoptotic resistance is likely crucial for harnessing ER stress to enhance
tumor growth.
disulphide bond formation and lipid desaturation require molecular oxygen. Nutrient deprivation, particularly of glucose and
glutamine, limits metabolic intermediates required for the hexosamine biosynthetic pathway (HBP). The HBP generates substrates for N-linked protein glycosylation, which is required for
successful ER protein folding (Huber et al., 2013). Proline starvation can also induce ER stress, potentially by inducing excessive
ROS accumulation (Sahu et al., 2016). Lastly, extracellular
acidosis can induce ER stress in a ROS-dependent manner,
possibly by driving lipid peroxidation-mediated chaperone
dysfunction (Xie et al., 2015). Whether acid-sensing G-proteincoupled-receptors (GPCRs) can trigger ER-stress-inducing
calcium fluxes and ROS generation in the TME remains to be
determined.
Clinical Sources
Multiple anticancer drugs induce potent ER stress responses
in vitro, and ER stress can facilitate anticancer drug efficacy or
the development of chemoresistance, depending on context
and tumor type. Paclitaxel, doxorubicin, the BRAF (V600E) inhibitor vemurafenib, and the epidermal growth factor receptor
(EGFR) inhibitor cetuximab potently induce eIF2a phosphorylation and downstream signaling, though the involvement of
PERK in some cases remains unclear (Jeon et al., 2015; Ma
et al., 2014; Pozzi et al., 2016). Several anthracyclines can also
trigger lethal ER stress by enhancing ROS levels and depleting
ER calcium stores, leading to PERK-dependent immunogenic
cell death (ICD) (Kepp et al., 2013). Additionally, the proteasome
inhibitor bortezomib induces ER stress and CHOP expression
but paradoxically reduces XBP1 protein accumulation by diminishing IRE1a-mediated XBP1 mRNA splicing and stabilizing the
dominant-negative unspliced XBP1 protein (Lee et al., 2003).
Mechanisms of ER-Stress-Mediated Tumor Progression
Irremediable ER protein-folding defects are often lethal, yet
tolerable levels of ER stress paradoxically facilitate multiple
mechanisms of tumor development. These include bolstering
viability under hypoxia and nutrient deprivation, enhancing metastatic spread by supporting EMT, tumor cell dormancy, and tumor-initiating cell (TIC) function and stimulating angiogenesis
(Figure 2). Many of these beneficial adaptations arise from intimate links between the ER stress response and fundamental
cell biological processes, such as autophagy and ER-mitochondrial crosstalk (Wang and Kaufman, 2014). As ER stress is a
common feature of aggressive cancers, understanding how
the UPR modulates disease is critical for identifying promising
new clinical strategies.
Cell Survival
ER stress dictates cell fate depending on context and signal
strength. Prolonged and severe pharmacological ER stress
can trigger caspase-mediated cell death through several
IRE1a- and PERK-dependent mechanisms. IRE1a-mediated
JNK activation represses anti-apoptotic BCL2 activity and enhances pro-apoptotic BIM function, thereby favoring cell death
(Wang and Kaufman, 2014). In parallel, RIDD de-represses proapoptotic caspase-2 and Txnip expression in MEFs and pancreatic b cells, respectively, by specifically cleaving microRNAs
miR-17, miR-34a, miR-96, and miR-125b (Lerner et al., 2012; Upton et al., 2012). ATF4 and CHOP accelerate protein synthesis by
upregulating tRNA synthetase expression, which elevates ROS
production from ER oxidative protein folding. Importantly, the
anti-oxidant butylated hydroxyanisole and silencing RPL24
significantly reduced cell death by reducing ROS and proteintranslation rates (Han et al., 2013). Excessive intratumoral ER
stress can also indirectly induce cancer cell death by enhancing
immunosurveillance (Kepp et al., 2013). Chromosomal tetraploidy enhances protein translation rates and induces ERstress-dependent calreticulin translocation to the plasma membrane, which serves as a phagocyte ‘‘eat me’’ signal to facilitate
ICD. Therefore, immunosurveillance selects against cancer cells
undergoing intense ER stress (Senovilla et al., 2012) and suggests that selectively exacerbating cancer cell ER stress could
enhance cell-intrinsic and immune-mediated tumor regression.
Mammalian cells have evolved multiple adaptive mechanisms
to limit pro-apoptotic UPR outputs. MEFs exposed to persistent
low-grade pharmacological ER stress resist subsequent ER insults, likely due to higher pro-survival Hspa5 mRNA stability
Cell 168, February 9, 2017 695
and reduced pro-apoptotic Ddit3 mRNA stability (Rutkowski
et al., 2006). Furthermore, STAT3 and NFkB, which can be activated by IRE1a and PERK, transcriptionally upregulate multiple
anti-apoptotic proteins, including BCL2 family members, the
caspase-8 inhibitor c-FLIP, MCL1, and inhibitor of apoptosis
proteins (IAP) (Grivennikov and Karin, 2010). Independently,
ATF4-induced miR-211 decreases DDIT3 expression by
enhancing histone methylation at the DDIT3 promoter (Chitnis
et al., 2012). Additionally, ATF6a-dependent p58(IPK) limits
apoptosis during oncogenic transformation by repressing
PERK activity (Huber et al., 2013). Consistent with these findings,
germline Ddit3 deletion enhanced lung lesion development in an
immunocompetent K-rasG12V-driven murine model of lung
cancer (Huber et al., 2013). BAX, BAK, PP2A, and GADD34
additionally modulate UPR signal strength, though how these
factors influence cell fate under chronic ER stress in the TME
is unknown (Hetz, 2012).
If cells successfully limit pro-apoptotic UPR outputs, ER stress
confers survival advantages during tumor progression in vivo.
Hypoxia and nutrient deprivation induce XBP1 splicing, which
sustains cell growth and viability in human breast cancer cell
lines in vitro and in vivo (Chen et al., 2014; Romero-Ramirez
et al., 2004). PERK-mediated NRF2 stabilization, glutathione
synthesis, and HO-1 upregulation collectively reduce cytotoxic
ROS levels to facilitate cancer cell growth (Bi et al., 2005; Dey
et al., 2015; Rouschop et al., 2013). Sublethal ER stress also enhances survival by sustaining autophagy, an intracellular recycling system charged with eliminating toxic cytosolic protein
aggregates and damaged organelles. The UPR and autophagy
are intimately linked, and pharmacological ER stress triggers autophagosome formation as an adaptive mechanism to remove
damaged ER and restrict ER expansion (Bernales et al., 2006).
Furthermore, PERK-mediated eIF2a phosphorylation is required
for LC3 lipidation, autophagy initiation, and survival in cells overexpressing aggregation-prone expanded polyglutamine 72
repeat (polyQ72) (Kouroku et al., 2007), whereas ATF4 and
CHOP upregulate numerous genes involved in autophagophore
formation and maturation, such as Atg5, Atg12, Atg16l1, and
Becn1 (B’chir et al., 2013). Invasive cancer cells require PERKmediated autophagy to resist anoikis, a form of cell death triggered by extracellular matrix (ECM) detachment. Substrate
detachment activates PERK-ATF4 signaling in human breast
cancer, fibrosarcoma, and colorectal adenocarcinoma cell lines,
and PERK-mediated autophagy limits damaging ROS accumulation (Avivar-Valderas et al., 2011; Dey et al., 2015). Accordingly,
silencing ATF4 strongly reduced fibrosarcoma lung metastasis in
a xenograft mouse model (Dey et al., 2015). Furthermore, human
breast ductal carcinomas exhibit higher PERK phosphorylation
than normal breast tissue, attesting to the physiological relevance of this mechanism (Avivar-Valderas et al., 2011).
ER-stress-mediated autophagy likely also supports therapeutic resistance. A wide variety of anticancer drugs can induce
autophagy, which can exert pro-survival or pro-apoptotic effects, depending on the drug used and tumor type (Sui et al.,
2013). Like PERK, IRE1a-JNK signaling can sustain autophagy
(Ogata et al., 2006), and this pathway facilitates sorafenib resistance in hepatocellular carcinoma cell lines (Shi et al., 2011). In
melanoma, immuno-histochemical analyses of paired pretreat696 Cell 168, February 9, 2017
ment and post-resistance tumor biopsies revealed a marked
increase in autophagy markers upon the development of vemurafenib resistance (Ma et al., 2014). BRAF (V600E) inhibition
induced ER-stress-dependent cytoprotective autophagy, which
was completely abrogated with the small-molecule PERK inhibitor GSK2606414 or by silencing EIF2AK3 (Ma et al., 2014). Critically, PERK inhibition alone had no effect on cell viability,
whereas simultaneously blocking BRAF (V600E) and PERK
induced apoptosis in chemoresistant cell lines. Vemurafenibmediated ER stress is also a therapeutic vulnerability that can
be exploited to sensitize chemoresistant melanoma to pharmacological ER-stress-induced cell death (Beck et al., 2013).
Thus, blocking UPR-mediated cytoprotective autophagy or
pharmacologically exacerbating chemotherapy-induced ER
stress can overcome chemoresistance. ER-stress-induced
autophagy can also be highly cytotoxic, as was recently shown
with the BiP inhibitor HA15. This compound induced apoptosis
in a variety of chemoresistant cancer cell lines in vitro and in vivo,
though the mechanism remains incompletely understood (Cerezo et al., 2016). In sum, optimal cancer growth and survival relies on carefully balanced UPR-signaling pathways that interact
with other cell processes, such as autophagy, to result in cancer
cell death or survival. Hence, depending on circumstances,
either inhibition or overactivation of UPR pathways can lead to
cell death.
Metastasis
Metastasis is a multi-step process in which cancer cells break
away from the primary tumor site, infiltrate the surrounding
ECM and stromal cell layers, enter the cardiovascular or
lymphatic circulatory systems, colonize foreign tissues, and
eventually grow into new tumor masses (Nieto et al., 2016).
The UPR contributes to multiple steps along this invasionmetastasis cascade. EMT facilitates initial stromal invasion and
upregulates extracellular matrix protein production to facilitate
migration and invasion. PERK buffers protein-folding stress during this increased secretory load and prevents anoikis during
EMT-induced loss of cell-cell contact (Dey et al., 2015; Feng
et al., 2014). Inducing EMT by silencing E-cadherin or overexpressing Twist strongly enhanced migration and tumorsphere
formation, which was inhibited by a small-molecule PERK inhibitor (Feng et al., 2014). Critically, ATF4 target gene expression
strongly correlated with an EMT gene signature in breast, colon,
gastric, lung, and mixed origin metastatic cancers. Consequently, pretreating cells with a PERK inhibitor or silencing
ATF4 dramatically reduced in vivo lung metastasis (Dey et al.,
2015; Feng et al., 2014). PERK also upregulates metastasisassociated LAMP3 to enhance migration and invasion in vitro
and in vivo (Mujcic et al., 2013). PERK is therefore a potential
therapeutic target to reduce EMT and invasiveness.
If invasive cells escape the stroma and successfully enter the
circulatory system, they are often deposited in inhospitable tissue microenvironments. Pioneer tumor cells adapt by entering
a p38-dependent program of anti-proliferative dormancy, which
persists until microenvironmental conditions improve. Dormant
cells are often quiescent and exhibit reduced metabolic rates,
which can insulate them from many anticancer drugs that rely
on active proliferation. Disseminated tumor cells in the bone
marrow of breast cancer patients exhibit high expression of
multiple ER chaperones, including BiP, which insulates these
cells from hypoxia and glucose deprivation (Bartkowiak et al.,
2010, 2015). Furthermore, comparative proteomic analyses of
highly proliferative T-HEp3 human squamous carcinoma cell
line and the D-HEp3 subclone, which becomes dormant in vivo
despite growing in vitro, revealed a p38-dependent program
sustaining high BiP expression and constitutive PERK phosphorylation. Dormancy-associated chemoresistance required both
BiP and PERK, as silencing HSPA5 or overexpressing a dominant-negative PERK variant sensitized D-HEp3 cells to doxorubicin and etoposide-mediated apoptosis (Ranganathan et al.,
2006). Subsequent studies identified constitutive ATF6a nuclear
translocation in the same dormant cell line, which was partially
dependent on p38 signaling. Though ATF6a was not required
for tumor cell growth in vitro, silencing ATF6 sensitized cells to
rapamycin treatment and reduced tumor nodule size in vivo
(Schewe and Aguirre-Ghiso, 2008). ATF6a knockdown reduced
pro-survival Rheb and mTOR expression in the dormant cell line,
though whether these genes are direct ATF6a transcriptional targets remains to be determined. Thus, multiple branches of the
UPR contribute to tumor cell dormancy during metastasis.
Even if environmental conditions become favorable for metastatic outgrowth, only TICs possess the necessary proliferative
capacity to generate clinically detectable tumor masses. We
recently identified the IRE1a-XBP1 pathway as a key regulator
of TIC function in human triple-negative breast cancer (TNBC).
Human basal-like breast cancer cell lines constitutively spliced
XBP1, with highest splicing in patient-derived CD44+CD24low
TICs. Silencing XBP1 potently inhibited mammosphere growth
in multiple TNBC cell lines and primary patient samples (Chen
et al., 2014). Inducible silencing of XBP1 in established TNBC
xenografts significantly reduced primary tumor growth, angiogenesis, secondary metastases, and tumor recurrence after
chemotherapy without enhancing tumor cell death. ChIP
sequencing and transcriptome analysis revealed that XBP1
and HIF1a cooperatively upregulated hypoxia response genes
to enhance TIC function. Critically, high expression of XBP1
target genes correlated with reduced overall survival in two independent TNBC patient cohorts, positioning XBP1 as an attractive clinical target in this malignancy. Taken together, all three
UPR branches collectively promote metastasis by sustaining invasion, dormancy, and tumor-initiating cell function.
Angiogenesis
Solid cancers require vascularization, often mediated by new
vessel growth, in order to supply sufficient oxygen and nutrients
for growth while removing potentially toxic waste buildup. UPR
induction stabilizes the VEGF mRNA via AMPK, though UPRAMPK crosstalk is poorly understood (Pereira et al., 2010).
PERK translationally upregulates the vessel growth and stabilization factors VCIP and PDGFRB. Consequently, K-Ras-transformed Eif2ak3/ MEFs exhibited extensive vascular hemorrhaging, failed microvasculature formation, and reduced tumor
growth in vivo (Blais et al., 2006). XBP1, ATF4, and ATF6a can
each transcriptionally upregulate Vegfa under hypoxia or
glucose deprivation by directly binding the Vegf promoter or intronic enhancers (Ghosh et al., 2010). IRE1a also sustains
expression of a broad array of pro-angiogenic cytokines,
including fibroblast growth factor 2 (FGF2), interleukin-6 (IL-6),
IL-8, and angiogenin, which likely contributes to the reduced
in vivo growth and neovascularization observed upon IRE1a or
XBP1 disruption in glioma and TNBC, respectively (Auf et al.,
2010; Chen et al., 2014). Interestingly, glioma cells adapt to
loss of IRE1a-mediated angiogenesis by enhancing mesenchymal differentiation and invasiveness to facilitate growth along
established blood vessels, suggesting IRE1a may exert tumortype-specific angiogenic functions. The IRE1a kinase domain,
but not the endoribonuclease domain, inhibited this vessel cooption, suggesting that vascularization can proceed independently of XBP1 or RIDD in certain cancer types (Jabouille et al.,
2015). Thus, IRE1a can therefore promote new vessel formation
at the expense of other vascularization modes, though whether
this occurs in other cancer types remains to be determined.
Like IRE1a, silencing EIF2AK3 strongly suppressed tumor
growth and vascularization in an orthotopic squamous cell carcinoma model, potentially due to PERK-mediated upregulation
of FGF2, vascular endothelial growth factor (VEGF), and IL-6
and suppression of the anti-angiogenic cytokines/chemokines
THBS1, CXCL14, and CXCL10 (Wang et al., 2012). Interestingly,
VEGF signaling directly activated PERK, IRE1a, and ATF6a in human umbilical vein endothelial cells (HUVECs) via a phospholipase Cg-mTORC1 signaling pathway. Silencing EIF2S1 and
ATF6 reduced VEGF-mediated AKT phosphorylation, cell survival, and neovascularization in vitro (Karali et al., 2014). Interestingly, silencing ERN1 did not inhibit these angiogenic processes
in vitro. Therefore, VEGF signaling and the UPR may occasionally engage in a positive feedback loop to sustain angiogenic
processes. In sum, tumors rely on cell-intrinsic ER stress
signaling and outward ER stress transmission to access oxygen
and nutrients in the blood.
ER Stress Responses in Tumor-Associated
Immune Cells
The microenvironment of most established tumors is formed by
stromal cells, including leukocytes, vascular cells, and fibroblasts, whose normal functions are actively co-opted by cancer
cells in order to promote malignant progression (Quail and
Joyce, 2013). For instance, leukocyte recruitment to tumor sites
can lead to unfavorable effects, such as the secretion of growth
factors enhancing cancer cell proliferation (Mantovani et al.,
2008), the induction of tumor vascularization through paracrine
mechanisms (De Palma et al., 2007), and the establishment of
complex immunosuppressive networks that restrain the protective function of cancer-reactive T cells (Crespo et al., 2013).
Whereas the role of sustained ER stress in influencing the phenotype of cancer cells has been extensively studied during the last
decade, the causes and consequences of ER stress in non-malignant cells that constitute the tumor microenvironment have
just begun to be characterized.
Cancer Cells Undergoing ER Stress Actively Modulate
Immune Cell Function
In vitro studies initially described some paracrine effects of ERstressed malignant cells on innate immune cell populations.
Pharmacological induction of ER stress prompted cancer cell
lines to release unknown soluble factors that induced upregulation of UPR markers and pro-inflammatory cytokines in
responder macrophages (Mahadevan et al., 2011). This process,
Cell 168, February 9, 2017 697
termed ‘‘transmissible ER stress,’’ was then shown to affect the
antigen-presenting capacity of bone-marrow-derived dendritic
cells (DCs) while provoking overexpression of immunosuppressive molecules like arginase (Mahadevan et al., 2012). These
studies suggested that ER-stressed cancer cells secrete factors
that actively modulate innate immune cell functions, but whether
tumor-infiltrating leukocytes indeed experience physiological ER
stress responses in vivo and whether this process contributed to
tolerance and/or immunosuppression in cancer hosts was unknown. Importantly, subsequent studies demonstrated that
administration of the ER stressor thapsigargin to tumor-bearing
mice accelerated cancer progression and stimulated the accumulation and immunosuppressive capability of myeloid-derived
suppressor cells (MDSCs), a process that could be alleviated
upon treatment of cancer hosts with chemical chaperones that
reduce ER stress (Lee et al., 2014).
Whereas sustained but controlled ER stress responses apparently endow cancer cells with greater immunomodulatory capacity, intense/lethal ER stress responses instigated by interventions, such as radiation or some chemotherapeutic agents, can
trigger ICD and protective anti-tumor immunity (Pol et al.,
2015). Malignant cells exposed to drugs of the anthracycline
family, for instance, experience irremediable ER stress characterized by ROS overproduction, increased cytoplasmic Ca2+
levels, and activation of ER stress sensors, such as PERK and
IRE1a, which can further promote activation of the inflammasome (Kepp et al., 2013; Lerner et al., 2012). Prior to evoking
cell death via induction of caspase-8, BAX, and BAK, these
agents trigger exposure of the ER-associated chaperone calreticulin on the cancer cell surface, which acts as a classical eat me
signal for neighboring immune cells (Kepp et al., 2013). Interestingly, eIF2a phosphorylation correlates with calreticulin expression in non-small-cell lung cancer (NSCLC), and this phenomenon is further associated with enhanced anti-cancer immune
responses and favorable prognosis (Fucikova et al., 2016). However, XBP1 was recently demonstrated to impede ICD in metastatic colorectal cancer cells exposed to epidermal growth factor
receptor blockers and chemotherapy (Pozzi et al., 2016). Plantderived polyphenol fractions have also been shown to induce
potent ICD and T-cell-dependent anti-tumor activity in preclinical models of melanoma and breast cancer (Gomez-Cadena
et al., 2016; Urueña et al., 2015), but whether ER stress response
factors mediate these effects remains to be determined.
p97 is an AAA family ATPase that promotes the egress of misfolded proteins from the ER to the cytosol for subsequent proteasomal-mediated degradation (Wolf and Stolz, 2012). Selective
p97 inhibitors have recently been found to trigger intense ER
stress responses, which interfere with autophagy and induce
cancer cell death (Magnaghi et al., 2013). Yet, it is unknown
whether this cytotoxic strategy can stimulate ICD or whether
systemic p97 inhibition impacts the optimal function of anti-tumor immune cells in cancer hosts. The magnitude of ER stress
in malignant cells therefore seems to define the development
of either immunosuppressive or immunogenic responses.
Intrinsic ER Stress Responses in Cancer-Associated
Immune Cells
Overexpression of ER stress markers in multiple cancer types
is associated with unfavorable prognosis and poor clinical
698 Cell 168, February 9, 2017
outcome (Chen et al., 2014; Dalton et al., 2013; Davies et al.,
2008; Matsuo et al., 2013; Shimizu et al., 2016). Though most
of these detrimental effects have been attributed to direct protumoral roles of ER stress in the cancer cell, evaluating whether ER
stress responses also operate in tumor stromal cells, such as
leukocytes, endothelial cells, or fibroblasts, to influence malignant progression has recently emerged as an area of active
research.
A function for the IRE1a-XBP1 branch of the UPR in cells of the
immune system has been well established (Bettigole and
Glimcher, 2015). IRE1a-XBP1 signaling is required for the
optimal differentiation of plasma cells, some dendritic cell populations, and eosinophils in cancer-free hosts (Bettigole et al.,
2015; Iwakoshi et al., 2007; Reimold et al., 2001). XBP1 expression in macrophages was necessary for optimal production of
IL-6 in response to Toll-like receptor (TLR) agonists, and mice
devoid of XBP1 in the hematopoietic system showed increased
bacterial burden in models of systemic Francisella tularensis
infection (Martinon et al., 2010). Neutrophils infiltrating acute
lung injury lesions exhibit XBP1 overactivation compared with
lung-resident neutrophils in naive mice. In this setting, XBP1
was necessary for optimal neutrophil granule secretion, and
ablation of XBP1-driven ER stress responses in these myeloid
cells relieved acute lung injury (Hu et al., 2015). Nevertheless,
the function of IRE1a-XBP1 signaling in cancer-associated
myeloid cells had not been explored.
Ovarian cancer is a highly aggressive and lethal malignancy
that subverts the normal function of host DCs as a key mechanism to suppress the development of protective immune responses (Conejo-Garcia et al., 2016; Scarlett et al., 2012). We
postulated that hostile conditions within the tumor itself could
trigger ER stress not only in cancer cells but also in immune cells
that reside in the same adverse milieu. We found that dysfunctional DCs commonly present in the ovarian cancer microenvironment in both humans and rodents demonstrated robust
expression of ER stress markers and sustained activation of
the IRE1a-XBP1 arm of the UPR, compared with DCs isolated
from non-tumor locations (Cubillos-Ruiz et al., 2015). Ovarian tumor-infiltrating DCs (tDCs) showed high levels of ROS that promoted intracellular lipid peroxidation and the consequential generation of byproducts, such as 4-hydroxynonenal (4-HNE). This
highly diffusible and reactive aldehyde modified several ER-resident chaperones and proteins in DCs, thereby disrupting ER homeostasis and triggering the UPR (Cubillos-Ruiz et al., 2015). Of
note, 4-HNE has been shown to promote vascular inflammation
and atherogenesis by provoking ER stress in endothelial cells
(Vladykovskaya et al., 2012). Numerous cytotoxic drugs induce
oxidative stress and subsequent 4-HNE generation (Velez
et al., 2011), suggesting that chemotherapy might promote
immunosuppressive ER stress in myeloid cells of the tumor
microenvironment. Supporting this notion, the status of lipid peroxidation at the time of tumor resection has been proposed as a
biomarker of disease recurrence in breast cancer patients (Herrera et al., 2014). Treatment with antioxidants that control ROS
overproduction, or hydrazine derivatives capable of sequestering 4-HNE, prevented the induction of ER stress in DCs
exposed to tumor-derived factors present in ovarian cancer ascites supernatants (Cubillos-Ruiz et al., 2015). Conditional
deletion of Xbp1 in DCs resulted in delayed ovarian cancer progression in various models of primary and metastatic ovarian
cancer, and these effects were mediated by the generation of
protective T cell responses (Cubillos-Ruiz et al., 2015). XBP1
overactivation in tDCs disrupted lipid biosynthetic processes
and stimulated aberrant accumulation of triglycerides, a process
that was associated with reduced tDC antigen-presenting capacity. Accordingly, extensive functional assays revealed that
tDCs lacking XBP1 abandoned their usual tolerogenic phenotype and became immunostimulatory cells in vivo and in situ.
Interestingly, uncontrolled lipid accumulation and the generation
of oxidized fatty acids have been demonstrated to be common
regulatory features of tumor-infiltrating myeloid cells (Herber
et al., 2010; Hossain et al., 2015; Ramakrishnan et al., 2014).
Consistent with the anti-tumoral effects elicited by tDCs lacking
IRE1a or XBP1, controlling receptor-mediated lipid uptake or inhibiting fatty acid oxidation can boost anti-cancer immunity by
enhancing myeloid cell function in the tumor microenvironment
(Herber et al., 2010; Hossain et al., 2015; Ramakrishnan
et al., 2014).
The key immunoregulatory role of aberrant IRE1a-XBP1
signaling in cancer-associated myeloid cells was supported by
Gabrilovich and colleagues (Condamine et al., 2016), who found
that overexpression of ER-stress-related gene markers and surface expression of the lectin-type oxidized LDL receptor-1
(LOX-1) could effectively distinguish high-density neutrophils
from low-density immunosuppressive polymorphonuclear
MDSCs (PMN-MDSCs). More importantly, pharmacological induction of ER stress using thapsigargin triggered LOX-1 upregulation in human neutrophils and simultaneously transformed them
into immunosuppressive cells, a process that could be prevented
by targeting IRE1a-XBP1 activation using selective IRE1a inhibitors, such as B-I09 (Condamine et al., 2016; Tang et al., 2014).
Tumor-associated macrophages promote malignant progression and support chemoresistance via multiple mechanisms
(Quail and Joyce, 2013). For instance, macrophage-derived cathepsins can enhance cancer cell invasion and angiogenesis in
the tumor microenvironment (Olson and Joyce, 2015). IL-4 had
been demonstrated to control Xbp1 expression in B cells, and
this process was required for optimal IL-6 synthesis by these
lymphoid cells (Iwakoshi et al., 2003). Interestingly, IL-4 was
recently found to synergize with IL-6 or IL-10 to trigger
IRE1a-XBP1 activation in macrophages via STAT6 and STAT3,
a process that promoted cathepsin secretion (Yan et al., 2016).
Intriguingly, IRE1a was shown to be required for STAT3 phosphorylation under IL-6 stimulation (Liu et al., 2015), implying a
potential positive feedback loop between these two factors.
Transient silencing or pharmacological inhibition of IRE1a in
bone-marrow-derived macrophages stimulated with IL-6 and
IL-4 impaired cathepsin secretion and blunted macrophagemediated cancer cell invasion in vitro (Yan et al., 2016). However,
genetic evidence is needed to ascertain whether IRE1a-XBP1
signaling controls cathepsin secretion by tumor-associated
macrophages in vivo. The IRE1a-XBP1 arm of the UPR therefore
emerges as a key modulator of tumor-associated myeloid cells,
such as DCs, neutrophils, MDSCs, and macrophages (Figure 3).
In addition to the protumoral role of the IRE1a-XBP1 arm in
myeloid cells, the UPR downstream effector CHOP also oper-
ates as a regulator of MDSC activity and turnover in tumors (Thevenot et al., 2014). CHOP was initially found to control the polarization of macrophages into ‘‘alternatively activated’’ cells and to
directly regulate pro-inflammatory cytokines, such as IL-23,
IL-1b, and IL-6 (Chen et al., 2009; Goodall et al., 2010; Oh
et al., 2012). Augmented CHOP expression was recently detected in MDSCs infiltrating mouse and human tumors, a process that directly correlated with the ability of MDSC to restrain
T cell responses (Condamine et al., 2014; Thevenot et al., 2014).
Ddit3-deficient hosts challenged with a variety of cancer types
demonstrated delayed tumor growth compared with their
wild-type counterparts, and these effects were mediated by
the induction of protective CD8+ T cells (Thevenot et al., 2014).
Notably, MDSCs isolated from tumor-bearing mice devoid of
CHOP exhibited reduced immunosuppressive activity toward
T cells due to defective expression of regulatory factors, such
as arginase (Thevenot et al., 2014). Similar to their activating effects on IRE1a-XBP1 in tDCs (Cubillos-Ruiz et al., 2015), endogenous ROS also provoked CHOP overexpression in tumor-associated MDSCs (Thevenot et al., 2014), suggesting a conserved
role for ROS in the induction of ER stress responses in cancerassociated myeloid cells.
Besides PERK, other UPR-independent kinases, such as HRI,
GCN2, and PKR, can regulate the eIF2a/ATF4/CHOP signaling
axis (Hetz et al., 2013). Thus, the precise upstream drivers of
enhanced CHOP activity in cancer-associated MDSCs remain
to be defined and characterized. In addition, it is unknown
whether ER-stress-induced PERK or ATF4 could play CHOPindependent roles that impact MDSCs or additional tumorinduced immunosuppressive mechanisms. Nonetheless, recent
studies indicate that ER stress may also control MDSC survival in
tumors (Condamine et al., 2014). UPR engagement was detected in tumor-infiltrating MDSCs and promoted their apoptosis
through tumor necrosis factor (TNF)-related apoptosis-induced
ligand receptor 2 (DR5) and caspase 8 activation (Condamine
et al., 2014), suggesting that treatment with DR5 agonists could
represent a potential strategy for targeting MDSCs in cancer.
Notably, tumor-infiltrating MDSCs devoid of CHOP demonstrated delayed apoptosis and prolonged survival rates, indicating that CHOP acts as a key regulator of MDSC turnover
in vivo (Thevenot et al., 2014).
Taken together, these recent findings suggest that ER stress
responses driven by IRE1a-XBP1 signaling and CHOP are
crucial modulators of myeloid cell activity and survival in tumors
(Figure 3). It remains unknown whether the ATF6a branch of the
UPR also contributes to myeloid cell dysfunction in cancer.
Furthermore, whether ER stress responses also operate in other
major cell types of the tumor microenvironment, such as T cells,
fibroblasts, and endothelial cells, remains to be tested.
Therapeutic Strategies to Control ER Stress Responses
in Cancer
Pharmacological Inhibitors
Sustained IRE1a-XBP1 signaling promotes cancer-cell-intrinsic
growth, metastasis, and chemoresistance, but its surprising role
as a key modulator of myeloid cell function in tumors emerges
as an attractive target for cancer immunotherapy. Whereas direct
pharmacological inhibition of nuclear XBP1 is difficult due to major
Cell 168, February 9, 2017 699
Figure 3. Effects of ER Stress in Cancer-Associated Myeloid Cells
IRE1a-XBP1 overactivation in tDCs is driven by lipid peroxidation byproducts like 4-HNE. This process disrupts their lipid metabolic homeostasis and cripples
antigen presentation to T cells, thereby impeding the development of protective immune responses. Neutrophils and MDSCs use IRE1a-XBP1 and CHOP,
respectively, to express factors, such as arginase, that actively suppress T cell function. It is unknown whether IRE1a-XBP1 activation is required for the
development of MDSCs in cancer hosts. IL-4 and IL-6 signaling triggers cathepsin expression in macrophages via IRE1a-XBP1 activation to promote cancer cell
invasion. It is plausible to speculate that other tumor-associated myeloid cells, besides macrophages, could use the ER stress response to secrete factors that
promote cancer cell survival and aggressiveness. Whether macrophages and/or cancer cells also use ER stress response factors to directly inhibit T cell function
within tumor masses is unknown.
technical limitations, targeting its upstream activator, IRE1a, represents a viable strategy. Indeed, the dual enzyme IRE1a is
amenable to small-molecule targeting, and two classes of direct
inhibitors have been identified. The first group of compounds
directly targets the IRE1a endoribonuclease domain, and some
examples of this class include toyocamycin (Ri et al., 2012),
STF-083010 (Papandreou et al., 2011), 4m8C (Cross et al., 2012),
MKC-3946 (Mimura et al., 2012), and B-I09 (Tang et al., 2014;
Figure 4). Notably, these direct IRE1a endonuclease inhibitors
were capable of blocking Xbp1 splicing without affecting IRE1a
phosphorylation or the PERK and ATF6a arms of the UPR. STF083010, MKC-3946, and toyocamycin have demonstrated therapeutic efficacy in multiple myeloma xenograft models, and B-I09
has been shown to control the aggressiveness of chronic lymphocytic leukemia cells in vivo (Tang et al., 2014). Importantly, daily
intraperitoneal administration of 4m8C substantially decreased
pathological joint swelling in the KBxN serum transfer murine
model of rheumatoid arthritis (Qiu et al., 2013). This suggests
that this class of inhibitors could be also used in cancer hosts to
modulate the function of intratumoral myeloid cells.
The second group of inhibitors targets the IRE1a kinase
domain in order to allosterically disrupt its endoribonuclease
function (Figure 4). A recent compound in this category is
KIRA6 (Ghosh et al., 2014), which reduced plasma glucose levels
and improved glucose tolerance in Ins2+/Akita mice that exhibit
chronic ER stress in pancreatic b cells (Ghosh et al., 2014). Moreover, intravitreal KIRA6 injection in the P23H transgenic rat
700 Cell 168, February 9, 2017
model of retinitis pigmentosa preserved photoreceptor viability
and function (Ghosh et al., 2014). Nevertheless, it has not been
determined whether treatment with IRE1a inhibitors fully recapitulates the biological effects of IRE1a genetic ablation. Developing novel IRE1a inhibitors with potent in vivo efficacy in the
tumor microenvironment could therefore be useful to directly
restrain cancer cell survival, metastasis, and chemoresistance
while eliciting protective anti-tumor immune responses via
myeloid cell reprogramming.
Because targeting CHOP in the nucleus using small-molecule
inhibitors would also involve major technical challenges,
restraining the activity of its upstream activators, PERK or
eIF2a, may represent a more practical approach (Figure 4).
GSK2606414 was the first reported PERK inhibitor (Axten
et al., 2012) and was found to be neuroprotective in mouse
models of prion disease (Moreno et al., 2013). Another ATPcompetitive inhibitor of PERK enzymatic activity, GSK2656157,
was shown to impede ER-stress-induced PERK autophosphorylation, eIF2a phosphorylation, and subsequent overexpression
of ATF4 and CHOP in multiple cell lines (Atkins et al., 2013). Oral
administration of GSK2656157 to mice impaired PERK autophosphorylation in the pancreas and compromised xenograft tumor growth in immunodeficient hosts (Atkins et al., 2013). However, further studies indicate that inhibition of PERK activity by
GSK2656157 does not always correlate with reduced eIF2a
phosphorylation and that this inhibitor fails to recreate the biological effects of PERK genetic inactivation (Krishnamoorthy
Figure 4. Therapeutic Strategies to Control
ER Stress Responses in Cancer
The IRE1a kinase domain can be inactivated using
compounds like KIRA6, leading to allosteric inhibition of the IRE1a RNase domain. Other compounds, such as 4m8c, MKC-3946, and B-I09, can
directly inhibit the IRE1a RNase domain to prevent
splicing of the Xbp1 mRNA. Treatment with these
IRE1a inhibitory compounds could be effective
to reduce tolerance to hypoxia, angiogenesis,
drug resistance, and metastatic capacity by cancer cells. These compounds could also be used
to reprogram the function of cancer-associated
myeloid cells, including macrophages, DCs, and
neutrophils. Furthermore, siRNA-loaded nanoparticles could be effective for selectively silencing
ERN1 or XBP1 in ovarian-cancer-associated DCs.
Novel genome-editing technologies could be
used to ablate IRE1a-XBP1 signaling in DCbased therapeutic vaccines for cancer. Recently
developed inhibitors of PERK and eIF2a could be
exploited to control CHOP overexpression by
MDSCs and relieve immunosuppression in the tumor microenvironment. However, CHOP-independent roles of PERK/eIF2a in MDSCs have not
been explored. Whether the ATF6a arm of the UPR
also operates as a modulator of immune cell
function in tumors has not been determined.
et al., 2014). The integrated stress response inhibitor (ISRIB) is a
symmetric bisglycolamide that renders cells resistant to eIF2a
phosphorylation, thereby blocking the activation of ATF4 and
the accumulation of CHOP during conditions of ER stress
(Sidrauski et al., 2015). Importantly, this compound showed significant in vivo effects by enhancing spatial and fear-associated
learning in rodents (Sidrauski et al., 2015). Whether GSK2656157
or ISRIB could modulate the function or survival of MDSCs in the
tumor microenvironment by impeding PERK/eIF2a-mediated
CHOP activation is yet to be tested. Given the importance of
PERK and IRE1a-XBP1 signaling in organ homeostasis of highly
secretory tissues, careful optimization of these inhibitory compounds for in vivo use is essential to minimize potential side effects and toxicity in treated hosts.
Ceapins, a new class of pyrazole amides, were recently demonstrated to specifically inhibit the ATF6a branch of the UPR by
blocking ATF6a processing and nuclear translocation in cells undergoing ER stress (Gallagher and Walter, 2016). Further optimization of Ceapins for in vivo use will hence be critical for determining the tumoricidal activity of these compounds alone or in
combination with other agents that ablate PERK and/or IRE1a
signaling in the tumor microenvironment. Additionally, developing
new pharmacological interventions capable of triggering lethal ER
stress and subsequent ICD selectively in malignant cells could
also be useful for eliciting robust anti-tumor immune responses.
Controlling Immune-Cell-Intrinsic ER Stress Responses
Gene-targeting strategies have proven effective for therapeutically disabling detrimental IRE1a-XBP1 signaling in DCs of can-
cer hosts (Cubillos-Ruiz et al., 2015). In
the ovarian cancer microenvironment,
DCs exhibit a remarkable phagocytic capacity that renders them exceptional targets for nanoparticle-mediated RNAi
therapeutics (Cubillos-Ruiz et al., 2009, 2012). Because ovarian
cancer metastasis and malignant ascites accumulation are
confined within the peritoneal cavity, administration of DC-targeting small interfering RNA (siRNA)-loaded nanocarriers in
this anatomical location represents a novel and feasible immunotherapeutic strategy. In preclinical models of metastatic
ovarian cancer, silencing Xbp1 or Ern1 using this approach
transformed tolerogenic tDCs into highly immunostimulatory
cells that extended host survival by evoking T-cell-mediated
anti-tumor immunity (Cubillos-Ruiz et al., 2015).
As a second strategy, we propose that IRE1a-XBP1
signaling could be genetically interrupted to enhance the efficacy of DC-based therapeutic vaccines in ovarian cancer,
which unfortunately have shown limited success in recent clinical trials (Kandalaft et al., 2013). In proof-of-concept experiments, we found that transferring Xbp1-deficient BMDCs intraperitoneally into mice bearing established ovarian cancer
significantly delayed tumor progression compared with infusion
of wild-type BMDCs (Cubillos-Ruiz et al., 2015). Notably, transplanted Xbp1-deficient DCs were dominantly immunostimulatory over the endogenous (wild-type) regulatory DCs residing
in the tumor microenvironment. Cutting-edge genome-editing
technologies, such as CRISPR/Cas9, zinc finger nucleases,
or TALENs (Gaj et al., 2013), should therefore enable precise
and efficient inactivation of XBP1 or ERN1 in DCs prior to
adoptive transfer (Figure 4), thereby protecting these transplanted DCs from the suppressive effects of aberrant ER stress
responses in the tumor microenvironment. Interestingly,
Cell 168, February 9, 2017 701
transplanted Ddit3-deficient MDSCs also show enhanced antigen-presenting capacity and potent T cell stimulatory capacity
(Thevenot et al., 2014), suggesting that the gene-targeting strategies described above may also be useful to re-program
MDSC function in tumors.
Conclusions
Tumors thrive under adverse conditions, such as hypoxia,
nutrient starvation, and oxidative stress, by adjusting their protein-folding capacity via the ER stress response pathway. Activation of multiple ER stress sensors has been demonstrated to
endow malignant cells with greater tumorigenic, metastatic,
and drug-resistant capacity. However, recent studies have uncovered a second mechanism by which abnormal ER stress responses promote malignant progression: by subverting the
protective function of innate immune cells in the tumor microenvironment to cripple the development of anti-tumor immunity. Harnessing the intrinsic ability of our immune system to
recognize and eliminate malignant cells represents an extraordinarily promising anti-cancer strategy, especially when combined with recently developed therapeutics, such as Gleevec,
that precisely target genetic driver mutations in the tumor
itself. Together, these two approaches offer the most exciting
opportunities for cancer treatment since the development of
chemotherapy. However, hostile microenvironmental conditions within aggressive solid tumors inhibit the optimal activity
of protective immune cells. Targeting immunosuppression and
re-programming immune cell function in the tumor microenvironment are fundamental requirements for developing
successful cancer immunotherapies. Abnormal ER stress responses emerge as critical regulators of immune cell function
in the tumor microenvironment and appear to integrate protumoral and immunosuppressive mechanisms in cancer hosts.
Therefore, identifying, understanding, and disabling the precise
molecular mechanisms by which ER stress inhibits the natural
function of innate immune cells in tumors could be a novel
approach to complement and enhance the efficacy of both
standard chemotherapies and evolving cancer immunotherapies, such as checkpoint blockade and adoptive T cell transfer in the clinic.
AUTHOR CONTRIBUTIONS
J.R.C.-R., S.E.B., and L.H.G. co-wrote and edited the manuscript.
ACKNOWLEDGMENTS
Our research was supported by the Irvington Institute Fellowship Program of
the Cancer Research Institute (J.R.C.-R.), the Ann Schreiber Mentored Investigator Award of the Ovarian Cancer Research Fund Alliance (J.R.C.-R.), the
Ovarian Cancer Academy Early-Career Investigator Award of the Department
of Defense (W81XWH-16-1-0438 to J.R.C.-R.), Stand Up to Cancer Innovative
Research Grant (SU2C-AACR-1RG-03-16 to J.R.C.-R.), the Clinic and Laboratory Integration Program of the Cancer Research Institute (J.R.C.-R.), Weill
Cornell Medical College Funds (L.H.G.), and NIH grant R01CA112663
(L.H.G.). We apologize to colleagues whose work was not cited in this review
due to space limitations. J.R.C.-R. and L.H.G. are co-founders of and scientific
advisors for Quentis Therapeutics. S.E.B. is co-founder and employee of
Quentis Therapeutics. L.H.G. also serves on the board of directors of and
holds equity in Bristol-Myers Squibb.
702 Cell 168, February 9, 2017
REFERENCES
Atkins, C., Liu, Q., Minthorn, E., Zhang, S.Y., Figueroa, D.J., Moss, K., Stanley,
T.B., Sanders, B., Goetz, A., Gaul, N., et al. (2013). Characterization of a novel
PERK kinase inhibitor with antitumor and antiangiogenic activity. Cancer Res.
73, 1993–2002.
Auf, G., Jabouille, A., Guérit, S., Pineau, R., Delugin, M., Bouchecareilh, M.,
Magnin, N., Favereaux, A., Maitre, M., Gaiser, T., et al. (2010). Inositolrequiring enzyme 1alpha is a key regulator of angiogenesis and invasion in malignant glioma. Proc. Natl. Acad. Sci. USA 107, 15553–15558.
Avivar-Valderas, A., Salas, E., Bobrovnikova-Marjon, E., Diehl, J.A., Nagi, C.,
Debnath, J., and Aguirre-Ghiso, J.A. (2011). PERK integrates autophagy and
oxidative stress responses to promote survival during extracellular matrix
detachment. Mol. Cell. Biol. 31, 3616–3629.
Axten, J.M., Medina, J.R., Feng, Y., Shu, A., Romeril, S.P., Grant, S.W., Li,
W.H., Heerding, D.A., Minthorn, E., Mencken, T., et al. (2012). Discovery of
7-methyl-5-(1-[3-(trifluoromethyl)phenyl]acetyl-2,3-dihydro-1H-indol-5-yl)-7Hpyrrolo[2,3-d]pyrimidin-4-amine (GSK2606414), a potent and selective first-inclass inhibitor of protein kinase R (PKR)-like endoplasmic reticulum kinase
(PERK). J. Med. Chem. 55, 7193–7207.
B’chir, W., Maurin, A.C., Carraro, V., Averous, J., Jousse, C., Muranishi, Y.,
Parry, L., Stepien, G., Fafournoux, P., and Bruhat, A. (2013). The eIF2a/ATF4
pathway is essential for stress-induced autophagy gene expression. Nucleic
Acids Res. 41, 7683–7699.
Bartkowiak, K., Effenberger, K.E., Harder, S., Andreas, A., Buck, F., Peter-Katalinic, J., Pantel, K., and Brandt, B.H. (2010). Discovery of a novel unfolded
protein response phenotype of cancer stem/progenitor cells from the bone
marrow of breast cancer patients. J. Proteome Res. 9, 3158–3168.
Bartkowiak, K., Kwiatkowski, M., Buck, F., Gorges, T.M., Nilse, L., Assmann,
V., Andreas, A., Müller, V., Wikman, H., Riethdorf, S., et al. (2015). Disseminated tumor cells persist in the bone marrow of breast cancer patients through
sustained activation of the unfolded protein response. Cancer Res. 75,
5367–5377.
Beck, D., Niessner, H., Smalley, K.S., Flaherty, K., Paraiso, K.H., Busch, C.,
Sinnberg, T., Vasseur, S., Iovanna, J.L., Drießen, S., et al. (2013). Vemurafenib
potently induces endoplasmic reticulum stress-mediated apoptosis in
BRAFV600E melanoma cells. Sci. Signal. 6, ra7.
Bernales, S., McDonald, K.L., and Walter, P. (2006). Autophagy counterbalances endoplasmic reticulum expansion during the unfolded protein
response. PLoS Biol. 4, e423.
Bertolotti, A., Zhang, Y., Hendershot, L.M., Harding, H.P., and Ron, D. (2000).
Dynamic interaction of BiP and ER stress transducers in the unfolded-protein
response. Nat. Cell Biol. 2, 326–332.
Bettigole, S.E., and Glimcher, L.H. (2015). Endoplasmic reticulum stress in immunity. Annu. Rev. Immunol. 33, 107–138.
Bettigole, S.E., Lis, R., Adoro, S., Lee, A.H., Spencer, L.A., Weller, P.F., and
Glimcher, L.H. (2015). The transcription factor XBP1 is selectively required
for eosinophil differentiation. Nat. Immunol. 16, 829–837.
Bi, M., Naczki, C., Koritzinsky, M., Fels, D., Blais, J., Hu, N., Harding, H., Novoa, I., Varia, M., Raleigh, J., et al. (2005). ER stress-regulated translation increases tolerance to extreme hypoxia and promotes tumor growth. EMBO J.
24, 3470–3481.
Blais, J.D., Addison, C.L., Edge, R., Falls, T., Zhao, H., Wary, K., Koumenis, C.,
Harding, H.P., Ron, D., Holcik, M., and Bell, J.C. (2006). Perk-dependent translational regulation promotes tumor cell adaptation and angiogenesis in
response to hypoxic stress. Mol. Cell. Biol. 26, 9517–9532.
Cerezo, M., Lehraiki, A., Millet, A., Rouaud, F., Plaisant, M., Jaune, E., Botton,
T., Ronco, C., Abbe, P., Amdouni, H., et al. (2016). Compounds triggering ER
stress exert anti-melanoma effects and overcome BRAF inhibitor resistance.
Cancer Cell 29, 805–819.
Chen, L., Jarujaron, S., Wu, X., Sun, L., Zha, W., Liang, G., Wang, X., Gurley,
E.C., Studer, E.J., Hylemon, P.B., et al. (2009). HIV protease inhibitor lopinavir-induced TNF-alpha and IL-6 expression is coupled to the unfolded protein
response and ERK signaling pathways in macrophages. Biochem. Pharmacol.
78, 70–77.
Anti-oncogenic role of the endoplasmic reticulum differentially activated by
mutations in the MAPK pathway. Nat. Cell Biol. 8, 1053–1063.
Chen, X., Iliopoulos, D., Zhang, Q., Tang, Q., Greenblatt, M.B., Hatziapostolou,
M., Lim, E., Tam, W.L., Ni, M., Chen, Y., et al. (2014). XBP1 promotes triplenegative breast cancer by controlling the HIF1a pathway. Nature 508,
103–107.
Dey, S., Sayers, C.M., Verginadis, I.I., Lehman, S.L., Cheng, Y., Cerniglia, G.J.,
Tuttle, S.W., Feldman, M.D., Zhang, P.J., Fuchs, S.Y., et al. (2015). ATF4-dependent induction of heme oxygenase 1 prevents anoikis and promotes
metastasis. J. Clin. Invest. 125, 2592–2608.
Chitnis, N.S., Pytel, D., Bobrovnikova-Marjon, E., Pant, D., Zheng, H., Maas,
N.L., Frederick, B., Kushner, J.A., Chodosh, L.A., Koumenis, C., et al. (2012).
miR-211 is a prosurvival microRNA that regulates chop expression in a
PERK-dependent manner. Mol. Cell 48, 353–364.
Feng, Y.X., Sokol, E.S., Del Vecchio, C.A., Sanduja, S., Claessen, J.H., Proia,
T.A., Jin, D.X., Reinhardt, F., Ploegh, H.L., Wang, Q., and Gupta, P.B. (2014).
Epithelial-to-mesenchymal transition activates PERK-eIF2a and sensitizes
cells to endoplasmic reticulum stress. Cancer Discov. 4, 702–715.
Condamine, T., Kumar, V., Ramachandran, I.R., Youn, J.I., Celis, E., Finnberg,
N., El-Deiry, W.S., Winograd, R., Vonderheide, R.H., English, N.R., et al. (2014).
ER stress regulates myeloid-derived suppressor cell fate through TRAILR-mediated apoptosis. J. Clin. Invest. 124, 2626–2639.
Fucikova, J., Becht, E., Iribarren, K., Goc, J., Remark, R., Damotte, D., Alifano,
M., Devi, P., Biton, J., Germain, C., et al. (2016). Calreticulin expression in human non-small cell lung cancers correlates with increased accumulation of
antitumor immune cells and favorable prognosis. Cancer Res. 76, 1746–1756.
Condamine, T., Dominguez, G.A., Youn, J.-I., Kossenkov, A.V., Mony, S., Alicea-Torres, K., Tcyganov, E., Hashimoto, A., Nefedova, Y., Lin, C., et al.
(2016). Lectin-type oxidized LDL receptor-1 distinguishes population of human polymorphonuclear myeloid-derived suppressor cells in cancer patients.
Sci. Immunol. 1, aaf8943.
Gaj, T., Gersbach, C.A., and Barbas, C.F., 3rd. (2013). ZFN, TALEN, and
CRISPR/Cas-based methods for genome engineering. Trends Biotechnol.
31, 397–405.
Conejo-Garcia, J.R., Rutkowski, M.R., and Cubillos-Ruiz, J.R. (2016). Stateof-the-art of regulatory dendritic cells in cancer. Pharmacol. Ther. 164,
97–104.
Corazzari, M., Rapino, F., Ciccosanti, F., Giglio, P., Antonioli, M., Conti, B., Fimia, G.M., Lovat, P.E., and Piacentini, M. (2015). Oncogenic BRAF induces
chronic ER stress condition resulting in increased basal autophagy and
apoptotic resistance of cutaneous melanoma. Cell Death Differ. 22, 946–958.
Crespo, J., Sun, H., Welling, T.H., Tian, Z., and Zou, W. (2013). T cell anergy,
exhaustion, senescence, and stemness in the tumor microenvironment.
Curr. Opin. Immunol. 25, 214–221.
Cross, B.C., Bond, P.J., Sadowski, P.G., Jha, B.K., Zak, J., Goodman, J.M.,
Silverman, R.H., Neubert, T.A., Baxendale, I.R., Ron, D., and Harding, H.P.
(2012). The molecular basis for selective inhibition of unconventional mRNA
splicing by an IRE1-binding small molecule. Proc. Natl. Acad. Sci. USA 109,
E869–E878.
Cubillos-Ruiz, J.R., Fiering, S., and Conejo-Garcia, J.R. (2009). Nanomolecular
targeting of dendritic cells for ovarian cancer therapy. Future Oncol. 5,
1189–1192.
Cubillos-Ruiz, J.R., Baird, J.R., Tesone, A.J., Rutkowski, M.R., Scarlett, U.K.,
Camposeco-Jacobs, A.L., Anadon-Arnillas, J., Harwood, N.M., Korc, M., Fiering, S.N., et al. (2012). Reprogramming tumor-associated dendritic cells in vivo
using miRNA mimetics triggers protective immunity against ovarian cancer.
Cancer Res. 72, 1683–1693.
Cubillos-Ruiz, J.R., Silberman, P.C., Rutkowski, M.R., Chopra, S., Perales-Puchalt, A., Song, M., Zhang, S., Bettigole, S.E., Gupta, D., Holcomb, K., et al.
(2015). ER stress sensor XBP1 controls anti-tumor immunity by disrupting dendritic cell homeostasis. Cell 161, 1527–1538.
Cullinan, S.B., Zhang, D., Hannink, M., Arvisais, E., Kaufman, R.J., and Diehl,
J.A. (2003). Nrf2 is a direct PERK substrate and effector of PERK-dependent
cell survival. Mol. Cell. Biol. 23, 7198–7209.
Dalton, L.E., Clarke, H.J., Knight, J., Lawson, M.H., Wason, J., Lomas, D.A.,
Howat, W.J., Rintoul, R.C., Rassl, D.M., and Marciniak, S.J. (2013). The endoplasmic reticulum stress marker CHOP predicts survival in malignant mesothelioma. Br. J. Cancer 108, 1340–1347.
Davies, M.P., Barraclough, D.L., Stewart, C., Joyce, K.A., Eccles, R.M., Barraclough, R., Rudland, P.S., and Sibson, D.R. (2008). Expression and splicing of
the unfolded protein response gene XBP-1 are significantly associated with
clinical outcome of endocrine-treated breast cancer. Int. J. Cancer 123, 85–88.
Gallagher, C.M., and Walter, P. (2016). Ceapins inhibit ATF6a signaling by
selectively preventing transport of ATF6a to the Golgi apparatus during ER
stress. eLife 5, e11880.
Ghosh, R., Lipson, K.L., Sargent, K.E., Mercurio, A.M., Hunt, J.S., Ron, D., and
Urano, F. (2010). Transcriptional regulation of VEGF-A by the unfolded protein
response pathway. PLoS ONE 5, e9575.
Ghosh, R., Wang, L., Wang, E.S., Perera, B.G., Igbaria, A., Morita, S., Prado,
K., Thamsen, M., Caswell, D., Macias, H., et al. (2014). Allosteric inhibition of
the IRE1a RNase preserves cell viability and function during endoplasmic reticulum stress. Cell 158, 534–548.
Gomez-Cadena, A., Urueña, C., Prieto, K., Martinez-Usatorre, A., Donda, A.,
Barreto, A., Romero, P., and Fiorentino, S. (2016). Immune-system-dependent
anti-tumor activity of a plant-derived polyphenol rich fraction in a melanoma
mouse model. Cell Death Dis. 7, e2243.
Goodall, J.C., Wu, C., Zhang, Y., McNeill, L., Ellis, L., Saudek, V., and Gaston,
J.S. (2010). Endoplasmic reticulum stress-induced transcription factor, CHOP,
is crucial for dendritic cell IL-23 expression. Proc. Natl. Acad. Sci. USA 107,
17698–17703.
Grivennikov, S.I., and Karin, M. (2010). Dangerous liaisons: STAT3 and
NF-kappaB collaboration and crosstalk in cancer. Cytokine Growth Factor
Rev. 21, 11–19.
Guzy, R.D., Hoyos, B., Robin, E., Chen, H., Liu, L., Mansfield, K.D., Simon,
M.C., Hammerling, U., and Schumacker, P.T. (2005). Mitochondrial complex
III is required for hypoxia-induced ROS production and cellular oxygen
sensing. Cell Metab. 1, 401–408.
Han, J., Back, S.H., Hur, J., Lin, Y.H., Gildersleeve, R., Shan, J., Yuan, C.L.,
Krokowski, D., Wang, S., Hatzoglou, M., et al. (2013). ER-stress-induced transcriptional regulation increases protein synthesis leading to cell death. Nat.
Cell Biol. 15, 481–490.
Harding, H.P., Zhang, Y., and Ron, D. (1999). Protein translation and folding
are coupled by an endoplasmic-reticulum-resident kinase. Nature 397,
271–274.
Hart, L.S., Cunningham, J.T., Datta, T., Dey, S., Tameire, F., Lehman, S.L., Qiu,
B., Zhang, H., Cerniglia, G., Bi, M., et al. (2012). ER stress-mediated autophagy
promotes Myc-dependent transformation and tumor growth. J. Clin. Invest.
122, 4621–4634.
Herber, D.L., Cao, W., Nefedova, Y., Novitskiy, S.V., Nagaraj, S., Tyurin, V.A.,
Corzo, A., Cho, H.I., Celis, E., Lennox, B., et al. (2010). Lipid accumulation and
dendritic cell dysfunction in cancer. Nat. Med. 16, 880–886.
De Palma, M., Murdoch, C., Venneri, M.A., Naldini, L., and Lewis, C.E. (2007).
Tie2-expressing monocytes: regulation of tumor angiogenesis and therapeutic
implications. Trends Immunol. 28, 519–524.
Herrera, A.C., Victorino, V.J., Campos, F.C., Verenitach, B.D., Lemos, L.T.,
Aranome, A.M., Oliveira, S.R., Cecchini, A.L., Simão, A.N., Abdelhay, E.,
et al. (2014). Impact of tumor removal on the systemic oxidative profile of patients with breast cancer discloses lipid peroxidation at diagnosis as a putative
marker of disease recurrence. Clin. Breast Cancer 14, 451–459.
Denoyelle, C., Abou-Rjaily, G., Bezrookove, V., Verhaegen, M., Johnson, T.M.,
Fullen, D.R., Pointer, J.N., Gruber, S.B., Su, L.D., Nikiforov, M.A., et al. (2006).
Hetz, C. (2012). The unfolded protein response: controlling cell fate decisions
under ER stress and beyond. Nat. Rev. Mol. Cell Biol. 13, 89–102.
Cell 168, February 9, 2017 703
Hetz, C., Chevet, E., and Harding, H.P. (2013). Targeting the unfolded protein
response in disease. Nat. Rev. Drug Discov. 12, 703–719.
Hollien, J., Lin, J.H., Li, H., Stevens, N., Walter, P., and Weissman, J.S. (2009).
Regulated Ire1-dependent decay of messenger RNAs in mammalian cells.
J. Cell Biol. 186, 323–331.
Hossain, F., Al-Khami, A.A., Wyczechowska, D., Hernandez, C., Zheng, L., Reiss, K., Valle, L.D., Trillo-Tinoco, J., Maj, T., Zou, W., et al. (2015). Inhibition of
fatty acid oxidation modulates immunosuppressive functions of myeloidderived suppressor cells and enhances cancer therapies. Cancer Immunol.
Res. 3, 1236–1247.
Hu, R., Chen, Z.F., Yan, J., Li, Q.F., Huang, Y., Xu, H., Zhang, X.P., and Jiang,
H. (2015). Endoplasmic reticulum stress of neutrophils is required for ischemia/
reperfusion-induced acute lung injury. J. Immunol. 195, 4802–4809.
Lerner, A.G., Upton, J.P., Praveen, P.V., Ghosh, R., Nakagawa, Y., Igbaria, A.,
Shen, S., Nguyen, V., Backes, B.J., Heiman, M., et al. (2012). IRE1a induces
thioredoxin-interacting protein to activate the NLRP3 inflammasome and
promote programmed cell death under irremediable ER stress. Cell Metab.
16, 250–264.
Leung-Hagesteijn, C., Erdmann, N., Cheung, G., Keats, J.J., Stewart, A.K., Reece, D.E., Chung, K.C., and Tiedemann, R.E. (2013). Xbp1s-negative tumor
B cells and pre-plasmablasts mediate therapeutic proteasome inhibitor resistance in multiple myeloma. Cancer Cell 24, 289–304.
Lin, J.H., Li, H., Yasumura, D., Cohen, H.R., Zhang, C., Panning, B., Shokat,
K.M., Lavail, M.M., and Walter, P. (2007). IRE1 signaling affects cell fate during
the unfolded protein response. Science 318, 944–949.
Liu, Y., Shao, M., Wu, Y., Yan, C., Jiang, S., Liu, J., Dai, J., Yang, L., Li, J., Jia,
W., et al. (2015). Role for the endoplasmic reticulum stress sensor IRE1a in liver
regenerative responses. J. Hepatol. 62, 590–598.
Huber, A.L., Lebeau, J., Guillaumot, P., Pétrilli, V., Malek, M., Chilloux, J., Fauvet, F., Payen, L., Kfoury, A., Renno, T., et al. (2013). p58(IPK)-mediated
attenuation of the proapoptotic PERK-CHOP pathway allows malignant progression upon low glucose. Mol. Cell 49, 1049–1059.
Lu, Y., Liang, F.X., and Wang, X. (2014). A synthetic biology approach identifies
the mammalian UPR RNA ligase RtcB. Mol. Cell 55, 758–770.
Iwakoshi, N.N., Lee, A.H., Vallabhajosyula, P., Otipoby, K.L., Rajewsky, K., and
Glimcher, L.H. (2003). Plasma cell differentiation and the unfolded protein
response intersect at the transcription factor XBP-1. Nat. Immunol. 4,
321–329.
Ma, X.H., Piao, S.F., Dey, S., McAfee, Q., Karakousis, G., Villanueva, J., Hart,
L.S., Levi, S., Hu, J., Zhang, G., et al. (2014). Targeting ER stress-induced autophagy overcomes BRAF inhibitor resistance in melanoma. J. Clin. Invest. 124,
1406–1417.
Iwakoshi, N.N., Pypaert, M., and Glimcher, L.H. (2007). The transcription factor
XBP-1 is essential for the development and survival of dendritic cells. J. Exp.
Med. 204, 2267–2275.
Magnaghi, P., D’Alessio, R., Valsasina, B., Avanzi, N., Rizzi, S., Asa, D., Gasparri, F., Cozzi, L., Cucchi, U., Orrenius, C., et al. (2013). Covalent and allosteric
inhibitors of the ATPase VCP/p97 induce cancer cell death. Nat. Chem. Biol. 9,
548–556.
Jabouille, A., Delugin, M., Pineau, R., Dubrac, A., Soulet, F., Lhomond, S., Pallares-Lupon, N., Prats, H., Bikfalvi, A., Chevet, E., et al. (2015). Glioblastoma
invasion and cooption depend on IRE1a endoribonuclease activity. Oncotarget 6, 24922–24934.
Jeon, Y.J., Khelifa, S., Ratnikov, B., Scott, D.A., Feng, Y., Parisi, F., Ruller, C.,
Lau, E., Kim, H., Brill, L.M., et al. (2015). Regulation of glutamine carrier proteins by RNF5 determines breast cancer response to ER stress-inducing chemotherapies. Cancer Cell 27, 354–369.
Mahadevan, N.R., Rodvold, J., Sepulveda, H., Rossi, S., Drew, A.F., and Zanetti, M. (2011). Transmission of endoplasmic reticulum stress and pro-inflammation from tumor cells to myeloid cells. Proc. Natl. Acad. Sci. USA 108,
6561–6566.
Mahadevan, N.R., Anufreichik, V., Rodvold, J.J., Chiu, K.T., Sepulveda, H.,
and Zanetti, M. (2012). Cell-extrinsic effects of tumor ER stress imprint myeloid
dendritic cells and impair CD8+ T cell priming. PLoS ONE 7, e51845.
Kandalaft, L.E., Powell, D.J., Jr., Chiang, C.L., Tanyi, J., Kim, S., Bosch, M.,
Montone, K., Mick, R., Levine, B.L., Torigian, D.A., et al. (2013). Autologous
lysate-pulsed dendritic cell vaccination followed by adoptive transfer of vaccine-primed ex vivo co-stimulated T cells in recurrent ovarian cancer. OncoImmunology 2, e22664.
Mantovani, A., Allavena, P., Sica, A., and Balkwill, F. (2008). Cancer-related
inflammation. Nature 454, 436–444.
Karali, E., Bellou, S., Stellas, D., Klinakis, A., Murphy, C., and Fotsis, T. (2014).
VEGF signals through ATF6 and PERK to promote endothelial cell survival and
angiogenesis in the absence of ER stress. Mol. Cell 54, 559–572.
Martinon, F., Chen, X., Lee, A.H., and Glimcher, L.H. (2010). TLR activation of
the transcription factor XBP1 regulates innate immune responses in macrophages. Nat. Immunol. 11, 411–418.
Kepp, O., Menger, L., Vacchelli, E., Locher, C., Adjemian, S., Yamazaki, T.,
Martins, I., Sukkurwala, A.Q., Michaud, M., Senovilla, L., et al. (2013). Crosstalk between ER stress and immunogenic cell death. Cytokine Growth Factor
Rev. 24, 311–318.
Matsuo, K., Gray, M.J., Yang, D.Y., Srivastava, S.A., Tripathi, P.B., Sonoda,
L.A., Yoo, E.J., Dubeau, L., Lee, A.S., and Lin, Y.G. (2013). The endoplasmic
reticulum stress marker, glucose-regulated protein-78 (GRP78) in visceral adipocytes predicts endometrial cancer progression and patient survival. Gynecol. Oncol. 128, 552–559.
Kitai, Y., Ariyama, H., Kono, N., Oikawa, D., Iwawaki, T., and Arai, H. (2013).
Membrane lipid saturation activates IRE1a without inducing clustering. Genes
Cells 18, 798–809.
Kouroku, Y., Fujita, E., Tanida, I., Ueno, T., Isoai, A., Kumagai, H., Ogawa, S.,
Kaufman, R.J., Kominami, E., and Momoi, T. (2007). ER stress (PERK/eIF2alpha
phosphorylation) mediates the polyglutamine-induced LC3 conversion, an
essential step for autophagy formation. Cell Death Differ. 14, 230–239.
Krishnamoorthy, J., Rajesh, K., Mirzajani, F., Kesoglidou, P., Papadakis, A.I.,
and Koromilas, A.E. (2014). Evidence for eIF2a phosphorylation-independent
effects of GSK2656157, a novel catalytic inhibitor of PERK with clinical implications. Cell Cycle 13, 801–806.
Marada, S., Stewart, D.P., Bodeen, W.J., Han, Y.G., and Ogden, S.K. (2013).
The unfolded protein response selectively targets active smoothened mutants.
Mol. Cell. Biol. 33, 2375–2387.
Mimura, N., Fulciniti, M., Gorgun, G., Tai, Y.T., Cirstea, D., Santo, L., Hu, Y.,
Fabre, C., Minami, J., Ohguchi, H., et al. (2012). Blockade of XBP1 splicing
by inhibition of IRE1a is a promising therapeutic option in multiple myeloma.
Blood 119, 5772–5781.
Moreno, J.A., Halliday, M., Molloy, C., Radford, H., Verity, N., Axten, J.M., Ortori, C.A., Willis, A.E., Fischer, P.M., Barrett, D.A., and Mallucci, G.R. (2013).
Oral treatment targeting the unfolded protein response prevents neurodegeneration and clinical disease in prion-infected mice. Sci. Transl. Med. 5,
206ra138.
Lee, A.H., Iwakoshi, N.N., Anderson, K.C., and Glimcher, L.H. (2003). Proteasome inhibitors disrupt the unfolded protein response in myeloma cells. Proc.
Natl. Acad. Sci. USA 100, 9946–9951.
Mujcic, H., Nagelkerke, A., Rouschop, K.M., Chung, S., Chaudary, N., Span,
P.N., Clarke, B., Milosevic, M., Sykes, J., Hill, R.P., et al. (2013). Hypoxic activation of the PERK/eIF2a arm of the unfolded protein response promotes
metastasis through induction of LAMP3. Clin. Cancer Res. 19, 6126–6137.
Lee, B.R., Chang, S.Y., Hong, E.H., Kwon, B.E., Kim, H.M., Kim, Y.J., Lee, J.,
Cho, H.J., Cheon, J.H., and Ko, H.J. (2014). Elevated endoplasmic reticulum
stress reinforced immunosuppression in the tumor microenvironment via
myeloid-derived suppressor cells. Oncotarget 5, 12331–12345.
Nakagawa, H., Umemura, A., Taniguchi, K., Font-Burgada, J., Dhar, D., Ogata,
H., Zhong, Z., Valasek, M.A., Seki, E., Hidalgo, J., et al. (2014). ER stress cooperates with hypernutrition to trigger TNF-dependent spontaneous HCC
development. Cancer Cell 26, 331–343.
704 Cell 168, February 9, 2017
Namba, T., Chu, K., Kodama, R., Byun, S., Yoon, K.W., Hiraki, M., Mandinova,
A., and Lee, S.W. (2015). Loss of p53 enhances the function of the endoplasmic reticulum through activation of the IRE1a/XBP1 pathway. Oncotarget
6, 19990–20001.
Nieto, M.A., Huang, R.Y., Jackson, R.A., and Thiery, J.P. (2016). Emt: 2016.
Cell 166, 21–45.
Obeng, E.A., Carlson, L.M., Gutman, D.M., Harrington, W.J., Jr., Lee, K.P., and
Boise, L.H. (2006). Proteasome inhibitors induce a terminal unfolded protein
response in multiple myeloma cells. Blood 107, 4907–4916.
Ogata, M., Hino, S., Saito, A., Morikawa, K., Kondo, S., Kanemoto, S., Murakami, T., Taniguchi, M., Tanii, I., Yoshinaga, K., et al. (2006). Autophagy is activated for cell survival after endoplasmic reticulum stress. Mol. Cell. Biol. 26,
9220–9231.
Oh, J., Riek, A.E., Weng, S., Petty, M., Kim, D., Colonna, M., Cella, M., and
Bernal-Mizrachi, C. (2012). Endoplasmic reticulum stress controls M2
macrophage differentiation and foam cell formation. J. Biol. Chem. 287,
11629–11641.
Olson, O.C., and Joyce, J.A. (2015). Cysteine cathepsin proteases: regulators
of cancer progression and therapeutic response. Nat. Rev. Cancer 15,
712–729.
Papandreou, I., Denko, N.C., Olson, M., Van Melckebeke, H., Lust, S., Tam, A.,
Solow-Cordero, D.E., Bouley, D.M., Offner, F., Niwa, M., and Koong, A.C.
(2011). Identification of an Ire1alpha endonuclease specific inhibitor with cytotoxic activity against human multiple myeloma. Blood 117, 1311–1314.
tion of glutathione synthesis and protection against ROS. Proc. Natl. Acad. Sci.
USA 110, 4622–4627.
Rutkowski, D.T., Arnold, S.M., Miller, C.N., Wu, J., Li, J., Gunnison, K.M., Mori,
K., Sadighi Akha, A.A., Raden, D., and Kaufman, R.J. (2006). Adaptation to ER
stress is mediated by differential stabilities of pro-survival and pro-apoptotic
mRNAs and proteins. PLoS Biol. 4, e374.
Sahu, N., Dela Cruz, D., Gao, M., Sandoval, W., Haverty, P.M., Liu, J., Stephan,
J.P., Haley, B., Classon, M., Hatzivassiliou, G., et al. (2016). Proline starvation
induces unresolved ER stress and hinders mTORC1-dependent tumorigenesis. Cell Metab. 24, 753–761.
Scarlett, U.K., Rutkowski, M.R., Rauwerdink, A.M., Fields, J., Escovar-Fadul,
X., Baird, J., Cubillos-Ruiz, J.R., Jacobs, A.C., Gonzalez, J.L., Weaver, J., et al.
(2012). Ovarian cancer progression is controlled by phenotypic changes in
dendritic cells. J. Exp. Med. 209, 495–506.
Schewe, D.M., and Aguirre-Ghiso, J.A. (2008). ATF6alpha-Rheb-mTOR
signaling promotes survival of dormant tumor cells in vivo. Proc. Natl. Acad.
Sci. USA 105, 10519–10524.
Schindler, A.J., and Schekman, R. (2009). In vitro reconstitution of ER-stress
induced ATF6 transport in COPII vesicles. Proc. Natl. Acad. Sci. USA 106,
17775–17780.
Senovilla, L., Vitale, I., Martins, I., Tailler, M., Pailleret, C., Michaud, M., Galluzzi, L., Adjemian, S., Kepp, O., Niso-Santano, M., et al. (2012). An immunosurveillance mechanism controls cancer cell ploidy. Science 337, 1678–1684.
Pereira, E.R., Liao, N., Neale, G.A., and Hendershot, L.M. (2010). Transcriptional and post-transcriptional regulation of proangiogenic factors by the
unfolded protein response. PLoS ONE 5, e12521.
Shen, J., Chen, X., Hendershot, L., and Prywes, R. (2002). ER stress regulation
of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of
Golgi localization signals. Dev. Cell 3, 99–111.
Pol, J., Vacchelli, E., Aranda, F., Castoldi, F., Eggermont, A., Cremer, I.,
Sautès-Fridman, C., Fucikova, J., Galon, J., Spisek, R., et al. (2015). Trial
watch: immunogenic cell death inducers for anticancer chemotherapy. OncoImmunology 4, e1008866.
Shi, Y.H., Ding, Z.B., Zhou, J., Hui, B., Shi, G.M., Ke, A.W., Wang, X.Y., Dai, Z.,
Peng, Y.F., Gu, C.Y., et al. (2011). Targeting autophagy enhances sorafenib
lethality for hepatocellular carcinoma via ER stress-related apoptosis. Autophagy 7, 1159–1172.
Pozzi, C., Cuomo, A., Spadoni, I., Magni, E., Silvola, A., Conte, A., Sigismund,
S., Ravenda, P.S., Bonaldi, T., Zampino, M.G., et al. (2016). The EGFR-specific
antibody cetuximab combined with chemotherapy triggers immunogenic cell
death. Nat. Med. 22, 624–631.
Shimizu, A., Kaira, K., Yasuda, M., Asao, T., and Ishikawa, O. (2016). Clinical
and pathological significance of ER stress marker (BiP/GRP78 and PERK)
expression in malignant melanoma. Pathol. Oncol. Res. Published online
August 8, 2016. http://dx.doi.org/10.1007/s12253-016-0099-9.
Qiu, Q., Zheng, Z., Chang, L., Zhao, Y.S., Tan, C., Dandekar, A., Zhang, Z., Lin,
Z., Gui, M., Li, X., et al. (2013). Toll-like receptor-mediated IRE1a activation as
a therapeutic target for inflammatory arthritis. EMBO J. 32, 2477–2490.
Shoulders, M.D., Ryno, L.M., Genereux, J.C., Moresco, J.J., Tu, P.G., Wu, C.,
Yates, J.R., 3rd, Su, A.I., Kelly, J.W., and Wiseman, R.L. (2013). Stress-independent activation of XBP1s and/or ATF6 reveals three functionally diverse
ER proteostasis environments. Cell Rep. 3, 1279–1292.
Quail, D.F., and Joyce, J.A. (2013). Microenvironmental regulation of tumor
progression and metastasis. Nat. Med. 19, 1423–1437.
Ramakrishnan, R., Tyurin, V.A., Veglia, F., Condamine, T., Amoscato, A., Mohammadyani, D., Johnson, J.J., Zhang, L.M., Klein-Seetharaman, J., Celis, E.,
et al. (2014). Oxidized lipids block antigen cross-presentation by dendritic cells
in cancer. J. Immunol. 192, 2920–2931.
Ranganathan, A.C., Zhang, L., Adam, A.P., and Aguirre-Ghiso, J.A. (2006).
Functional coupling of p38-induced up-regulation of BiP and activation of
RNA-dependent protein kinase-like endoplasmic reticulum kinase to drug
resistance of dormant carcinoma cells. Cancer Res. 66, 1702–1711.
Reimold, A.M., Iwakoshi, N.N., Manis, J., Vallabhajosyula, P., SzomolanyiTsuda, E., Gravallese, E.M., Friend, D., Grusby, M.J., Alt, F., and Glimcher,
L.H. (2001). Plasma cell differentiation requires the transcription factor
XBP-1. Nature 412, 300–307.
Ri, M., Tashiro, E., Oikawa, D., Shinjo, S., Tokuda, M., Yokouchi, Y., Narita, T.,
Masaki, A., Ito, A., Ding, J., et al. (2012). Identification of toyocamycin, an agent
cytotoxic for multiple myeloma cells, as a potent inhibitor of ER stress-induced
XBP1 mRNA splicing. Blood Cancer J. 2, e79.
Romero-Ramirez, L., Cao, H., Nelson, D., Hammond, E., Lee, A.H., Yoshida,
H., Mori, K., Glimcher, L.H., Denko, N.C., Giaccia, A.J., et al. (2004). XBP1 is
essential for survival under hypoxic conditions and is required for tumor
growth. Cancer Res. 64, 5943–5947.
Rouschop, K.M., Dubois, L.J., Keulers, T.G., van den Beucken, T., Lambin, P.,
Bussink, J., van der Kogel, A.J., Koritzinsky, M., and Wouters, B.G. (2013).
PERK/eIF2a signaling protects therapy resistant hypoxic cells through induc-
Sidrauski, C., McGeachy, A.M., Ingolia, N.T., and Walter, P. (2015). The small
molecule ISRIB reverses the effects of eIF2a phosphorylation on translation
and stress granule assembly. eLife 4, e05033.
Signer, R.A., Magee, J.A., Salic, A., and Morrison, S.J. (2014). Haematopoietic
stem cells require a highly regulated protein synthesis rate. Nature 509, 49–54.
Sui, X., Chen, R., Wang, Z., Huang, Z., Kong, N., Zhang, M., Han, W., Lou, F.,
Yang, J., Zhang, Q., et al. (2013). Autophagy and chemotherapy resistance: a
promising therapeutic target for cancer treatment. Cell Death Dis. 4, e838.
Tam, A.B., Mercado, E.L., Hoffmann, A., and Niwa, M. (2012). ER stress activates NF-kB by integrating functions of basal IKK activity, IRE1 and PERK.
PLoS ONE 7, e45078.
Tang, C.H., Ranatunga, S., Kriss, C.L., Cubitt, C.L., Tao, J., Pinilla-Ibarz, J.A.,
Del Valle, J.R., and Hu, C.C. (2014). Inhibition of ER stress-associated
IRE-1/XBP-1 pathway reduces leukemic cell survival. J. Clin. Invest. 124,
2585–2598.
Thevenot, P.T., Sierra, R.A., Raber, P.L., Al-Khami, A.A., Trillo-Tinoco, J., Zarreii, P., Ochoa, A.C., Cui, Y., Del Valle, L., and Rodriguez, P.C. (2014). The
stress-response sensor chop regulates the function and accumulation of
myeloid-derived suppressor cells in tumors. Immunity 41, 389–401.
Upton, J.P., Wang, L., Han, D., Wang, E.S., Huskey, N.E., Lim, L., Truitt, M.,
McManus, M.T., Ruggero, D., Goga, A., et al. (2012). IRE1a cleaves select
microRNAs during ER stress to derepress translation of proapoptotic Caspase-2. Science 338, 818–822.
Cell 168, February 9, 2017 705
Urano, F., Wang, X., Bertolotti, A., Zhang, Y., Chung, P., Harding, H.P., and
Ron, D. (2000). Coupling of stress in the ER to activation of JNK protein kinases
by transmembrane protein kinase IRE1. Science 287, 664–666.
Urueña, C., Gomez, A., Sandoval, T., Hernandez, J., Li, S., Barreto, A., and
Fiorentino, S. (2015). Multifunctional T Lymphocytes generated after therapy
with an antitumor gallotanin-rich normalized fraction are related to primary tumor size reduction in a breast cancer model. Integr. Cancer Ther. 14, 468–483.
Vaupel, P., Kallinowski, F., and Okunieff, P. (1989). Blood flow, oxygen and
nutrient supply, and metabolic microenvironment of human tumors: a review.
Cancer Res. 49, 6449–6465.
Velez, J.M., Miriyala, S., Nithipongvanitch, R., Noel, T., Plabplueng, C.D.,
Oberley, T., Jungsuwadee, P., Van Remmen, H., Vore, M., and St Clair, D.K.
(2011). p53 regulates oxidative stress-mediated retrograde signaling: a novel
mechanism for chemotherapy-induced cardiac injury. PLoS ONE 6, e18005.
Vladykovskaya, E., Sithu, S.D., Haberzettl, P., Wickramasinghe, N.S., Merchant, M.L., Hill, B.G., McCracken, J., Agarwal, A., Dougherty, S., Gordon,
S.A., et al. (2012). Lipid peroxidation product 4-hydroxy-trans-2-nonenal
causes endothelial activation by inducing endoplasmic reticulum stress.
J. Biol. Chem. 287, 11398–11409.
Vogelstein, B., Papadopoulos, N., Velculescu, V.E., Zhou, S., Diaz, L.A., Jr.,
and Kinzler, K.W. (2013). Cancer genome landscapes. Science 339,
1546–1558.
Volmer, R., van der Ploeg, K., and Ron, D. (2013). Membrane lipid saturation
activates endoplasmic reticulum unfolded protein response transducers
through their transmembrane domains. Proc. Natl. Acad. Sci. USA 110,
4628–4633.
Wang, M., and Kaufman, R.J. (2014). The impact of the endoplasmic reticulum
protein-folding environment on cancer development. Nat. Rev. Cancer 14,
581–597.
706 Cell 168, February 9, 2017
Wang, Y., Alam, G.N., Ning, Y., Visioli, F., Dong, Z., Nör, J.E., and Polverini,
P.J. (2012). The unfolded protein response induces the angiogenic switch in
human tumor cells through the PERK/ATF4 pathway. Cancer Res. 72,
5396–5406.
Wolf, D.H., and Stolz, A. (2012). The Cdc48 machine in endoplasmic reticulum
associated protein degradation. Biochim. Biophys. Acta 1823, 117–124.
Xie, W.Y., Zhou, X.D., Li, Q., Chen, L.X., and Ran, D.H. (2015). Acid-induced
autophagy protects human lung cancer cells from apoptosis by activating
ER stress. Exp. Cell Res. 339, 270–279.
Yaari-Stark, S., Shaked, M., Nevo-Caspi, Y., Jacob-Hircsh, J., Shamir, R., Rechavi, G., and Kloog, Y. (2010). Ras inhibits endoplasmic reticulum stress in
human cancer cells with amplified Myc. Int. J. Cancer 126, 2268–2281.
Yamamoto, K., Takahara, K., Oyadomari, S., Okada, T., Sato, T., Harada, A.,
and Mori, K. (2010). Induction of liver steatosis and lipid droplet formation in
ATF6alpha-knockout mice burdened with pharmacological endoplasmic reticulum stress. Mol. Biol. Cell 21, 2975–2986.
Yan, D., Wang, H.W., Bowman, R.L., and Joyce, J.A. (2016). STAT3 and STAT6
signaling pathways synergize to promote cathepsin secretion from macrophages via IRE1a activation. Cell Rep. 16, 2914–2927.
Yoshida, H., Matsui, T., Yamamoto, A., Okada, T., and Mori, K. (2001). XBP1
mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor. Cell 107, 881–891.
Young, R.M., Ackerman, D., Quinn, Z.L., Mancuso, A., Gruber, M., Liu, L.,
Giannoukos, D.N., Bobrovnikova-Marjon, E., Diehl, J.A., Keith, B., and Simon,
M.C. (2013). Dysregulated mTORC1 renders cells critically dependent on
desaturated lipids for survival under tumor-like stress. Genes Dev. 27,
1115–1131.
Leading Edge
Review
Primary, Adaptive, and Acquired Resistance
to Cancer Immunotherapy
Padmanee Sharma,1,* Siwen Hu-Lieskovan,2 Jennifer A. Wargo,3 and Antoni Ribas2,*
1Department of Genitourinary Medical Oncology and Immunology,The University of Texas MD Anderson Cancer Center, Houston,
TX 77030, USA
2Department of Medicine, Division of Hematology-Oncology, University of California, Los Angeles and the Jonsson Comprehensive
Cancer Center, Los Angeles, CA 90095, USA
3Department of Melanoma Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
*Correspondence: padsharma@mdanderson.org (P.S.), aribas@mednet.ucla.edu (A.R.)
http://dx.doi.org/10.1016/j.cell.2017.01.017
SUMMARY
Cancer immunotherapy can induce long lasting responses in patients with metastatic cancers of a
wide range of histologies. Broadening the clinical applicability of these treatments requires an
improved understanding of the mechanisms limiting cancer immunotherapy. The interactions between the immune system and cancer cells are continuous, dynamic, and evolving from the initial
establishment of a cancer cell to the development of metastatic disease, which is dependent on immune evasion. As the molecular mechanisms of resistance to immunotherapy are elucidated, actionable strategies to prevent or treat them may be derived to improve clinical outcomes for patients.
Introduction
Metastatic cancers remain an incurable disease for the great
majority of patients, as the intrinsic genomic instability common
to all cancers facilitates the escape from cytotoxic or targeted
therapies. The recent breakthroughs in the understanding of tumor immune biology and the development of newer generation
of cancer immunotherapies have opened a brand new chapter
in the war against cancer. This change in landscape is based
on the discovery of cancer immune checkpoints and the success of checkpoint inhibitors, as well as the advances in technology to generate genetically modified immune cells (Miller
and Sadelain, 2015). The focus of treatment has shifted from
the tumor itself to the host’s immune system, to mobilize immune cells to recognize and eventually eliminate the cancer
cells. A hallmark of immunotherapy is the durability of responses, most likely due to the memory of the adaptive immune
system, which translates into long-term survival for a subset of
patients.
Early efforts to harness the immune system in cancer control,
pioneered by Dr. William B. Coley in the 1890s (Coley, 1910),
were overlooked due to the lack of consistency in response
and were soon overwhelmed by the development of more effective treatments such as radiotherapy and chemotherapy. However, investigations continued to unravel and elucidate the interactions between the immune system and cancer cells. The
concept of cancer immunosurveillance, which was proposed
by Paul Ehrlich (Ehrlich, 1956) and enriched by Burnet and
Thomas (Burnet, 1971) in the 1950s, stated that the emergence
of malignant cells is a frequent event but is suppressed by the
host’s natural immunity, that cancer develops when this immunity is weakened, and that lymphocytes are responsible for this
process. Finally, the cancer immune-editing concept was eluci-
dated by Schreiber and colleagues in 2002 (Dunn et al., 2002),
recognizing a dual role of the host’s immunity, both as an
extrinsic tumor suppressor and a facilitator of tumor growth
and progression, acting across three sequential phases—elimination, equilibrium and escape—through constant interactions
between tumor cells, immune cells, and the tumor microenvironment. Importantly, host immune responses and tumor genomics
are tightly related, as illustrated by the notion that neoantigens
arising from genomic mutations may shape immune responses
(Schumacher and Schreiber, 2015); however, these responses
may prove ineffective against a heterogeneous and evolving tumor microenvironment.
The process of T cell activation involves antigen presentation
by the major histocompatibility complex (MHC) molecules on
the antigen-presenting cells (APCs) to the corresponding T cell
receptor (TCR) on naive T cells. The interaction of costimulatory
molecules CD28 and B7 is required for full activation, which is
tightly regulated by inhibitory checkpoints to avoid collateral
damage and autoimmunity. The CTLA-4 receptor on activated
effector T cells and regulatory T cells (Tregs) was discovered in
the 1980s (Brunet et al., 1987). Seminal work by James Allison
and colleagues showed that CTLA-4 competes with CD28 for
B7 ligands and inhibits proliferation and IL-2 secretion by
T cells (Krummel and Allison, 1995) and that CTLA-4 blocking antibodies could treat tumors in immune competent animal models
(Leach et al., 1996). Subsequent clinical testing resulted in the
approval of ipilimumab for treatment of advanced melanoma in
2011, the first in class CTLA-4 checkpoint inhibitor approved
by the US Food and Drug Administration (FDA) (Hodi et al.,
2010; Robert et al., 2011). Pooled data from clinical trials of ipilimumab confirmed durable clinical responses, depicted by a
plateau in the survival curve beginning around year 3, that lasted
Cell 168, February 9, 2017 ª 2017 Elsevier Inc. 707
Table 1. Terminology for Different Resistance Mechanisms to Immunotherapy
Term
Description
primary resistance
A clinical scenario where a cancer does not respond to an immunotherapy strategy. The mechanistic basis of
lack of response to immunotherapy may include adaptive immune resistance.
adaptive immune
resistance
A mechanism of resistance where a cancer is recognized by the immune system but it protects itself by adapting
to the immune attack. Given the evolving nature of the immune/cancer cell interaction, this could clinically manifest
as primary resistance, mixed responses or acquired resistance.
acquired resistance
A clinical scenario in which a cancer initially responded to immunotherapy but after a period of time it relapsed
and progressed.
10 years or more in a subset of approximately 21% of patients
(Schadendorf et al., 2015). In 2015, ipilimumab was also
approved by the FDA as adjuvant therapy for locally advanced
melanoma. Due to enhanced immune responses, possibly during early stages of T cell activation, significant immune-related
toxicities have been observed, but most can be managed by systemic steroid therapy.
Another checkpoint receptor expressed by activated T cells,
programed death 1 (PD-1), was cloned in 1992 (Ishida et al.,
1992), and subsequently its ligand PD-L1 was characterized
(Dong et al., 1999; Freeman et al., 2000). PD-L1 expression
can be constitutive or induced in many tumors to evade immune
attack. Since PD-L1 expression can be induced by IFNg, which
is expressed during an active anti-tumor immune response, it
has been referred to as a mechanism of adaptive immune resistance (Table 1). Antibodies blocking the PD-1 and PD-L1 inhibitory axis can unleash activated tumor-reactive T cells and have
been shown in clinical trials to induce durable anti-tumor responses in increasing numbers of tumor histologies, including
the tumor types that are not traditionally considered immunotherapy sensitive (Okazaki et al., 2013; Zou et al., 2016). This
led to the approval of two anti-PD1 antibodies (pembrolizumab
and nivolumab) and one anti-PD-L1 antibody (atezolimumab)
for the treatment of advanced melanoma, non-small-cell lung
cancer, renal cell carcinoma, head and neck squamous carcinoma, Hodgkin’s lymphoma, and bladder cancer. Currently
there are over ten anti-PD-1 and anti-PD-L1 antibodies in various
stages of clinical testing in many different tumor types. Interestingly, there have been thousands of patients receiving PD-1
blockade therapy thus far, with similar immune related toxicities
as observed for anti-CTLA-4 but with generally lower frequency,
possibly because the PD-1 and PD-L1 checkpoint may act later
in the T cell response, resulting in a more restricted T cell reactivity toward tumor cells, with the majority of patients tolerating
treatment well (Larkin et al., 2015b). Due to the non-overlapping
mechanism of action of anti-CTLA4 and anti-PD1 antibodies
(Das et al., 2015; Gubin et al., 2014), clinical testing of the combination of these two classes of checkpoint inhibitors showed
improved clinical response (up to 60%) in melanoma at the
expense of significantly increased frequency of toxicities (Larkin
et al., 2015a). The combination of CTLA4 and PD-1 and PD-L1
checkpoint blockade has been approved as front line therapy
for advanced melanoma patients and is being tested in other tumor types with different dose levels and intervals of anti-CTLA4
to reduce toxicity.
Cell-based immunotherapy was pioneered by many investigators, including Alex Fefer, Phil Greenberg, Zelig Eshhar, Steven
708 Cell 168, February 9, 2017
Rosenberg, and colleagues in the 1980s, inspired by the correlation of the number of tumor infiltrating lymphocytes (TILs) and
survival in some cancers. This process required TILs to be isolated from the patient’s surgical specimen, expanded in vitro,
and re-infused back to the lymphocyte-depleted patient. In
these studies, sufficient TILs could not be isolated or expanded
from tumors of approximately 50%–60% of patients, which
limited the number of patients who could be treated. For patients
who could be treated with the expanded TILs, the reported
response rate was 50% for melanoma, including 20% complete
responses, and 95% of these complete responders had more
than 5 years of survival (Rosenberg et al., 2011). This approach,
however, requires large surgical samples, experienced academic centers, and tumors enriched with anti-tumor T cells,
which is a rare event for most tumor types. The recent advance
of gene transfer technologies and T cell engineering has enabled
more versatile approaches, including adoptive cell transfer (ACT)
of the patient’s peripheral T cells that are genetically modified
to target cancer specific antigens, via physiological TCRs or
chimeric antigen receptors (CARs) (Sadelain, 2016; Yang and
Rosenberg, 2016). TCRs are usually cloned from TILs that are
reactive to specific cancer antigens with no or very limited
expression in normal adult tissue but are widely expressed by
cancer cells. Such TCRs recognize tumor antigen presented in
the context of the MHC. Clinical success has been documented
(Yee et al., 2015). The TCR approach allows intracellular antigen
targets but is MHC restricted and can be subject to treatment
failure for tumors that have downregulated their MHC surface
expression. CAR technology was first developed by Eshhar
et al., 1993, who genetically engineered T cells with chimeric
genes, linking single chain antibodies (scFv) targeting tumor
cell surface antigens to intracellular signaling adaptors for
TCR: in the first generation, to the T cell specific activating z
chain of the CD3 complex. Subsequent modification with costimulatory molecules CD28 (second generation) and 4-1BB
(third generation) has enabled the expansion of T cells while retaining function upon repeated antigen exposure. CAR T cells do
not require MHC restriction and can be engineered to enhance
T cell function. Recent clinical success with CD19 targeting
CAR to treat CD19+ B cell malignancy has shown great success,
with a remarkable 90% complete remission in a cohort of 30 patients with relapsed or refractory pediatric acute lymphoblastic
leukemia (ALL), and two thirds of these patients remained in
remission after 6 months (Maude et al., 2014). The biggest challenge facing the field of ACT is the identification of target tumor
antigens that are not expressed by normal tissues, both to maximize specificity and efficacy and to minimize toxicity (Fesnak
Table 2. Mechanisms of Primary and Adaptive Resistance to
Immunotherapy
Mechanism
Examples
absence of antigenic
proteins
low mutational burden
lack of viral antigens
lack of cancer-testis antigens
overlapping surface proteins
absence of antigen
presentation
deletion in TAP
deletion in B2M
silenced HLA
genetic T cell
exclusion
MAPK oncogenic signaling
stabilized b-catenin
mesenchymal transcriptome
oncogenic PD-L1 expression
insensibility
to T cells
mutations in interferon gamma
pathway signaling
absence of
T cells
lack of T cells with tumor
antigen-specific TCRs
Figure 1. Clinical Scenarios of Primary, Adaptive, and Acquired
Resistance to Immunotherapy
inhibitory immune
checkpoints
VISTA, LAG-3, TIM-3
(A) Patient’s tumor is resistant to immunotherapy with no active immune
response.
(B) Patient’s tumor is resistant to immunotherapy; active anti-tumor immune
response, but turned off by checkpoints or other adaptive resistance mechanisms.
(C) Patient has an initial response to immunotherapy but later progressed;
heterogeneous population and selection of resistant clones that were present
before treatment started.
(D) Patient has an initial response to immunotherapy but later progressed; true
acquired resistance during the immunotherapy.
immunosuppressive
cells
TAMs, Tregs
tumor cell
intrinsic
tumor cell
extrinsic
et al., 2016). A commonly seen toxicity in ACT therapy is cytokine
release syndrome, which can be life-threatening and requires
prompt management with steroids and IL-6 receptor antibody
(tocilizumab).
Despite the unprecedented durable response rates observed
with cancer immunotherapies, the majority of patients do not
benefit from the treatment (primary resistance), and some
responders relapse after a period of response (acquired resistance). Several common cancer types have shown very low frequency of response (breast, prostate, and colon cancers), and
heterogeneous responses have been seen even between
distinct tumors within the same patient (Figure 1). For the purposes of this review, we have categorized primary, adaptive,
and acquired resistance as described in Table 1, in keeping
with the most typical conceptualization for practicing clinicians.
However, in considering resistance mechanisms to immunebased therapies, it is important to remember that the immune
response is dynamic and constantly evolving in each patient,
either as a result of the patient’s own environmental and genetic
factors or as a result of treatment interventions, including surgery, chemotherapy, radiation therapy, and immunotherapy.
Anti-tumor immune responses that are ongoing throughout the
course of a patient’s disease may be affected by many of these
factors, and the establishment of resistance mechanisms relevant to immunotherapeutic failure may pre-date immunotherapy
challenge. Without recourse to detailed immune and tumor
characterization, these resistance mechanisms can be divided,
clinically, into those that prevent a patient from ever responding
to an immunotherapy or those that facilitate relapse after an
initial response. Thus, although resistance to immunotherapies
may manifest at different times, in many cases, similar or overlapping mechanisms enable tumor cells to evade anti-tumor immune responses. We discuss known resistance mechanisms
and provide rationale for combination therapies to overcome
resistance.
Primary and Adaptive Resistance to Immunotherapy
Patients who have primary resistance to checkpoint inhibitors
do not respond to the initial therapy. Ongoing studies indicate
that both tumor-cell-intrinsic and tumor-cell-extrinsic factors
contribute to the resistance mechanisms (Table 2). The most
straightforward reason why a tumor would not respond to immune checkpoint therapy or ACT is lack of recognition by
T cells because of absence of tumor antigens (Gubin et al.,
2014). Alternatively, cancer cells may have tumor antigens but
develop mechanisms to avoid presenting them on the surface
restricted by MHC, due to alterations in the antigen-presenting
machinery (such as proteasome subunits or transporters associated with antigen processing), beta-2-microglobulin (B2M), or
MHC itself (Marincola et al., 2000; Sucker et al., 2014). B2M is
required for HLA class I folding and transport to the cell surface,
and its genetic deficiency leads to lack of CD8 T cell recognition
(Figures 2 and 3).
Tumor-Cell-Intrinsic Factors for Primary and Adaptive
Resistance
Tumor-cell-intrinsic factors that contribute to immunotherapy
resistance include expression or repression of certain genes
and pathways in tumor cells that prevent immune cell infiltration
or function within the tumor microenvironment. These mechanisms may exist at the time of initial presentation, highlighting
primary resistance mechanisms, or these mechanisms may
Cell 168, February 9, 2017 709
Figure 2. Known Intrinsic Mechanisms of
Resistance to Immunotherapy
(A) Intrinsic factors that lead to primary or adaptive
resistance including lack of antigenic mutations,
loss of tumor antigen expression, loss of HLA
expression, alterations in antigen processing machinery, alterations of several signaling pathways
(MAPK, PI3K, WNT, IFN), and constitutive PD-L1
expression.
(B) Intrinsic factors that are associated with acquired resistance of cancer, including loss of
target antigen, HLA, and altered interferon
signaling, as well as loss of T cell functionality.
evolve later, highlighting adaptive resistance mechanisms. Multiple tumor-intrinsic mechanisms have recently been identified
and include (1) signaling through the mitogen-activated protein
kinase (MAPK) pathway and/or loss of PTEN expression, which
enhances PI3K signaling, (2) expression of the WNT/b-catenin
signaling pathway, (3) loss of interferon-gamma (IFNg) signaling
pathways, and (4) lack of T cell responses as result of loss of tumor antigen expression.
Oncogenic signaling through the MAPK pathway results in the
production of VEGF and IL-8, among many other secreted proteins, which have known inhibitory effects on T cell recruitment
and function (Liu et al., 2013). Similarly, loss of PTEN, which enhances PI3K signaling and is a common phenomenon across
several cancers, including 30% of melanomas, was found to
be associated with resistance to immune checkpoint therapy
(Peng et al., 2016). PTEN loss in tumors of the Cancer Genome
Atlas (TCGA) melanoma dataset correlated with significantly
decreased gene expression of IFNg, granzyme B, and CD8+
T cell infiltration; importantly, the frequency of PTEN deletions
and mutations was higher in non-T-cell-inflamed tumors as
compared to T-cell-inflamed tumors. In a murine model, PTENknockout tumors were less susceptible to adoptive cell therapy
than PTEN-expressing tumors.
The potential of oncogenic signaling pathways to induce T cell
exclusion from cancers has also been described through the
stabilization of b -catenin resulting in constitutive WNT signaling
(Spranger et al., 2015). In a murine model, tumors with elevated
b-catenin lacked a subset of dendritic cells (DCs) known as
CD103+ DCs, due to decreased expression of CCL4, a chemokine
710 Cell 168, February 9, 2017
that attracts CD103+ DCs. In addition, murine tumors lacking b-catenin responded
effectively to immune checkpoint therapy
whereas b-catenin-positive tumors did
not. Non-T-cell-inflamed human melanoma tumors, which lacked T cells and
CD103+ DCs in the tumor microenvironment, had significantly higher expression
of tumor intrinsic b-catenin signaling genes.
Cancer cells that constitutively express
immunosuppressive cell surface ligands
like PD-L1 may actively inhibit anti-tumor
T cell responses. A genetic amplification
of a locus in chromosome 9 that contains
the genes for the two ligands of PD-1 (PDL1 and PD-L2) and the interferon gamma
receptor signaling molecule Janus kinase 2 (JAK2) is termed the
PDJ amplicon (Ansell et al., 2015; Green et al., 2010; Rooney
et al., 2015). PDJ is amplified in the malignant Reed-Sternberg
cells in Hodgkin’s disease, and anti-PD-1 therapy results in
objective responses in over 80% of patients with chemotherapy-refractory Hodgkin’s disease (Ansell et al., 2015). Other
mechanisms that have been described as leading to constitutive
PD-L1 expression by cancer cells include PTEN deletions or
PI3K and/or AKT mutations (Lastwika et al., 2016; Parsa et al.,
2007), EGFR mutations (Akbay et al., 2013), MYC overexpression (Casey et al., 2016), CDK5 disruption (Dorand et al., 2016),
and an increase in PD-L1 transcripts stabilized by truncation of
the 30 UTR of this gene (Kataoka et al., 2016). It is currently unclear whether constitutive PD-L1 expression resulting from these
oncogenic signaling processes results in decreased or increased
likelihood of responding to anti-PD-1 and PD-L1 therapy, but it
may indeed result in lack of response to other cancer immunotherapy strategies by actively inhibiting anti-tumor T cells.
The interferon-gamma pathway is emerging as a key player
in primary, adaptive, and acquired resistance to checkpoint
blockade therapy (Gao et al., 2016; Pardoll, 2012; Ribas, 2015;
Shin et al., 2016; Zaretsky et al., 2016). It has both favorable
and detrimental effects on anti-tumor immune responses. Interferon-gamma produced by tumor-specific T cells that have
recognized their cognate antigen on cancer cells or APCs induces an effective anti-tumor immune response through (1)
enhanced tumor antigen presentation that occurs as a result of
increased expression of proteins, such as MHC molecules,
involved in antigen presentation, (2) recruitment of other immune
Figure 3. Known Extrinsic Mechanisms of
Resistance to Immunotherapy
This includes CTLA-4, PD1, and other immune
checkpoints, T cell exhaustion and phenotype
change, immune suppressive cell populations
(Tregs, MDSC, type II macrophages), and cytokine
and metabolite release in the tumor microenvironment (CSF-1, tryptophan metabolites, TGF-b,
adenosine). Abbreviations are as follows:
APC, antigen-presenting cells; MHC, major histocompatibility complex; TCR, T cell receptor;
Treg, regulatory T cell; MDSC, myeloid-derived
suppressor cell; Mɸ II, type II macrophage.
cells, and (3) direct anti-proliferative and pro-apoptotic effects
on tumor cells (Platanias, 2005). But continuous interferongamma exposure can lead to immunoediting of cancer cells, resulting in immune escape (Benci et al., 2016; Shankaran et al.,
2001). One mechanism by which cancer cells could escape the
effects of interferon gamma is by downregulating or mutating
molecules involved in the interferon gamma signaling pathway,
which goes through the interferon gamma receptor chains
JAK1 and/or JAK2 and the signal transducer and activators of
transcription (STATs) (Darnell et al., 1994). In cell line and animal
models, mutations or epigenetic silencing of molecules in the
interferon receptor signaling pathway results in loss of the antitumor effects of interferon gamma (Dunn et al., 2005; Kaplan
et al., 1998). Analysis of tumors in patients who did not respond
to therapy with the anti-CTLA-4 antibody ipilimumab revealed an
enriched frequency of mutations in the interferon gamma
pathway genes interferon gamma receptor 1 and 2 (IFNGR1
and IFNGR2), JAK2, and interferon regulatory factor 1 (IRF1)
(Gao et al., 2016). Any of these mutations would prevent
signaling in response to interferon gamma and give an advantage to the tumor cells escaping from T cells, thereby resulting
in primary resistance to anti-CTLA-4 therapy. Mutations in this
pathway would additionally result in lack of PD-L1 expression
upon interferon gamma exposure, thereby resulting in cancer
cells that would be genetically negative for inducible PD-L1
expression. In such a scenario, blocking PD-L1 or PD-1 with
therapeutic antibodies would not be useful, and these would
be patients who are primary resistant to anti-PD-1 therapy
(Shin and Ribas, 2015; Shin et al., 2016).
An additional cancer-cell-intrinsic mechanism of primary
resistance to immunotherapy is expression of a certain set of
genes that were found to be enriched in tumors from patients
who did not respond to anti-PD-1 therapy, termed innate anti-PD-1 resistance
signature, or IPRES (Hugo et al., 2016).
These genes that lead to lack of response
are related to mesenchymal transformation, stemness, and wound healing and
are preferentially expressed by cancers
that seldom respond to PD-1 blockade
therapy, such as pancreatic cancer.
Epigenetic modification of the DNA
in cancer cells may lead to changes
in gene expression of immune-related
genes, which can impact antigen processing, presentation, and immune evasion (Karpf and Jones,
2002; Kim and Bae, 2011). Therefore, demethylating agents
may enable re-expression of immune related genes, with potential for therapeutic impact, especially in the setting of combination treatment with immunotherapy. (Héninger et al., 2015). Histone deacetylase inhibitors led to increased expression of MHC
and tumor-associated antigens, which synergized with ACT
therapy to improve anti-tumor responses in a murine melanoma
model (Vo et al., 2009). Similarly, in a lymphoma model, hypomethylating agents were found to increase CD80 expression on tumor cells, with aconcomitant increase in tumor-infiltrating CD8+
T cells (Wang et al., 2013). These pre-clinical data indicate the
potential to reverse the epigenetic changes in cancer cells,
which may enable enhanced immune recognition and response
to immunotherapy.
Tumor-Cell-Extrinsic Factors for Primary and Adaptive
Resistance
Tumor-cell-extrinsic mechanisms that lead to primary and/or
adaptive resistance involve components other than tumor cells
within the tumor microenvironment, including Tregs, myeloid
derived suppressor cells (MDSCs), M2 macrophages, and other
inhibitory immune checkpoints, which may all contribute to inhibition of anti-tumor immune responses.
Tregs, which can be identified by expression of the FoxP3
transcription factor, have a central role in maintaining selftolerance (Rudensky, 2011). The existence of suppressor
T cells that could downregulate immune responses of antigen-specific T cells was first identified nearly four decades
ago in thymectomized, lethally irradiated, bone-marrow-reconstituted mice (Gershon and Kondo, 1970). Tregs are known to
suppress effector T cell (Teff) responses by secretion of certain
Cell 168, February 9, 2017 711
inhibitory cytokines, such as IL-10, IL-35, and TGF-b, or by
direct cell contact (Oida et al., 2003; Sakaguchi et al., 2008;
Sundstedt et al., 2003). Published data indicate that many human tumors are infiltrated by Tregs (Chaudhary and Elkord,
2016; Ormandy et al., 2005; Woo et al., 2002). A vast number
of murine studies have shown that the depletion of Treg cells
from the tumor microenvironment can enhance or restore
anti-tumor immunity (Linehan and Goedegebuure, 2005; Viehl
et al., 2006). In murine models, response to anti-CTLA-4 therapy was shown to be associated with an increase in the ratio
of Teffs to Tregs (Quezada et al., 2006). This shift in the ratio
of Teffs to Tregs was found to be a result of both an increase
in Teffs and depletion of Tregs in a murine tumor model (Simpson et al., 2013). These data suggest that tumors for which
immunotherapy is unable to increase Teffs and/or deplete
Tregs to increase the ratio of Teffs to Tregs are likely to be
resistant to treatment, either initially or during the relapsed
disease setting. However, it is possible that tumor-infiltrating
Tregs may co-exist with other immune cells, indicating a potentially immune-responsive tumor. A retrospective study of patients treated with anti-CTLA-4 reported that a high baseline
expression of FoxP3+ Tregs in the tumor was associated with
better clinical outcomes (Hamid et al., 2011). Additional studies
are ongoing to determine the impact of tumor-infiltrating Tregs
on clinical outcomes for patients who receive treatment with
immunotherapy agents.
Myeloid-derived suppressor cells (MDSCs) have emerged as
major regulators of immune responses in various pathological
conditions, including cancer. MDSCs were initially defined in
murine models and were characterized by the expression of
CD11b (CR3A or integrin aM) and Gr-1 markers (Bronte et al.,
1998; Talmadge and Gabrilovich, 2013). Human MDSCs express
markers such as CD11b+and CD33+ but are mostly negative for
HLA-DR and lineage-specific antigens (Lin), including CD3,
CD19, and CD57. Monocytic MDSCs are HLA-DR, CD11b+,
CD33+, and CD14+ and granulocytic MDSCs are HLA-DR,
CD11b+, CD33+, CD15+; however, mature monocytes express
HLA-DR (Wesolowski et al., 2013). MDSCs have been implicated
in promoting angiogenesis, tumor cell invasion, and metastases
(Yang et al., 2004; Yang et al., 2008). Furthermore, clinical findings have shown that the presence of MDSCs correlates with
reduced survival in human cancers, including breast cancer
and colorectal cancer (Solito et al., 2011). Reports suggest that
the presence of MDSCs in the tumor microenvironment correlates with decreased efficacy of immunotherapies, including immune checkpoint therapy (Meyer et al., 2014), adoptive T cell
therapy (Kodumudi et al., 2012), and DC vaccination (Laborde
et al., 2014). Therefore, eradicating or reprogramming MDSCs
could enhance clinical responses to immunotherapy. Indeed, in
melanoma, breast cancer, and head and neck murine tumor
models, selective inactivation of macrophage PI3Kg synergized
with immune checkpoint inhibitors to promote tumor regression
and increase survival (De Henau et al., 2016; Kaneda et al.,
2016). In one study, the investigators demonstrated that mice
lacking PI3Kg or tumor-bearing mice treated with PI3Kg inhibitors (TG100-115 or IPI-549) had reduced tumor growth, which
was associated with enhanced expression of pro-inflammatory
cytokines and inhibition of immune-suppressive factors in the tu712 Cell 168, February 9, 2017
mors (Kaneda et al., 2016). Moreover, genes and proteins associated with immune activation were upregulated in macrophages
that were treated with PI3Kg inhibitors or those from mice lacking PI3Kg. These data established PI3Kg as a molecular switch
that regulates macrophage function. The investigators also
demonstrated that a PI3Kg inhibitor (TG100-115) plus anti-PD1 led to improved tumor rejection and survival of tumor-bearing
mice (Kaneda et al., 2016). In a second study, tumor-bearing
mice treated with triple-combination therapy, a PI3Kg inhibitor
(IPI-549) plus anti-CTLA-4 and anti-PD-1, had improved tumor
regression and long-term survival as compared to dual therapy
with anti-CTLA-4 plus anti-PD-1 (De Henau et al., 2016). These
pre-clinical studies highlight inhibitors of PI3Kg as a therapeutic
potential for combination strategies with immune checkpoint
therapy in cancer patients.
Tumor-associated macrophages (TAMs) are another subset of
cells that seem to affect responses to immunotherapy. TAMs
include both M1 macrophages, which are involved in promoting
anti-tumor immunity, and the M2 macrophages, which possess
pro-tumorigenic properties (Chanmee et al., 2014). M1 and M2
macrophages can be distinguished based on the differential
expression of transcription factors and surface molecules and
the disparities in their cytokine profile and metabolism (Biswas
and Mantovani, 2010; Hu et al., 2016). Clinical studies have
shown an association between higher frequencies of TAMs
and poor prognosis in human cancers (Hu et al., 2016). In a
chemically induced mouse model of lung adenocarcinoma,
depletion of TAMs reduced tumor growth as a result of downregulation of M2 and/or TAM recruitment, possibly due to the
inactivation of CCL2 and/or CCR2 signaling (Fritz et al., 2014).
Likewise, depletion of M2 macrophages in various murine tumor
models, including cutaneous T cell lymphoma (Wu et al., 2014),
colon cancer, lung cancer, breast cancer (Luo et al., 2006),
and melanoma (Ries et al., 2014; Ruffell et al., 2014; Tham
et al., 2015), have shown similar results. Several reports have
discussed the role of macrophages in mediating therapeutic
resistance in cancer (De Palma and Lewis, 2013; Ruffell et al.,
2014; Ruffell and Coussens, 2015). Reports suggest that macrophages can directly suppress T cell responses through programmed death-ligand 1 (PD-L1) in hepatocellular carcinoma
(Kuang et al., 2009) and B7-H4 in ovarian carcinoma (Kryczek
et al., 2006). To overcome the potential resistance mechanism
of macrophages, investigators tested blockade of CSF-1R,
a receptor for macrophage-colony stimulating growth factor,
in a murine model of pancreatic cancer and demonstrated
decreased frequencies of TAMs, with subsequent increase in
interferon production and restrained tumor progression. Importantly, neither PD-1 nor CTLA-4 blockade could significantly
reduce tumor growth in the murine model, results that were
similar to findings from single agent studies in patients with
pancreatic cancer (Le et al., 2013; Zhu et al., 2014). However,
CSF1R blockade in combination with either an antibody against
PD-1 or CTLA-4, in addition to gemcitabine, led to improved
tumor regression (Zhu et al., 2014). These data suggest that
CSF-1R blockade induced reduction of TAMs, which enabled
response to immune checkpoint therapy. Similarly, in a melanoma model, CSF-1R inhibitor was shown to synergize with
ACT therapy (Mok et al., 2014). Several early phase clinical trials
are underway to test the combination of CSF-1R inhibition with
checkpoint inhibitors (Table 3).
The immune response is dynamic and signals that enhance
anti-tumor immune responses also tend to turn on inhibitory
genes and pathways in order to tightly regulate the immune
response. For example, initial T cell activation, via TCR signaling
and CD28 co-stimulation, eventually leads to increased expression of the inhibitory CTLA-4 immune checkpoint (Leach et al.,
1996). Similarly, effector T cell responses such as increased
IFNg production leads to increased expression of the PD-L1
protein on multiple cell types, including tumor cells, T cells and
macrophages, which can engage the PD-1 receptor on T cells
to suppress anti-tumor immunity (Chen, 2004; Dong et al.,
2002). Apart from this, IFNg may additionally promote the
expression of immunosuppressive molecules such as indolaimine-2, 3-deoxygenase (IDO), a tryptophan-metabolizing
enzyme that can contribute to peripheral tolerance and can
have a direct negative effect on effector T cell function (Gajewski
et al., 2013). Similarly, carcinoembryonic antigen cell adhesion
molecule-1 (CEACAM1), seems to be another inhibitory molecule that is induced by IFNg (Takahashi et al., 1993), (GrayOwen and Blumberg, 2006). Therapeutic antibodies blocking
CEACAM1 (Ortenberg et al., 2012) and TIM-3 have resulted in
enhanced anti-tumor immune responses (Pardoll, 2012; Sakuishi et al., 2010). A recent study in an immunocompetent
mouse model of lung adenocarcinoma demonstrated that recurrent tumors after anti-PD-1 treatment were due to increased
expression of TIM-3 on T cells. Notably, anti-PD-1 plus antiTIM-3 led to improved responses in the tumor bearing mice.
Similarly, two lung cancer patients who developed recurrent disease after anti-PD-1 treatment were found to have increased
TIM-3 expression on T cells (Koyama et al., 2016).
Immune suppressive cytokines are often released by tumor or
macrophages for local suppression of anti-tumor immune responses. Transforming growth factor b (TGF-b) is a cytokine
that plays important roles in angiogenesis and immunosuppression by stimulating Tregs (Lebrun, 2012). Increased levels of
TGF-b are associated with poor prognosis in multiple tumor
types (Lin and Zhao, 2015; Massagué, 2008). Preclinical models
have shown synergy combining TGF-b receptor kinase inhibitor
with anti-CTLA-4, which led to anti-tumor responses in a melanoma model (BRAFV600EPTEN/) (Hanks et al., 2014). Another
pre-clinical study consisting of radiation therapy combined
with TGF-b inhibition also demonstrated anti-tumor responses
(Vanpouille-Box et al., 2015). Adenosine was shown to inhibit
T cell proliferation and cytotoxic function via the A2A receptor
on T cells (Zhang et al., 2004) as well as to promote metastasis
via the A2B receptor on tumor cells (Mittal et al., 2016). In addition, CD73 is the enzyme that dephosphorylates adenosine
monophosphate (AMP) to form adenosine, thus also suppressing immune function and promoting tumor cell metastasis (Stagg
et al., 2010), and also stimulates angiogenesis (Allard et al.,
2014). High expression of CD73 is associated with poor prognosis in different cancer types (Leclerc et al., 2016; Loi et al.,
2013; Turcotte et al., 2015). CD73 is also a potential biomarker
for anti-PD-1 therapy, with high expression limiting anti-PD-1 efficacy, which can be rescued by concomitant A2A blockade
(Beavis et al., 2015).
Specific chemokines and chemokine receptors are important
for trafficking of MDSCs and Tregs to the tumor. For example, tumors secrete ligands CCL5, CCL7, and CXCL8, bind to their receptors CCR1 or CXCR2 expressed on subtypes of MDSCs
(Highfill et al., 2014), and attract MDSCs in the tumor microenvironment. Inhibitors of these chemokine receptors could abrogate immune evasion and improve anti-tumor T cell responses.
CCR4 is highly expressed by Tregs in the blood and tumors (Sugiyama et al., 2013), and anti-CCR4 inhibits Treg recruitment as
well as promotes antibody-dependent cell-mediated cytotoxicity (ADCC), further reducing the Treg population (Chang
et al., 2012). CXCR4 is a receptor for the chemokine CXCL12,
which has been shown to promote an immunosuppressive tumor
microenvironment through several mechanisms, including Treg
localization (Gil et al., 2014).
Acquired Resistance to Immunotherapy
A hallmark of cancer immunotherapy has been the induction
of long lasting tumor responses. However, with higher activity
and broader use of immunotherapies, the denominator of patients with a tumor response has increased and the chances
of finding patients who responded for a period of time and
then progressed, termed acquired resistance, increases. It is
becoming clear that approximately one fourth to one third of patients with metastatic melanoma who have objective responses
to checkpoint blockade therapy with anti-CTLA-4 or anti-PD-1
will relapse over time, even despite receiving continued therapy
(Schachter et al., 2016). The potential mechanisms of relapse
include loss of T cell function, lack of T cell recognition by downregulation of tumor antigen presentation, and development of
escape mutation variants in the cancer (Figures 2 and 3). There
is evidence that each of these mechanisms can lead to acquired
resistance to checkpoint inhibitor therapy or ACT.
If the anti-tumor T cells change their functional phenotype
and stop exerting their cytotoxic activity, then a patient who responded to immunotherapy may develop a tumor relapse even
if everything else continues to be the same. Acquired resistance
to TCR-engineered ACT is rather frequent, with high initial antitumor response followed by a high frequency of tumor relapses
within months. This has been evident with the ACT of T cells expressing TCRs to melanosomal antigens (MART-1, gp100) and
to cancer testis antigens (NY ESO-1) (Chodon et al., 2014; Morgan et al., 2006; Robbins et al., 2011). By studying how the TCR
transgenic T cells change their functionality after ACT to humans,
it has been reported that the initial highly cytolytic profile when
administered shifts over time to a Th2-type cytokine release
and lack of cytotoxic functions in late time points when recovered from patients at the time of tumor relapse (Ma et al.,
2013; Ma et al., 2011).
It was already well documented by the 1990s that some
patients who initially respond to cancer immunotherapies with
IL-2 or TIL ACT might develop acquired resistance through
loss of the shared component of all HLA class I molecules,
B2M, which leads to absence of surface expression of HLA
class I (D’Urso et al., 1991; Restifo et al., 1996). B2M is required
for HLA class I folding and transport to the cell surface, and
its genetic deficiency would lead to lack of CD8 T cell recognition. This mechanism of acquired resistance has also been
Cell 168, February 9, 2017 713
Table 3. Examples of Combination Therapies Being Developed to Overcome Resistance to Cancer Immunotherapy
Broad Approach
Specific Approach
Examples in Clinical Testing
combination checkpoint
blockade
anti-PD-1/L1 plus anti-CTLA4
Durvalumab + tremelimumab
Nivolumab + ipilimumab
Pembrolizumab + ipilimumab
anti-PD-1 plus anti-PD-L1
MEDI0680 + durvalumab
PDR001 + FAZ053
anti-PD-1/L1 plus anti-TIM 3
Nivolumab + TSR022
PDR001 + MBG453
anti-PD-1/L1 plus anti-LAG 3
Nivolumab + BMS 986016
PDR001 + LAG525
Pembrolizumab + IMP321
REGN2810 + REGN3767
anti-PD-1/L1 plus anti-41BB/CD137
Avelumab + utomilumab
Nivolumab + urelumab
Pembrolizumab + utomilumab
anti-CTLA4 plus anti-OX40
anti-PD-1/L1 plus anti-OX40
anti-CTLA4 plus Anti-PD-1/L1 plus anti-OX40
anti-41BB/CD137 plus anti-OX40
Atezolimumab + MOXR0916 ± bevacizumab
Avelumab + PF-04518600
Durvalumab + MEDI0562
Pembrolizumab + GSK3174998
Tremelimumab + durvalumab + MEDI6469
Tremelimumab + MEDI0562
Utomilumab + PF-04518600
anti-CTLA4 plus anti-CD40
anti-PD-1/L1 plus anti-CD40
Atezolimumab + RO7009789
Tremelimumab + CP870893
anti-PD-1/L1 plus anti-GITR
Nivolumab + BMS986156
PDR001 + GWN323
anti-PD-1/L1 plus anti-ICOS
Nivolumab + JTX-2011
anti-CTLA-4 plus IDO inhibitors
anti-PD-1/L1 plus IDO inhibitors
Atezolizumab + GDC0919
Ipilimumab + epacadostat
Ipilimumab + indoximid
Nivolumab + BMS986205
Pembrolizumab+ epacadostat
anti-PD-1/L1 plus A2AR inhibitors or anti-CD73
Atezolizumab + CPI-444
Durvalumab + MEDI9447
PDR001+ PBF509
anti-PD-1/L1 plus TGFb inhibitors
Nivolumab + LY2157299
PDR001 + NIS793
anti-PD-1/L1 plus CXCR4 inhibitors
Nivolumab + ulocuplumab
Durvalumab + LY2510924
anti-PD-1/L1 plus CCR4 inhibitors
Nivolumab + mogamulizumab
anti-PD-1/L1 plus anti-CD27
Nivolumab + varlilumab
Atezolizumab + varlilumab
anti-PD-1/L1 plus CD122-biased cytokine
Nivolumab + NKTR-214
anti-PD-1/L1 plus yeast-derived soluble b-glucan
Pembrolizumab + Imprime PGG
checkpoint blockade plus
immune-stimulatory agents
checkpoint blockade plus
metabolic modulators
checkpoint blockade plus
other immune modulators
checkpoint blockade plus
macrophage inhibitors
anti-PD-1/L1 plus anti- TRAIL-DR5
Nivolumab + DS-8273a
anti-PD-1/L1 plus glutaminase inhibitor
Nivolumab + CB839
anti-PD-1/L1 plus IAP inhibitor
PDR001 + LCL161
anti-CTLA4 plus CSF1R inhibitors
anti-PD-1/L1 plus CSF1R inhibitors
Durvalumab + Pexidartinib (PLX3397)
Durvalumab + LY3022855
Nivolumab + FPA008
Pembrolizumab + Pexidartinib
PDR001 + BLZ945
Tremelimumab + LY3022855
(Continued on next page)
714 Cell 168, February 9, 2017
Table 3.
Continued
Broad Approach
Specific Approach
Examples in Clinical Testing
checkpoint blockade plus
injectable therapies
anti-CTLA-4 plus oncolytic viruses
anti-PD-1/L1 plus oncolytic viruses
Ipilimumab + Talimogene Laherparepvec
Nivolumab + Talimogene Laherparepvec
Pembrolizumab + DNX2401
Pembrolizumab + Talimogene Laherparepvec
anti-CTLA4 plus TLR agonists
anti-PD-1/L1 plus TLR agonists
Ipilimumab + MGN1703
Pembrolizumab + CMP001
Pembrolizumab + SD101
Tremelimumab + PF-3512676
checkpoint blockade plus
cancer vaccines
anti-CTLA4 plus DC vaccine
anti-PD-1/L1 plus DC vaccine
anti-PD-1/L1 plus peptide vaccine
anti-PD-1/L1 plus neoantigen vaccine
Durvalumab + ADXS11-001
Durvalumab + TPIV200/huFR-1
Ipilimumab + GVAX
Nivolumab + GVAX + CRS207
Nivolumab + CIMAvax
Nivolumab+ CV301
Nivolumab + NEO-PV-01
Nivolumab + Viagenpumatucel-L (HS-110)
Pembrolizumab + ADXS31-142
Durvalumab ± tremelimumab + IMCgp100
checkpoint blockade plus
adoptive cell transfer (ACT)
anti-CTLA4 plus ACT
anti-PD-1/L1 plus ACT
anti-PD-1/L1 plus anti-CD137 plus ACT
Atezolimuamb + KTE-C19
Ipilimumab + NYESO TCR ACT
Nivolumab + NYESO TCR ACT
Nivolumab + urelumab + TIL ACT
Pembrolizumab + TIL ACT
Ipilimumab + modified CD8 T cell ACT
Pembrolizumab + modified CD8 T cell ACT
checkpoint blockade plus
targeted therapies
anti-CTLA4 plus BRAF+MEK inhibitors
anti-CTLA4 plus VEGF inhibitors
anti-PD-1/L1 plus BRAF+MEK inhibitors
anti-PD-1/L1 plus EGFR inhibitors
anti-PD-1/L1 plus VEGF inhibitors
anti-PD-1/L1 plus PI3K delta inhibitor
Atezolizumab + bevacizumab versus sunitinib
Atezolizumab + trametinib
Atezolizumab + vemurafenib ± cobimetinib
Durvalumab + ensartinib (ALK inhibitor)
Durvalumab + gefitinib
Durvalumab + trametinib ± dabrafenib
Ipilimumab + bevacizumab
Ipilimumab + dabrafenib ± trametinib
Ipilimumab +vemurafenib
Nivolumab + sunitinib or pazopanib
Nivolumab + trametinib ± dabrafenib
PDR001 + sorafenib
Pembrolizumab + dabrafenib + trametinib
Pembrolizumab + lenalidomide
Pembrolizumab + nintedarnib
Pidilizumab + lenalidomide
Tremelimumab + sunitinib
Nivolumab + SYM004
anti-PD-1/L1 plus PARP inhibitors
Atezolizumab + Veliparib
Durvalumab + olaparib
BGB-A317 + BGB-290
checkpoint blockade plus
radiation therapy (RT)
anti-PD-1/L1 plus mTOR inhibitor
PDR001 + everolimus
anti-PD-1/L1 plus pan RAF inhibitor
PDR001 + LXH254
anti-PD-1/L1 plus glutaminase inhibitor
Nivolumab + CB839
anti-CTLA4 plus RT
anti-PD-1/L1 plus RT
anti-CTLA4 plus Anti-PD-1/L1 plus RT
Atezolizumab + stereotactic radiation therapy
Pembrolizumab + cisplatin/radiotherapy
Pembrolizumab + sterotactic body radiotherapy
Pembrolizumab + hypofractionated radiotherapy
(Continued on next page)
Cell 168, February 9, 2017 715
Table 3.
Continued
Broad Approach
Specific Approach
Examples in Clinical Testing
checkpoint blockade plus
chemotherapy
anti-CTLA4 plus chemotherapy
anti-PD-1/L1 plus chemotherapy
anti-CTLA4 plus Anti-PD-1/L1 plus chemotherapy
Atezolizumab + carboplatin/paclitaxel
Atezolizumab + carboplatin/gemcitabine
Durvalumab + paclitaxel
Ipilimumab +carboplatin/paclitaxel
Ipilimumab +dacarbazine
Nivolumab + platinum doublets
Pembrolizumab + carbo/paclitaxel or carbo/pemetrexed
checkpoint blockade plus
epigenetic modifications
anti-PD-1/L1 plus histone deacetylase inhibitors
anti-PD-1/L1 plus hypomethylating agents
Azacitidine + entinostat followed by nivolumab
Atezolizumab + azacitidine
Nivolumab + RRX001
Pembrolizumab + CC486
Pembrolizumab + CC486 + romidepsin
Pembrolizumab + romidepsin
Pembrolizumab + vorinostat + tamoxifen
PDR001 + panobinostat
checkpoint blockade plus
NK activation
anti-CTLA4 plus anti-KIR
anti-PD-1/L1 plus anti-KIR
Ipilimumab + lirilumab
Nivolumab + lirilumab
documented in a case of late acquired resistance to anti-PD-1
therapy, where the resistant cells had a new and homozygous
truncating mutation in B2M, leading to lack of surface expression
of HLA class I (Zaretsky et al., 2016). In two other cases of tumor
relapse, there were copy-number-neutral loss-of-function mutations in JAK1 or JAK2, concurrent with loss of heterozygosity
due to deletion of the wild-type allele, which were absent in the
baseline biopsies. These mutations allowed the cancer cells to
escape from the anti-proliferative effects of interferon gamma
(Zaretsky et al., 2016). Additional evidence of loss of antigenpresenting machinery leading to acquired resistance to cancer
immunotherapy is provided by a case of a patient with metastatic
colorectal carcinoma who responded to TIL ACT. The therapeutic TIL recognized mutated KRAS G12D presented by
HLA-C*08:02, resulting in an objective tumor response for
9 months, followed by an isolated relapse in a lesion that had
lost HLA-C*08:02 in chromosome 6 (Tran et al., 2016). Therefore,
acquired resistance to anti-PD-1 therapy and ACT could be
mediated through genetic mechanisms that altered antigen-presenting machinery and interferon gamma signaling.
Because anti-tumor T cells are specific for cancer cells that
express their cognate antigen, it is possible that cancers may
develop acquired resistance through decreased expression or
mutations in these tumor antigens. Data suggest that anti-tumor
T cells turned on by checkpoint blockade therapy primarily
recognize mutational neoantigens (Schumacher and Schreiber,
2015; van Rooij et al., 2013). Therefore, genetic deletions,
mutations, or epigenetic changes that would lead to loss of
expression of these mutational neoantigens presented by MHC
molecules might result in acquired resistance to checkpoint
blockade therapy. However, thus far there has not been evidence of such mechanisms in the clinic. CAR T cells are also antigen-specific, but they rely on the whole protein expression
on the cancer cell surface. In some cases of patients with ALL
who responded initially to CD19 CAR T cell ACT, it has been
documented that the epitope in the CD19 protein sequence
that is recognized by the CAR can be selectively deleted at pro716 Cell 168, February 9, 2017
gression (Ruella et al., 2016) and that preexisting alternatively
spliced CD19 isoforms might predispose to acquired resistance
(Sotillo et al., 2015). Therefore, there is evidence from the clinic
that loss of the target of the anti-tumor T cells can result in progression to cancer immunotherapy.
This yin and yang of the immune response, which results in immune editing and eventually immune escape, is clearly a factor
as we administer immunotherapeutic agents and attempt to
drive anti-tumor immune responses, which may encounter a
multitude of inhibitory pathways, either during initial treatment
or at the time of relapsed disease. Additional inhibitory immune
checkpoints that are often expressed in the tumor microenvironment include LAG-3, TIGIT, VISTA, and many more that are being identified in ongoing studies (Topalian et al., 2015). Several
clinical trials are currently underway to test antibodies against
these inhibitory pathways, both as monotherapy and combination therapy strategies (Anderson et al., 2016; Sharma and Allison, 2015). To date, the combination of anti-CTLA-4 (ipilimumab)
plus anti-PD-1 (nivolumab) has demonstrated improved clinical
outcomes as compared to monotherapy, and this combination
was recently FDA-approved for patients with metastatic melanoma (Larkin et al., 2015a). We will need data from ongoing
and future clinical trials to determine whether combination
therapies targeting other inhibitory pathways, either as doublets
or triplets in concurrent or sequential treatment strategies,
will effectively overcome the resistance mechanisms that act
to regulate immune responses and provide additional clinical benefit.
Monitoring Resistance Mechanisms
There are significant efforts underway to identify reliable predictive biomarkers of response and resistance to checkpoint
inhibitors in baseline tumor biopsies in patients on immune
checkpoint blockade. To date, the best predictive biomarkers
identified include total tumor mutational load (Roszik et al.,
2016; Snyder et al., 2014), as well as markers of an effective
immune infiltrate within a tumor signifying a ‘‘hot’’ tumor
Figure 4. Schema for Analysis of Baseline and Longitudinal Tumor,
Blood, and Other Samples
(A) Baseline assessment of the tumor microenvironment typically involves
molecular analysis for mutational load, driver mutations, and gene expression,
with immune profiling including analysis of CD8+ T cells, PD-L1 expression,
and T cell clonality.
(B) Longitudinal evaluation of fresh serial human specimens (tumor, blood,
serum, and microbiome) during treatment (at pre-treatment, early-on-treatment, and progression time points) allows for deep analysis to unveil potential
mechanisms of therapeutic resistance.
microenvironment, typified by an increased number of CD8+
cytotoxic T lymphocytes in proximity to PD-L1-positive cells
(Taube et al., 2014; Tumeh et al., 2014). Mutational load is highly
relevant, given that tumors with a higher mutational load exhibit
higher levels of neoantigens capable of inducing anti-tumor immune responses, translating into a higher likelihood of response
to immune checkpoint blockade across several cancer types
(Rizvi et al., 2015; Snyder et al., 2014; Van Allen et al., 2015).
In addition to genomic markers and immune regulatory gene
expression profiles (Hugo et al., 2016), immune markers in pretreatment biopsies, including the density and distribution of
CD8+ T lymphocytes, PD-L1 expression, and T cell clonality
(Taube et al., 2014; Tumeh et al., 2014), have also been associated with differential responses to immune checkpoint blockade,
although significant limitations exist when each of these biomarkers is assessed in isolation. Integrative approaches incorporating analysis of several of these features have also been
developed, such as the cancer immunogram, which incorporates analysis of seven distinct features within the tumor
microenvironment: tumor sensitivity to immune effectors, tumor
foreignness, general immune status, immune cell infiltration,
absence of checkpoint molecule expression, absence of soluble
inhibitors such as interleukin-1 and interleukin-6, and absence
of inhibitory tumor metabolism (Blank et al., 2016). These efforts
are critical and will ultimately contribute to more personalized
treatment strategies for cancer immunotherapy.
An emerging strategy in elucidating mechanisms of response
and resistance to immune checkpoint blockade involves the
assessment of longitudinal tumor samples throughout the
course of treatment. This approach is powerful because it transcends conventional analysis of static time points and seeks to
identify superior predictive biomarkers by assessing dynamic responses to cancer treatment. Such an approach has been employed to better understand response and resistance to immune
checkpoint blockade (Chen et al., 2016; Hugo et al., 2016; Madore et al., 2015; Tumeh et al., 2014) and has yielded important
information that would not have been elucidated through analysis of static unpaired biopsies. A key example is in a recent
report describing immune markers in longitudinal tumor samples
of patients on immune checkpoint blockade, demonstrating that
although pre-treatment markers were largely non-predictive, immune markers in early-on-treatment samples were highly predictive of treatment response (Chen et al., 2016). In addition to this,
resistance mechanisms were identified via pairwise comparison
of gene expression profiles in pre- to on-treatment tumor samples of responders versus non-responders, including defects in
interferon signaling as well as antigen processing and presentation (Chen et al., 2016). This approach is currently under-utilized
but is gaining traction in light of advantages over assessment of
static baseline biomarkers (Figure 4), as well as an increasing
need to better understand responses to a growing number of
immunotherapeutic approaches. However nuances exist with regard to immune monitoring in the tumor microenvironment
(Wargo et al., 2016), and an appreciation of the importance of
concurrent monitoring in the peripheral blood is growing, though
the ideal assays to perform are still being elucidated.
Overcoming Resistance to Immunotherapy
On the basis of insights gained (Hugo et al., 2016; Snyder et al.,
2014; Van Allen et al., 2015), efforts are currently underway to
derive actionable strategies to combat therapeutic resistance
to immunotherapy. This includes fundamental efforts to transform immunologically ‘‘cold’’ tumors into ‘‘hot’’ tumors through
the use of several approaches (Corrales et al., 2015; Holmgaard
et al., 2013; Tang et al., 2016) and also involves tactics to either
enhance endogenous T cell function (Gubin et al., 2014; Hodi
et al., 2010; Miller et al., 2002; Redmond et al., 2007; Ribas
et al., 2015; Weber et al., 2015) or to adoptively transfer antigen-specific T lymphocytes via ex vivo expansion of tumor-infiltrating lymphocytes (Rosenberg et al., 2011) or via administration of antigen-specific engineered T cells (via transduction
with CARs or TCRs) (Beatty et al., 2014; Kalos et al., 2011).
Though some of these approaches involve treatment with
drugs as monotherapy (including monoclonal antibodies), the
majority of contemporary approaches focus on combination
strategies in an effort to overcome resistance associated
with treatment with single-pronged efforts (Table 3) (Hicklin
et al., 1998; Moon et al., 2014; Ninomiya et al., 2015). A prime
example of enhanced efficacy with combination therapy is the
use of combined therapy with blocking antibodies against two
key immune checkpoints, CTLA-4 and PD-1, which results in
significantly higher response rates to therapy and improved survival in patients with metastatic melanoma (Larkin et al., 2015a;
Postow et al., 2015; Wolchok et al., 2013). The rationale for this
combination approach is several fold, as blocking several checkpoints on anergized tumor-specific T cells has been shown to be
more efficacious (Berrien-Elliott et al., 2013; Curran et al., 2010;
Redmond et al., 2014; Spranger et al., 2014) and CTLA-4
blockade may itself facilitate the conversion of a tumor
Cell 168, February 9, 2017 717
microenvironment from ‘‘cold’’ to ‘‘hot’’ (Simpson et al., 2013).
Indeed, each of these checkpoint inhibitors has been shown to
have both overlapping and unique effects on tumor-specific
T cells (Gubin et al., 2014), substantiating the use of these in
combination. Numerous other strategies combining immune
modulation of the tumor microenvironment with immune checkpoint inhibitor therapy are currently being tested in clinical trials
(Puzanov et al., 2016) (NCT02263508, NCT02626000,
NCT02565992, NCT02043665, NCT02501473). Vaccine strategies against identified neoantigen epitopes are also being combined with immunotherapeutic approaches, though mature data
are not available regarding efficacy.
Another combination strategy with strong clinical and pre-clinical rationale involves the use of molecularly targeted therapy in
conjunction with immunotherapy. The most extensively studied
cancer type treated with this strategy is melanoma, though the
concept is now being widely extended across solid and liquid tumors. The rationale for combining these treatments is that treatment with molecularly targeted therapy can have a substantial
effect on anti-tumor immunity and potential synergy when used
with immunotherapy (Homet Moreno et al., 2015; Hu-Lieskovan
et al., 2015; Koya et al., 2012). Perhaps most illustrative of this is
oncogenic BRAF in melanoma. Though treatment with BRAFtargeted therapy alone provides limited durable disease control
(Chapman et al., 2011; Hauschild et al., 2012), it is associated
with favorable effects in the tumor microenvironment, including
increased antigen (Boni et al., 2010) and HLA expression (Bradley et al., 2015), increased T cell infiltrate, reduced immunosuppressive cytokines (Frederick et al., 2013; Wilmott et al., 2012),
and improved T cell function (Comin-Anduix et al., 2010). Thus,
treatment with molecularly targeted therapy may indeed help
convert a ‘‘cold’’ microenvironment to a ‘‘hot’’ one, with resultant
increased expression of PD-L1 via the phenomenon of adaptive
resistance (Taube et al., 2012), further supporting a multi-modality treatment approach. Emerging strategies to enhance responses to immunotherapy are being developed based on novel
insights into T cell and overall immune function. Examples of
this include insights into metabolic reprogramming of T cells to
enhance therapeutic responses (Buck et al., 2016; Chang and
Pearce, 2016) and via modulation of the gut microbiome to
augment responses to cancer immunotherapy (Sivan et al.,
2015; Vétizou et al., 2015).
Complexities exist when attempting to validate these combination strategies given that the extent of possible combinations far
outnumbers the human and technical resources available. There
is an urgent need to test these combinations in appropriate preclinical models and expedite clinical translation through novel approaches to clinical trial design. In addition, we need to have a
deep understanding of the kinetics of the immune response to
each of these agents in isolation as well as in combination in order
to narrow the search space of biologically promising and optimal
combination strategies. Immune responses to targeted agents
may be short-lived (Cooper et al., 2014), thus proper timing and
sequence of therapy must be strongly considered.
Conclusions
Great advances have occurred in the field of cancer immunotherapy as a result of elegant research work conducted to eluci718 Cell 168, February 9, 2017
date the mechanisms that regulate anti-tumor T cell responses,
including eventual translation of these concepts to the clinic.
This has allowed the rational design and clinical development
of treatment strategies that might result in tumor regression
and long-term survival for patients with metastatic cancer. However, the benefit, to date, has been limited to a minority of patients with certain cancer types. In addition, as a result of more
successful immunotherapy treatments, we now have a significant subset of patients who initially respond but eventually
relapse. Bringing clinical benefit to the majority of patients requires a complete understanding of the mechanisms that would
lead to an effective anti-tumor response and the different tumorcell-intrinsic and -extrinsic factors that would result in primary,
adaptive, and acquired resistance to immunotherapy. Elucidation of these mechanisms will reveal important clues as to the
next steps that need to be taken to potentially overcome resistance to immunotherapy.
ACKNOWLEDGMENTS
This research was supported by the Parker Institute for Cancer Immunotherapy,
NIH/NCI grants R01 CA1633793 (P.S.), R35 CA197633, and P01 CA168585
(A.R.), Cancer Prevention Research in Texas (CPRIT) grant RP120108 (P.S.),
the Ressler Family Fund, the Garcia-Corsini Family Fund, the Samuels Family
Fund, and the Grimaldi Family Fund (A.R.), and a Stand Up To Cancer-Cancer
Research Institute Cancer Immunology Dream Team Translational Research
Grant (SU2C-AACR-DT1012; P.S. and A.R.). S.H.-L. was supported by a Career
Development Award from the American Society of Clinical Oncology (ASCO), a
Tower Cancer Research Foundation Grant, a Dr. Charles Coltman Fellowship
Award from the Hope Foundation, and a UCLA KL2 Award.
REFERENCES
Akbay, E.A., Koyama, S., Carretero, J., Altabef, A., Tchaicha, J.H., Christensen, C.L., Mikse, O.R., Cherniack, A.D., Beauchamp, E.M., Pugh, T.J., et al.
(2013). Activation of the PD-1 pathway contributes to immune escape in
EGFR-driven lung tumors. Cancer Discov. 3, 1355–1363.
Allard, B., Turcotte, M., Spring, K., Pommey, S., Royal, I., and Stagg, J. (2014).
Anti-CD73 therapy impairs tumor angiogenesis. Int. J. Cancer 134,
1466–1473.
Anderson, A.C., Joller, N., and Kuchroo, V.K. (2016). Lag-3, Tim-3, and TIGIT:
Co-inhibitory Receptors with Specialized Functions in Immune Regulation.
Immunity 44, 989–1004.
Ansell, S.M., Lesokhin, A.M., Borrello, I., Halwani, A., Scott, E.C., Gutierrez, M.,
Schuster, S.J., Millenson, M.M., Cattry, D., Freeman, G.J., et al. (2015). PD-1
blockade with nivolumab in relapsed or refractory Hodgkin’s lymphoma.
N. Engl. J. Med. 372, 311–319.
Beatty, G.L., Haas, A.R., Maus, M.V., Torigian, D.A., Soulen, M.C., Plesa, G.,
Chew, A., Zhao, Y., Levine, B.L., Albelda, S.M., et al. (2014). Mesothelin-specific chimeric antigen receptor mRNA-engineered T cells induce anti-tumor
activity in solid malignancies. Cancer Immunol. Res. 2, 112–120.
Beavis, P.A., Slaney, C.Y., Milenkovski, N., Henderson, M.A., Loi, S., Stagg, J.,
Kershaw, M.H., and Darcy, P.K. (2015). CD73: A potential biomarker for antiPD-1 therapy. OncoImmunology 4, e1046675.
Benci, J.L., Xu, B., Qiu, Y., Wu, T.J., Dada, H., Twyman-Saint Victor, C.,
Cucolo, L., Lee, D.S., Pauken, K.E., Huang, A.C., et al. (2016). Tumor interferon
signaling regulates a multigenic resistance program to immune checkpoint
blockade. Cell 167, 1540–1554.
Berrien-Elliott, M.M., Jackson, S.R., Meyer, J.M., Rouskey, C.J., Nguyen, T.L.,
Yagita, H., Greenberg, P.D., DiPaolo, R.J., and Teague, R.M. (2013). Durable
adoptive immunotherapy for leukemia produced by manipulation of multiple
regulatory pathways of CD8+ T-cell tolerance. Cancer Res. 73, 605–616.
Biswas, S.K., and Mantovani, A. (2010). Macrophage plasticity and interaction
with lymphocyte subsets: cancer as a paradigm. Nat. Immunol. 11, 889–896.
Blank, C.U., Haanen, J.B., Ribas, A., and Schumacher, T.N. (2016). Cancer
immunology. The ‘‘cancer immunogram.’’. Science 352, 658–660.
Boni, A., Cogdill, A.P., Dang, P., Udayakumar, D., Njauw, C.N., Sloss, C.M.,
Ferrone, C.R., Flaherty, K.T., Lawrence, D.P., Fisher, D.E., et al. (2010). Selective BRAFV600E inhibition enhances T-cell recognition of melanoma without
affecting lymphocyte function. Cancer Res. 70, 5213–5219.
Bradley, S.D., Chen, Z., Melendez, B., Talukder, A., Khalili, J.S., RodriguezCruz, T., Liu, S., Whittington, M., Deng, W., Li, F., et al. (2015). BRAFV600E
co-opts a conserved MHC class I internalization pathway to diminish antigen
presentation and CD8+ T-cell recognition of melanoma. Cancer Immunol.
Res. 3, 602–609.
Bronte, V., Wang, M., Overwijk, W.W., Surman, D.R., Pericle, F., Rosenberg,
S.A., and Restifo, N.P. (1998). Apoptotic death of CD8+ T lymphocytes after
immunization: induction of a suppressive population of Mac-1+/Gr-1+ cells.
J. Immunol. 161, 5313–5320.
Brunet, J.F., Denizot, F., Luciani, M.F., Roux-Dosseto, M., Suzan, M., Mattei,
M.G., and Golstein, P. (1987). A new member of the immunoglobulin superfamily–CTLA-4. Nature 328, 267–270.
Buck, M.D., O’Sullivan, D., Klein Geltink, R.I., Curtis, J.D., Chang, C.H., Sanin,
D.E., Qiu, J., Kretz, O., Braas, D., van der Windt, G.J., et al. (2016). Mitochondrial Dynamics Controls T Cell Fate through Metabolic Programming. Cell 166,
63–76.
Burnet, F.M. (1971). Immunological surveillance in neoplasia. Transplant. Rev.
7, 3–25.
Casey, S.C., Tong, L., Li, Y., Do, R., Walz, S., Fitzgerald, K.N., Gouw, A.M.,
Baylot, V., Gütgemann, I., Eilers, M., and Felsher, D.W. (2016). MYC regulates
the antitumor immune response through CD47 and PD-L1. Science 352,
227–231.
Chang, C.H., and Pearce, E.L. (2016). Emerging concepts of T cell metabolism
as a target of immunotherapy. Nat. Immunol. 17, 364–368.
Chang, D.-K., Sui, J., Geng, S., Muvaffak, A., Bai, M., Fuhlbrigge, R.C., Lo, A.,
Yammanuru, A., Hubbard, L., Sheehan, J., et al. (2012). Humanization of an
anti-CCR4 antibody that kills cutaneous T-cell lymphoma cells and abrogates
suppression by T-regulatory cells. Mol. Cancer Ther. 11, 2451–2461.
Chanmee, T., Ontong, P., Konno, K., and Itano, N. (2014). Tumor-associated
macrophages as major players in the tumor microenvironment. Cancers
(Basel) 6, 1670–1690.
Chapman, P.B., Hauschild, A., Robert, C., Haanen, J.B., Ascierto, P., Larkin,
J., Dummer, R., Garbe, C., Testori, A., Maio, M., et al.; BRIM-3 Study Group
(2011). Improved survival with vemurafenib in melanoma with BRAF V600E
mutation. N. Engl. J. Med. 364, 2507–2516.
Chaudhary, B., and Elkord, E. (2016). Regulatory T Cells in the Tumor Microenvironment and Cancer Progression: Role and Therapeutic Targeting. Vaccines
(Basel) 4.
Chen, L. (2004). Co-inhibitory molecules of the B7-CD28 family in the control of
T-cell immunity. Nat. Rev. Immunol. 4, 336–347.
Chen, P.L., Roh, W., Reuben, A., Cooper, Z.A., Spencer, C.N., Prieto, P.A.,
Miller, J.P., Bassett, R.L., Gopalakrishnan, V., Wani, K., et al. (2016). Analysis
of Immune Signatures in Longitudinal Tumor Samples Yields Insight into Biomarkers of Response and Mechanisms of Resistance to Immune Checkpoint
Blockade. Cancer Discov. 6, 827–837.
Chodon, T., Comin-Anduix, B., Chmielowski, B., Koya, R.C., Wu, Z., Auerbach, M., Ng, C., Avramis, E., Seja, E., Villanueva, A., et al. (2014). Adoptive
transfer of MART-1 T-cell receptor transgenic lymphocytes and dendritic
cell vaccination in patients with metastatic melanoma. Clin. Cancer Res. 20,
2457–2465.
Coley, W.B. (1910). The treatment of inoperable sarcoma by bacterial toxins
(the mixed toxins of the Streptococcus erysipelas and the Bacillus prodigiosus). Proc. R. Soc. Med. 3(Surg Sect), 1–48.
Comin-Anduix, B., Chodon, T., Sazegar, H., Matsunaga, D., Mock, S., Jalil, J.,
Escuin-Ordinas, H., Chmielowski, B., Koya, R.C., and Ribas, A. (2010). The
oncogenic BRAF kinase inhibitor PLX4032/RG7204 does not affect the viability
or function of human lymphocytes across a wide range of concentrations. Clin.
Cancer Res. 16, 6040–6048.
Cooper, Z.A., Juneja, V.R., Sage, P.T., Frederick, D.T., Piris, A., Mitra, D., Lo,
J.A., Hodi, F.S., Freeman, G.J., Bosenberg, M.W., et al. (2014). Response
to BRAF inhibition in melanoma is enhanced when combined with immune
checkpoint blockade. Cancer Immunol. Res. 2, 643–654.
Corrales, L., Glickman, L.H., McWhirter, S.M., Kanne, D.B., Sivick, K.E.,
Katibah, G.E., Woo, S.R., Lemmens, E., Banda, T., Leong, J.J., et al. (2015).
Direct activation of STING in the tumor microenvironment leads to potent
and systemic tumor regression and immunity. Cell Rep. 11, 1018–1030.
Curran, M.A., Montalvo, W., Yagita, H., and Allison, J.P. (2010). PD-1 and
CTLA-4 combination blockade expands infiltrating T cells and reduces regulatory T and myeloid cells within B16 melanoma tumors. Proc. Natl. Acad. Sci.
USA 107, 4275–4280.
D’Urso, C.M., Wang, Z.G., Cao, Y., Tatake, R., Zeff, R.A., and Ferrone, S.
(1991). Lack of HLA class I antigen expression by cultured melanoma cells
FO-1 due to a defect in B2m gene expression. J. Clin. Invest. 87, 284–292.
Darnell, J.E., Jr., Kerr, I.M., and Stark, G.R. (1994). Jak-STAT pathways and
transcriptional activation in response to IFNs and other extracellular signaling
proteins. Science 264, 1415–1421.
Das, R., Verma, R., Sznol, M., Boddupalli, C.S., Gettinger, S.N., Kluger, H.,
Callahan, M., Wolchok, J.D., Halaban, R., Dhodapkar, M.V., and Dhodapkar,
K.M. (2015). Combination therapy with anti-CTLA-4 and anti-PD-1 leads to
distinct immunologic changes in vivo. J. Immunol. 194, 950–959.
De Henau, O., Rausch, M., Winkler, D., Campesato, L.F., Liu, C., Cymerman,
D.H., Budhu, S., Ghosh, A., Pink, M., Tchaicha, J., et al. (2016). Overcoming
resistance to checkpoint blockade therapy by targeting PI3Kg in myeloid cells.
Nature 539, 443–447.
De Palma, M., and Lewis, C.E. (2013). Macrophage regulation of tumor responses to anticancer therapies. Cancer Cell 23, 277–286.
Dong, H., Zhu, G., Tamada, K., and Chen, L. (1999). B7-H1, a third member of
the B7 family, co-stimulates T-cell proliferation and interleukin-10 secretion.
Nat. Med. 5, 1365–1369.
Dong, H., Strome, S.E., Salomao, D.R., Tamura, H., Hirano, F., Flies, D.B.,
Roche, P.C., Lu, J., Zhu, G., Tamada, K., et al. (2002). Tumor-associated
B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion.
Nat. Med. 8, 793–800.
Dorand, R.D., Nthale, J., Myers, J.T., Barkauskas, D.S., Avril, S., Chirieleison,
S.M., Pareek, T.K., Abbott, D.W., Stearns, D.S., Letterio, J.J., et al. (2016).
Cdk5 disruption attenuates tumor PD-L1 expression and promotes antitumor
immunity. Science 353, 399–403.
Dunn, G.P., Bruce, A.T., Ikeda, H., Old, L.J., and Schreiber, R.D. (2002). Cancer immunoediting: from immunosurveillance to tumor escape. Nat. Immunol.
3, 991–998.
Dunn, G.P., Bruce, A.T., Sheehan, K.C., Shankaran, V., Uppaluri, R., Bui, J.D.,
Diamond, M.S., Koebel, C.M., Arthur, C., White, J.M., and Schreiber, R.D.
(2005). A critical function for type I interferons in cancer immunoediting. Nat.
Immunol. 6, 722–729.
Ehrlich, P. (1956). Collected Papers in Four Volumes Including A Complete
Bibliography (Pergamon Press).
Eshhar, Z., Waks, T., Gross, G., and Schindler, D.G. (1993). Specific activation
and targeting of cytotoxic lymphocytes through chimeric single chains consisting of antibody-binding domains and the gamma or zeta subunits of the
immunoglobulin and T-cell receptors. Proc. Natl. Acad. Sci. USA 90, 720–724.
Fesnak, A.D., June, C.H., and Levine, B.L. (2016). Engineered T cells:
the promise and challenges of cancer immunotherapy. Nat. Rev. Cancer 16,
566–581.
Frederick, D.T., Piris, A., Cogdill, A.P., Cooper, Z.A., Lezcano, C., Ferrone,
C.R., Mitra, D., Boni, A., Newton, L.P., Liu, C., et al. (2013). BRAF inhibition
is associated with enhanced melanoma antigen expression and a more
favorable tumor microenvironment in patients with metastatic melanoma.
Clin. Cancer Res. 19, 1225–1231.
Cell 168, February 9, 2017 719
Freeman, G.J., Long, A.J., Iwai, Y., Bourque, K., Chernova, T., Nishimura, H.,
Fitz, L.J., Malenkovich, N., Okazaki, T., Byrne, M.C., et al. (2000). Engagement
of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to
negative regulation of lymphocyte activation. J. Exp. Med. 192, 1027–1034.
Fritz, J.M., Tennis, M.A., Orlicky, D.J., Lin, H., Ju, C., Redente, E.F., Choo,
K.S., Staab, T.A., Bouchard, R.J., Merrick, D.T., et al. (2014). Depletion of
tumor-associated macrophages slows the growth of chemically induced
mouse lung adenocarcinomas. Front. Immunol. 5, 587.
Gajewski, T.F., Schreiber, H., and Fu, Y.X. (2013). Innate and adaptive immune
cells in the tumor microenvironment. Nat. Immunol. 14, 1014–1022.
Gao, J., Shi, L.Z., Zhao, H., Chen, J., Xiong, L., He, Q., Chen, T., Roszik, J.,
Bernatchez, C., Woodman, S.E., et al. (2016). Loss of IFN-g pathway genes
in tumor cells as a mechanism of resistance to anti-CTLA-4 therapy. Cell
167, 397–404.
Gershon, R.K., and Kondo, K. (1970). Cell interactions in the induction of tolerance: the role of thymic lymphocytes. Immunology 18, 723–737.
Gil, M., Komorowski, M.P., Seshadri, M., Rokita, H., McGray, A.J., Opyrchal,
M., Odunsi, K.O., and Kozbor, D. (2014). CXCL12/CXCR4 blockade by oncolytic virotherapy inhibits ovarian cancer growth by decreasing immunosuppression and targeting cancer-initiating cells. J. Immunol. 193, 5327–5337.
Gray-Owen, S.D., and Blumberg, R.S. (2006). CEACAM1: contact-dependent
control of immunity. Nat. Rev. Immunol. 6, 433–446.
Green, M.R., Monti, S., Rodig, S.J., Juszczynski, P., Currie, T., O’Donnell, E.,
Chapuy, B., Takeyama, K., Neuberg, D., Golub, T.R., et al. (2010). Integrative
analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma
and primary mediastinal large B-cell lymphoma. Blood 116, 3268–3277.
Gubin, M.M., Zhang, X., Schuster, H., Caron, E., Ward, J.P., Noguchi, T., Ivanova, Y., Hundal, J., Arthur, C.D., Krebber, W.J., et al. (2014). Checkpoint
blockade cancer immunotherapy targets tumour-specific mutant antigens.
Nature 515, 577–581.
Hamid, O., Schmidt, H., Nissan, A., Ridolfi, L., Aamdal, S., Hansson, J., Guida,
M., Hyams, D.M., Gómez, H., Bastholt, L., et al. (2011). A prospective phase II
trial exploring the association between tumor microenvironment biomarkers
and clinical activity of ipilimumab in advanced melanoma. J. Transl. Med.
9, 204.
Hanks, B.A., Holtzhausen, A., Evans, K., Heid, M., and Blobe, G.C. (2014).
Combinatorial TGF-beta}signaling blockade and anti-CTLA-4 antibody immunotherapy in a murine BRAFV600E-PTEN-/-transgenic model of melanoma.
ASCO Annual Meeting Proceedings.
Hauschild, A., Grob, J.J., Demidov, L.V., Jouary, T., Gutzmer, R., Millward, M.,
Rutkowski, P., Blank, C.U., Miller, W.H., Jr., Kaempgen, E., et al. (2012).
Dabrafenib in BRAF-mutated metastatic melanoma: a multicentre, open-label,
phase 3 randomised controlled trial. Lancet 380, 358–365.
Héninger, E., Krueger, T.E., and Lang, J.M. (2015). Augmenting antitumor immune responses with epigenetic modifying agents. Front. Immunol. 6, 29.
Hicklin, D.J., Wang, Z., Arienti, F., Rivoltini, L., Parmiani, G., and Ferrone, S.
(1998). beta2-Microglobulin mutations, HLA class I antigen loss, and tumor
progression in melanoma. J. Clin. Invest. 101, 2720–2729.
Highfill, S.L., Cui, Y., Giles, A.J., Smith, J.P., Zhang, H., Morse, E., Kaplan,
R.N., and Mackall, C.L. (2014). Disruption of CXCR2-mediated MDSC tumor
trafficking enhances anti-PD1 efficacy. Sci. Transl. Med. 6.
Hodi, F.S., O’Day, S.J., McDermott, D.F., Weber, R.W., Sosman, J.A., Haanen,
J.B., Gonzalez, R., Robert, C., Schadendorf, D., Hassel, J.C., et al. (2010).
Improved survival with ipilimumab in patients with metastatic melanoma.
N. Engl. J. Med. 363, 711–723.
Hu, W., Li, X., Zhang, C., Yang, Y., Jiang, J., and Wu, C. (2016). Tumor-associated macrophages in cancers. Clin. Transl. Oncol. 18, 251–258.
Hu-Lieskovan, S., Mok, S., Homet Moreno, B., Tsoi, J., Robert, L., Goedert, L.,
Pinheiro, E.M., Koya, R.C., Graeber, T.G., Comin-Anduix, B., and Ribas, A.
(2015). Improved antitumor activity of immunotherapy with BRAF and MEK
inhibitors in BRAF(V600E) melanoma. Sci. Transl. Med. 7, 279ra41.
Hugo, W., Zaretsky, J.M., Sun, L., Song, C., Moreno, B.H., Hu-Lieskovan, S.,
Berent-Maoz, B., Pang, J., Chmielowski, B., Cherry, G., et al. (2016). Genomic
and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic
Melanoma. Cell 165, 35–44.
Ishida, Y., Agata, Y., Shibahara, K., and Honjo, T. (1992). Induced expression
of PD-1, a novel member of the immunoglobulin gene superfamily, upon programmed cell death. EMBO J. 11, 3887–3895.
Kalos, M., Levine, B.L., Porter, D.L., Katz, S., Grupp, S.A., Bagg, A., and June,
C.H. (2011). T cells with chimeric antigen receptors have potent antitumor
effects and can establish memory in patients with advanced leukemia. Sci.
Transl. Med. 3, 95ra73.
Kaneda, M.M., Messer, K.S., Ralainirina, N., Li, H., Leem, C.J., Gorjestani, S.,
Woo, G., Nguyen, A.V., Figueiredo, C.C., Foubert, P., et al. (2016). PI3Kg is a
molecular switch that controls immune suppression. Nature 539, 437–442.
Kaplan, M.H., Wurster, A.L., and Grusby, M.J. (1998). A signal transducer and
activator of transcription (Stat)4-independent pathway for the development of
T helper type 1 cells. J. Exp. Med. 188, 1191–1196.
Karpf, A.R., and Jones, D.A. (2002). Reactivating the expression of methylation
silenced genes in human cancer. Oncogene 21, 5496–5503.
Kataoka, K., Shiraishi, Y., Takeda, Y., Sakata, S., Matsumoto, M., Nagano, S.,
Maeda, T., Nagata, Y., Kitanaka, A., Mizuno, S., et al. (2016). Aberrant PD-L1
expression through 30 -UTR disruption in multiple cancers. Nature 534,
402–406.
Kim, H.J., and Bae, S.C. (2011). Histone deacetylase inhibitors: molecular
mechanisms of action and clinical trials as anti-cancer drugs. Am. J. Transl.
Res. 3, 166–179.
Kodumudi, K.N., Weber, A., Sarnaik, A.A., and Pilon-Thomas, S. (2012).
Blockade of myeloid-derived suppressor cells after induction of lymphopenia
improves adoptive T cell therapy in a murine model of melanoma. J. Immunol.
189, 5147–5154.
Koya, R.C., Mok, S., Otte, N., Blacketor, K.J., Comin-Anduix, B., Tumeh, P.C.,
Minasyan, A., Graham, N.A., Graeber, T.G., Chodon, T., and Ribas, A. (2012).
BRAF inhibitor vemurafenib improves the antitumor activity of adoptive cell
immunotherapy. Cancer Res. 72, 3928–3937.
Koyama, S., Akbay, E.A., Li, Y.Y., Herter-Sprie, G.S., Buczkowski, K.A.,
Richards, W.G., Gandhi, L., Redig, A.J., Rodig, S.J., Asahina, H., et al.
(2016). Adaptive resistance to therapeutic PD-1 blockade is associated with
upregulation of alternative immune checkpoints. Nat. Commun. 7, 10501.
Krummel, M.F., and Allison, J.P. (1995). CD28 and CTLA-4 have opposing
effects on the response of T cells to stimulation. J. Exp. Med. 182, 459–465.
Kryczek, I., Zou, L., Rodriguez, P., Zhu, G., Wei, S., Mottram, P., Brumlik, M.,
Cheng, P., Curiel, T., Myers, L., et al. (2006). B7-H4 expression identifies
a novel suppressive macrophage population in human ovarian carcinoma.
J. Exp. Med. 203, 871–881.
Kuang, D.M., Zhao, Q., Peng, C., Xu, J., Zhang, J.P., Wu, C., and Zheng, L.
(2009). Activated monocytes in peritumoral stroma of hepatocellular carcinoma foster immune privilege and disease progression through PD-L1.
J. Exp. Med. 206, 1327–1337.
Holmgaard, R.B., Zamarin, D., Munn, D.H., Wolchok, J.D., and Allison, J.P.
(2013). Indoleamine 2,3-dioxygenase is a critical resistance mechanism in antitumor T cell immunotherapy targeting CTLA-4. J. Exp. Med. 210, 1389–1402.
Laborde, R.R., Lin, Y., Gustafson, M.P., Bulur, P.A., and Dietz, A.B. (2014).
Cancer Vaccines in the World of Immune Suppressive Monocytes (CD14(+)
HLA-DR(lo/neg) Cells): The Gateway to Improved Responses. Front. Immunol.
5, 147.
Homet Moreno, B., Mok, S., Comin-Anduix, B., Hu-Lieskovan, S., and Ribas,
A. (2015). Combined treatment with dabrafenib and trametinib with immunestimulating antibodies for BRAF mutant melanoma. OncoImmunology 5,
e1052212.
Larkin, J., Chiarion-Sileni, V., Gonzalez, R., Grob, J.J., Cowey, C.L., Lao, C.D.,
Schadendorf, D., Dummer, R., Smylie, M., Rutkowski, P., et al. (2015a). Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma.
N. Engl. J. Med. 373, 23–34.
720 Cell 168, February 9, 2017
Larkin, J., Lao, C.D., Urba, W.J., McDermott, D.F., Horak, C., Jiang, J., and
Wolchok, J.D. (2015b). Efficacy and safety of nivolumab in patients with
BRAF V600 mutant and BRAF wild-type advanced melanoma: a pooled analysis of 4 clinical trials. JAMA Oncol. 1, 433–440.
Miller, R.E., Jones, J., Le, T., Whitmore, J., Boiani, N., Gliniak, B., and Lynch,
D.H. (2002). 4-1BB-specific monoclonal antibody promotes the generation
of tumor-specific immune responses by direct activation of CD8 T cells in a
CD40-dependent manner. J. Immunol. 169, 1792–1800.
Lastwika, K.J., Wilson, W., 3rd, Li, Q.K., Norris, J., Xu, H., Ghazarian, S.R., Kitagawa, H., Kawabata, S., Taube, J.M., Yao, S., et al. (2016). Control of PD-L1
expression by oncogenic activation of the AKT-mTOR pathway in non-small
cell lung cancer. Cancer Res. 76, 227–238.
Mittal, D., Sinha, D., Barkauskas, D., Young, A., Kalimutho, M., Stannard, K.,
Caramia, F., Haibe-Kains, B., Stagg, J., and Khanna, K.K. (2016). Adenosine
2B receptor expression on cancer cells promotes metastasis. Cancer Res. 76.
Le, D.T., Lutz, E., Uram, J.N., Sugar, E.A., Onners, B., Solt, S., Zheng, L.,
Diaz, L.A., Jr., Donehower, R.C., Jaffee, E.M., and Laheru, D.A. (2013). Evaluation of ipilimumab in combination with allogeneic pancreatic tumor cells
transfected with a GM-CSF gene in previously treated pancreatic cancer.
J. Immunother. 36, 382–389.
Leach, D.R., Krummel, M.F., and Allison, J.P. (1996). Enhancement of antitumor immunity by CTLA-4 blockade. Science 271, 1734–1736.
Lebrun, J.-J. (2012). The dual role of TGFb in human cancer: from tumor suppression to cancer metastasis. ISRN Mol. Biol. 2012, 381428.
Leclerc, B.G., Charlebois, R., Chouinard, G., Allard, B., Pommey, S., Saad, F.,
and Stagg, J. (2016). CD73 expression is an independent prognostic factor in
prostate cancer. Clin. Cancer Res. 22, 158–166.
Lin, R.L., and Zhao, L.J. (2015). Mechanistic basis and clinical relevance of the
role of transforming growth factor-b in cancer. Cancer Biol. Med. 12, 385–393.
Mok, S., Koya, R.C., Tsui, C., Xu, J., Robert, L., Wu, L., Graeber, T.G., West,
B.L., Bollag, G., and Ribas, A. (2014). Inhibition of CSF-1 receptor improves
the antitumor efficacy of adoptive cell transfer immunotherapy. Cancer Res.
74, 153–161.
Moon, E.K., Wang, L.C., Dolfi, D.V., Wilson, C.B., Ranganathan, R., Sun, J.,
Kapoor, V., Scholler, J., Pure, E., Milone, M.C., et al. (2014). Multifactorial
T-cell hypofunction that is reversible can limit the efficacy of chimeric antigen
receptor-transduced human T cells in solid tumors. Clin. Cancer Res. 20,
4262–4273.
Morgan, R.A., Dudley, M.E., Wunderlich, J.R., Hughes, M.S., Yang, J.C.,
Sherry, R.M., Royal, R.E., Topalian, S.L., Kammula, U.S., Restifo, N.P., et al.
(2006). Cancer regression in patients after transfer of genetically engineered
lymphocytes. Science 314, 126–129.
Linehan, D.C., and Goedegebuure, P.S. (2005). CD25+ CD4+ regulatory
T-cells in cancer. Immunol. Res. 32, 155–168.
Ninomiya, S., Narala, N., Huye, L., Yagyu, S., Savoldo, B., Dotti, G., Heslop,
H.E., Brenner, M.K., Rooney, C.M., and Ramos, C.A. (2015). Tumor indoleamine 2,3-dioxygenase (IDO) inhibits CD19-CAR T cells and is downregulated
by lymphodepleting drugs. Blood 125, 3905–3916.
Liu, C., Peng, W., Xu, C., Lou, Y., Zhang, M., Wargo, J.A., Chen, J.Q., Li, H.S.,
Watowich, S.S., Yang, Y., et al. (2013). BRAF inhibition increases tumor infiltration by T cells and enhances the antitumor activity of adoptive immunotherapy
in mice. Clin. Cancer Res. 19, 393–403.
Oida, T., Zhang, X., Goto, M., Hachimura, S., Totsuka, M., Kaminogawa, S.,
and Weiner, H.L. (2003). CD4+CD25- T cells that express latency-associated
peptide on the surface suppress CD4+CD45RBhigh-induced colitis by a TGFbeta-dependent mechanism. J. Immunol. 170, 2516–2522.
Loi, S., Pommey, S., Haibe-Kains, B., Beavis, P.A., Darcy, P.K., Smyth, M.J.,
and Stagg, J. (2013). CD73 promotes anthracycline resistance and poor
prognosis in triple negative breast cancer. Proc. Natl. Acad. Sci. USA 110,
11091–11096.
Okazaki, T., Chikuma, S., Iwai, Y., Fagarasan, S., and Honjo, T. (2013). A rheostat for immune responses: the unique properties of PD-1 and their advantages for clinical application. Nat. Immunol. 14, 1212–1218.
Luo, Y., Zhou, H., Krueger, J., Kaplan, C., Lee, S.H., Dolman, C., Markowitz,
D., Wu, W., Liu, C., Reisfeld, R.A., and Xiang, R. (2006). Targeting tumor-associated macrophages as a novel strategy against breast cancer. J. Clin. Invest.
116, 2132–2141.
Ma, C., Fan, R., Ahmad, H., Shi, Q., Comin-Anduix, B., Chodon, T., Koya, R.C.,
Liu, C.C., Kwong, G.A., Radu, C.G., et al. (2011). A clinical microchip for evaluation of single immune cells reveals high functional heterogeneity in phenotypically similar T cells. Nat. Med. 17, 738–743.
Ma, C., Cheung, A.F., Chodon, T., Koya, R.C., Wu, Z., Ng, C., Avramis, E.,
Cochran, A.J., Witte, O.N., Baltimore, D., et al. (2013). Multifunctional T-cell analyses to study response and progression in adoptive cell transfer immunotherapy. Cancer Discov. 3, 418–429.
Madore, J., Vilain, R.E., Menzies, A.M., Kakavand, H., Wilmott, J.S., Hyman, J.,
Yearley, J.H., Kefford, R.F., Thompson, J.F., Long, G.V., et al. (2015). PD-L1
expression in melanoma shows marked heterogeneity within and between patients: implications for anti-PD-1/PD-L1 clinical trials. Pigment Cell Melanoma
Res. 28, 245–253.
Marincola, F.M., Jaffee, E.M., Hicklin, D.J., and Ferrone, S. (2000). Escape of
human solid tumors from T-cell recognition: molecular mechanisms and functional significance. Adv. Immunol. 74, 181–273.
Massagué, J. (2008). TGFbeta in Cancer. Cell 134, 215–230.
Maude, S.L., Frey, N., Shaw, P.A., Aplenc, R., Barrett, D.M., Bunin, N.J., Chew,
A., Gonzalez, V.E., Zheng, Z., Lacey, S.F., et al. (2014). Chimeric antigen receptor T cells for sustained remissions in leukemia. N. Engl. J. Med. 371,
1507–1517.
Meyer, C., Cagnon, L., Costa-Nunes, C.M., Baumgaertner, P., Montandon, N.,
Leyvraz, L., Michielin, O., Romano, E., and Speiser, D.E. (2014). Frequencies of
circulating MDSC correlate with clinical outcome of melanoma patients
treated with ipilimumab. Cancer Immunol. Immunother. 63, 247–257.
Miller, J.F., and Sadelain, M. (2015). The journey from discoveries in fundamental immunology to cancer immunotherapy. Cancer Cell 27, 439–449.
Ormandy, L.A., Hillemann, T., Wedemeyer, H., Manns, M.P., Greten, T.F., and
Korangy, F. (2005). Increased populations of regulatory T cells in peripheral
blood of patients with hepatocellular carcinoma. Cancer Res. 65, 2457–2464.
Ortenberg, R., Sapir, Y., Raz, L., Hershkovitz, L., Ben Arav, A., Sapoznik, S.,
Barshack, I., Avivi, C., Berkun, Y., Besser, M.J., et al. (2012). Novel immunotherapy for malignant melanoma with a monoclonal antibody that blocks
CEACAM1 homophilic interactions. Mol. Cancer Ther. 11, 1300–1310.
Pardoll, D.M. (2012). The blockade of immune checkpoints in cancer immunotherapy. Nat. Rev. Cancer 12, 252–264.
Parsa, A.T., Waldron, J.S., Panner, A., Crane, C.A., Parney, I.F., Barry, J.J.,
Cachola, K.E., Murray, J.C., Tihan, T., Jensen, M.C., et al. (2007). Loss of tumor
suppressor PTEN function increases B7-H1 expression and immunoresistance in glioma. Nat. Med. 13, 84–88.
Peng, W., Chen, J.Q., Liu, C., Malu, S., Creasy, C., Tetzlaff, M.T., Xu, C.,
McKenzie, J.A., Zhang, C., Liang, X., et al. (2016). Loss of PTEN promotes
resistance to T cell-mediated immunotherapy. Cancer Discov. 6, 202–216.
Platanias, L.C. (2005). Mechanisms of type-I- and type-II-interferon-mediated
signalling. Nat. Rev. Immunol. 5, 375–386.
Postow, M.A., Chesney, J., Pavlick, A.C., Robert, C., Grossmann, K.,
McDermott, D., Linette, G.P., Meyer, N., Giguere, J.K., Agarwala, S.S., et al.
(2015). Nivolumab and ipilimumab versus ipilimumab in untreated melanoma.
N. Engl. J. Med. 372, 2006–2017.
Puzanov, I., Milhem, M.M., Minor, D., Hamid, O., Li, A., Chen, L., Chastain, M.,
Gorski, K.S., Anderson, A., Chou, J., et al. (2016). Talimogene Laherparepvec
in Combination With Ipilimumab in Previously Untreated, Unresectable Stage
IIIB-IV Melanoma. J. Clin. Oncol. 34, 2619–2626.
Quezada, S.A., Peggs, K.S., Curran, M.A., and Allison, J.P. (2006). CTLA4
blockade and GM-CSF combination immunotherapy alters the intratumor
balance of effector and regulatory T cells. J. Clin. Invest. 116, 1935–1945.
Redmond, W.L., Gough, M.J., Charbonneau, B., Ratliff, T.L., and Weinberg,
A.D. (2007). Defects in the acquisition of CD8 T cell effector function after
Cell 168, February 9, 2017 721
priming with tumor or soluble antigen can be overcome by the addition of an
OX40 agonist. J. Immunol. 179, 7244–7253.
Sakaguchi, S., Yamaguchi, T., Nomura, T., and Ono, M. (2008). Regulatory
T cells and immune tolerance. Cell 133, 775–787.
Redmond, W.L., Linch, S.N., and Kasiewicz, M.J. (2014). Combined targeting
of costimulatory (OX40) and coinhibitory (CTLA-4) pathways elicits potent
effector T cells capable of driving robust antitumor immunity. Cancer Immunol.
Res. 2, 142–153.
Sakuishi, K., Apetoh, L., Sullivan, J.M., Blazar, B.R., Kuchroo, V.K., and Anderson, A.C. (2010). Targeting Tim-3 and PD-1 pathways to reverse T cell exhaustion and restore anti-tumor immunity. J. Exp. Med. 207, 2187–2194.
Restifo, N.P., Marincola, F.M., Kawakami, Y., Taubenberger, J., Yannelli, J.R.,
and Rosenberg, S.A. (1996). Loss of functional beta 2-microglobulin in metastatic melanomas from five patients receiving immunotherapy. J. Natl. Cancer
Inst. 88, 100–108.
Ribas, A. (2015). Adaptive Immune Resistance: How Cancer Protects from Immune Attack. Cancer Discov. 5, 915–919.
Ribas, A., Puzanov, I., Dummer, R., Schadendorf, D., Hamid, O., Robert, C.,
Hodi, F.S., Schachter, J., Pavlick, A.C., Lewis, K.D., et al. (2015). Pembrolizumab versus investigator-choice chemotherapy for ipilimumab-refractory melanoma (KEYNOTE-002): a randomised, controlled, phase 2 trial. Lancet Oncol.
16, 908–918.
Ries, C.H., Cannarile, M.A., Hoves, S., Benz, J., Wartha, K., Runza, V., ReyGiraud, F., Pradel, L.P., Feuerhake, F., Klaman, I., et al. (2014). Targeting tumor-associated macrophages with anti-CSF-1R antibody reveals a strategy
for cancer therapy. Cancer Cell 25, 846–859.
Rizvi, N.A., Hellmann, M.D., Snyder, A., Kvistborg, P., Makarov, V., Havel, J.J.,
Lee, W., Yuan, J., Wong, P., Ho, T.S., et al. (2015). Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung
cancer. Science 348, 124–128.
Robbins, P.F., Morgan, R.A., Feldman, S.A., Yang, J.C., Sherry, R.M., Dudley,
M.E., Wunderlich, J.R., Nahvi, A.V., Helman, L.J., Mackall, C.L., et al. (2011).
Tumor regression in patients with metastatic synovial cell sarcoma and melanoma using genetically engineered lymphocytes reactive with NY-ESO-1.
J. Clin. Oncol. 29, 917–924.
Robert, C., Thomas, L., Bondarenko, I., O’Day, S., Weber, J., Garbe, C.,
Lebbe, C., Baurain, J.F., Testori, A., Grob, J.J., et al. (2011). Ipilimumab plus
dacarbazine for previously untreated metastatic melanoma. N. Engl. J. Med.
364, 2517–2526.
Rooney, M.S., Shukla, S.A., Wu, C.J., Getz, G., and Hacohen, N. (2015). Molecular and genetic properties of tumors associated with local immune cytolytic activity. Cell 160, 48–61.
Rosenberg, S.A., Yang, J.C., Sherry, R.M., Kammula, U.S., Hughes, M.S.,
Phan, G.Q., Citrin, D.E., Restifo, N.P., Robbins, P.F., Wunderlich, J.R., et al.
(2011). Durable complete responses in heavily pretreated patients with metastatic melanoma using T-cell transfer immunotherapy. Clin. Cancer Res. 17,
4550–4557.
Roszik, J., Haydu, L.E., Hess, K.R., Oba, J., Joon, A.Y., Siroy, A.E., Karpinets,
T.V., Stingo, F.C., Baladandayuthapani, V., Tetzlaff, M.T., et al. (2016). Novel
algorithmic approach predicts tumor mutation load and correlates with immunotherapy clinical outcomes using a defined gene mutation set. BMC Med.
14, 168.
Rudensky, A.Y. (2011). Regulatory T cells and Foxp3. Immunol. Rev. 241,
260–268.
Ruella, M., Barrett, D.M., Kenderian, S.S., Shestova, O., Hofmann, T.J.,
Perazzelli, J., Klichinsky, M., Aikawa, V., Nazimuddin, F., Kozlowski, M.,
et al. (2016). Dual CD19 and CD123 targeting prevents antigen-loss relapses
after CD19-directed immunotherapies. J. Clin. Invest. 126, 3814–3826.
Ruffell, B., and Coussens, L.M. (2015). Macrophages and therapeutic resistance in cancer. Cancer Cell 27, 462–472.
Schachter, J. R.A., Long GV, Arance A, Grob JJ, Mortier L, Daud A, Carlino MS,
McNeil CM, Lotem M, Larkin J, Lorigan P, Neyns B, Blank CU, Petrella TM, Hamid O, Zhou H, Ebbinghaus S, Ibrahim N, Robert C (2016). Pembrolizumab
versus ipilimumab for advanced melanoma: final overall survival analysis of
KEYNOTE-006. Clin Oncol 34 (suppl; abstr 9504).
Schadendorf, D., Hodi, F.S., Robert, C., Weber, J.S., Margolin, K., Hamid, O.,
Patt, D., Chen, T.-T., Berman, D.M., and Wolchok, J.D. (2015). Pooled analysis
of long-term survival data from phase II and phase III trials of ipilimumab in unresectable or metastatic melanoma. J. Clin. Oncol. 33.
Schumacher, T.N., and Schreiber, R.D. (2015). Neoantigens in cancer immunotherapy. Science 348, 69–74.
Shankaran, V., Ikeda, H., Bruce, A.T., White, J.M., Swanson, P.E., Old, L.J.,
and Schreiber, R.D. (2001). IFNgamma and lymphocytes prevent primary
tumour development and shape tumour immunogenicity. Nature 410,
1107–1111.
Sharma, P., and Allison, J.P. (2015). Immune checkpoint targeting in cancer
therapy: toward combination strategies with curative potential. Cell 161,
205–214.
Shin, D.S., and Ribas, A. (2015). The evolution of checkpoint blockade as a
cancer therapy: what’s here, what’s next? Curr. Opin. Immunol. 33, 23–35.
Shin, D.S., Zaretsky, J.M., Escuin-Ordinas, H., Garcia-Diaz, A., Hu-Lieskovan,
S., Kalbasi, A., Grasso, C.S., Hugo, W., Sandoval, S., Torrejon, D.Y., et al.
(2016). Primary Resistance to PD-1 Blockade Mediated by JAK(1/2) Mutations.
Cancer Discov.
Simpson, T.R., Li, F., Montalvo-Ortiz, W., Sepulveda, M.A., Bergerhoff, K.,
Arce, F., Roddie, C., Henry, J.Y., Yagita, H., Wolchok, J.D., et al. (2013). Fcdependent depletion of tumor-infiltrating regulatory T cells co-defines
the efficacy of anti-CTLA-4 therapy against melanoma. J. Exp. Med. 210,
1695–1710.
Sivan, A., Corrales, L., Hubert, N., Williams, J.B., Aquino-Michaels, K., Earley,
Z.M., Benyamin, F.W., Lei, Y.M., Jabri, B., Alegre, M.L., et al. (2015).
Commensal bifidobacterium promotes antitumor immunity and facilitates
anti-PD-L1 efficacy. Science 350, 1084–1089.
Snyder, A., Makarov, V., Merghoub, T., Yuan, J., Zaretsky, J.M., Desrichard,
A., Walsh, L.A., Postow, M.A., Wong, P., Ho, T.S., et al. (2014). Genetic basis
for clinical response to CTLA-4 blockade in melanoma. N. Engl. J. Med. 371,
2189–2199.
Solito, S., Falisi, E., Diaz-Montero, C.M., Doni, A., Pinton, L., Rosato, A.,
Francescato, S., Basso, G., Zanovello, P., Onicescu, G., et al. (2011). A human
promyelocytic-like population is responsible for the immune suppression
mediated by myeloid-derived suppressor cells. Blood 118, 2254–2265.
Sotillo, E., Barrett, D.M., Black, K.L., Bagashev, A., Oldridge, D., Wu, G., Sussman, R., Lanauze, C., Ruella, M., Gazzara, M.R., et al. (2015). Convergence of
Acquired Mutations and Alternative Splicing of CD19 Enables Resistance to
CART-19 Immunotherapy. Cancer Discov. 5, 1282–1295.
Spranger, S., Koblish, H.K., Horton, B., Scherle, P.A., Newton, R., and Gajewski, T.F. (2014). Mechanism of tumor rejection with doublets of CTLA-4, PD-1/
PD-L1, or IDO blockade involves restored IL-2 production and proliferation of
CD8(+) T cells directly within the tumor microenvironment. J. Immunother.
Cancer 2, 3.
Spranger, S., Bao, R., and Gajewski, T.F. (2015). Melanoma-intrinsic b-catenin
signalling prevents anti-tumour immunity. Nature 523, 231–235.
Ruffell, B., Chang-Strachan, D., Chan, V., Rosenbusch, A., Ho, C.M., Pryer, N.,
Daniel, D., Hwang, E.S., Rugo, H.S., and Coussens, L.M. (2014). Macrophage
IL-10 blocks CD8+ T cell-dependent responses to chemotherapy by suppressing IL-12 expression in intratumoral dendritic cells. Cancer Cell 26,
623–637.
Stagg, J., Divisekera, U., McLaughlin, N., Sharkey, J., Pommey, S., Denoyer,
D., Dwyer, K.M., and Smyth, M.J. (2010). Anti-CD73 antibody therapy
inhibits breast tumor growth and metastasis. Proc. Natl. Acad. Sci. USA
107, 1547–1552.
Sadelain, M. (2016). Chimeric antigen receptors: driving immunology towards
synthetic biology. Curr. Opin. Immunol. 41, 68–76.
Sucker, A., Zhao, F., Real, B., Heeke, C., Bielefeld, N., Mabetaen, S., Horn, S.,
Moll, I., Maltaner, R., Horn, P.A., et al. (2014). Genetic evolution of T-cell
722 Cell 168, February 9, 2017
resistance in the course of melanoma progression. Clin. Cancer Res. 20,
6593–6604.
Sugiyama, D., Nishikawa, H., Maeda, Y., Nishioka, M., Tanemura, A.,
Katayama, I., Ezoe, S., Kanakura, Y., Sato, E., Fukumori, Y., et al. (2013).
Anti-CCR4 mAb selectively depletes effector-type FoxP3+CD4+ regulatory
T cells, evoking antitumor immune responses in humans. Proc. Natl. Acad.
Sci. USA 110, 17945–17950.
Sundstedt, A., O’Neill, E.J., Nicolson, K.S., and Wraith, D.C. (2003). Role for IL10 in suppression mediated by peptide-induced regulatory T cells in vivo.
J. Immunol. 170, 1240–1248.
Takahashi, H., Okai, Y., Paxton, R.J., Hefta, L.J., and Shively, J.E. (1993). Differential regulation of carcinoembryonic antigen and biliary glycoprotein by
gamma-interferon. Cancer Res. 53, 1612–1619.
Talmadge, J.E., and Gabrilovich, D.I. (2013). History of myeloid-derived suppressor cells. Nat. Rev. Cancer 13, 739–752.
Tang, D.Y., Ellis, R.A., and Lovat, P.E. (2016). Prognostic Impact of Autophagy
Biomarkers for Cutaneous Melanoma. Front. Oncol. 6, 236.
Taube, J.M., Anders, R.A., Young, G.D., Xu, H., Sharma, R., McMiller, T.L.,
Chen, S., Klein, A.P., Pardoll, D.M., Topalian, S.L., and Chen, L. (2012). Colocalization of inflammatory response with B7-h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape.
Sci. Transl. Med. 4, 127ra37.
Taube, J.M., Klein, A., Brahmer, J.R., Xu, H., Pan, X., Kim, J.H., Chen, L.,
Pardoll, D.M., Topalian, S.L., and Anders, R.A. (2014). Association of PD-1,
PD-1 ligands, and other features of the tumor immune microenvironment
with response to anti-PD-1 therapy. Clin. Cancer Res. 20, 5064–5074.
Vo, D.D., Prins, R.M., Begley, J.L., Donahue, T.R., Morris, L.F., Bruhn, K.W., de
la Rocha, P., Yang, M.-Y., Mok, S., Garban, H.J., et al. (2009). Enhanced antitumor activity induced by adoptive T-cell transfer and adjunctive use of the histone deacetylase inhibitor LAQ824. Cancer Res. 69, 8693–8699.
Wang, L.X., Mei, Z.Y., Zhou, J.H., Yao, Y.S., Li, Y.H., Xu, Y.H., Li, J.X., Gao,
X.N., Zhou, M.H., Jiang, M.M., et al. (2013). Low dose decitabine treatment induces CD80 expression in cancer cells and stimulates tumor specific cytotoxic
T lymphocyte responses. PLoS ONE 8, e62924.
Wargo, J.A., Reddy, S.M., Reuben, A., and Sharma, P. (2016). Monitoring immune responses in the tumor microenvironment. Curr. Opin. Immunol. 41,
23–31.
Weber, J.S., D’Angelo, S.P., Minor, D., Hodi, F.S., Gutzmer, R., Neyns, B., Hoeller, C., Khushalani, N.I., Miller, W.H., Jr., Lao, C.D., et al. (2015). Nivolumab
versus chemotherapy in patients with advanced melanoma who progressed
after anti-CTLA-4 treatment (CheckMate 037): a randomised, controlled,
open-label, phase 3 trial. Lancet Oncol. 16, 375–384.
Wesolowski, R., Markowitz, J., and Carson, W.E., 3rd. (2013). Myeloid derived
suppressor cells - a new therapeutic target in the treatment of cancer.
J. Immunother. Cancer 1, 10.
Wilmott, J.S., Long, G.V., Howle, J.R., Haydu, L.E., Sharma, R.N., Thompson,
J.F., Kefford, R.F., Hersey, P., and Scolyer, R.A. (2012). Selective BRAF inhibitors induce marked T-cell infiltration into human metastatic melanoma. Clin.
Cancer Res. 18, 1386–1394.
Wolchok, J.D., Kluger, H., Callahan, M.K., Postow, M.A., Rizvi, N.A., Lesokhin,
A.M., Segal, N.H., Ariyan, C.E., Gordon, R.-A., Reed, K., et al. (2013). Nivolumab plus ipilimumab in advanced melanoma. N. Engl. J. Med. 369, 122–133.
Tham, M., Khoo, K., Yeo, K.P., Kato, M., Prevost-Blondel, A., Angeli, V., and
Abastado, J.P. (2015). Macrophage depletion reduces postsurgical tumor
recurrence and metastatic growth in a spontaneous murine model of melanoma. Oncotarget 6, 22857–22868.
Woo, E.Y., Yeh, H., Chu, C.S., Schlienger, K., Carroll, R.G., Riley, J.L., Kaiser,
L.R., and June, C.H. (2002). Cutting edge: Regulatory T cells from lung cancer
patients directly inhibit autologous T cell proliferation. J. Immunol. 168,
4272–4276.
Topalian, S.L., Drake, C.G., and Pardoll, D.M. (2015). Immune checkpoint
blockade: a common denominator approach to cancer therapy. Cancer Cell
27, 450–461.
Wu, X., Schulte, B.C., Zhou, Y., Haribhai, D., Mackinnon, A.C., Plaza, J.A.,
Williams, C.B., and Hwang, S.T. (2014). Depletion of M2-like tumor-associated
macrophages delays cutaneous T-cell lymphoma development in vivo.
J. Invest. Dermatol. 134, 2814–2822.
Tran, E., Robbins, P.F., Lu, Y.C., Prickett, T.D., Gartner, J.J., Jia, L., Pasetto,
A., Zheng, Z., Ray, S., Groh, E.M., et al. (2016). T-Cell Transfer Therapy Targeting Mutant KRAS in Cancer. N. Engl. J. Med. 375, 2255–2262.
Tumeh, P.C., Harview, C.L., Yearley, J.H., Shintaku, I.P., Taylor, E.J., Robert,
L., Chmielowski, B., Spasic, M., Henry, G., Ciobanu, V., et al. (2014). PD-1
blockade induces responses by inhibiting adaptive immune resistance. Nature
515, 568–571.
Turcotte, M., Spring, K., Pommey, S., Chouinard, G., Cousineau, I., George, J.,
Chen, G.M., Gendoo, D.M., Haibe-Kains, B., Karn, T., et al. (2015). CD73 is
associated with poor prognosis in high-grade serous ovarian cancer. Cancer
Res. 75, 4494–4503.
Van Allen, E.M., Miao, D., Schilling, B., Shukla, S.A., Blank, C., Zimmer, L.,
Sucker, A., Hillen, U., Geukes Foppen, M.H., Goldinger, S.M., et al. (2015).
Genomic correlates of response to CTLA-4 blockade in metastatic melanoma.
Science 350, 207–211.
van Rooij, N., van Buuren, M.M., Philips, D., Velds, A., Toebes, M., Heemskerk,
B., van Dijk, L.J., Behjati, S., Hilkmann, H., El Atmioui, D., et al. (2013). Tumor
exome analysis reveals neoantigen-specific T-cell reactivity in an ipilimumabresponsive melanoma. J. Clin. Oncol. 31, e439–e442.
Vanpouille-Box, C., Diamond, J.M., Pilones, K.A., Zavadil, J., Babb, J.S., Formenti,
S.C., Barcellos-Hoff, M.H., and Demaria, S. (2015). TGFb is a master regulator of radiation therapy-induced antitumor immunity. Cancer Res. 75, 2232–2242.
Vétizou, M., Pitt, J.M., Daillère, R., Lepage, P., Waldschmitt, N., Flament, C.,
Rusakiewicz, S., Routy, B., Roberti, M.P., Duong, C.P., et al. (2015). Anticancer
immunotherapy by CTLA-4 blockade relies on the gut microbiota. Science
350, 1079–1084.
Viehl, C.T., Moore, T.T., Liyanage, U.K., Frey, D.M., Ehlers, J.P., Eberlein, T.J.,
Goedegebuure, P.S., and Linehan, D.C. (2006). Depletion of CD4+CD25+
regulatory T cells promotes a tumor-specific immune response in pancreas
cancer-bearing mice. Ann. Surg. Oncol. 13, 1252–1258.
Yang, J.C., and Rosenberg, S.A. (2016). Chapter Seven—Adoptive T-Cell
Therapy for Cancer. In Advances in Immunology, D.S. Robert, ed. (Academic
Press), pp. 279–294.
Yang, L., DeBusk, L.M., Fukuda, K., Fingleton, B., Green-Jarvis, B., Shyr, Y.,
Matrisian, L.M., Carbone, D.P., and Lin, P.C. (2004). Expansion of myeloid immune suppressor Gr+CD11b+ cells in tumor-bearing host directly promotes
tumor angiogenesis. Cancer Cell 6, 409–421.
Yang, L., Huang, J., Ren, X., Gorska, A.E., Chytil, A., Aakre, M., Carbone, D.P.,
Matrisian, L.M., Richmond, A., Lin, P.C., and Moses, H.L. (2008). Abrogation of
TGF beta signaling in mammary carcinomas recruits Gr-1+CD11b+ myeloid
cells that promote metastasis. Cancer Cell 13, 23–35.
Yee, C., Lizee, G., and Schueneman, A.J. (2015). Endogenous T-Cell Therapy:
Clinical Experience. Cancer J. 21, 492–500.
Zaretsky, J.M., Garcia-Diaz, A., Shin, D.S., Escuin-Ordinas, H., Hugo, W., HuLieskovan, S., Torrejon, D.Y., Abril-Rodriguez, G., Sandoval, S., Barthly, L.,
et al. (2016). Mutations Associated with Acquired Resistance to PD-1
Blockade in Melanoma. N. Engl. J. Med. 375, 819–829.
Zhang, H., Conrad, D.M., Butler, J.J., Zhao, C., Blay, J., and Hoskin, D.W.
(2004). Adenosine acts through A2 receptors to inhibit IL-2-induced tyrosine
phosphorylation of STAT5 in T lymphocytes: role of cyclic adenosine 30 ,50 monophosphate and phosphatases. J. Immunol. 173, 932–944.
Zhu, Y., Knolhoff, B.L., Meyer, M.A., Nywening, T.M., West, B.L., Luo, J.,
Wang-Gillam, A., Goedegebuure, S.P., Linehan, D.C., and DeNardo, D.G.
(2014). CSF1/CSF1R blockade reprograms tumor-infiltrating macrophages
and improves response to T-cell checkpoint immunotherapy in pancreatic
cancer models. Cancer Res. 74, 5057–5069.
Zou, W., Wolchok, J.D., and Chen, L. (2016). PD-L1 (B7-H1) and PD-1 pathway
blockade for cancer therapy: Mechanisms, response biomarkers, and combinations. Sci. Transl. Med. 8.
Cell 168, February 9, 2017 723
Leading Edge
Review
The Principles of Engineering Immune Cells
to Treat Cancer
Wendell A. Lim1,* and Carl H. June2,*
1Howard Hughes Medical Institute, Department of Cellular and Molecular Pharmacology, UCSF Center for Systems and Synthetic Biology,
University of California, San Francisco, San Francisco, CA 94158, USA
2Center for Cellular Immunotherapies, the Department of Pathology and Laboratory Medicine at the Perelman School of Medicine,
and the Parker Institute for Cancer Immunotherapy, University of Pennsylvania, Philadelphia, PA 19104, USA
*Correspondence: wendell.lim@ucsf.edu (W.A.L.), cjune@exchange.upenn.edu (C.H.J.)
http://dx.doi.org/10.1016/j.cell.2017.01.016
Chimeric antigen receptor (CAR) T cells have proven that engineered immune cells can serve as a
powerful new class of cancer therapeutics. Clinical experience has helped to define the major challenges that must be met to make engineered T cells a reliable, safe, and effective platform that can
be deployed against a broad range of tumors. The emergence of synthetic biology approaches for
cellular engineering is providing us with a broadly expanded set of tools for programming immune
cells. We discuss how these tools could be used to design the next generation of smart T cell precision therapeutics.
Introduction
The emergence of engineered T cells as a form of cancer therapy
marks the beginning of a new era in medicine, providing a transformative way to combat complex diseases such as cancer.
Within the past few years, clinical trials using T cells engineered
to recognize B cell cancers have shown high rates of response
(70%–90%) and durability of response that are unprecedented
in acute (Brentjens et al., 2013; Maude et al., 2014; Turtle
et al., 2016) and chronic leukemia (Kalos et al., 2011). In 2017,
we expect to see the first approved engineered T cell therapies
coming to market. While poised to revolutionize cancer therapy,
the optimism about T cell cancer therapies remains tempered by
concerns about safety and off-target toxicity, as well as the
development of resistance. Meanwhile, the field also awaits a
clear demonstration of clinical efficacy in solid tumors. The developments in this field over the coming years—in the areas of
safety, reliability, and efficacy against solid tumors—will ultimately determine how disruptive this new modality can be in
the broader battle against cancer.
Living Therapies Can Uniquely Perform Complex
Sensing and Response Functions
Engineered T cells are part of a much broader explosion in
immuno-oncology, but what perhaps makes the