вход по аккаунту


Anti-Aquaporin-4 Antibodies in Neuromyelitis Optica: How to Prove

код для вставки
Anti-Aquaporin-4 Antibodies in
Neuromyelitis Optica: How to Prove
their Pathogenetic Relevance?
The discovery of a specific autoantibody response in
neuromyelitis optica (NMO) patients, which selectively targets
astrocytic end feet at the glia limitans1 and which is directed
against the water channel aquaporin-4 (AQP-4) was a
milestone in defining this disease entity and profoundly
changed our view regarding its pathogenesis.3 Indirect
evidence, coming from clinical observations and pathology
strongly suggest that these autoantibodies play a major role
in driving the disease process. The pathological hallmark of
NMO lesions is a very selective and characteristic deposition
of immunoglobulins and complement on astrocytes at the
glia limitans, which is associated with destruction and loss of
glial fibrillary acidic protein and AQP-4 positive astrocytes in
fresh lesions4–6 followed by demyelination and global tissue
destruction. As in other antibody-mediated diseases,
granulocytes, and in particular eosinophils, are a major
component of the inflammatory infiltrates.4 Furthermore, the
distribution of the lesions in the brain and spinal cord of
NMO patients correlates with the extent of regional AQP-4
expression.6,7 Most importantly, therapies, which eliminate
antibodies (plasma exchange),8 or which target Blymphocytes (anti-CD20 antibodies),9 are at least in part
effective in NMO patients. Based on this evidence NMO is
now considered an antibody-mediated autoimmune disease,
however, direct proof of the pathogenic potential of AQP-4
antibodies is so far lacking.
To firmly establish the pathogenic role of
autoantibodies, it has to be shown in vivo that
pathological changes, which are identical or at least
similar to NMO, can be induced by the transfer of
such antibodies in an experimental model. To achieve
this, several essential requirements have to be fulfilled.
1) The expression of the target epitope of the
autoantibodies has to be similar between humans
and the animal species, in which transfer is
performed. So far, the molecular configuration of
the epitope, which is recognized by NMO
associated AQP-4 autoantibodies is unknown.10
Although potential binding sites have been
identified through the knowledge obtained from
the crystal structure of AQP-4, it is not clear yet
The International MS Journal 2008; 15: 75–78
whether these antibodies recognize a linear or a
conformational epitope. Thus the simple approach,
comparing the amino acid sequence of potential
binding sites between humans and rodents, is not
possible. However, it has been shown that some,
but not all NMO sera bind to rodent astrocytes in
tissue sections in a similar extent as they do on
cell lines, transfected with human AQP-4.1,11 This
indicates that disease transfer into rodents (rats
or mice) may be feasible with NMO-IgG fractions
from some, but not all patients.
2) The target epitope has to be accessible in vivo in
the normal brain and spinal cord for antibodies,
present in the extracellular space. This means that
the epitope has to be expressed on the extracellular
surface of cells or on molecules present in the
extracellular matrix. This requirement seems to be
fulfilled for NMO/AQP-4 antibodies. Sera from
NMO patients bind to the surface of cells,
transfected with human AQP-411 and human NMO
antibodies may cross link and cause internalization
of AQP-4 after binding to cells, which express the
target antigen.12 In addition, we have recently seen
by immune electron microscopy that some, but not
all human NMO sera selectively bind to the surface
of astrocytes in normal rat tissue.13
3) Antibodies have to reach the central nervous
system (CNS) in concentrations, which are
sufficient to induce disease or lesions. The brain
tissue is shielded from circulating proteins by the
blood–brain barrier. The equilibrium for IgG
molecules between the cerebrospinal fluid (CSF),
which reflects the brain extracellular space, and
the blood is in the range of 1:500. Thus, antibody
titres within the extracellular compartments of the
nervous system, which are pathogenetically
relevant are only reached when either the titres in
the blood are enormously high, or in areas which
either lack a blood–brain barrier or where blood–
brain barrier function is disturbed by concomitant
pathology. As an example, circulating antibodies
against myelin oligodendrocyte glycoprotein
(MOG), which are highly efficient to augment
antibodies, the transfer of disease and lesions by
AQP-4 antibodies or human NMO serum alone,
demyelination in lesions of autoimmune
encephalomyelitis (EAE), do not induce disease or
the transfer of NMO serum into animals with
T-cell–mediated brain inflammation and the disease
brain lesions in the absence of T-cell–mediated
inflammation.14 This situation may be different for
AQP-4 antibodies. Aquaporin-4 is highly expressed
induction by active sensitization of susceptible
animals with AQP-4.
in areas which either lack an endothelial
blood–brain barrier (circumventricular organs)7 or
where blood–brain barrier permeability is higher
compared to other regions of the CNS, such as
the optic nerve head.15 These areas are thus
potential target sites for pathology after
intravenous transfer of antibodies alone.
4) Activated effector mechanisms have to be present
to destroy target cells, opsonized by
autoantibodies. When antibodies against MOG
or sera from EAE animals were injected directly
into the CSF even in high concentration, pathology
in the spinal cord was restricted to demyelination
of single fibres at the root entry zones. However,
when the antibodies were injected either with
complement16 or with gamma-interferon,17 massive
demyelination was present within the spinal cord.
This implies that either complement activation or
the activation of effector cells, such as
macrophages and microglia, are required to
promote antibody mediate tissue injury in the
CNS. It is likely, but not proven so far, that the
same rules apply for antibodies against AQP-4.
A very efficient way to induce blood–brain
barrier damage and simultaneously provide activated
effector mechanisms, such as complement and
activated macrophages in lesions of the CNS, is to
induce brain inflammation by a T-cell–mediated
autoimmune reaction. As shown in different EAE
models circulating autoantibodies against MOG are
very potent to induce antibody-mediated
demyelination, when they enter the CNS tissue at
sites of T-cell–mediated inflammation.14
Consequences for the Design of Future
Experimental Strategies to Unravel the
Pathogenic Role of AQP-4 Antibodies
Based on the theoretical requirements discussed
above three different experimental strategies can be
followed to establish the pathogenic role of NMO
1) Antibody transfer alone. The basic requirement for
an antibody transfer to be successful is that the
pathogenic epitope is recognized in the species in
which the transfer is performed and that its
expression is similar to that found in humans. In
addition, large amounts of antibody will be
necessary to reach serum titres, which are
comparable with those seen in humans. Thus, IgG
fractions of human serum or, even better, affinity
purified specific NMO/AQP-4 antibodies have to
be prepared. In addition, the binding properties of
purified antibodies to rodent tissue in vitro and in
vivo have to be tested prior to transfer. A single
antibody transfer may be insufficient and therefore
repeated IgG injections may become necessary.
This will be counteracted by the induction of an
antibody response directed against human IgG in
the host, which may neutralize the injected
antibodies and may induce serum sickness.
Despite these potential difficulties and problems,
such an experiment may be successful. Particular
attention should be paid to lesions in
circumventricular organs and the optic nerve head,
since in both regions the blood–brain barrier is
less tight than in other regions of the CNS, and
AQP-4 is highly expressed at these sites.
Antibodies may either functionally disturb water
homeostasis in astrocytes or may even induce
complement-mediated astrocytic damage. Whether
endogenous complement in rats and mice is
sufficient to do so is uncertain.
2) Transfer of NMO/AQP-4 IgG into animals with
acute T-cell–mediated EAE. This strategy in the
short term is the most likely to succeed. Its basic
principles have been well established in previous
experimental investigations of the pathogenic role
of antibodies directed against brain and spinal
cord antigen.14 A transient and monophasic
inflammatory disease of the CNS is induced by the
passive transfer of T-lymphocyte lines against a
CNS antigen. At the onset of the inflammatory
The International MS Journal 2008; 15: 75–78
disease, which occurs between Days 3 and 4 after
T-cell transfer in rats, the respective IgG
preparations are systemically injected. When T-cell
lines are not available, inflammation can also be
induced by active sensitization with an
encephalitogenic peptide; however, the more
pronounced variability of EAE in an active model
may mask additional effects, induced by
antibodies, in particular when they are subtle.
One could also expect disease augmentation by
the antibodies – if they are pathogenic – resulting
in increased clinical scores in comparison with
animals injected with a control human IgG
fraction. Detailed pathological analysis can be
done to determine similarities and differences of
the lesions between the experimental model and
NMO in humans. When such a model is
established many different neurophysiological,
immunopathological, immunological and
therapeutic studies can be performed in a very
standardized manner.
A key question for this experimental design is the
choice of the target antigen for T-cell–mediated
inflammation. Many different brain proteins,
including myelin proteins as well as astrocytic or
neuronal proteins, can induce a T-cell response,
which gives rise to inflammation in the brain or
spinal cord. The topographical distribution of
inflammatory lesions, however, is determined by the
degree of expression of the respective proteins in
the CNS.18 For a study, as described above, the
ideal T-cell–mediated inflammation should be
directed against AQP-4 itself. So far, however, it is
unknown, whether AQP-4 is able to mount a T-cell
response at all and, if yes, whether this T-cell
response is able to induce CNS inflammation.10
However, studies with MOG antibodies clearly show
that the antigen specificity of the T-cell response does
not necessarily have to be directed against the same
target antigen as the antibody response. For
pragmatic reasons it is thus sufficient to select T-cells,
which either target the spinal cord (such as myelin
basic protein reactive T-cells) or the spinal cord and
optic nerves (such as MOG reactive T-cells). In both
situations it should be possible to see, whether
NMO/AQP-4 antibodies mediate additional specific
tissue injury.
The International MS Journal 2008; 15: 75–78
Key Points
• Clinical and pathological features of
neuromyelitis optica are consistent with a
pathogenic role of antibodies against AQP-4
• The chance to induce NMO lesions with AQP-4
antibodies alone is low, since the intact
blood–brain barrier does not allow them to
reach the brain in sufficient concentrations
• Circulating AQP-4 antibodies are likely to induce
NMO lesions, when inflammatory lesions in the
brain are induced by other means, such as for
instance by encephalitogenic T-cells
• Whether AQP-4 itself is able to induce an
encephalitogenic T-cell response is currently
3) Induction of NMO in animals by active
sensitization. In principle, AQP-4 is immunogenic
in experimental animals, as documented by the
availability of AQP-4 antibodies raised in rabbits
and mice.19 It can thus be expected that active
sensitization of specific strains of rodents will
induce both a T-cell, as well as an antibody
response against AQP-4. What animal strains,
however, will be suitable is at present difficult to
predict. The induction of a T-cell response depends
upon the recognition of the antigenic peptide,
embedded within the major histocompatability
complex (MHC) groove on antigen presenting
cells. It is thus critical that the protein contains
amino acid stretches, which can bind to the
respective strain-dependent MHC molecules.
Good tools in bioinformatics can be used to, in
part, predict whether a certain MHC haplotype
will be able to induce a specific T-cell response
to a given protein ( or Similarly, straindependent epitope selection also relates to the
induction of autoantibodies. As an example,
Bourquin et al21 showed that in the course of
active sensitization with MOG, C57BL/6 mice
mount an antibody response which is entirely
directed against non-pathogenic linear MOG
epitopes, whereas SJL mice predominantly produce
autoantibodies against a conformational epitope
of MOG, accessible on the surface of MOG
transfected cells or oligodendrocytes, which can
induce demyelination in vivo. It is thus
unpredictable which strains of rodents will produce
a pathogenic autoantibody response against
AQP-4. For these reasons, the induction of NMO
in experimental animals by active sensitization will
either occur by chance (finding with a large
portion of luck the right animal species and strain)
or only after very extensive and labourious
immunological studies.
The immunopathology of NMO lesions supports the
concept that autoantibodies against AQP-4 are
pathogenic. Experimental tools are available to test
this hypothesis and it is expected that they will
provide unequivocal evidence in the near future.
NMO may, however, turn out not to be a pure
antibody-mediated disease. Brain inflammation,
induced by other mechanisms, may be instrumental to
allow AQP-4 antibodies to reach their target antigen
in the nervous system and to induce tissue injury.
M Bradl and H Lassmann
Vienna, Austria
Address for correspondence:
Professor Dr Hans Lassmann
Center for Brain Research
Medical University of Vienna
Spitalgasse 4, A-1090, Vienna
Phone: +431 4277 62811
Fax: +431 4277 9628
1. Lennon VA, Wingerchuk DN,
Kryzer TJ, et al. A serum
autoantibody marker of
neuromyelitis optica: distinction
from multiple sclerosis. Lancet
2004; 364: 2106–2112.
2. Lennon VA, Kryzer TJ, Pittock SJ,
et al. IgG marker of optic-spinal
multiple sclerosis binds to the
aquaporin-4 water channel.
J Exp Med 2005; 202:
3. Wingerchuck DM, Lennon VA,
Pittock SJ, et al. Revised
diagnostic criteria for
neuromyelitis optica. Neurology
2006; 66: 1485–1489.
4. Lucchinetti CF, Mandler R,
McGavern D, et al. A role for
humoral mechanisms in the
pathogenesis of Devic’s
neuromyelitis optica. Brain
2002; 125: 1450–1461.
5. Misu T, Fujihara K, Kakita A,
et al. Loss of aquaporin-4 in
lesions of neuromyelitis optica:
distinction from multiple sclerosis.
Brain 2007; 130: 1224–1234.
6. Roemer SF, Parisi JE, Lennon VA,
et al. Distinct pattern of
aquaporin-4 expression in
neuromyelitis optica lesions.
Brain 2007; 130: 1194–1205.
7. Pittock SJ, Lennon VA, Krecke K,
et al. Brain abnormalities in
neuromyelitis optica. Arch
Neurol 2006; 63: 390–396.
8. Watanabe S, Nakashima I,
Misu T, et al. Therapeutic
efficacy of plasma exchange in
NMO-positive patients with
neuromyelitis optica. Mult Scler
2007; 13: 128–132.
9. Cree BA, Lamb S, Morgan K,
et al. An open label study of the
effects of rituximab in
neuromyelitis optica. Neurology
2005; 64: 1270–1272.
10.Graber DJ, Levy M, Kerr D,
et al. Neuromyelitis optica
pathogenesis and aquaporin 4.
J Neuroinflammation 2008; 5:
11.Takahashi T, Fujihara K,
Nakashima I, et al.
Establishment of a new sensitive
assay for anti-human aquaporin4 antibody in neuromyelitis
optica. Tohoku J Exp Med
2006; 210: 307–313.
12.Hinson SR, Pittock SJ,
Lucchinetti CF, et al. Pathogenic
potential of IgG binding to
water channel extracellular
domain in neuromyelitis optica.
Neurology 2007; 68:
13.Lassmann H. Personal
14.Linington C, Bradl M,
Lassmann H, et al. Augmentation
of demyelination in rat acute
allergic encephalomyelitis by
circulating mouse monoclonal
antibodies directed against a
myelin/ oligodendrocyte
glycoprotein. Am J Pathol 1988;
130: 443–454.
15.Tso MO, Shih CY, McLean IW.
Is there a blood brain barrier at
the optic nerve head? Arch
Ophthalmol 1975; 93:
16.Lassmann H, Stemberger H,
Kitz K, et al. In vivo
demyelinating activity of sera
from animals with chronic
experimental allergic
encephalomyelitis: Antibody
nature of the demyelinating
factor and the role of
complement. J Neurol Sci 1983;
59: 123–137.
17.Vass K, Heininger K, Schäfer B,
et al. Interferon-gamma
potentiates antibody mediated
demyelination in vivo. Ann
Neurol 1992; 32: 189–206.
18.Berger T, Weerth S, Kojima K,
et al. Experimental autoimmune
encephalomyelitis: The antigen
specificity of T-lymphocytes
determines the topography of
lesions in the central and
peripheral nervous system. Lab
Invest 1997; 76: 355–364.
19.Nagy G, Szekeres G, Kvell K,
et al. Development and
characterization of a
monoclonal antibody family
against aquaporin 1 and
aquaporin 4. Pathol Oncol Res
2002; 8: 115–124. or
21.Bourquin C, Schubart A,
Tobollik S, et al. Selective
unresponsiveness to
conformational B cell epitopes
of the myelin oligodendrocyte
glycoprotein in H-2b mice.
J Immunol 2003; 171:
The International MS Journal 2008; 15: 75–78
Без категории
Размер файла
90 Кб
Пожаловаться на содержимое документа