вход по аккаунту


Criteria for the tracer kinetic measurement of cerebral protein synthesis in humans with positron emission tomography.

код для вставкиСкачать
Criteria for the Tracer Kmetic Measurement
of Cerebral Protein Synthesis in Humans
with Positron Emission Tomography
M. E. Phelps, PhD,"I J. R. Barrio, PhD,*j: S-C. Huang, DSc,"I R. E. Keen, BA,"? H. Chugmi, MD,?
and J. C. Mazziotta, MD, P h D W
The principles and initial results of the use of PET to measure the local cerebral metabolic rate for protein synthesis
(ICRPS) in humans are described. The labeling of leucine, phenylalanine, and methionine in the carboxyl position
provides a strategy (selective position labeling) for discriminating between the incorporation of these amino acids into
proteins and metabolic oxidation. In metabolic oxidation the label is removed from tissue through decarboxylation.
The resulting labeled carbon dioxide is diluted by the tissue carbon dioxide pool, cleared from cerebral tissue by blood
flow, and subsequently ventilated by the lungs. This approach also provides a plasma input function that is free of other
labeled amino acids produced through systemic reactions, such as those that occur for methionine labeled in the methyl
group. The measured ICRPS is in good agreement with values determined by Smith and Sokoloff by autoradiographic
and biochemical assay techniques, as are the measured kinetic rate constants of bidirectional transport, incorporation
into proteins, and metabolism, as determined in monkeys and humans using L-leucine labeled with carbon-I 1 in
position 1 (L-{l-"C}leucine) with PET. The tissue leucine precursor pool exhibits a rapid turnover rate (1.5 to 2
minutes), while the metabolic pathway has a half-time (about 18 minutes) that is close to the radioactive half-life of
carbon-I 1. The dietary state was found to affect the branching ratio of ICRPS/metabolism, with a fasted value of 0.4 and
carbohydrate feed values ranging up to 1.7. The principle of the method appears sound, and a first-order model
provides good fits to data, but much more work is required to determine and validate the model structure and to
optimize the study conditions and estimation criteria.
Phelps ME, Barrio JR, Huang S-C, Keen RE,Chugani H, Mazziotta JC: Criteria for the tracer kinetic
measurement of cerebral protein synthesis in humans with positron emission tomography.
Ann Neurol 15(suppl):S192-S202, 1984
The development of positron emission tomography
(PET) 1251 has provided a method for the application
of tracer kinetic techniques to the measurement of local hemodynamics and biochemical processes in humans. Within this context a number of tracer methods
for PET have been developed for the assessment of
cerebral blood flow, cerebral glucose and oxygen metabolism, and cerebral blood volume, as well as the
assessment of facilitated and passive blood-brain barrier and cell membrane transport. Recently, initial results from a variety of new techniques have also been
reported, such as ligand assays for receptor binding,
tissue pH, and protein synthesis. These approaches and
applications have been reviewed elsewhere { 16, 291.
The development of different types of biochemical
assays with this new technology results from the
numerous variables we would like to be able to measure to advance our knowledge of normal and abnormal cerebral functions. Whereas cerebral blood flow,
along with glucose and oxygen metabolism, forms the
basis of the supply-demand balance of cerebral function, there are other exogenous substrates and complex
sets of endogenous biochemical reactions that are involved in both normal cerebral function and the alterations that occur in cerebral disorders. Thus, there is a
need to develop an array of biochemical probes to provide more comprehensive investigative ability for
studying the human brain. Our expectations will, however, be limited by practical considerations that dictate
which measurements are feasible. The choice of measurements will also depend on the mechanisms that are
most important in a particular condition of cerebral
function or dysfunction.
This paper reports the criteria and initial investigations of a technique for estimating local cerebral rates
for protein synthesis (ICRPS) using carbon- 1I-labeled
amino acids and PET [24]. The specific amino acids
being investigated in this work are L-leucine, L-
From the *Division of Biophysics, Department of Radiological Sciences, and the ?Department of Neurology, Reed Neuroiogic Research Institute, UCLA School of Medicine, and the $Laboratory of
Nuclear Medicine, Los Angeles, CA 00024.
Address reprint requests to Dr Phelps, Division of Biophysics, Department of Radiological Sciences, UCLA School of Medicine, Los
Angeles, CA 90024.
phenylalanine, and L-methionine. The purpose of this
study is to develop a method for assessing potential
alterations in the lCRPS associated with such processes
as cerebral development, regeneration and repair in
response to injury or disease, plasticity, changes induced by drugs and hormones, degenerative disorders,
neoplastic lesions, and possibly learning and memory.
It should be pointed out that the term protein synthesis is a general term that has become widely used
and accepted. The specific process being measured,
however, is the incorporation rate of a specific amino
acid into proteins.
Tracer Kinetic Method
The following are desired properties for a tracer kinetic
model for measuring the ICRPS:
1. High blood-brain barrier permeability of the
labeled amino acid.
2. Rapid clearance of the labeled amino acid from
plasma to a low steady-state value, so that it does not
pose substantial delays in reaching near-steady-state
conditions. The only labeled compound in the
plasma during the study time should be the amino
acid under study.
3. Rapid turnover rate of tissue amino acid precursor
pool. For example, the precursor pool turnover
time should be much less than the radioactive halflife and should be short enough so the incorporation of amino acids into proteins can be considered
to be in nearly a steady state within a reasonable
study time. In addition, the precursor pool turnover
rate should be rapid, since it usually determines the
temporal response time of the tracer kinetic
4. Supply of the tissue precursor pool from plasma
only, since only the exogenous utilization rate can
be measured with PET. The existence of substantial
alternative sources will produce errors in the
method. Thus, since amino acids are highly compartmentalized in the brain, the large acid-soluble
fraction and the breakdown of proteins should supply only minor fractions of the amino acid to the
precursor pool of protein synthesis within the time
frame of the experiment.
5. An approach that can accurately account for or
eliminate the metabolic pathway for the labeled
amino acid in cerebral tissue.
6. Retention of the radioactive label in a large and/or
slow turnover pool(s) of proteins to provide an environment where the end product of the desired reaction sequence is trapped. This criterion allows the
use of an operational equation with favorable characteristics similar to those of the deoxygiucose
model [18, 26, 31, 331.
7 . Sufficient activity must be transported and retained
Table 1. Brain Uptake Index of L-An~inoAcid. in the Rat"
Concentration (nrnol)
Amino Acid
Hydrogen-3labeled water
0.02 1
Brain Uptake
? 2
40 ? 2
38 t 2
22 ? 2
20 ? 1
16 t 3.2
11.7 .t 0.4
8.1 ? 1.6
7.5 5 0.5
4.7 t 0.5
3.21 ? 0.26
2.77 ? 0.79
2.53 t 0.21
From Oldendod r221.
"Injections were with the specific iabeied amino acid in Ringer's
irrigation buffered to a pH of 7.56 without any other amino acids.
Errors are standard deviations.
in the tissue to provide acceptable statistical accuracy for image reconstruction and for parameter estimation with the tracer kinetic model.
8. A chemical synthesis technique for producing
sufficient quantities (usually 10 to 30 mCi) of the
labeled L-amino acid.
These properties of selected amino acids will be examined in the sections that follow. However, it must be
remembered that tracer kinetic models are best structured from direct biochemical assay studies. Once a
model is properly structured and validated, the tracer
kinetic studies can provide accurate local in vivo estimates of the reaction rates, measurements of the kinetic rate constants, assessments of conditions (i.e., abnormal ,or extreme states) under which the model
assumptions may break down, and ways to correct for
Blood-Brain Barrier Transport
Any tracer kinetic method for in vivo studies must first
address the issue of transport through the blood-brain
barrier. Table 1 shows an estimate of the unidirectional
transport of various essential and nonessential amino
acids as determined in the rat by Oldendorf and Szabo
[22, 231. The L-isomers of the essential amino acids
exhibit high extraction rates, consistent with the
facilitated transport provided by their stereospecific
carrier system. It should be noted that the values
Phelps et al: Measurement of Cerebral Protein Synthesis in Humans
shown in Table 1 are estimates of maximum unidirectional transport, since they are from a saline solution
without any competitively transported amino acids
present [22, 231. Moreover, these values are indicative
of unidirectional transport, not of the net tissue extraction rate in steady-state conditions. For example, under
normal conditions the majority of the plasma amino
acids are transported into and out of tissue (unidirectional transport is typically 10 to 30%, while net extraction is usually only 1 to 35%); this level represents
the supply and transport reserve that exists for almost
all exogenous substrates used by the brain. The obvious exception to this relationship is oxygen, for which
the unidirectional and net extraction rates are equal
under normal conditions 130). This exception probably
reflects to some degree the rate-controlling position of
oxygen in the oxidative metabolism of the brain; it also
reflects that the reserve is provided by oxyhemoglobin
in the blood.
The amino acids selected in this work are L-leucine,
L-phenylalanine, and L-methionine, all of which have
high potential extraction rates through the blood-brain
barrier (see Table 1). With this requirement satisfied,
we now examine the biochemical sequence of reactions
in the tissue.
Selective Labeling Approach
After the amino acid has been transported through the
blood-brain barrier, the highly specific cytosolic
aminoacyl transfer RNA (tRNA) synthetases catalyze
the esterification of the L-amino acids to their corresponding tRNAs to form the aminoacyl tRNA. This
energy-requiring process of activating the amino acid is
necessary for the formation of the polypeptide chain in
the ribosome in a process by which the aminoacyl
tRNA is recognized by the messenger RNA coding
system. It is obvious that the specificity of the
aminoacyl tRNA synthetases contributes to the
specificity of the biosynthetic process, and the chance
of error under isolated cellular conditions is less than 1
in 10,000, although it appears to increase with age
In addition to their incorporation into proteins,
these amino acids are substrates for various metabolic
pathways. These alternative reactions can lower or
even eliminate the specificity of the tracer technique
for estimating the ICRPS, since the technique is capable of separating chemical reaction products only by
differences in kinetic rates through known pathways.
This problem is dealt with by using the L-isomer
labeled in the carboxyl position (or position I), as originally proposed by Smith, Sokoloff and co-workers 132,
341. This strategy of selective position labeling is illustrated in Figure 1 for L-leucine labeled with carbon-1 1
in position 1 (L-[ l-"C)leucine). In the metabolic pathway L-leucine is reversibly transaminated with aS194 Annals of Neurology
~ L U
co2 *- .---------------
Fig 1 . Simplzjied version of the pathways of leucine. Alternative
pathways are incorporution i n proteins and metabolism through
the transamination and decarboxylation steps. The figure illustrates how carbon-11 is eliminated in the metabolic pathway at
the decarboxylation step as labeled carbon dioxide (COZ) that d;ffuses out of the tissue and is removed by cerebral bloodflow. Leu
= Leucine; a-KG = alpha-ketoglutaricacid: GLU= glutamic
acid: NAD+ = nicotinamide adenine dinucleotide: NADH =
reduced form of NAD'; CoA = coenzyme A; BCKAD =
branched-chain ketoacid dehydrogenase.
ketoglutarate to form a-ketoisocaproate, which is then
oxidatively decarboxylated in mitochondria C9, 3 71. At
this step the labeled carbon is removed as carbon dioxide (C02) that is diluted in the large CO2-bicarbonate
pool in the brain, which is constantly generated,
primarily from carbohydrate metabolism. Labeled COJ
is then removed from tissue by diffusion and blood
flow, with only a minor fraction retained by C 0 2
fixation. The labeled C 0 2 is also removed from the
plasma by ventilation in the lungs. This action essentially produces a shunt to remove the label from the
metabolic pathway, both in the brain and throughout
the body. Since the rate of this process is faster in the
brain than the turnover of the labeled amino acid in the
protein pool, the ratio of label in proteins to that in the
free amino acid or in the products of the metabolic
pathway increases with time (see Fig 7).
This strategy of labeling in the carboxyl position is
also effective for phenylalanine and merhionine, as
shown in Figures 2 and 3. While the brain is not known
to contain phenylalanine hydroxylase, it does contain
tyrosine hydroxylase, for which phenylalanine is a substrate, although with a lower affinity than tyrosine.
Methionine has also been labeled in the methyl
group with carbon-11 by Comar and associates 1101,
and its kinetics have been investigated with PET by
Bustany and co-workers [71. With the label in the
methyl position the end products of both incorporation
into proteins and metabolism are retained in cerebral
tissue. As shown in Figure 3, methionine is demethylated and the carbon-1 1 is transferred to various methyl
acceptors (neurotransmitters, metabolic intermediates,
Supplement to Volume 15, 1984
PKllTt I N
w k w ALA A TYH ---L DOPA
( " C o p ) k4
Precursor Pool
p-bH P H t N Y L P Y K U V I C A C I D
1 1 C-Protein
Fig 2. Pathways for incorporation ofphenylalanine into proteins
and for metabolism.As discussed in Figure 1, the carbon-1 I label
in the carboxylposition of L-phenylalanine is removed in the metabolic pathway by decarboxylation to labeled carbon dioxide
(COz),which is cleared by dgfusion and bloodflow. PHENYL
ALA = phenylalanine; TYR = tyrosine; NOR-EPI = norepinephrine; EPI = epinephrine; p-OH = para hydroxy. Other
abbreviations as defined i n Figure 1.
Fig 3. Pathways for incorporation of methionine into proteins
and for metabolism.AJdiscussed in F i g w e 1, the carbon-1 1 label
in the carboxylposition of L-methionine is remozjed in the metabolic pathway & decarboxylation t o labeled carbon dioxide
(COz).If L-methionine 1s labeled in the methyl position with
carbon-1 I , the label is remvved by methyl acceptors, as shown in
the figure. These labeled methylated compounds are then retained
in the brain, making it difficult to discriminate between the
pathways of protein incorporation and metabolism by tracer kinetic analyses (see text). ATP = adenosine triphasphate; CoA
= coenzyme A.
etc.) that are retained in the brain. Thus, kinetic studies
with PET will probably be unable to differentiate between those two processes and will be able only to
provide estimates of the sum of these two reaction
pathways. Some studies have indicated that normally
the fraction of methionine metabolized is small (P.
Bustany, private communication, 1983). However,
even if this finding is true for the whole brain, local
variations may be of greater magnitude.
Although there are questions regarding the advis-
ability of using methionine labeled in the methyl position with carbon-11 and a tracer kinetic model that
ignores the metabolic pathway, as employed by Bustany and associates 171, interesting results have been
achieved by Bustany and associates in studies of
normal individuals and patients with Alzheimer's disease and schizophrenia. The validity of this assumption
and the occurrence of other labeled amino acids in
plasma during measurement of lCRPS is now under
investigation by Bustany and associates (private communication, 1983) using direct biochemical assay techniques in baboons. In any case, the principle of labeling
methionine in the carboxyl position is designed to reduce this metabolic effect.
Tracer Kinetic Model
As stated previously, data from direct biochemical
studies are required to structure a tracer kinetic model
properly. These studies provide such information as
the temporal sequence of the specific activity (labeled
to unlabeled concentrations) for the substrate in plasma
as well as for the substrate and reaction products in
tissue after intravenous or arterial injection of the
tracer. With this information, along with information
from other biochemical techniques, we can configure a
compartmental description of the process, as shown in
Figure 4 for L-[l-"C]leucine. One of the difficulties
existing presently is that while much is known about
the biochemical reaction sequences and total tissue
Phelps et al: Measurement of Cerebral Protein Synthesis in Humans
concentrations of reactants and products, much less is
known about their kinetic rates in gross tissue or in its
compartmental components. For example, the leucine
tissue precursor pool is not the large acid-soluble pool
but rather only a small compartmentally separated portion. Very little is known about the concentration and
turnover rate of this pool, the degree to which this pool
communicates with the plasma pool and the endogenous leucine, the rates of formation and breakdown of
the various proteins into which leucine is incorporated,
and the degree to which leucine from protein breakdown returns to the precursor pool as opposed to being incorporated in the large acid-soluble pool. Thus,
there are many perplexing issues regarding how to formulate a model properly and how to estimate the inaccuracies that can result from assumptions of the model
under different conditions.
Although a tracer kinetic model for PET must, by
nature of the data, be fairly simple, we nevertheless
must work to understand and justify the simplifying
assumptions. For example, the first-order model shown
in Figure 4 assumes, among other things, that (1) the
tissue precursor pool is supplied only by plasma, (2)
proteins are not broken down to allow the release of
labeled or unlabeled leucine back to the precursor pool
or to the bloodstream ro any substantial degree during
the time of the study (i.e., over a 40 to 80 minute
period), and (3) the metabolic pathway is essentially
unidirectional for the labeled leucine. The validity of
these assumptions must be determined by biochemical
assay and tracer studies in animals. Models for human
studies will probably require these types of data to be
acquired from investigations in primates to assure that
data from rats are valid or to accommodate species
differences. The combination of tracer, autoradiographic, biochemical, and tomographic assay techniques is probably required to approach these questions efficiently on both a global and a local basis and to
validate the model as applied to humans with PET.
Because of the similarity of the biochemical sequences of L-leucine, L-phenylalanine, and L-methionine labeled in the carboxyl position, the model in
Figure 4 is also used as a starting point for phenylalanine and methionine. However, each amino acid
must be investigated specifically to validate the accuracy of the model.
Amino Acid Analogs
A number of amino acid analogs (such as p halogenophenylalanine, ethionine, and norleucine)
have been investigated and have been shown to be
substrates for specific aminoacyl rRNA synthetases
with varying degrees of affinity relative to the natural
amino acids. Para-chlorophenylalanine has been shown
to be a substrate for phenylalanine tRNA synthetase in
the rat brain [IS]. Para-fluorophenylalanine is an a
S196 Annals of Neurology
priori attractive amino acid for PET, because it could
be labeled with fluorine-18, which has a longer half-life
than carbon-I 1 (110 minutes versus 20 minutes) [15].
It is activated as well as phenylalanine with Escherichiu
coli tRNA synthetases 1111. However, in rabbit reticulocytes, the active synthetase showed less affinity
for p-fluorophenylalanine than for phenylalanine [ 11.
Subsequently, p-fluorophenylalanine was considered
unsuitable for quantitative studies in the brain, because
50% of the compound was hydroxylated to tyrosine,
both in vivo and in vitro [13, 191.
At present the lack of information on the kinetics of
membrane transport, incorporation into proteins, competitive enzyme inhibition, and metabolism of these
amino acid analogs has limited their consideration for
PET. Further biochemical studies are required to
define the feasibility and specificity of these analogs.
Chemical Synthesis of L-E 1-' 'CILeucine,
Phenylalanine, and Methionine
The most effective approach to labeling these amino
acids in the carboxyl position with carbon-1 1 is to use
the Bucherer-Streker reaction, as initially reported by
Washburn and colleagues [351. The resultant labeled
amino acids, however, contain the admixture of D- and
L-isomers. Therefore, a rapid and efficient technique
for isolating the L-isomer is required. The D- and Lisomers can be separated by high-pressure liquid
chromatography using a chiral mobil-phase and reversed-phase column [14, 361.Resolution is excellent,
and the system resolves basic and acidic as well as neutral amino acids.
Alternatively, isolation of L-amino acids can be
achieved by treating the enantiomeric mixture with Damino acid oxidase, an enzyme that deaminates the Damino acid without affecting the L form { 5 , 6, 81.
D-amino acid oxidase is most effective for neutral
amino acids [12]. The D-isomers of alanine, methionine, and proline are among the best substrates for Damino acid oxidase, whereas those of phenylalanine,
leucine, isoleucine, valine, serine, and tryptophan are
oxidized somewhat less rapidly. However, even with
these latter amino acids the reaction is complete
within minutes.
Barrio and associates 141 recently used immobilized
D-amino acid oxidase (co-immobilized with catalase) to
prepare L-[ 1-' 'C]leucine. This technique has now been
extensively used at UCLA and has been found to be
very convenient, providing good yields (40 to 50 mCi)
of high purity (>99%), pyrogen-free L-[ 1-"C]leucine
141. These enzyme columns, when properly stored, can
be used repeatedly over periods of several months
without appreciable reductions in yields [4]. This approach has also been used by Barrio and associates [31
to prepare L-{1-"Cfphenylalanine and rnethionine in
good yields.
Volume 1 5 , 1984
TIME (sec)
Fig 5 . Example of the plasma clearance curve after the intravenous injection of L-leucine labeled with carbon-I I in position 1 ,
Circles indicate the total carbon-1 I activity in the plasma after
the removal of lubeled carbon dioxide. Triangles indicate the
amounts of labeled proteins in the plasma as a function of time.
The differencebetween the values in the curves dejined by circles
and triangles is the labeled leucine plasma concentration curve.
Note that labeled leucine rapidly declines t o a low steady-state
cjalue, and labeled proteins rapidly appear in the plasma andprogressively become a major portion of the total activity in the
plasma. Analysis by high-pressure liquid chromatography
showed that there were no other labeled products except for carbon
dioxide over a 60 minute examination time. CPM = counts per
Kinetic Studies with ~-[l-”C]Leucine
Kmetic studies were performed in 5 rhesus monkeys
and 8 human subjects with PET subsequent to intravenous injection of 15 to 30 mCi of ~-[l-~’C]leucine.
Scans were taken with the NeuroECAT {I71 in the
low-resolution mode (11.5 mm), with initial scans
every 60 seconds for 10 minutes, every 3 minutes for
30 minutes, and every 10 minutes for 50 minutes.
Arterial blood samples (from the radial artery in humans) were taken every 10 to 15 seconds for the first
2 minutes, with the time interval progressively
lengthened throughout the study. Plasma was separated and was acidified with perchloric acid to remove
the labeled CO2. Separation of labeled plasma Lleucine and identification of any other labeled products
in the plasma were performed using high-pressure liquid chromatography. Cold plasma leucine concentrations were also determined by this method. Blood samples were taken before, during, and at the end of each
experiment for blood gases, pH, and glucose determiPhelps
et al:
nations. The plasma input function and tomographic
data were analyzed using least squares estimation criteria with an operational equation of the tracer kinetic
model shown in Figure 4.
Rhesus monkeys were also injected with an intracarotid bolus of nitrogen-1 3-labeled L-leucine (L[13N]leucine) 121 and 1 hour later with a bolus of L[I-”C]leucine, and time activity curves were recorded
for both runs with a collimated probe positioned over
the brain. Data were recorded every 0.1 second for
the first 2 minutes, and the times were progressively
lengthened throughout the 1-hour duration of each
Figure 5 shows a typical L-[ 1-’‘Clleucine arterial
plasma curve after intravenous injection of 20 mCi of
this labeled amino acid. Note the desired rapid decline
to a low, relatively constant plasma concentration and
the rapid occurrence of labeled proteins, probably albumin from the liver. N o other labeled product was
identified in the plasma except for COz.This outcome
illustrates another important aspect of labeling in the
carboxyl position. Not only does this approach provide
a “shunt” to eliminate carbon-11 from the metabolic
pathway in the brain, but it also appears to eliminate the occurrence of other labeled amino acids
in the plasma that have been synthesized elsewhere in
the body. For example, with methionine labeled in the
methyl position, other labeled amino acids (e.g., serine)
appear in the plasma [2 11. Their presence complicates
the input function, the tissue kinetic data, and the
Figure 6 illustrates the time activity curves from the
serial intracarotid bolus injections of L-[’ 3N]leucine
and then L-{ I-”C}leucine. Notice that the unidirectional extraction (about 30%) is the same for the two
Measurement of Cerebral Protein Synthesis in Humans
100 I
TIME ( S e c )
compounds, as it should be, but the subsequent tissue
clearance is quite different. The nitrogen-1 3 activity
clears very slowly (with approximately a 900 minute
half-time), because of two mechanisms: (1) incorporation of labeled L-leucine into proteins and (2) transfer
of the nitrogen-13 to the L-glutamate pool (see Fig 1)
via the transamination reaction catalyzed by the pyridoxal phosphate-dependent L-leucine transaminase.
The L-glutamate clears slowly, because of dilution in
the large L-glutamate tissue pool and because of its low
blood-brain barrier permeability 1281 (see Table 1).
The biphasic nature of the carbon-I 1 tissue clearance curve results from (1) the metabolic pathway with
the decarboxylation reaction that yields carbon- 11labeled COz, which clears from the tissue, and (2) the
retained carbon-11 in labeled leucine that has been
incorporated into proteins that turn over very slowly.
These studies illustrate general kinetic features that are
consistent with the pathways shown in Figure 1 and
with the model shown in Figure 4.
Figure 7 illustrates the results of a kinetic study in a
normal subject. The study shows agood fit of the tracer
kinetic model to the measured carbon-1 1 tissue concentration and to the calculated temporal course of the
label in the leucine precursor pool, the proteins, and
the metabolic pathway. Results of the kinetic analysis
are shown in Table 2 , and the image quality of a typical
PET study with L-[l-"C)leucine is shown in Figure 8.
These data, along with the standard error of the fit,
illustrate that the model fits the data quite well. The
numerical values of the rate constants, pool turnover
rates, and cortical ICRPS are in reasonable agreement
with initial estimates in adult rats and monkeys, which
were arrived at by direct biochemical and quantitative
autoradiographic analysis using L-{ l-'*C]leucine f241,
S198 Annals of Neurology
Fig 6. Time activity ruwe.( subsequent to the intracarotid bolus
injection of L-leucine labeled with nitrogen-13 (circles) and Lleucine labeled with carbon-1 1 in position 1 (solid line) in the
rhesus monkey. Note that the temporal courses are identical initially, since extraction i.r independent of the label or its position.
Subsequently the curves diverge. The N-13-labeled products are
trapped and retained in the tissue (see text). In contrast, a portion of the activity in the leucine labeled with carbon-11 in the
carboxylposition clears out of the tissue through the reaction of
the metabolic pathway. This uction results in decarboxylation
and clearance of the activity as labeled carbon dioxide. The residual retained portion of the activity results from the incorporation
of leucine into the proteins.
(L. Sokoloff, private communication, 1983). Due to
species-dependent variables, only approximate comparisons can be made. The values in adult rhesus monkeys were also in good agreement with those estimated
in humans with PET [241.
Table 2 shows that the precursor pool for tissue
leucine does have the desired rapid turnover rate (a
half-time of about 1.5 to 2 minutes). The rate of clearance through the metabolic pathway has a half-time of
about the radioactive half-life of carbon-11, which limits how long one can wait to diminish background from
this origin. Another factor that affects the level of background activity from metabolism is the branching fraction (the ratio of lCRPS/metabolism) through these
competitive pathways. Results are shown in states in
which subjects (1) fast for about 8 hours, ( 2 ) fast for
about 8 hours and are then given a load dose of oral
glucose ( I to 2 hours before the study), and (3) are fed
a carbohydrate meal 1 to 2 hours before the study plus
a supplement of oral glucose 30 minutes to 1 hour
prior to the study. Studies performed in the fasted state
Supplement to Volume 15, 1984
Fig 7. Results from a kinetic study in a normal human subject
after the intravenous injection of L-leucine labeled with carbon11 in position I , as measured with PET. The region is about a
square centimeter area of the cortex. Circles indicate the actual
measured tissue concentration with the tomograph. Squares,
triangles, and diamonds zndicate the concentrationsof the tracer
in the indicated compartments,as calculated from fitting the tissue kinetic data with the model shown in Figure 4. Note that in
this study b o r n a subjectfed a carbohydratemed with a supplement of oral glucose (see text), the ratio of label in proteins compared to that i n the precursor pool and in the metabolic pathway
is dominant for times Later than about 20 minutes after intravenous injection.
Table 2. Cortical Protein Synthesis in HumanJ
Rate Constants
Glucose Synthesis
(mmol) (nmollmidgm)
CPSRlMet (ymol)
0.033 0.032
0.037 0.036
0.39 0.056
0.26 0.044
0.041 0.078
0.31 0.066
0.31 0.10
Half-time (min)
"All k values are given in min-
'Studies in two separate subjects in the same dietary state but having different levels of plasma leucine (shown in the Table) and glucose (top
subject = 98 mgllO0 ml; bottom subject = 130 mg/IOO ml). Note changes in transport constants and ICRPSiMer ratio.
Kinetic data: k,, k, = bidirectional transport;
incorporation in proteins.
metabolic rate.
cerebral protein synthesis rate; Met
metabolism; k4
clearance of labeled carbon dioxide from tissue to blood;
Phelps et al: Measurement of Cerebral Protein Synthesis in Humans
demonstrated that most of the labeled leucine was
taken through the metabolic pathway (a low branching
fraction). Because glucose and pyruvate had been
shown to inhibit leucine metabolism in peripheral muscle and heart [ 7 ] , fasted subjects were given a glucose
load prior to the study. However, the branching fraction remained low (see Table 2). This low level appears
to have resulted from the fact that the acute glucose
load stimulated the release of insulin, since the plasma
glucose level was relatively low (80 to 90 mR/lOO ml);
that is, at the time of the L-leucine study, the subject
showed a normal response to a glucose tolerance test.
When a more natural approach was taken by giving the
subject a carbohydrate meal with supplemental glucose, the branching fraction increased and the plasma
glucose was maintained at higher levels (100 to 130
mg/lOO ml). The kinetic study shown in Figure 7 was
done under this more favorable condition. One can see
that at 40 minutes the amount of carbon-11 in proteins
in cortical gray matter regions is about 85% of the total
radioactivity in brain. This level is contrasted to the
fasted state, in which it is only about 40% of the total
level at this time. The rate constants in these different
states are quite different, as would be expected, owing
to the different plasma amino acid concentrations and
rates of metabolism and protein synthesis.
Future Perspectives
Although the results shown are preliminary and the
model contains a number of unproved assumptions,
the principle of the approach appears sound, and many
of the criteria described in earlier sections appear to be
met to a reasonable degree. The strategy of labeling in
Fig 8. Examp1es of images in a noma1 human subject 40 minutes after intravenous injection of the L-leucine labeled with carbon-11 i n position 1 (bottom row). Top row shows transmission
images recorded with the NeuroECAT for attenuation corvection
and identificationof the outer boundary of the head. Images from
left to right are the level of the cerebral fissures, the basal gungliu
and thalamus, and the cerebellum. Note the high d u e s of the
cortex and subcortical nuclei, active areas apparentlyfrom subthulumic nuclei, and low activity in the caudate nucleus.
Identification of the cerebellar cortex, vemzis, and dentate nuclei is
also seen. Images contain from 800.000 to I .3 million counts.
the carboxyl position improves specificity of the measurement of 1CRPS. Of course, as with all exogenous
substrates, the utilization rate reflects a very distributed
set of functional processes. The 20 minute half-life of
carbon-11 and the similar pool turnover rate for the
metabolic pathway pose some limitations in the degree
to which background from the latter can be reduced
(e.g., the signdnoise ratio, which is the ratio of activity
in labeled proteins to that in the metabolic pathway, is
increasing with about the same half-time as the signd
noise ratio decrease due to radioactive decay). However, the initial studies indicate that this level of
background noise is acceptable, particularly if other
developments can be applied to yield improved signal/
noise characteristics in the study.
The following are some approaches to improving the
technique for assessing the lCRPS in humans with
1. Biochemical and tracer assay studies in animals
should be done to define better the structure and
S200 Annals of Neurology Supplement to Volume 15, 1984
Fig 9. Autoradiographs after the intravenous injection of phenylalanine, leucine, and methionine labeled with carbon-14 in
the carboxyl position. Examples are from three d$,fment rats and
illustrate a similar distribution of carbon-I4 tissue concentration
at 60 minutes after injection. In initial autoradiographic studies. only minor differences have been noted among results from the
three amino acids in regard t o the tissue concentration normalized
to the amount injected, washed versus unwashed tissue activity
concentrations, and the anatomical distributions. However, more
&tailed autoradiographicand tissue biochemical assay studies as
a function of time are required to discriminate among these
amino acidsfor the measurement of protein synthesis.
assumptions of the tracer kinetic model. Because of
potential species differences, however, a combination of these approaches and tracer kinetic studies
with PET in humans will be required to understand
better the translation of the method to humans.
2. Improved understanding of the structure and assumptions of the model, along with good initial estimates of the kinetic parameters, will allow the application of statistical techniques to develop parameter
estimates based on optimization criteria (maximum
likelihood, weighting functions, etc.).
3. Other amino acids should be examined to determine if the specificity and signdnoise ratio of the
technique can be improved (e.g., by means of more
favorable transport, smaller and more rapid tissue
precursor pool turnover rates, higher branching
fractions, and faster turnover rates for the metabolic
pathway). Initial studies with carbon-1 1 and carbon14 carboxyl-labeled L-leucine, L-phenylalanine, and
L-methionine are being performed at UCLA with
PET and autoradiography, respectively. Although it
is too early to determine a preference, initial studies
show that these three amino acids have similar qualitative characteristics (Fig 9).
4. The effects of dietary states and plasma concenrrations of the various competitive amino acids must
be better understood to allow improved study protocols.
5. Improvements in spatial and temporal resolution
and overall image signalhoise ratio of positron tomographs are needed, such as with system designs
based on the Signal Amplification Technique E27).
Supported in part by Contract DE-AM03-76-SF000 12 from the
Department of Energy and Grants 1-R01-MH37916-01 from
the National Institute of Mental Health and POI-NS15654
from the National Institute of Neurological and Communicative
Disorders and Stroke. Dr Mazziotta is the recipient of Teacher/
Investigator Award 1K07-0058801-NSPA from the National Institute of Neurological and Communicative Disorders and Stroke.
The authors thank Dr MacDonald and his cyclotron sraft Joann
Carson, Ron Sumida, Larry Pang, Tony Ricci, and Jerry Low for their
assistance; and Maureen Kinney for preparing the manuscript.
1. Arnstein HRV, Richmond MR: Utilization of p-fluorophenylalanine for protein synthesis by the phrnylalanineincorporation system from rabbit reticulocytes. Biochem J
91:340-346, 1964
2. Barrio JR, Baumgartner FJ, H e m e E, et al: Synthesis and
myocardial kinetics of N-13 and C-11 labeled branchedchain L-amino acids. J Nucl Med 24937-944, 1983
3. Barrio JR, Keen RE, Chugani H, et al: ~-[l-C-l1]Phenylalaninc
Phelps et al: Measurement of Cerebral Protein Synthesis in Humans
for determination of cerebral protein synthesis in man with PET.
J Nucl Med 24:70, 1983
4. Barrio JR, Keen RE, Ropchan JR, et al: LEl-"C]Leucine: routine
synthesis by enzymatic resolution. J Nucl Med 24:5 15-521,
5. Bright HJ, Porter DJT Flavoprotein oxidases. D-Amino acid
oxidase. In Boyer P D (ed): The Enzymes, Vol 12. 3rd edition.
New York, Academic, 1975, pp 445-456
6.Buse MG, Biggers JF, Friderici KH, et al: Oxidation of branched
chain amino acids by isolated hearts and diaphragms of the rat. J
Biol Chem 247:8085-8096, 1972
7. Bustany P, Henry JF, Sargent T , e t al: Local brain protein metabolism in dementia and schizophrenia: in vivo studies with "C-Lmethionine and positron emission tomography. In Heiss W-D,
Phelps ME (eds): Positron Emission Tomography of the Brain.
New York, Springer-Verlag, 1983, pp 208-211
8.Casey DL, Digenis GA, Wesner DA, et al: Preparation and
preliminary tissue studies of optically active "C D- and Lphenylalanine. Int J Appl Radiat Isor 32:325-330, 1981
9. Chaplin ER, Goldberg AL, Diamond I: Leucine oxidation in
brain slices and nerve endings. J Neurochem 26:701-707,
10. Comar D, Cartron JC, Maziere M, et al: Labelling and metabolism of methionine-methyl "C. Eur J Nucl Med 1:11-14,
11. Conway TW, Lansford EM, Shive W: Purification and substrate
specificity of a phenylalanine-activating enzyme from Escherichza
rolz 9723. J Biol Chem 237:2850-2854, 1962
12. Dixon M, Kleppe K D-Amino acid oxidase. I. Dissociation and
recombination of the holoenzyme. Biochim Biophys Acta
96:357-367, 1965
13. Gal EM: Synthetic p-halogenophenylalanines and protein synthesis in the brain. Ciba Found Symp 22:343-359, 1974
14. Gil-Av E, Tishbee A, Hare PE: Resolution of underivatized
amino acids by reversed-phase chromatography. J Am Chem
Soc 102:5115-5117, 1980
15. Goulding RW, Palmer AJ: The preparation of fluorine-I8
labeled p-fluorophenylalanine for clinical use. Int J Appl Radiat
Isot 23:113-137, 1972
16. Heiss W-D, Phelps ME (eds): Positron Emission Tomography of
the Brain. New York, Springer-Verlag, 1983
17. Hoffman EJ, Phelps ME, Huang SC: Performance evaluation of
a positron tomograph designed for brain imaging. J Nucl Med
24:245-257, 1983
18. Huang SC, Phelps ME, Hoffman EJ, e t al: Noninvasive determination of local cerebral metabolic rate of glucose in man. Am J
Physiol 238:E69-E82, 1980
19. Kaufman S:Enzyme conversion of 4-fluorophenylalaninc to tyrosine. Biochim Biophys Acta 51:619-621, 1961
20. Lehninger AL: Biochemistry. 2nd edition. New York, Worth,
1975, pp 933-935
21. Lundquist H, Langstrom B, Stalnacke CG: 11-C-Labeled
metabolites in plasma after 11-C-methionine i.v. administration.
Presented at VII Nobel Conference on Metabolism of the Hu-
S202 Annals of N e u r o l o g y
man Brain Studied with Positron Emission Tomography, Stockholm, Sweden, 1983
22. Oldendorf WH: Brain uptake of radiolabeled amino acids,
amines and hexoses after arterial injection. Am J Physiol
22 1:1629-1639, 197 1
23. Oidendorf W H , Szabo J: Amino acid assignment to one of three
blood-brain barrier amino acid carriers. Am J Physiol 230:9498, 1976
24. Phelps ME, Barrio JR, Huang SC, et al: The measurement of
local cerebral protein synthesis in man with positron computed
tomography (PCT) and ["C]~-leucine. J Nucl Med 2 3 6 , 1982
25. Phelps ME, Hoffman EJ, Mullani NA, et al: Application of
annihilation coincidence detection to transaxial reconstruction
tomography. J Nucl Med 16:210-224, 1975
26. Phelps ME, Huang SC, Hoffman EJ, e t al: Tomographic measurement of local cerebral glucose metabolic rate in humans with
validation of method. Ann
Neurol 6:371-388, 1979
27. Phelps ME, Huang SC, Hoffman EJ, et al: An analysis of signal
amplification using small detectors in positron emission tomography. J Comput Assist Tumogr 6:551-565, 1982
28. Phelps ME, Huang SC, Kuhl DE, et al: Cerebral extraction of N13 ammonia: its dependence on cerebral blood flow and capillary permeability. Stroke 12:607-619, 1981
29. Phelps ME, Mazziotta JC, Huang SC: The study of cerebral
function with positron computed tomography. J Cerebral Blood
Flow Metab 2:113-162, 1982
30. Raichle ME, Grubb RLJ, Eichling JO, e t d:Measurement of
brain oxygen utilization with radioactive 0-15: experimental
verification. J Appl Physiol 40:638-64 1, 1976
31. Reivich M, Kuhl D, Wolf A, e t al: The ["F)fluorodeoxyglucose
method for the mesurement of local cerebral glucose utilization
in man. Circ Res 44:127-137, 1979
32. Smith CB, Davidsen L, Deibler G , et al: A method for the
determination of local rates of protein synthesis in brain. Trans
Am Soc Neurochem 1194,1980
33. Sokoloff L, Reivich M, Kennedy C, et al: The [C14]deoxyglucose method for the measurement of local cerebral
glucose utilization: theory, procedure and normal values in the
conscious and anesthetized albino rat. J Neurochem 28:897916, 1977
34. Sokoloff L, Smith C: Biochemical principles for the measurement of metabolic rates in vivo. In Heiss W-D, Phelps ME (eds):
Positron Emission Tomography of the Brain. New York,
Springer-Verlag, 1983, pp 2-18
35. Washburn LC, Sun TT, Byrd BL, et al: High-level production of
"C carboxyl-labeled amino acids. In Sorenson JA (ed):
Radiopharmaceuticals, Vol 2. New York, Society of Nuclear
Medicine, 1979, pp 767-717
36. Washburn LC, Sun IT,Byrd BL, er al: Production of ~ - { l l1C}valine by HPLC resolution. J Nucl Med 23:29-33, 1982
37. Williamson JR, Walajtys-Rode E, Coil KE: Effects of branched
chain alpha-ketoacids on the metabolism of isolated rat liver
cells. J Biol Chem 254:11511-11520, 1979
S u p p l e m e n t to Volume 15, 1984
Без категории
Размер файла
1 245 Кб
measurements, synthesis, protein, tomography, trace, emissions, kinetics, criterias, positron, human, cerebral
Пожаловаться на содержимое документа