| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Radiology, Department of Medicinal Chemistry, Department of Neurosurgery, and The Graduate Program in Neuroscience, University of Minnesota, Minneapolis, Minnesota 55455, USA
1Correspondence: Department of Radiology (Mayo Box 292), University of Minnesota. Academic Health Center, 420 Delaware St. SE, Minneapolis, MN 55455, USA. E-mail: efang001{at}maroon.tc.umn.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
Key Words: choline acetyltransferase • positron emission tomography • single photon emission computed tomography • radiotracer
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
, 2)
. Basically,
PET differs from SPECT in that the former uses coincidence detection to
capture the signal emitted by the annihilation of a positron, as the
two resulting annihilation photons diverge at 180° relative to each
other. Traditionally, the coincidence detection technique has provided
higher spatial resolution for PET; however, recent advances in SPECT
technology have narrowed the gap between the two technologies. In both
SPECT and PET, the desired radioligand is synthesized by introduction
of a selected radionuclide into a precursor molecule. For PET scanning,
the radionuclides carbon-11, nitrogen-13, and fluorine-18 are commonly
used. The most common radionuclides used in SPECT are iodine-123 and
technetium-99m. In vivo validation of radioligands generally
takes the form of biodistribution and radioligand blocking studies,
kinetic and metabolite analysis. In the biodistribution study, the
distribution of the tracer in the organ of interest is examined over
time and compared with the distribution of the molecular target
determined by independent methods. To facilitate data analysis, regions
of low or insignificant target density are generally used as reference
regions, which explains the frequent reference to parameters such
as the striatum:cerebellum and striatum:cortex ratios in this
review. Blocking studies are used to show that the radioligand binds to
the site in question, whereas the kinetic assessment is used to
adequately describe the in vivo behavior of the ligand. A
major shortcoming of the radiotracer methodology is its inability to
distinguish between chemical species bearing identical radionuclides.
Radiotracer metabolism is therefore an important consideration in the
study of living systems with PET or SPECT. The current review focuses on the development of radioligands and in vivo imaging of the vesicular acetylcholine transporter (VAChT) and vesicular monoamine transporter (VMAT). Although the main thrust of the review is in vivo imaging, a portion of the discussion is devoted to ligand development and studies performed with 125I- and tritium-labeled ligands because the latter provide the foundation on which in vivo investigations are built. Because the results of an imaging experiment can be profoundly influenced by the radiotracer used, the author chose to review the available in vivo data with reference to specific radiotracers. This arrangement provides the reader with a better appreciation of developments in the area and a grasp of controversial issues and their origins.
|
PROGRESS IN THE DESIGN, SYNTHESIS, AND PHARMACOLOGICAL CHARACTERIZATION OF VAChT LIGANDS |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
, 4)
. The subsequent identification (5)
of
a unique binding site for (-)-[3H]vesamicol in
Torpedo synaptic vesicles heightened interest in the study
of the underlying pharmacology and biology of the binding protein.
Although (-)-vesamicol was shown to inhibit uptake of acetylcholine
(ACh) by cholinergic synaptic vesicles, it was not clear from the early
studies whether vesamicol acts directly on VAChT or on some other
protein associated with ACh storage. Nevertheless, the availability of
(-)-[3H]vesamicol quickly led to attempts to
map the distribution of this binding site in mammalian brain. In rat
brain sections (6)
, the radioligand was found to saturate
a single class of binding sites with moderately high affinity.
Densitometric analysis of autoradiographs obtained from rat brain
sections treated with (-)-[3H]vesamicol
revealed that the binding sites were heterogeneously distributed in the
mammalian brain. Specific binding of the radioligand was highest in the
diagonal band of Broca and olfactory tubercle, high to moderate in the
caudate-putamen, nucleus accumbens, and cerebral cortex, and low in the
cerebellum. Moreover, after unilateral transsection of the left
fimbria, specific binding of (-)-[3H]vesamicol
was decreased significantly as was the level of choline
acetyltransferase (ChAT) (7)
. These findings led to the
suggestion that the (-)-[3H]vesamicol binding
site is associated with cholinergic nerve terminals.
Since then the (-)-[3H]vesamicol binding site
has been shown to reside on VAChT, a critical component of neuronal
cholinergic machinery. VAChT has been cloned from several sources
including the nematode Caenorhabditis elegans, three species
of Torpedo, Drosophila, mouse, rat, and human
(8
9
10
11
; reviewed in ref 12
). ACh accumulation
and vesamicol binding are expressed by a single polypeptide, suggesting
that all of the essential components of the VAChT reside in that
polypeptide. Thus, the terms vesamicol receptor (VR) and VAChT are
synonymous. The gene for VAChT is embedded within the first intron of
the gene for ChAT in all of the species examined, suggesting that the
expression of ChAT is tightly coupled to that of VAChT (11
, 13
, 14)
. In several studies of the rat brain, staining for VAChT
protein, vacht mRNA, ChAT protein, and chat mRNA
has unequivocally revealed that VAChT is localized to synaptic vesicles
within cholinergic terminals (11
, 13
, 14
15
16
17
18
19)
. In fact,
VAChT has been used to confirm certain aspects of the cholinergic
system whose existence had been a subject of controversy. Consequently,
VAChT is now firmly established as a reliable cholinergic marker. This
review summarizes the development of VAChT ligands for in
vivo imaging in the past decade.
Although the distribution of (-)-[3H]vesamicol
in rat brain paralleled the distribution of other cholinergic markers,
experimentally induced loss of cholinergic innervation to the
hippocampus resulted in a mismatch between reductions in
(-)-[3H]vesamicol binding and ChAT activity.
According to Marien et al. (6)
, ChAT was reduced by
61 ± 7% whereas (-)-[3H]vesamicol
binding was reduced only by 33 ± 6%. In subsequent
investigations, Ruberg et al. (20)
and Kish et al.
(21)
observed a similar mismatch in the rat cortex after
unilateral lesioning of the nucleus basalis. Furthermore, in the brains
of deceased Alzheimer’s disease (AD) patients, Kish et al.
(21)
and Holley et al. (22)
found that
although ChAT activity was reduced by 60–80% in the neocortex and
amygdala, (-)-[3H]vesamicol binding was either
unchanged or only slightly reduced. To explain the apparent discrepancy
between ChAT activity and [3H]vesamicol
binding, these earlier workers advanced the following hypotheses:
1) cholinergic neurons that survive the lesion may
compensate by expressing higher levels of synaptic vesicles;
2) one segment of surviving cholinergic terminals may
express inadequate levels of ChAT and acetylcholine esterase (AChE);
3) degenerating (ChAT-deficient) cholinergic terminals may
exhibit up-regulation of VR; and 4) a substantial fraction
of [3H]vesamicol binding sites may be
associated with non-neuronal elements. The latter hypothesis was
subsequently proved by the demonstration that vesamicol displays
nanomolar affinity for 
receptors (23)
. The marginal
selectivity of (-)-[3H]vesamicol has led to a
vigorous effort to develop more selective VAChT ligands. A number of
such ligands have indeed been developed and some of the more prominent
ones are shown in Fig. 1
.
|
|
IN VITRO AND EX VIVO EVALUATION OF VAChT LIGANDS |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
(-)-[3H]Vesamicol
As described above, early studies reported a mismatch between the
loss of (-)-[3H]vesamicol binding and ChAT
activity in AD and in experimentally induced lesions in the rodent.
Since the publication of these studies, other investigators have
reported their findings regarding the binding of
(-)-[3H]vesamicol in human disease and in
animal models. Wenk and Mobley (24)
found that
(-)-[3H]vesamicol binding was unchanged in the
rat cortex after experimentally induced cholinergic cell loss, whereas
ChAT activity declined by 51–68%. On the other hand, in Rett
syndrome, both ChAT activity and
(-)-[3H]vesamicol accumulation declined
significantly in the putamen and thalamus. However, the decline in ChAT
activity in these regions was twofold greater the reduction in
(-)-[3H]vesamicol binding.
(+)-[125I]MIBT
(+)-[125I]MIBT binds to a single
high-affinity binding site (Kd=4.4±0.7 nM) in
monkey striatum and occipital cortex (25)
. Age- and
disease-related changes in VAChT density have been reported with this
ligand. Efange at al. (26)
have shown that after
unilateral destruction of the nigrostriatal dopaminergic pathway,
ipsilateral striatal cholinergic neurons become refractory to the
disinhibitory effects of dopamine D2 receptor blockade. Thus, in
6-OHDA-lesioned animals pretreated systemically with the D2 antagonist
spiperone, (+)-[125I]MIBT levels in the
ipsilateral striatum increased by only 23% over control values whereas
those of the contralateral side increased by 87%. These investigators
also report that the effects of haloperidol and S-(-)-eticlopride on
striatal (+)-[125I]MIBT accumulation are age
related (27)
. In young adult rats pretreated with
haloperidol or S-(-)-eticlopride,
(+)-[125I]MIBT accumulation in the striatum
increased by 35 and 66%, respectively, over the controls. However, in
aged animals haloperidol failed to increase radiotracer accumulation in
the striatum, and S-(-)-eticlopride was only half as effective as in
the young animals. The authors conclude that aging and Parkinson’s
disease (PD) are associated with a decline in striatal cholinergic
functional reserve (vide infra).
In a study of human brain postmortem (28)
,
(+)-[125I]MIBT binding in the temporal cortex
was shown to be sensitive to changes in cholinergic density. The study
used brain tissue from 4 young adults (29–40 years), 9 aged controls
with no history of neurological disorder and 6 aged adults (67–85
years) with a history of AD, confirmed histopathologically by the
presence of neuritic plaques and neurofibrillary tangles.
(+)-[125I]MIBT binding was found to decrease
with age (Bmax: 31.2±6.3 pmol/g in young adults vs. 17.0±2.0 pmol/g
in aged adults, P<0.005) and with neuropathology (Bmax:
9.4±1.6 pmol/g in AD patients), reflecting a decline of 46% and 45%,
respectively. On the other hand, measurements of ChAT activity in the
three study groups revealed a 37% decline with age and a 67% decline
with disease. Thus, (+)-[125I]MIBT binding
correlates with ChAT activity (r=0.72). However, the
mismatch between ChAT levels and VAChT density persists, although it is
smaller than that observed for
(-)-[3H]vesamicol.
(-)-[125I]IBVM
The ability of this radioligand to detect cholinergic lesions has
been verified in the rodent. In control animals receiving an
intravenous (i.v.) injection of
(-)-[125I]IBVM, the distribution of
radioactivity, as determined by autoradiography, was found to match the
distribution of cholinergic terminals in the brain (29)
.
Seven days after treatment with the cholinergic immunotoxin 192
IgG-saporin, (-)-[125I]IBVM binding was
reduced by 50–60% in the cortex and hippocampus. In addition, the
reduction in (-)-[125I]IBVM binding was
correlated (r=0.78, P<0.02) with decrements in
AChE. No changes in either radiotracer binding or AChE were observed in
the cerebellum. Therefore, (-)-[125I]IBVM
appears to be suitable for detecting changes in cholinergic terminal
density in vivo.
|
IN VIVO IMAGING OF VAChT |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
of this compound has been
reported. (+)-[123I]MIBT was administered in a
female baboon by i.v. bolus, followed by constant infusion over 6–7 h.
Stable levels of radioactivity were observed in the striatum, occipital
cortex, and cerebellum after 3 h. Injection of pentazocine at
4 h postinjection induced a dose-dependent decline of
radioactivity in the occipital cortex and cerebellum, suggesting
contribution from receptors. The striatum was unaffected by this
treatment. These observations agree with earlier findings in the rat;
however, they are clearly at variance with studies of
(+)-[123I]MIBT in postmortem human brain. In
any case, the apparent involvement of receptors in the in
vivo binding of (+)-[123I]MIBT has
tempered enthusiasm for the use of this ligand in SPECT imaging.
(+)-[18F]FBT
A number of PET studies have used this radioligand. In one study
(31)
of young adult rhesus monkeys,
(+)-[18F]FBT was found to bind reversibly and
to distribute heterogeneously in the brain. Clearance of the
radiotracer was slow in the basal ganglia relative to the cortex and
cerebellum, resulting in a progressive increase in the
striatum:cerebellum ratio. At 180 min after radiotracer injection, the
value of this ratio approached 2. Although the cerebral cortex contains
a much higher density of cholinergic terminals than the cerebellum,
comparable levels of the radiotracer were observed in both structures
throughout the duration of the study. Since (+)-FBT displays 100-fold
higher affinity for VAChT than for 1 and 2 receptors, the authors
concluded that the poor signal:noise ratio obtained from the cortex was
not due to receptors but to one or more of the following factors:
the low density of cholinergic terminals in the primate cortex, partial
volume averaging caused by the low spatial resolution of the PET
scanner, or the relatively high lipophilicity of
(+)-[18F]FBT which results in a high level of
nonspecific binding. Exclusion of receptors as a contributing
factor appears to be partly supported by the observation that
(+)-[18F]FBT accumulation is unaffected by
pretreatment with potent 1- and
2-selective ligands (R. H. Mach et al.,
unpublished results). Whatever the source of this discrepancy,
(+)-[18F]FBT appears to be unsuitable for
studying cholinergic function in the primate cortex. However, the
ligand is useful for probing cholinergic mechanisms in the basal
ganglia. Recent work in aged rhesus monkeys demonstrates that
(+)-[18F]FBT accumulation in the basal ganglia
declines with age (R. H. Mach et al., unpublished results).
However, the decrease was not observed in every aged subject. Indeed,
whereas some aged subjects displayed significant reductions in
(+)-[18F]FBT accumulation, others were found to
accumulate the radiotracer at levels comparable to those observed in
young adults (Fig. 2
). This observation led the authors to conclude that the susceptibility
of the cholinergic system to the aging process is subject to individual
variation. Future studies that link the decline in
(+)-[18F]FBT accumulation with a diminution in
cognitive function should help to establish the latter as a useful
tracer for studying cholinergic function in vivo.
(+)-[18F]FBT may also be useful for studying
cholinergic function in the primate spinal cord as suggested by PET
studies in the rhesus monkey. Cholinergic activation in the spinal cord
induced by morphine is accompanied by an increase in
(+)-[18F]FBT accumulation within this structure
(33
, 34)
. The effect is reversed by naloxone and is
independent of blood flow. These data strongly suggest that
1) (+)-[18F]FBT binds to spinal
cholinergic neurons in vivo and 2) specific
binding of (+)-[18F]FBT in this structure is
dependent on the level of cholinergic activity. The latter study
extends earlier observations with this and other VAChT ligands and
provides additional support for the use of pharmacologic activation
strategies for in vivo imaging of VAChT. Taken together, the
foregoing suggests that (+)-[18F]FBT is a
potentially useful tracer for studying cholinergic function in
vivo. Evaluation of this radioligand in humans under baseline and
pharmacologic activation conditions may provide further insight into
the utility of this agent.
|
(-)-[18F]NEFA
Preliminary evaluation of (-)-[18F]NEFA
(35)
as a PET ligand was conducted in male cynomolgus
monkeys. Radiotracer distribution in this study was compared to that of
the potent VAChT ligand (-)-ABV. After i.v. administration, (-)-
[18F]NEFA accumulated rapidly in the primate
brain. Non-uniform radiotracer egress from key brain regions
subsequently resulted in a heterogeneous distribution that mirrors the
density of cholinergic terminals in the brain. At 45 min postinjection,
the highest levels of radiotracer were observed in the basal ganglia,
whereas moderate and low levels were found in the cortex and
cerebellum, respectively. No such late-phase heterogeneity was observed
with the inactive enantiomer (+)-[18F]NEFA.
Selective accumulation of (-)-[18F]NEFA was
abolished by pretreatment with the prototypical VAChT ligand vesamicol;
however, the latter compound could not displace bound
(-)-[18F]NEFA.
(-)-[18F]NEFA accumulation was similarly
unaffected by pretreatment with the ligand pentazocine, thus
eliminating receptors as a factor in the in vivo
accumulation of this radioligand. In contrast, pretreatment with the
dopamine D2 antagonist (and ligand) haloperidol selectively delayed
(-)-[18F]NEFA egress from the basal ganglia,
thereby resulting in a two- to threefold increase in
striatum:cerebellum and cortex:cerebellum ratios.
(-)-[18F]NEFA therefore appears to bind
selectively to VAChT sites in the primate brain. Moreover, the effects
of dopamine D2 receptor blockade suggest that the binding of
(-)-[18F]NEFA is dependent on the level of
striatal cholinergic activity.
Human studies
(-)-[18F]NEFA
Only one PET study (36)
has been reported for this
radioligand. The study used three subjects aged 60–65 years; two were
controls and one suffered from AD. As previously observed in cynomolgus
monkeys, high levels of radiotracer were observed in the striata of the
healthy subjects. The striatal concentration remained constant for the
duration of the study (60 min). However, the authors could find no
demonstrable binding of radiotracer in the human cortex. Similar
results were obtained from the single AD subject. To explain the poor
retention of (-)-[18F]NEFA in the human
cortex, the authors suggest that the kinetics of this radiotracer in
the striatum are different from those of the cortex. This claim is
supported by the finding that in rat brain slices, dissociation of
(-)-[3H]aminobenzovesamicol (ABV, the
immediate precursor for (-)-[18F]NEFA) is 30
times faster from cortex than striatum (35)
. Since
(-)-[18F]NEFA is poorly retained in the
cortex, it was not surprising that these investigators could detect no
differences between the healthy controls and the AD subject. Although
(-)-[18F]NEFA is derived from the potent and
highly selective VAChT ligand ABV, additional studies are needed to
determine whether this tracer is suitable for in vivo
imaging.
(-)-[123I]IBVM Two SPECT studies have been
reported for this compound. In the first study
(37)
, carried out to evaluate the kinetics and
distribution of (-)-[123I]IBVM in humans, tracer
distribution was heterogeneous in the human brain. Kinetic analysis
favored a 3-compartment model. Estimates of the rate constants for
tracer binding ranged from 0.102 min-1 in the striatum to
0.013 min-1 in the cortex. (-)-[123I]IBVM
was metabolized to the extent of 20% within 10 min after i.v.
injection. By 1 h postinjection, the level of parent compound in
the plasma was only 30%. The striatal tracer level increased steadily
during the first 4 h sampling period. In contrast, tracer levels
in the cortex decreased rapidly during the same period, reaching 50%
of their initial value within 80 min. Four hours after
(-)-[123I]IBVM injection, radiotracer levels in the
striatum were fourfold higher than those in the cerebellum and in all
regions of the cortex examined (frontal, temporal, parietal, and
occipital). At 22 h postinjection the striatum:cortex and
striatum:cerebellum ratios had increased to 6:1. As these ratios
compared favorably with those published previously for ChAT, the
authors concluded that SPECT imaging of VAChT with
(-)-[123I]IBVM is a reliable measure of cholinergic
integrity. In a follow-up study of normal aging (22–91 years,
n=36), AD (n=22), and PD
(n=15), Kuhl et al. (38)
observed a 3.7%
per decade reduction in cortical (-)-[123I]IBVM binding.
However, in AD these investigators observed an inverse relationship
between cortical (-)-[123I]IBVM binding and the severity
of the disease. With age of onset of less than 65 years, radiotracer
binding was reduced by 30–40% throughout the cerebral cortex and
hippocampus. On the other hand, for subjects with age of onset greater
than 65 years, reductions in (-)-[123I]IBVM were
restricted to the temporal cortex and hippocampus. In nondemented PD
subjects, (-)-[123I]IBVM binding was reduced only in the
parietal and occipital cortex, whereas demented PD subjects displayed
extensive reductions in (-)-[123I]IBVM binding similar
to those observed in early-onset AD. As postmortem studies of AD
patients had reported that ChAT is reduced by 50–80% in moderate to
severe AD, the latter study provided yet another example of the
apparent discordance between VAChT and ChAT. The source of the apparent
discordance between these two otherwise tightly linked cholinergic
markers thus remains largely unexplained.
|
The cholinergic reserve strategy |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
ligand,
new VAChT ligands have been routinely screened for site binding. In
some of the earlier studies of radiolabeled vesamicol analogs,
investigators observed that the accumulation of radiolabeled VAChT
ligands within the mammalian striatum was enhanced by pretreatment with
the sigma ligand (and dopamine antagonist) haloperidol. Upon further
investigation, the seemingly paradoxical enhancement was found to be
mediated by blockade of striatal dopamine D2 receptors. Dopaminergic
control of striatal cholinergic neurons is exerted via D2 receptors
found on these neurons. Blockade of these receptors results in
disinhibition of striatal cholinergic activity that manifests as an
increase in the release of ACh (39
, 40)
, probably via the
recruitment of synaptic vesicles from a reserve pool into an
active pool. Since the enhanced accumulation of radiolabeled VAChT
ligands appeared to parallel the facilitated release of ACh in the
striatum, Ingvar et al. (35)
suggested that the binding of
VAChT ligands is sensitive to the level of striatal cholinergic
activity. Further investigation of this link has led to the development
of a pharmacologic activation strategy for in vivo VAChT
imaging that is similar to the ‘stress test’ used routinely in
nuclear cardiology. Within the context of this strategy, striatal
cholinergic activity is presumed to vary between a basal and maximum
level. Between these two extremes exists a dynamic range that reflects
the functional reserve of the cholinergic system. To determine the
magnitude of this reserve, striatal cholinergic neurons are exposed to
a dopamine D2 antagonist to increase their level of activity. The
increased cholinergic activity is then measured as enhanced
accumulation of a radiolabeled VAChT ligand within the striatum. A key
underlying assumption of this method is that the extent of dopamine
antagonist-mediated enhancement in cholinergic activity is proportional
to the cholinergic reserve. While the technique is currently undergoing
extensive investigation in nonhuman primates, preliminary studies in
rodents suggest that the cholinergic reserve as measured by
radiolabeled VAChT ligands declines with aging (27)
and
after destruction of the nigrostriatal dopaminergic pathway
(26)
. Because pharmacological activation may provide
important clues regarding tissue/organ function, the cholinergic
reserve may yet prove to have important correlates in central
cholinergic function. |
DISCORDANCE BETWEEN VAChT and ChAT IN CENTRAL CHOLINERGIC DYSFUNCTION: REAL OR IMAGINED? |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
. 42)
. Reduced cholinergic function is also associated with other
neurological disorders, including Parkinson’s disease with dementia
(43)
, olivopontocerebellar atrophy (OPCA)
(44)
, and amyotrophic lateral sclerosis (45)
.
A putative cholinergic marker would therefore be expected to show the
appropriate changes when examined in these disorders and/or in animal
models thereof. Because ChAT is generally regarded as the gold standard
for cholinergic markers and because of the unique anatomic and
functional arrangement of the cholinergic gene locus, changes in VAChT
expression in disease states have invariably been compared with
corresponding alterations in ChAT. Coordinate expression of VAChT and
ChAT has been observed in the sympathetic superior cervical ganglion
(46)
, in cultured sympathetic neurons (47)
,
and in a murine septal cell line (48)
after treatment with
cytokines or neurotrophic factors. VAChT and ChAT also appear to be
coordinately expressed after experimentally induced injury. Matsuura et
al. (49)
report that after unilateral transsection of the
hypoglossal nerve in adult rats, the levels of vacht and
chat mRNA were dramatically reduced within 1 wk. Thereafter,
the expression of both markers recovered slowly. Gilmor et al.
(50)
report similar observations after either axotomy or
immunotoxin-induced lesions of the rodent septohippocampal pathway.
Finally, maternal exposure to lead also results in a decline in the
expression of both cholinergic markers (51)
. Coordinate
expression of VAChT and ChAT appears to extend to idiopathic
neuropathology as revealed by a recent report. In postmortem human
brain tissue, Gilmor et al. (52)
found that the decrease
in the number of VAChT- and ChAT-immunopositive neurons in the nucleus
basalis highly correlates with the severity of dementia. In an
examination of subjects with no cognitive impairment, mild cognitive
impairment, and early-stage AD, there were no significant reductions in
either marker, suggesting that VAChT and ChAT are preserved in early
AD. The tight functional link between the two markers thus appears to
survive the early stages of AD. If coordinate regulation is the norm at the cholinergic locus, what then is the source of the disparity between ChAT and VAChT in human neurological disorders and in animal models? [We note here that the disparity in question is that observed between ChAT enzyme activity and VAChT density as measured by radioligands.] As indicated previously, earlier workers had attributed the disparity to one or more factors, including compensatory mechanisms and poor radioligand selectivity. Since those studies predate our current knowledge of VAChT and the cholinergic locus, a re-examination of the explanations advanced by these workers appears to be warranted at this time:
1. A substantial fragment of [3H]vesamicol binding
sites may be associated with non-neuronal elements. Indeed,
subsequent studies have shown that vesamicol displays nanomolar
affinity for binding sites. Pharmacological characterization of
[3H]vesamicol in monkey brain clearly reveals
that this ligand displays poor selectivity for VAChT over binding
sites (25)
. As amply demonstrated with
[3H]vesamicol and other VAChT ligands
(30
, 31
, 53)
, the loss of selectivity becomes particularly
evident in brain regions, such as cortex, containing a high density of
binding sites.
2. Cholinergic neurons that survive the lesion may compensate by
expressing higher levels of synaptic vesicles (and concomitantly higher
levels of VAChT). Evidence from rapid autopsy material collected
from patients with Alzheimer’s disease suggests that surviving
cholinergic terminals in the cortex display marked hyperactivity
(54)
, suggesting compensatory up-regulation. Although the
increase in cholinergic activity was inferred from increased
sodium-dependent, high-affinity choline uptake, it is reasonable to
expect that both ChAT and VAChT would also be up-regulated under these
conditions in order to sustain the heightened level of cholinergic
activity. The latter view is consistent with the available evidence,
which strongly suggests that the tight functional link between ChAT and
VAChT is maintained in both normal and pathological conditions.
Consequently, without some evidence of functional decoupling between
ChAT and VAChT, one can find little support for the above claim at this
time.
3. One segment of surviving cholinergic terminals may express
inadequate levels of ChAT and AChE. Previous studies (55
, 56)
have suggested that the extent of reduction of ChAT in the
cerebral cortex exceeds the magnitude of neuronal loss in the nucleus
basalis. However, VAChT levels were not measured in these studies. As a
result, this claim remains largely unverified.
4. Degenerating (ChAT-deficient) cholinergic terminals may
exhibit up-regulation of VAChT. Although the majority of studies
suggest that VAChT and ChAT are coordinately regulated under normal
conditions and in neuropathology, at least one study (57)
has provided evidence of differential regulation of these markers
during development. Consequently, one cannot completely discount the
possibility that such functional decoupling also occurs in certain
phases of the neuropathology.
Given the high affinity of vesamicol for binding sites (and its
marginal selectivity for VAChT), it now appears that the failure of
this ligand to detect reductions in VAChT density is largely due to
competition from binding sites. This view is supported by the
observation that the extent of the mismatch between ChAT activity and
VAChT density varies with both the nature of the VAChT ligand and the
brain region examined. Thus, with more selective VAChT ligands, the
mismatch between ChAT and VAChT appears to be smaller (vide
supra). However, the discrepancy between these reliable
cholinergic two markers has not been completely eliminated. Therefore,
additional investigation is clearly indicated.
Since the discovery of vesamicol, several analogs have been synthesized
and characterized. One of the analogs, 4-aminobenzovesamicol (or ABV),
is the most potent VAChT ligand (Ki = 6.5±0.5 pM)
known (58)
. The compound is significantly more potent
(58)
and selective than vesamicol (23)
.
Because the high affinity of (-)-[3H]vesamicol
for 1 and 2 receptors complicates the interpretation of studies
conducted with this radioligand in tissues other than purified
Torpedo synaptic vesicles, we believe it is now time to
replace (-)-[3H]vesamicol with the more
selective ABV. Presently, 4- (3', 4'-3H) aminobenzovesamicol can be obtained
from American Radiolabeled Chemicals, Inc. (St. Louis, Mo.).
|
THE VESICULAR MONOAMINE TRANSPORTER |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
). Whereas VMAT1 is localized in chromaffin granules,
VMAT2 is found almost exclusively in neurons. The two forms of VMAT are
also distinguished by their sensitivity to inhibitors; thus, reserpine
inhibits the function of VMAT1 and VMAT2 with comparable potency,
whereas tetrabenazine (TBZ) exhibits 10-fold greater potency for VMAT2
over VMAT1. Reserpine and tetrabenazine are the most selective
inhibitors of vesicular monoamine transport. As a result, these two
compounds have provided the starting point for the development of
radioligands for studying VMAT both in vitro and in
vivo. This review focuses on tetrabenazine and its analogs
(Fig. 3
), as little work has been done on reserpine analogs as radiotracers for
in vivo imaging. An earlier review (60)
of
radiotracers for the in vivo imaging of VMAT2 has been
published.
|
Radioligands for studying VMAT
-[2-3H]Dihydrotetrabenazine
([3H]TBZOH)
TBZOH is one of the major metabolites of TBZ. Like the
parent, TBZOH is a potent inhibitor of VMAT2.
[3H]TBZOH was the first radiotracer for VMAT2.
Pharmacological characterization of this radioligand revealed a single
class of high-affinity binding sites on the chromaffin granule membrane
(reviewed in ref 61
). The radioligand displayed low
nonspecific binding. Blockade of vesicular monoamine transport was
highly correlated with occupancy of [3H]TBZOH
binding sites, leading to the conclusion that
[3H]TBZOH binding sites are associated with
VMAT2. Although both substrates and inhibitors of vesicular monoamine
transport can displace [3H]TBZOH from its
binding site, the former are significantly less potent than the latter.
In addition, [3H]TBZOH binding is independent
of the electrochemical membrane gradient. These characteristics have
lent appeal to the use of radiolabeled TBZOH and/or its analogs as
imaging agents.
In the rat, specific binding of [3H]TBZOH
appears to be unaffected by medium-term treatment with a number of
pharmacologic agents that influence dopaminergic neurotransmission
(62)
. This led to the conclusion that reliable measures of
VMAT2 density may be obtained from subjects undergoing long-term
pharmacotherapy with these agents. However, in ovariectomized rats,
treatment with progesterone for 21 days resulted in a reduction in
[3H]TBZOH binding in the striatum and nucleus
accumbens that was matched by a decline in VMAT2 mRNA expression in the
substantia nigra pars compacta and dorsal raphe nuclei
(63)
. Estradiol treatment had similar effects on
[3H]TBZOH binding. However, with estradiol VMAT
mRNA expression was altered only in the nucleus accumbens. VMAT2
density may thus be affected by some pharmacologic treatments.
Pharmacological characterization of [3H]TBZOH
binding (64)
in the human brain postmortem revealed a
single class of high-affinity sites (Kd=7 nM). The
distribution of [3H]TBZOH binding sites was
heterogeneous in the mesencephalon of control subjects and consistent
with the distribution obtained from tyrosine hydroxylase
immunohistochemistry. The highest densities of
[3H]TBZOH binding were observed in the
substantia nigra pars compacta, locus coeruleus, and dorsal raphe
nucleus. Moderate and low densities were found in the ventral tegmental
area and substantia nigra par reticulata, respectively. In Parkinsonian
subjects, [3H]TBZOH binding declined
drastically in all regions examined except in the substantia nigra pars
reticulata, thereby supporting the use of radiolabeled TBZOH for
in vivo imaging.
2-[125/123I]iodovinyldihydrotetrabenazine
Radioiodinated [125I]IV-TBZOH
(65)
was developed for imaging VMAT with SPECT.
[125I]IV-TBZOH binds to a single class of
high-affinity sites in rat striatal membranes. Both TBZ and TBZOH are
potent inhibitors of [125I]IV-TBZOH binding;
however, compounds such as reserpine, dopamine, norepinephrine, and
serotonin were found to be poor inhibitors of
[125I]IV-TBZOH binding. Like TBZOH,
[125I]IV-TBZOH displays low nonspecific
binding. [125I]IV-TBZOH distributes
heterogeneously in rat brain slices, showing high densities in brain
regions that are rich in monoaminergic innervation. Unfortunately, due
to its high degree of lipophilicity,
[125I]IV-TBZOH displays poor accumulation in
the mammalian brain in vivo. The compound is therefore
unsuitable for in vivo imaging.
[11C]Tetrabenazine ([11C]TBZ)
This radiotracer was developed for in vivo imaging of
VMAT with PET. After i.v. injection in mice,
[11C]TBZ accumulates rapidly in the mouse
(66)
. Differential tracer egress from regions of high and
low monoaminergic terminal density later results in a pattern of
distribution that is consistent with central monoaminergic innervation.
Accordingly, the highest levels of [11C]TBZ
were observed in the striatum, whereas moderate to low tracer levels
were observed in the cortex and cerebellum, respectively. Specific
binding of [11C]TBZ was demonstrated by
coadministration of the radiotracer with unlabeled TBZ. Consistent with
previous reports on unlabeled TBZ, analysis of rat blood samples showed
that [11C]TBZ was converted principally to -
and ß-[11C]dihydrotetrabenazine. At 15 min
postinjection, these metabolites accounted for up to 57% of the total
radioactivity in the blood. Imaging experiments in the monkey revealed
a similar pattern of tracer distribution. After MPTP-induced unilateral
lesion of the nigrostriatal pathway, [11C]TBZ
binding in the ipsilateral striatum was abolished; radiotracer
accumulation in the contralateral striatum and cerebral cortex was
unaffected (67)
. Therefore,
[11C]TBZ appears to be a suitable tracer for
studying central nervous system (CNS) disorders characterized by
changes in monoaminergic terminal density. Data analysis may, however,
be complicated by the presence of the metabolite
-[11C]TBZOH, a potent inhibitor of VMAT2.
[11C]Methoxytetrabenazine ([11C]MTBZ)
In preliminary studies in mice, [11C] MTBZ
displayed a distribution profile similar to that described for
[11C]TBZ. The specific binding of
[11C]MTBZ in rat brain sections was blocked by
reserpine, but was unaffected by dopamine, bromocriptine,
N-methylspiperone, haloperidol, nomifensine, L-deprenyl, and
desipramine (68)
. After unilateral 6-OHDA-induced lesion
of the median forebrain bundle, specific binding of
[11C]MTBZ declined by 62% in the ipsilateral
striatum, whereas the contralateral striatum was unaffected. The
unilateral reduction in [11C]MTBZ binding was
highly correlated with the loss of tyrosine hydroxylase-positive cells
in the substantia nigra pars compacta (68)
.
To further validate the use of this radiotracer for studying CNS
disorders associated with changes in monoaminergic terminals, Kilbourn
et al. (69)
evaluated this tracer in the tottering mouse,
a model of generalized epilepsy. Fluorescence histochemistry
measurements of monoamine levels in this model have previously revealed
two- to threefold increases in norepinephrine levels in striatum,
hippocampus, and cerebellar cortex, and a moderate increase of this
neurotransmitter in the occipital cortex, relative to control animals.
Dopamine levels in striatum appear to be unchanged. In all brain
regions examined, [11C]MTBZ concentrations were
increased significantly relative to controls and correlated highly with
independently determined measures of norepinephrine.
[11C]MTBZ binding is thus sensitive to both
reductions and increases in monoaminergic innervation.
The distribution of [11C]MTBZ has also been
evaluated in normal human volunteers (70)
. Rapid tracer
accumulation in the brain is quickly followed by clearance from all
brain regions. Tracer efflux is slowest in regions of high VMAT
density. As a result, regions such as the caudate and putamen are
clearly distinguished from the cerebellum 45 min after radiotracer
injection. Consideration of kinetic parameters obtained from a
2-compartment model also suggests that
[11C]MTBZ is suitable for imaging VMAT in
vivo with PET.
(+)--[11C]Dihydrotetrabenazine
([11C]TBZOH)
Production of [11C]TBZOH, a potent VMAT
inhibitor, tends to complicate the interpretation of data obtained from
PET studies of [11C]TBZ. Development of
[11C]TBZOH was therefore spurred by the need to
simplify kinetic analysis in PET studies of VMAT2. Two modes of
administration have been investigated for
[11C]TBZOH: 1) bolus injection; and
2) infusion to equilibrium between brain and blood
(71
, 72)
. With the infusion method, total and specific
distribution volumes correlated well with the specific distribution of
VMAT2 determined by in vitro methods. With the bolus method,
regional estimates of radiotracer density also correlated highly with
in vitro values. However, in regions of high radiotracer
density, this mode of administration tends to overestimate VMAT2
density by 10–15%. A subsequent study concluded that excellent
estimates of VMAT2 can be obtained by a simple protocol that combines a
loading bolus injection, followed by continuous infusion. Detailed
kinetic analysis of dynamic PET data suggests that with a 3-compartment
model [11C]TBZOH can provide excellent measures
of VMAT2 density in the human brain from a single PET study
(73)
. Other investigators have confirmed that PET imaging
with [11C]TBZOH can provide reproducible
in vivo measurements of VMAT density (74)
.
However, these authors find that the cortex serves as a more reliable
reference region than the cerebellum. Since neither region is
completely devoid of monoaminergic terminals, data analysis must
proceed with caution no matter which of the two regions is used as
reference.
In vivo Imaging of VMAT in neuropathology
PET imaging with [11C]TBZOH has been used
to assess the status of monoaminergic innervation in a number of
neurological/neuropsychiatric disorders.
Immunochemical analysis of VMAT2 protein density in postmortem human
tissue reveals marked reductions in the caudate and putamen in PD
(75)
. The distribution of VMAT parallels that of the
plasma membrane dopamine transporter (DAT) in normal subjects.
Moreover, reductions in VMAT2 density parallel those observed for DAT,
a phenotypic marker of dopaminergic neurons. PET imaging with
[11C]TBZOH confirms the decline in
nigrostriatal dopaminergic innervation (76)
. In this
study, the authors found that the specific binding of
[11C]TBZOH was reduced with advancing age.
Specific binding of the radiotracer declined at the rate of 7.7% per
decade. In PD patients, specific binding of
[11C]TBZOH declined by 61% and 43% in the
putamen and caudate nucleus, respectively. There was no overlap between
age-related decline in [11C]TBZOH binding and
decline resulting from neuropathology. The rate of age-related decline
reported in this study agrees closely with that obtained from human
postmortem tissue with the dopamine reuptake inhibitor
[3H]GBR 12935. Therefore,
[11C]TBZOH appears to be a reliable tracer for
studying CNS disorders associated with changes in monoaminergic
innervation.
The binding and distribution of [11C]TBZOH
have been assessed in severe alcoholism (77)
. In
this study, seven severely alcoholic men without Wernicke-Korsakoff
disease were compared with an equal number of male control males of
similar ages. Both the blood-to-brain transfer rate and the specific
binding of [11C]TBZOH were reduced in the
caudate and putamen; however, the reductions in these two parameters
only reached significance in the putamen. Therefore, PET scanning with
[11C]TBZOH suggests measurable damage to
striatal monoaminergic terminals in severe chronic alcoholism.
In multiple system atrophy (MSA), recent studies with
[11C]TBZOH suggest neurochemical correlates for
different forms of the disease (78
, 79)
. In a study
comparing normal controls with subjects afflicted with MSA, specific
[11C]TBZOH binding declined by 61 and 58% in
the caudate and putamen, respectively. When the same controls were
compared with subjects afflicted with sporadic OPCA, specific
radiotracer binding in the caudate and putamen were reduced by only 26
and 24%, respectively. A subsequent PET study with
[11C]TBZOH (79)
compared 7 normal
control subjects with sporadic OPCA patients, and patients diagnosed
with MSA characterized by 1) predominantly Parkinsonian
features (MSA-P) and 2) principally cerebellar dysfunction
(MSA-C). The authors report a reduction in mean blood-to-brain transfer
rate in the putamen of all three patient groups and in the cerebellar
hemispheres of the MSA-C and sOPCA groups. No change was observed in
the cerebellar hemispheres of the MSA-P group. Moreover, a significant
negative correlation was observed between striatal tracer accumulation
and the severity of Parkinsonism. Similarly, the cerebellar
blood-to-brain transfer rate was negatively correlated with the
intensity of cerebellar dysfunction.
Expression of VMAT2 and DAT is highly correlated in the normal brain.
However, preliminary studies with [11C]TBZOH
suggest that this concordance is lost after chronic cocaine use
(80)
. In a postmortem study of 15 chronic cocaine users
and an equal number of matched controls, the authors found that
striatal DAT binding sites were increased significantly with chronic
cocaine use. Moreover, the increase in DAT was correlated with the
severity of cocaine use. In contrast, VMAT2 density (measured with
[3H]TBZOH) declined slightly (
11%).
Postmortem studies of human cocaine overdose victims have examined
VMAT2 expression with a combination of radioligand
([125I]IV-TBZOH) binding and
immunoautoradiographic techniques (81)
. Quantitative
autoradiography failed to show significant differences in VMAT2 density
between control and cocaine overdose victims. Similar results were
obtained from the immunocytochemical technique. Previous studies in the
rat have shown that VMAT2 density is unaffected by chronic cocaine use
(82)
. Dysregulation of DAT and VMAT2 may therefore be
characteristic of neuropathology. Indeed, such dysregulation may also
underlie the discordance between DAT density and
[11C]TBZOH binding observed in Tourette’s
syndrome (83)
.
Given the apparent sensitivity of [11C]TBZOH to changes in VMAT2 density, it is reasonable to conclude that the latter offers promise as a clinical tool for diagnosis of neurological disorders associated with changes in monoaminergic innervation.
|
SUMMARY |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
1 and 2 receptors.
From these studies, numerous potent and highly selective VAChT ligands
have been developed. Though these appear to show promise in the
detection of cholinergic lesions in vivo, data
interpretation has been complicated by a recurring mismatch between
measurements of VAChT density and ChAT activity. Since VAChT has been
clearly established as a reliable cholinergic marker, the mismatch does
not question the validity of VAChT as a target. The apparent
discordance between the two cholinergic markers does, however, calls
for additional investigations that are aimed at validating newly
developed tracers and their method of use.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPROGRESS IN THE DESIGN,...IN VITRO AND EX...IN VIVO IMAGING OF...The cholinergic reserve strategyDISCORDANCE BETWEEN VAChT and...THE VESICULAR MONOAMINE...SUMMARYREFERENCES |
|---|
