Directory UMM :Data Elmu:jurnal:B:Biological Psichatry:Vol48.Issue8.2000:

Imaging Serotonergic Neurotransmission in Depression:
Hippocampal Pathophysiology May Mirror Global
Brain Alterations
Masahiro Fujita, Dennis S. Charney, and Robert B. Innis
The recent development of [carbonyl-11C]WAY-100635
for serotonin (5-HT)1A and [18F]setoperone and [18F]altanserin for 5-HT2A positron emission tomography receptor imaging has allowed studies of 5-HT neurotransmission in depressive disorders. The hippocampus is likely to
be an important brain structure in the pathophysiology of
depression because it may mediate both cognitive deficits
and hypercortisolemia found in this disorder. Decreased
5-HT1A binding was reported in the medial temporal
cortex, which receives dense 5-HT innervation, and also
throughout neocortical regions. Because the 5-HT1A antagonist pindolol may hasten antidepressant effects of
selective serotonin reuptake inhibitor medications, its
receptor occupancy has been measured in both presynaptic and postsynaptic sites. The results are controversial but
suggest that pindolol has preferential occupancy of somatodendritic autoreceptors in the raphe. The results of
5-HT2A receptors are mixed, with one showing a significant decrease in the right orbitoinsular cortex and three
not detecting a significant change. The disparate findings
in patients with depression almost certainly reflect the
heterogeneity of the disorder, and we highlight the utility
of the hippocampus as a useful target region not only to
compare depressed subjects with healthy subjects but also

to correlate findings with cognitive function and activity of
the limbic– hypothalamic–pituitary axis system. Biol
Psychiatry 2000;48:801– 812 © 2000 Society of Biological Psychiatry
Key Words: Stress, corticosteroids, pindolol, emission
tomography, neurogenesis, drug effect

Introduction

A

bnormalities in several neurotransmission systems,
including dopamine, serotonin (5-hydroxytryptamine; 5-HT), norepinephrine, acetylcholine, and substance

From the Departments of Psychiatry (MF, RBI) and Pharmacology, Yale University
School of Medicine and VA Connecticut, West Haven (RBI), and Program on
Mood and Anxiety Disorders, National Institute of Mental Health, Bethesda,
Maryland (DSC).
Address reprint requests to Masahiro Fujita, M.D., Ph.D., VA Connecticut
Healthcare System/116A2, 950 Campbell Avenue, West Haven CT 06516.
Received March 9, 2000; revised May 30, 2000; accepted June 23, 2000.

© 2000 Society of Biological Psychiatry

P, may be relevant to the pathophysiology of depression.
The 5-HT system has been the most extensively studied, in
part because of the therapeutic efficacy of selective serotonin reuptake inhibitor (SSRI) medications. The hypothesis that depression is caused by low 5-HT transmission
was derived from the findings of largely peripheral measures, such as low plasma L-tryptophan, reduced 5-HT
content and serotonin transporters in blood platelets, and
low CSF levels of the 5-HT metabolite 5-hydroxyindoleacetic acid (5-HIAA; reviewed in Maes and Meltzer
1995). These results have been replicated with sufficient
consistency to justify direct measurements in brain tissue.
With the development of tracers for 5-HT receptors and
transporters, several postmortem studies have been performed in brain tissue from suicide victims, a portion of
which had depressive disorders. As reviewed previously
(Staley et al 1998), the results of these studies vary
significantly. One potential source of this variability was
that both suicidal attempt and depression may be relevant
to the alterations in 5-HT system, and, in some studies,
retrospective case histories of depression were missing or
inaccurate.

Brain imaging studies of the 5-HT system in wellcharacterized living patients will be useful both to assess
abnormalities and to delineate the potential roles of 5-HT
neurotransmission in specific pathophysiologic features of
the disorder (e.g., patients with and without suicidal
features or those with and without cognitive deficits).
Among more than a dozen 5-HT receptor subtypes, tracers
suitable for PET imaging have been developed for 5-HT1A
([carbonyl-11C]WAY-100635 (N-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-N-(2-pyridinyl)cyclohexane carboxamide trihydrochloride)) and 5-HT2A ([18F]setoperone
and [18F]altanserin) receptors. This article will review the
ongoing PET studies of 5-HT1A and 5-HT2A receptors as
well as preliminary single photon emission computed
tomography (SPECT) investigations of serotonin transporter (SERT). Because the hippocampus is likely to play
an important role in the pathophysiology of depression
(including cognitive dysfunction and hypercortisolemia),
these studies will be reviewed with the orientation that this
0006-3223/00/$20.00
PII S0006-3223(00)00960-4

802

BIOL PSYCHIATRY
2000;48:801– 812

brain structure will be a useful target to examine abnormalities and covariance of symptomatology in this
disorder.

Stress, Corticosteroids, and the
Hippocampal 5-HT1A Receptor
Serotonin 1A receptors are densely distributed in the
hippocampus. Increased activity of the hypothalamic–
pituitary–adrenal (HPA) axis has been consistently reported in severe depression, and recent animal studies
suggest that interactions among stress, corticosteroids, and
hippocampal 5-HT1A receptor play a critical role in
depression.

Limbic–Hypothalamic–Pituitary–Adrenal Axis
Function in Depression
Because mood disorders are frequently associated with
high levels of the stress hormone cortisol, abnormalities in
the system regulating its secretion may be pathogenic in

the disorder. The secretion of cortisol is regulated by the
limbic– hypothalamic–pituitary–adrenal (LHPA) axis. The
neurons in the medial parvocellular division of the paraventricular nucleus of the hypothalamus (mpPVN) synthesize corticotropin-releasing hormone (CRH), which stimulates the secretion of adrenocorticotropin hormone
(ACTH) from the anterior pituitary corticotropes; ACTH
activates synthesis and release of glucocorticoids such as
corticosterone and cortisol from the adrenal cortex. The
hippocampus provides negative feedback to this HPA axis.
The removal of the hippocampus reduces but does not
eliminate the efficacy of cortisol inhibition (Sapolsky et al
1990). The hippocampus is distinguished from other
feedback sites, including the hypothalamus and pituitary,
by the high expression of both mineralocorticoid (MR)
and glucocorticoid (GR) receptors (Jacobson and Sapolsky
1991), enabling it to modulate the HPA axis over a wide
range of corticosteroid levels. The increased response of
the LHPA axis to stress after hippocampal damage or
antagonism of hippocampal GR (Feldman and Conforti
1980; Sapolsky et al 1984) suggests these receptors also
contribute significantly to HPA axis regulation.
Decades of research have led to the conclusion that a

subgroup of depressed patients exhibit hypercortisolemia
due to LHPA dysregulation. Evidence supporting this
assertion includes, but is not limited to, dexamethasone
nonsuppression (Carroll et al 1981; Rush et al 1996),
abnormal 24-hour secretion of plasma cortisol (Murphy
1991; Sachar et al 1973), and elevated secretion of 24-hour
urinary free cortisol (Anton 1987; Rothschild et al 1993).
Hypercortisolemia may be associated with more severe,
recurrent illness with specific neuropsychologic impair-

M. Fujita et al

ments (Force 1987; Nemeroff and Evans 1984; Rothschild
et al 1993).
The mechanism underlying the hypercortisolemia of
depression is multifactoral. Possible causes include 1)
increased secretion of CRH from mpPVN, 2) enhanced
pituitary secretion of ACTH, 3) enhanced adrenal secretion of cortisol, 4) decreased sensitivity to negative feedback at the level of the hypothalamus, pituitary, or
hippocampus (Krishnan et al 1991; McAllister-Williams
et al 1998; Nemeroff 1996; Nemeroff et al 1992).

5-HT, Corticosteroids, and Hippocampal
Neurogenesis
Studies in laboratory animals have shown that glucocorticoids, which are released during stress, are associated with
damage to neurons of the hippocampus, (McEwen et al
1992; Sapolsky 1996; Sapolsky et al 1990). Monkeys who
died spontaneously following exposure to severe stress
were found on autopsy to have multiple gastric ulcers,
consistent with exposure to chronic stress, and hyperplastic adrenal cortices, consistent with sustained cortisol
release. These monkeys also had damage to the CA3
subfield of the hippocampus (Uno et al 1989). Studies in
a variety of animal species indicate that stress and increased glucocorticoid exposure results in decreased hippocampal dendritic branching (Watanabe et al 1992;
Woolley et al 1990), alterations in hippocampal synaptic
terminal structure (Magarinos et al 1997), and a loss of
hippocampal neurons (Uno et al 1990).
Contrary to prior beliefs, neurogenesis (i.e., cell division with the production of new neurons) is known to
occur in the adult as well as the developing brain. Gould
and colleagues demonstrated neurogenesis initially in
rodent hippocampus (Cameron et al 1993). Neurogenesis
was subsequently demonstrated in nonhuman primate

(Gould et al 1998) and even human hippocampus (Eriksson et al 1998). Stress has been shown to reduce neurogenesis in the CA3 hippocampal region of adult rodents
and primates (Gould et al 1998). Furthermore, reduction of
corticosteroids by adrenalectomy restored neurogenesis in
the hippocampus of aged rats (Cameron and McKay
1999). The implication of these findings for both depression and stress disorders is that the associated elevation of
cortisol may not only damage existing neurons, but also
decrease the production of new neurons. In addition, a
vicious cycle could be initiated in which cortisol-induced
damage of the hippocampus further reduces the ability of
this structure to inhibit additional release of cortisol.
Both postsynaptic 5-HT1A receptor in the hippocampus
and LHPA axis function have been proposed to play an
important role in the pathophysiology of mood disorders.
Strong evidence suggests that the interactions between

Imaging Serotonergic System in Depression

corticosteroids and the 5-HT1A receptor are robust and
may have clinical relevance (Lopez et al 1998; Meijer and
de Kloet 1998). Several groups have reported that adrenalectomy causes increases in 5-HT1A receptor binding

and gene expression (Chalmers et al 1994; Mendelson and
McEwen 1990). Acute and chronic uncontrollable stress
increased plasma corticosterone and decreased hippocampal 5-HT1A receptor binding and 5-HT1A mRNA levels
(Lopez et al 1998; Lopez et al 1999; Maines et al 1999).
Furthermore, the elimination of the stress-induced increase
in corticosteroids by adrenalectomy prevented the decrease in hippocampal 5-HT1A receptor levels (Lopez et al
1998). These results suggest that the 5-HT1A downregulation observed after chronic stress is mediated, at least in
part, by increases in plasma corticosteroids; however,
there is also a report showing that long-term exposure to
high corticosterone decreased 5-HT induced hyperpolarization but did not decrease 5-HT1A mRNA levels (Karten
et al 1999).
The mechanisms underlying the effect of stress-induced
corticosteroid downregulation of 5-HT1A receptors may
relate to the localization of MR, GR, and 5-HT1A receptors
in the hippocampus. The highest concentration of 5-HT1A
mRNA expression occurs in the granular layer of the
dentate gyrus, which also contains high densities of MR
and GR mRNA. In fact, areas positive for 5-HT1A receptors overlap with 95% of those positive for MR mRNA
(Lopez et al 1998). Although the close association between hippocampal 5-HT1A, GR, and MR mRNA does not
absolutely confirm the colocalization of these receptors in

the same neurons, the data are consistent with the electrophysiologic studies indicating that these receptors are
colocalized in the pyramidal cells of the CA1 subfield
(Joels et al 1991). Corticosteroids probably decrease
5-HT1A receptors via interactions with MR. Under physiologic conditions, the association of corticosteroids with
MR tonically suppresses the responsiveness of postsynaptic 5-HT1A receptors, probably mediated by both receptor
downregulation and suppression of 5-HT1A receptor mediated hyperpolarization of CA1 pyramidal neurons (Meijer et al 1997). Further, Gould and colleagues have
recently presented preliminary data that 5-HT stimulates
the production of new hippocampal granule cell neurons
through 5-HT1A receptors (Jacobs et al 1998; Radley et al
1998) and that chronic antidepressant administration increases neurogenesis in adult rats (Malberg et al 1999).
The animal studies reviewed above suggest that stress
or elevated cortisol associated with depression would
decrease neuronal density and neurogenesis. In fact, MRI
studies of the hippocampus have reported decreased size
in patients with depression and that the decrease was
correlated with the lifetime duration of depressive symptoms (Sheline et al 1999). Furthermore, patients with

BIOL PSYCHIATRY
2000;48:801– 812

803

hippocampal atrophy may tend to be those who are
treatment resistant (Shah et al 1998). Therefore, all of
these studies can be combined into a single hypothesis that
a subgroup of patients with depression would be characterized by hypercortisolemia, reduced hippocampal volume, reduced hippocampal 5-HT1A receptor density, and
treatment resistance. Receptor imaging studies in depression will be reviewed within this hypothetical framework.

The Postsynaptic 5-HT1A Receptor in Depression:
Postmortem Studies
Based on the animal studies and the hypothesis described
above, 5-HT1A receptor density in hippocampus is expected to be decreased in depression, particularly in
patients with high cortisol levels. The results of postmortem studies have not been consistent, however. Two
groups did not find a change in 5-HT1A density in
hippocampus in suicide victims with major depression
using [3H]8-hydroxy-2-[di-n-propyl-amino]tetralin (8OH-DPAT) autoradiography (Lowther et al 1997; Stockmeier et al 1997). One group reported a significantly
decreased hippocampal 5-HT1A mRNA, decreased hippocampal MR mRNA, and a decreased hippocampal
MR/GR ratio in suicide victims with a history of major
depression, changes similar to those found in animals
subjected to chronic unpredictable stress (Lopez et al
1998).
A few groups also studied the density of 5-HT1A
receptors in frontal cortex of suicide victims using [3H]8OH-DPAT. Two did not find a change in receptor density
(Lowther et al 1997; Stockmeier et al 1997), whereas two
found an increased receptor density (Matsubara et al 1991;
Arango et al 1995), one of which found an increase only
in nonviolent suicide (Matsubara et al 1991). Only one of
these studies confirmed an ongoing episode of depression
at the time of suicide, however (Stockmeier et al 1997).
The increased 5-HT1A receptor densities reported by the
two groups may not necessarily contradict the hypothesis
regarding hippocampal pathology in depression. Regulation of 5-HT1A receptors in brain may be regionally
selective. For example, hippocampal 5-HT1A receptors,
because their colocalization with MR and GR (Joels et al
1991), may be more sensitive to circulating corticosteroids, whereas receptors in the prefrontal cortex may be
less responsive to corticosteroids and more responsive to
local 5-HT concentrations.
Results of postmortem studies of suicide victims should
not be equated with the pathophysiology of depression.
Some suicide victims have primary psychiatric diagnoses
other than depression (e.g., schizophrenia and personality
disorder). Furthermore, abnormalities in the 5-HT system
may have a unique pathophysiologic role in suicidality,

804

BIOL PSYCHIATRY
2000;48:801– 812

which is distinct from, but interactive with, that of depression. For example, lower 5-HIAA levels were reported in
more medically damaging suicide attempts (Mann et al
1996). Because of the difficulty of retrospectively determining diagnoses and symptomatology in suicide victims,
PET studies of living patients more easily provide complete characterizations, which may be assessed relative to
in vivo neurochemical measurements.

The Postsynaptic 5-HT1A Receptor in Depression:
PET Studies
Positron emission tomography studies focusing on 5-HT1A
receptor have been hampered by the lack of an appropriate
tracer. A major advance was the development of the highly
selective antagonist WAY-100635 (Fletcher et al 1994).
This tracer was subsequently labeled with [11C] at either
the O-methyl or carbonyl positions. Carbonyl labeling is
preferable, although more difficult, because it does not
produce radiolabeled lipophilic metabolites which enter
the brain and increase nondisplaceable activity (Osman et
al 1998). On the other hand, the O-methyl labeled tracer
does produce a radiolabeled lipophilic metabolite that
enters the brain and binds to 5-HT1A and a1-adrenoreceptors (Osman et al 1996).
One potential problem of [carbonyl-11C]WAY-100635
is the generation of an active metabolite, [carbonyl11
C]desmethyl-WAY-100635, which also binds to 5-HT1A
receptors (Osman et al 1998; Pike et al 1998). However,
the production of the desmethyl metabolite appears to be
negligibly small in human (Osman et al 1998; Pike et al
1998). In fact, the desmethyl compound itself (i.e., [carbonyl-11C]desmethyl-WAY-100635) can be used as a
5-HT1A receptor probe (Pike et al 1998). Although [carbonyl-11C]WAY-100635 seems well suited as a PET
tracer, improved ligands are under development, including
[18F]FCWAY (fluoro-cyclohexyl WAY analog; Lang et al
1999), which showed high specific-to-nondisplaceable
ratio and suitable pharmacokinetic characteristics for PET
studies in nonhuman primates (Carson et al 2000) and has
technical advantages of the longer lived 18F nuclide (T1/2
110 min) compared with 11C (T1/2 20 min). Because of the
short half-life of [carbonyl-11C]WAY-100635, the reliability of the late time point data (even 60 min after
injection) may be poor, particularly the measurement of
the radiotracer’s concentration in plasma. The use of the
longer physical half-life of 18F should ameliorate these
difficulties.
Quantitation of PET receptor imaging often requires
concurrent analysis of the parent tracer in arterial plasma.
Errors in these plasma measurements confound the receptor outcome values and may be particularly problematic
for short-lived isotopes such as 11C. Lammertsma has

M. Fujita et al

developed a reference tissue model for the analysis of
neuroreceptor levels that does not require plasma measurements (Lammertsma and Hume 1996). This method uses a
brain region devoid of receptors (e.g., cerebellum) as a
substitute for arterial plasma levels of the radiotracer. An
extension of this method by Gunn et al (1998) provides an
image in which individual pixel values correlate with
receptor densities. Images of groups (e.g., depressed patients vs. healthy subjects) can then be analyzed by a
pixel-based method, such as statistical parametric mapping
(SPM), to survey all regions and pixels of the brain. Both
the tissue reference model and its extension to pixelwise
measurements were developed in part to analyze 5-HT1A
receptor images and are certainly significant advances in
PET image analysis. Nevertheless, controversy still exists
as to whether these methods can be accurately applied to
[carbonyl-11C]WAY-100635 images because the cerebellum may not reflect nondisplaceable uptake or may show
significant differences between subjects or between
groups. The report from a recent meeting on PET imaging
of the 5-HT1A receptor provides a discussion of these
technical issues (http://www.ki.se/org/way/).
The first 5-HT1A receptor imaging studies in depressed
patients were performed at Hammersmith Hospital, where
the tracer [carbonyl-11C]WAY-100635 was developed
(Sargent et al 1999). To date, two complete studies have
been published using a reference tissue analytic method:
one from Hammersmith Hospital (Sargent et al 2000) and
one from University of Pittsburgh (Drevets et al 1999).
Sargent et al (2000) found moderate reductions (about
10%) of 5-HT1A receptor levels throughout brain including medial temporal cortex (hippocampus and amygdala)
in unmedicated depressed patients. Drevets et al (1999)
found a greater 27% decrease of binding potential in
medial temporal cortex and a 42% decrease in the dorsal
raphe of seven unmedicated depressed patients (Table 1).
The decrease in mesiotemporal cortex was not correlated
with plasma cortisol nor with 24-hour urinary free cortisol
levels. There was a difference in the magnitude of decreases between these two studies, and there was also a
difference in subject population: Drevets et al included
patients with a history of bipolar disorder, whereas Sargent
et al did not. Sargent et al (2000) also studied the effects
of SSRI treatment. The reduced levels of 5-HT1A receptors
were not altered in 10 of the subjects, who were scanned
again during SSRI treatment. Half of these 10 patients
were considered SSRI responders, but the resulting sample
size was probably inadequate to assess potential normalization of receptor levels with remission of depression.
Although the decreased binding was expected from the
hypothesis described above, several confounding factors
in the PET findings must be considered. Postmortem
studies used the agonist [3H]8-OH-DPAT, which predom-

Imaging Serotonergic System in Depression

BIOL PSYCHIATRY
2000;48:801– 812

805

Table 1. Binding at 5-HT Receptors Measured in Positron Emission Tomography
Receptor
1A

Author
Drevets et al 1999
Sargent et al 1999, 2000

2A

Biver et al 1997
Attar-Levy et al 1999
Meltzer et al 1999
Meyer et al 1999b

inantly binds to the high-affinity state of the receptor,
whereas the PET studies used an antagonist [carbonyl11
C]WAY-100635, which binds to both high- and lowaffinity states. The affinity state is regulated by coupling
and uncoupling to a G-protein. The extent of G-protein
coupling in vivo is unknown under normal conditions, and
the level of one G-protein subunit was reported to be
decreased in prefrontal cortex of depressed suicide victims, with the hippocampus not examined (Pacheco et al
1996). Therefore, changes in G-protein coupling to the
5-HT1A receptor may be a source of the discrepancy
between some postmortem studies and these PET studies.
Furthermore, in vivo extracellular levels of 5-HT may
modify radioligand binding by either blocking access of
the tracer to the receptor or modifying its cellular trafficking, which is known to exist for most G-protein coupled
receptors (Bloch et al 1999). One study showed a lower
level of total tissue 5-HT levels in the right hippocampus
of suicide victims with depression (Cheetham et al 1989),
whereas another did not (Owen et al 1986). The influence
of extracellular 5-HT levels on the PET measurement of
[carbonyl-11C]WAY-100635 binding are currently under
investigation and not fully understood (Parsey et al 1999);
however, if depression is associated with low 5-HT
neurotransmission (and low synaptic 5-HT concentrations), then the decreased uptake of [carbonyl-11C]WAY100635 in depressed patients could not be explained by
5-HT blockade of the receptor and, thereby reduced
binding access of the radiotracer. The effects of agonists
and antagonists on cellular trafficking are complex and
may well modify the ability of the tracer to bind to an
internalized and potentially modified receptor. Nevertheless, with these caveats in mind, the first two 5-HT1A
receptor imaging studies in depression both reported the
predicted decrease in receptor binding, but in a global and
not regionally selective manner.

The Presynaptic 5-HT1A Receptor
The presynaptic 5-HT1A receptors in the raphe nuclei
regulate the release of 5-HT through a negative feedback
mechanism. Blier and de Montigny (1994) have hypothe-

Brain region

Findings

Mesiotemporal cortex
Raphe
Cortices including limbic areas
Raphe
Rt. orbitoinsular cortex
Cortices
Cortices including mesiotemporal
Prefrontal cortex

27% decrease
42% decrease
11% decrease
14% decrease
17% decrease
No change
No change
No change

sized that abnormally increased function of the somatodendritic 5-HT1A receptor has an etiologic role in depression and that a common mechanism of action of several
antidepressant treatments is the downregulation of presynaptic 5-HT1A receptor function. That is, the inhibition of
5-HT reuptake by SSRIs does not lead to increased 5-HT
levels in terminal synapses until an adaptive desensitization of inhibitory 5-HT1A autoreceptors has occurred. In a
series of studies, De Montigny, Blier, and colleagues have
demonstrated that the antidepressant mechanism of action
of several classes of antidepressant medications are related
to downregulation of dorsal raphe 5-HT1A receptors,
resulting in a net increase in forebrain 5-HT transmission
(Blier and de Montigny 1998).
Consistent with this hypothesis, the concomitant use of
a 5-HT1A receptor antagonist with antidepressants may
hasten medication response. Recent therapeutic studies
have evaluated whether coadministration of pindolol (an
antagonist at both b-adrenoreceptors and 5-HT1A receptors) and an SSRI will hasten antidepressant drug response. Some studies showed accelerated therapeutic response by concurrent use of pindolol (Bordet et al 1998;
Perez et al 1997; Zanardi et al 1997), but others did not
(Berman et al 1997; Perez et al 1999). This discrepancy
may indicate that concurrent use of pindolol is effective in
only a subgroup of patients, which has not yet been clearly
identified.

Positron Emission Tomography Studies of
Presynaptic 5-HT1A Receptors
As mentioned above, two studies with [carbonylC]WAY-100635 have reported decreased presynaptic
5-HT1A receptor levels in the raphe region of depressed
patients. Sargent et al (2000) reported a 14% decrease;
Drevets et al (1999) reported a 42% decrease (Table 1).
Both of these in vivo results are contrary to the 5% to 30%
increase (the increase was dependent on the rostral-tocaudal level) of postmortem 5-HT1A receptor binding in
dorsal raphe reported by Stockmeier et al (1998). As
described in the section on postsynaptic receptor imaging,
difference of tracers (agonist vs. antagonist) and the
11

806

M. Fujita et al

BIOL PSYCHIATRY
2000;48:801– 812

Table 2. Receptor Occupancy by Pindolol Measured in Positron Emission Tomography
Occupancy
Author
Andree et al 1999
Martinez et al 2000

Rabiner et al 2000

Dose of pindolol (mg)

n

Presynaptic (%)

Postsynaptic (%)

Postsynaptic regions

10
7.5 (4 hours after a dose)
7.5 (10 hours after a dose)
30
5
10
20

3
8
8
8
3
4
3

17
40
38
64
10
37
39

24
18
12
42
5
13
46

Frontal and temporal cortices
Cortices

influence of extracellular 5-HT must be considered as
sources of this discrepancy. In fact, tissue 5-HT levels in
midbrain are more than twice as that in terminal regions
(Dewar et al 1992). Furthermore, in vivo imaging methods
such as PET have intrinsic limitations in their abilities to
measure structures as small as the raphe nuclei. Decreased
5-HT1A receptor binding measured with PET may actually
reflect a decrease in the number of 5-HT neurons and
shrinkage of raphe without change in the density of the
receptors per remaining neuron. In the opposite direction,
a postmortem study reported a 32% significant increase in
neuronal density without change in the volume of dorsal
raphe in suicide victims primarily with major depression
(Underwood et al 1999). If substantiated, this increased
neuronal density may help explain the increased number
of 5-HT1A receptors in the postmortem study of Stockmeier et al (1998).
Positron emission tomography receptor imaging would
seem particularly well suited to measure 5-HT1A occupancy by pindolol and the hypothesis that pindolol has
higher affinity/occupancy of presynaptic (i.e., raphe) than
postsynaptic (i.e., hippocampal) sites. 5-HT1A receptor
occupancy by pindolol in human subjects has been measured by three groups. Following oral administration,
receptor occupancy was dose related. Occupancies in the
presynaptic site were as follows: 17 to 40% for 7.5 to 10
mg and 39 to 64% for 20 to 30 mg; and in postsynaptic
regions 12 to 24% for 7.5 to 10 mg and 42 to 46% for 20
to 30 mg (Table 2). Using [carbonyl-11C]WAY-100635,
two of the three studies reported that pindolol had higher
occupancy of presynaptic than postsynaptic sites (Andree
et al 1999; Martinez et al 2000; Rabiner et al 2000).
Rabiner et al showed a higher occupancy of pindolol at
presynaptic receptors at a dose of 10 mg (37% vs. 13%)
but not at the doses of 5 and 20 mg given orally 2 hours
before the PET study (Table 2; Rabiner et al 2000)
indicating that the appropriate dose range of pindolol is
small. Using a larger sample size, Martinez et al showed
higher occupancy at the presynaptic site by both 7.5 and
30 mg given after 1 week of treatment with 7.5 mg q.d.
The difference in the occupancy in pre- and postsynaptic

Cortices

sites was greater after 7.5 mg. The results by these two
groups may indicate that differential occupancies are
achieved by a relatively low dose.

The 5-HT2A Receptor
5-HT2A Receptor Density
As reviewed previously (Staley et al 1998), more than 10
studies have reported the densities of 5-HT2A receptors in
prefrontal cortex of suicide victims. Many studies used the
antagonist [3H]ketanserin, which has only moderate (2- to
140-fold, most findings are close to the lower end)
selectivity for the 5-HT2A vs. 5-HT2C receptor. About half
of the studies did not detect a significant change; the other
half showed significantly increased binding; and only one
detected a significant decrease. As in other 5-HT markers,
however, the potentially differential effects of suicide and
depression were difficult to discriminate in these postmortem studies.
Preliminary SPECT and PET 5-HT2A receptor imaging
studies were performed in patients with depression using
[2-123I]ketanserin and [11C]-N-methyl spiperone. These
tracers suffer from major limitations of high nonspecific
uptake or no selectivity relative to dopamine D2 receptors.
Recently, several PET studies of depression have been
performed using a new generation of tracers, [18F]setoperone and [18F]altanserin. The former has 100 times greater
affinity for 5-HT2A than 5-HT2C receptors but only 10- to
50-fold higher affinity for 5-HT2 compared with D2
receptors. Even this 10- to 50-fold selectivity may be
inadequate in the striatum, where the density of D2
receptors is so much higher than that of 5-HT2 receptors.
Therefore, [18F]setoperone is useful to study cortex where
the density of D2 receptors is low. On the other hand,
[18F]altanserin has more than 100-fold selectivity relative
to dopamine D2 receptors but only 20-fold selectivity
relative to 5-HT2C receptors (Tan et al 1999). Even this
modest selectivity relative to 5-HT2C receptors may be
adequate for quantitative studies, when combined with the
relatively low density of these sites in neocortex. One

Imaging Serotonergic System in Depression

BIOL PSYCHIATRY
2000;48:801– 812

807

Table 3. The Effects of Antidepressant Treatment on 5-HT2A Receptors Measured in Positron
Emission Tomography
Author
Attar-Levy et al 1999
Yatham et al 1999

Medication
Clomipramine 26 6 6 days
Desipramine 3– 4 weeks

limitation of [18F]altanserin is that it produces radioactive
lipophilic metabolites, which probably cross the blood–
brain barrier. Intersubject variability in peripheral metabolism may introduce different levels of lipophilic metabolites in brain and may affect measures of specific-tonondisplaceable ratio, in which cerebellum is used as a
receptor-poor region. These metabolites have been partially characterized (Baldwin et al 1998). Furthermore, a
deuterated analog of [18F]altanserin has been synthesized
to decrease production of these metabolites (Tan et al
1997, 1999), and a method using bolus plus constant
infusion has been instituted to correct for the presence of
any residual radiolabeled metabolites that have entered the
brain (van Dyck et al, in press).
Using [18F]altanserin, a significant reduction in tracer
uptake was detected in right posterolateral orbitofrontal
and anterior insular cortices in medication-free depressed
patients (Biver et al 1997); however, three studies using
[18F]altanserin (Meltzer et al 1999) and [18F]setoperone
(Attar-Levy et al 1999; Meyer et al 1999b) did not detect
a significant change in medication-free depressed patients
(Table 1). Unfortunately, some of these studies suffer from
major methodologic limitations. By applying large volumes of interest (VOIs), Meyer et al (1999b) and AttarLevy et al (1999) may have overlooked small regions with
significant changes. In fact, postmortem studies indicated
that changes in receptor density often are modest in
magnitude and anatomically restricted to only one or two
Brodmann areas (Arango et al 1997). Therefore, pixelbased analysis such as SPM may be necessary to detect
these changes. Some readers may doubt the results by
Biver et al because they initially analyzed relative changes
in receptor density by applying global normalization and
by using only brain data without correction for plasma
tracer levels; however, they did post hoc VOI analyses for
the region detected by SPM and corrected the data using
the uptake in cerebellum. Therefore, the significant decrease they detected may be true. Nevertheless, lipophilic
metabolites of [18F]altanserin may enter the brain and
contribute activity in nondisplaceable compartment.
Therefore, the measurement using cerebellum as a reference region is not justified if there is a systematic
difference in the metabolism of [18F]altanserin between
depressed patients and healthy subjects.
In summary, the results of 5-HT2A receptor imaging in
depression report either decreased levels or no change, in

Decrease of binding (%)

Brain region

19 (18 –20)
Significant decrease

Cortices
Cortices

apparent contradiction to postmortem studies that have
often found increased 5-HT2A receptor levels. Limitations
in technical aspects of the 5-HT2A receptor PET studies
justify caution in the generalizability of these findings and
support the utility of studies with larger sample sizes to
examine potential correlates of depressive symptomatology and subtypes with the imaging results.

The Effect of Antidepressants on 5-HT2A Receptors
The 4- to 6-week time lag between initiating antidepressant
treatment and clinical response suggest that alterations in the
5-HT system occurring during this period are relevant to the
therapeutic mechanism of the medications. Animal studies
have shown that administration of tricyclic antidepressants
and monoamine oxidase inhibitors cause a downregulation in
the number of 5-HT2 receptors (Cowen 1990; Eison and
Mullins 1996), whereas the results with SSRIs are mixed
(Beasley et al 1992; Cadogan et al 1993; Eison and Mullins
1996). Two PET [18F]setoperone studies were conducted to
investigate changes in 5-HT2A receptors before and after 3 to
4 weeks of desipramine (Yatham et al 1999) or clomipramine
(Attar-Levy et al 1999) treatment in depressed patients. Both
of these studies showed significant decreases in tracer binding in almost all cortical regions (Table 3), which is consistent with prior animal experiments. The PET studies have
significant methodologic limitations, however. The analysis
by Yatham et al (1999) was based on the assumption of no
substantial change in global binding, although the results
showed such global alterations with desipramine treatment.
In addition, because the analysis used global normalization in
SPM, the results did not indicate a decrease in absolute
quantity of binding potential but a decrease relative to the
binding potential in the whole brain. Potential changes in the
clearance of tracer were not taken into account in the
measurement by Attar-Levy et al (1999). Because the affinity
of 5-HT at 5-HT2A receptors is low compared with that at
5-HT1A receptors, potential changes of extracellular 5-HT
levels induced by antidepressants probably did not affect the
measurements. In fact, a single oral administration of 20 mg
paroxetine did not change the binding of [18F]setoperone
(Meyer et al 1999a).

The 5-HT Transporter
Because the 5-HT transporter in platelet has been considered a model of its counterpart in brain, many studies have

808

M. Fujita et al

BIOL PSYCHIATRY
2000;48:801– 812

been conducted to compare [3H]imipramine and [3H]paroxetine binding between depressed patients and healthy
subjects. Although the results were mixed, the majority
(65–75%) of studies have confirmed reductions in platelet
SERT levels in patients with major depression. A metaanalysis has shown this effect to be robustly significant
across studies, including those examining high-affinity
binding sites (Ellis and Salmond 1994). Consistent with
findings of decreased platelet SERT binding, and as
reviewed previously (Staley et al 1998), two postmortem
studies have reported a significantly decreased 5-HT
transporter levels in hypothalamus.
Using [123I]b-CIT and SPECT, the binding potential in
hypothalamus and midbrain was measured in patients.
Medication-free depressed patients showed a significant
decrease of 18% (Malison et al 1998). A significant
decrease was also detected in midbrain in alcoholism, and
the decrease was significantly correlated with depressive
symptoms (Heinz et al 1998). Therefore, decreased SERT
levels in the midbrain may be related to depressive
symptomatology.
A significant limitation of [123I]b-CIT is its lack of
selectivity for serotonin versus dopamine transporters (i.e.,
SERT vs. DAT). Although studies in nonhuman primates
suggest that the majority of activity detected in the
“midbrain” region is associated with SERT (Laruelle et al
1993), a more selective tracer should be used in patients
with depression. In addition, [123I]b-CIT does not show
measurable specific binding in cortical regions. Therefore,
efforts have been focused on the development of improved
PET and SPECT tracers selective for SERT and capable of
measuring the relatively low cortical levels above the
nonspecific uptake of the tracer. The following tracers are
recently developed: [11C] and [18F]McN5652 (Suehiro et
al 1993, 1996), [11C]-labeled cocaine analogs with high
selectivity (Helfenbein et al 1999), [123I] and [76Br]nitroquipazine (Jagust et al 1996; Lundkvist et al 1999),
[123I]IDAM (Kung et al 1999), and [123I]ODAM (Acton et
al 1999). Most of these tracers suffer from some limitations such as slow kinetics, insufficient specific-to-nondisplaceable ratio, or need for additional validation studies.
Among them, [123I]IDAM is perhaps the most promising
at the moment. Nevertheless, none of these tracers have
been shown to provide measurable specific binding in
cortical regions.

Conclusions
The development of suitable PET tracers for 5-HT1A and
5-HT2A receptors has recently allowed studies to further
elucidate probable abnormalities of serotonergic neurotransmission in depression. These studies have consistently shown decreased 5-HT1A receptor binding in pre-

and postsynaptic brain areas in depressed patients, but
much more mixed results in cortical 5-HT2A receptor
levels. The latter finding may well be related to the
heterogeneity of depressive disorders and could reflect
selected symptoms present in subsets of patients (e.g.,
suicidality). Preliminary studies have also reported decreased SERT binding in the brain stem of depressed
patients, but the tracer ([123I]b-CIT) also binds to dopamine transporters and therefore is not selective for SERT.
Although preliminary, these studies are clearly promising
and justify further efforts. For example, study of larger
sample sizes would help determine whether these brain
markers are associated with subtypes of the disorder or
with specific clinical symptoms. New tracers would also
be useful, including probes selective for SERT or entirely
new tracers (e.g., glucocorticoid, CRH or substance P
receptors) of likely relevance to depressive pathophysiology. Finally, because of the documented abnormalities in
the LHPA axis in depression (including hypercortisolemia
and cognitive deficits), the hippocampus is likely to be a
valuable target area for several of these studies. PET
studies of molecular alterations in the hippocampus can be
connected meaningfully to pathophysiologic features of
the disorder. The changes in this region might mirror
global changes and can be used either as a model for or an
important contrast to changes in other brain regions.
This work was supported by funds from the National Institute of Mental
Health (Grant Nos. MH58620 and MH30929) and the Department of
Veterans Affairs (Research Enhancement Award Program in
Depression).
The authors thank Drs. E.A. Rabiner (Hammersmith Hospital) and M.
Laruelle (Columbia University) for helpful discussions.
Aspects of this work were presented at the conference “Depression in
the Twenty-First Century: New Insights into Drug Development and
Neurobiology,” February 21–22, 2000, Dana Point, California. The
conference was sponsored by the Society of Biological Psychiatry
through an unrestricted educational grant provided jointly by Pharmacia
& Upjohn and Janssen Pharmaceutica.

References
Acton PD, Mu M, Plossl K, Hou C, Siciliano M, Zhuang ZP, et
al (1999): Single-photon emission tomography imaging of
serotonin transporters in the nonhuman primate brain with
[123I]ODAM. Eur J Nucl Med 26:1359 –1362.
Andree B, Thorberg SO, Halldin C, Farde L (1999): Pindolol
binding to 5-HT1A receptors in the human brain confirmed
with positron emission tomography. Psychopharmacology
(Berl) 144:303–305.
Anton RF (1987): Urinary free cortisol in psychotic depression.
Biol Psychiatry 22:24 –34.
Arango V, Underwood MD, Gubbi AV, Mann JJ (1995):
Localized alterations in pre- and postsynaptic serotonin binding sites in the ventrolateral prefrontal cortex of suicide
victims. Brain Res 688:121–133.

Imaging Serotonergic System in Depression

Arango V, Underwood MD, Mann JJ (1997): Postmortem
findings in suicide victims. Implications for in vivo imaging
studies. Ann N Y Acad Sci 836:269 –287.
Attar-Levy D, Martinot JL, Blin J, Dao-Castellana MH, Crouzel
C, Mazoyer B, et al (1999): The cortical serotonin2 receptors
studied with positron-emission tomography and [18F]-setoperone during depressive illness and antidepressant treatment with clomipramine. Biol Psychiatry 45:180 –186.
Baldwin RM, Tan P-Z, van Dyck CH, Al-Tikriti M, Amici L,
Roth B, et al (1998): Radiometabolites of [18F]altanserin in
rats and humans: 4-(p-[18F]fluorobenzoyl)piperidine and
[18F]altanserinol. J Labeled Compounds Radiopharm 41:
133–135.
Beasley CM, Masica DN, Potvin JH (1992): Fluoxetine: A
review of receptor and functional effects and their clinical
implications. Psychopharmacology (Berl) 107:1–10.
Berman RM, Darnell AM, Miller HL, Anand A, Charney DS
(1997): Effect of pindolol in hastening responses to fluoxetine
in the treatment of major depression: a double-blind placebocontrolled trial. Am J Psychiatry 154:37– 43.
Biver F, Wikler D, Lotstra F, Damhaut P, Goldman S, Mendlewicz J (1997): Serotonin 5-HT2 receptor imaging in major
depression: Focal changes in orbito-insular cortex. Br J
Psychiatry 171:444 – 448.
Blier P, de Montigny C (1994): Current advances and trends in
the treatment of depression. Trends Pharmacol Sci 15:220 –
226.
Blier P, de Montigny C (1998): Possible serotonergic mechanisms underlying the antidepressant and anti-obsessive-compulsive disorder responses. Biol Psychiatry 44:313–323.
Bloch B, Dumartin B, Bernard V (1999): In vivo regulation of
intraneuronal trafficking of G protein-coupled receptors for
neurotransmitters. Trends Pharmacol Sci 20:315–319.
Bordet R, Thomas P, Dupuis B (1998): Effect of pindolol on
onset of action of paroxetine in the treatment of major
depression: Intermediate analysis of a double-blind, placebocontrolled trial. Reseau de Recherche et d’Experimentation
Psychopharmacologique. Am J Psychiatry 155:1346 –1351.
Cadogan AK, Marsden CA, Tulloch I, Kendall DA (1993):
Evidence that chronic administration of paroxetine or fluoxetine enhances 5-HT2 receptor function in the brain of the
guinea pig. Neuropharmacology 32:249 –256.
Cameron HA, McKay RD (1999): Restoring production of
hippocampal neurons in old age. Nat Neurosci 2:894 – 897.
Cameron HA, Woolley CS, McEwen BS, Gould E (1993):
Differentiation of newly born neurons and glia in the dentate
gyrus of the adult rat. Neuroscience 56:337–344.
Carroll BJ, Feinberg M, Greden JF, Tarika J, Albala AA, Haskett
RF, et al (1981): A specific laboratory test for the diagnosis
of melancholia. Standardization, validation, and clinical utility. Arch Gen Psychiatry 38:15–22.
Carson RE, Lang L, Watabe H, Der MG, Adams HR, Jagoda E,
et al (2000): PET evaluation of [18F]FCWAY, an analog of
the 5-HT1A antagonist WAY 100635. Nucl Med Biol 27:493–
497.
Chalmers DT, Lopez JF, Vazquez DM, Akil H, Watson SJ
(1994): Regulation of hippocampal 5-HT1A receptor gene
expression by dexamethasone. Neuropsychopharmacology
10:215–222.

BIOL PSYCHIATRY
2000;48:801– 812

809

Cheetham SC, Crompton MR, Czudek C, Horton RW, Katona
CL, Reynolds GP (1989): Serotonin concentrations and turnover in brains of depressed suicides. Brain Res 502:332–340.
Cowen PJ (1990): A role for 5-HT in the action of antidepressant
drugs. Pharmacol Ther 46:43–51.
Dewar KM, Grondin L, Carli M, Lima L, Reader TA (1992):
[3H]paroxetine binding and serotonin content of rat cortical
areas, hippocampus, neostriatum, ventral mesencephalic tegmentum, and midbrain raphe nuclei region following pchlorophenylalanine and p-chloroamphetamine treatment.
J Neurochem 58:250 –257.
Drevets WC, Frank E, Price JC, Kupfer DJ, Holt D, Greer PJ, et
al (1999): PET imaging of serotonin 1A receptor binding in
depression. Biol Psychiatry 46:1375–1387.
Eison AS, Mullins UL (1996): Regulation of central 5-HT2A
receptors: A review of in vivo studies. Behav Brain Res
73:177–181.
Ellis PM, Salmond C (1994): Is platelet imipramine binding
reduced in depression? A meta-analysis. Biol Psychiatry
36:292–299.
Eriksson PS, Perfilieva E, Bjork-Eriksson T, Alborn AM, Nordborg C, Peterson DA, et al (1998): Neurogenesis in the adult
human hippocampus. Nat Med 4:1313–1317.
Feldman S, Conforti N (1980): Participation of the dorsal
hippocampus in the glucocorticoid feedback effect on adrenocortical activity. Neuroendocrinology 30:52–55.
Fletcher A, Cliffe IA, Forster EA, Bill DJ, Y R (1994): A
pharmacological profile of WAY-100635, a potent and selective 5-HT1A receptor antagonist. Br J Pharmacol 112:91P.
Force AT (1987): The dexamethasone suppression test: An
overview of its current status in psychiatry. The APA Task
Force on Laboratory Tests in Psychiatry. Am J Psychiatry
144:1253–1262.
Gould E, Tanapat P, McEwen BS, Flugge G, Fuchs E (1998):
Proliferation of granule cell precursors in the dentate gyrus of
adult monkeys is diminished by stress. Proc Natl Acad Sci U
S A 95:3168 –3171.
Gunn RN, Sargent PA, Bench CJ, Rabiner EA, Osman S, Pike
VW, et al (1998): Tracer kinetic modeling of the 5-HT1A
receptor ligand [carbonyl-11C]WAY-100635 for PET. Neuroimage 8:426 – 440.
Heinz A, Ragan P, Jones DW, Hommer D, Williams W, Knable
MB, et al (1998): Reduced central serotonin transporters in
alcoholism. Am J Psychiatry 155:1544 –1549.
Helfenbein J, Sandell J, Halldin C, Chalon S, Emond P, Okubo
Y, et al (1999): PET examination of three potent cocaine
derivatives as specific radioligands for the serotonin transporter. Nucl Med Biol 26:491– 499.
Jacobs BL, Tanapat P, Reeves AJ, Gould E (1998): Serotonin
stimulates the production of new hippocampal granule neurons via the 5HT1A receptor in the adult rat. Soc Neurosci
Abstr 24:1992.
Jacobson L, Sapolsky R (1991): The role of the hippocampus in
feedback regulation of the hypothalamic-pituitary-adrenocortical axis. Endocr Rev 12:118 –134.
Jagust WJ, Eberling JL, Biegon A, Taylor SE, VanBrocklin HF,
Jordan S, et al (1996): Iodine-123-5-iodo-6-nitroquipazine:
SPECT radiotracer to image the serotonin transporter. J Nucl
Med 37:1207–1214.

810

BIOL PSYCHIATRY
2000;48:801– 812

Joels M, Hesen W, de Kloet ER (1991): Mineralocorticoid
hormones suppress serotonin-induced hyperpolarization of rat
hippocampal CA1 neurons. J Neurosci 11:2288 –2294.
Karten YJ, Nair SM, van Essen L, Sibug R, Joels M (1999):
Long-term exposure to high corticosterone levels attenuates
serotonin responses in rat hippocampal CA1 neurons. Proc
Natl Acad Sci U S A 96:13456 –13461.
Krishnan KR, Doraiswamy PM, Lurie SN, Figiel GS, Husain
MM, Boyko OB, et al (1991): Pituitary size in depression.
J Clin Endocrinol Metab 72:256 –259.
Kung MP, Hou C, Oya S, Mu M, Acton PD, Kung HF (1999):
Characterization of [123I]IDAM as a novel single-photon
emission tomography tracer for serotonin transporters. Eur
J Nucl Med 26:844 – 853.
Lammertsma AA, Hume SP (1996): Simplified reference tissue
model for PET receptor studies. Neuroimage 4:153–158.
Lang L, Jagoda E, Schmall B, Vuong BK, Adams HR, Nelson
DL, et al (1999): The development of F-18 labeled, reversible
5-HT1A silent antagonists with polar metabolites. J Med
Chem 42:1576 –1586.
Laruelle M, Baldwin RM, Malison RT, Zea-Ponce Y, Zoghbi SS,
Al-Tikriti MS, et al (1993): SPECT imaging of dopamine and
serotonin transporters with [123I]b-CIT: Pharmacological
characterization of brain uptake in nonhuman primates. Synapse 13:295–309.
Lopez JF, Chalmers DT, Little KY, Watson SJ (1998): Regulation of serotonin1A, glucocorticoid, and mineralocorticoid
receptor in rat and human hippocampus: implications for the
neurobiology of depression. Biol Psychiatry 43:547–573.
Lopez JF, Liberzon I, Vazquez DM, Young EA, Watson SJ
(1999): Serotonin 1A receptor messenger RNA re