Opinion

Serotonin and depression:

pathophysiological mechanism or
marketing myth?
Philip J. Cowen
Neurosciences Building, Warneford Hospital, Old Road, Headington, Oxford OX3 7JX, UK

The notion that impaired serotonin (5-HT) function can

lead to clinical depression has a long history but is still
controversial. Some have argued that the 5-HT hypothesis has been misused by the pharmaceutical industry to
promote a simplistic biological model of depression to
market selective serotonin reuptake inhibitors (SSRIs) to
medical practitioners and the public. By contrast, there is
now substantial evidence that unmedicated depressed
patients have abnormalities in brain 5-HT function; however, the relation of these abnormalities to the clinical
syndrome is unclear. The best evidence that 5-HT contributes to the pathophysiology of depression comes
from studies of tryptophan depletion, which show that
lowering brain 5-HT levels can induce acute symptomatic relapse in recovered depressed patients. Clarification of the mechanism of this effect will enable an
understanding of how impaired 5-HT activity contributes
to the subjective experience of depression.
Serotonin and depression
The serotonin (5-HT) hypothesis of depression is >40 years
old [1]. At its simplest the hypothesis proposes that diminished activity of 5-HT pathways plays a causal role in the
pathophysiology of depression. Part of the evidence for the
original 5-HT hypothesis was the finding that tricyclic
antidepressants inhibited the reuptake of 5-HT as well
as noradrenaline and, thereby, presumably enhanced
brain 5-HT activity in depressed patients. The development of selective serotonin reuptake inhibitors (SSRIs)
further suggested that increased 5-HT function could be
sufficient to ameliorate depressive symptoms in depressed
patients [2].
As well as being moderately useful antidepressants,
SSRIs have become a cultural phenomenon and the source
of a relentless commentary in both the lay and scientific
media, much of it highly critical. For example, the coverage
by the British Medical Journal at one point became so
demented that the journal achieved the near-impossible
feat of having to apologize publicly to Eli Lilly [3]. Indeed,
much of the criticism has been directed at the pharmaceutical industry, which has been accused of overpromoting
SSRIs while at the same time concealing their equivocal
therapeutic efficacy [4] and serious adverse effects, particularly a tendency to aggravate suicidal ideation and behaviour [5].
Corresponding author: Cowen, P.J. (phil.cowen@psych.ox.ac.uk).

Although these important issues are not the focus of the

present article, they are relevant in that current refutations of the 5-HT hypothesis of depression often arise
as part of an attack on the pharmaceutical industry, which
stands accused of using this notion as a marketing ploy.
Thus, Lacasse and Leo [6] argue that the success of directto-consumer advertising by SSRI manufacturers hinges on
the claim that SSRI treatment corrects a 5-HT chemical
imbalance in depression and they suggest that convincing
evidence for this view is nonexistent. David Healy, another
trenchant critic of the pharmaceutical industry, expresses
a similar view of the 5-HT hypothesis, This is an idea that
has never any more evidence to support it than the uric
acid diathesis had [7], thereby giving his early work in this
area rather short shrift [8]. (For those interested in the
history of medicine, the many clinical manifestations of
faulty uric acid diathesis were discussed in the January
edition of the British Medical Journal (1888), Ref. [9]).
Here, I aim to summarize the evidence for 5-HT dysfunction in depression, focusing on the more reliable experimental findings and new developments in methodology. I
also will attempt to describe how changes in brain 5-HT
might be linked to some of the psychological abnormalities
in emotional expression and regulation apparent in acute
depression. I think at this point we all can agree that the
efficacy of 5-HT-promoting agents as clinically useful antidepressants, although an important issue, does not logically provide an argument for low 5-HT activity having a
role in causing depression.
Evidence for 5-HT abnormalities in depression
At the time Alec Coppen proposed the 5-HT hypothesis of
depression [1], direct investigation of the living human
brain was not possible and a variety of indirect biochemical
measures were employed to support the notion that
depressed patients had abnormalities in brain 5-HT function (Table 1). From this early stage it was apparent that
assessing 5-HT abnormalities in depressed patients was a
challenging task for several reasons, particularly the
inconsistency of findings from different laboratories as well
as the problem of taking into account biochemical effects of
current and previous antidepressant treatment. The most
reliable evidence of 5-HT abnormalities from this era in
depressed patients was diminished uptake of 5-HT by
blood platelets [10] and lowered levels in plasma of the
amino acid precursor of 5-HT, tryptophan [11].

0165-6147/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tips.2008.05.004 Available online 26 June 2008

433

Opinion

Trends in Pharmacological Sciences Vol.29 No.9

Table 1. Abnormalities in 5-HT mechanisms in depressed patients

5-HT measure
Platelet 5-HT uptake [10]
Platelet imipramine binding [34]
Platelet 5-HT2A receptor binding [35]
Plasma tryptophan [11]
CSF 5-HIAA [36]
Prolactin response to 5-HT-reuptake blockers [12]
Brain 5-HT1A receptor binding (PET) [15,16]
Brain 5-HT2A receptor binding (PET/SPET) [19]
Brain 5HTT binding (PET/SPET) [20]

Abnormality reported in depressed patients

Diminished
Low
Increased
Low
Low
Diminished
Low
Low, increased and normal
Low, increased and normal

The development of neuroendocrine challenge techniques enabled assessments of brain 5-HT function to be
carried out in unmedicated depressed patients. These
investigations yielded more reliable evidence of abnormalities in brain 5-HT function, albeit in hypothalamic
regions [12]. The most consistent changes were found by
using 5-HT-reuptake inhibitors; acute drug challenge is
reliably associated with diminished prolactin release in
unmedicated depressed patients [12,13]. This suggests
that depressed patients have impaired 5-HT-mediated
neuroendocrine responses. This could be due to diminished
synaptic 5-HT levels or reduced sensitivity of postsynaptic
5-HT receptors in hypothalamic regions. It is not clear,
however, whether this abnormality is specific for depression, as opposed to, for example, anxiety disorders. In
addition, the prolactin response to intravenous administration of the SSRI citalopram remains blunted even when
patients have recovered and withdrawn from drug treatment [13] (Figure 1). This suggests that impaired 5-HTmediated prolactin release might be associated with
vulnerability to depression rather than the acute symptomatic state.
More modern investigations have employed brain-imaging techniques, in particular ligand imaging in conjunction with positron emission tomography (PET) and single
photon emission tomography (SPET) [14]. This enables
more direct investigation of 5-HT receptors in the living
human brain. These data have provided consistent and
convincing evidence that the binding density of 5-H1A
receptors shows a widespread decrease in depressed
patients [15,16]. However, as with the prolactin response

Figure 1. Mean (SEM) prolactin response, measured as area under the curve
(AUC), to citalopram (10 mg intravenously) in healthy controls (labelled as
controls) (n = 16), unmedicated depressed subjects (labelled as acute) (n = 13)
and unmedicated recovered depressed subjects (labelled as recov) (n = 16). Data
are taken, with permission, from Ref. [13].

434

Reliability of finding
High
Modest
Poor
High
Weak, might be stronger in suicidal subgroup
High (also diminished in recovered patients)
High (also diminished in recovered patients)
Poor
Poor

to citalopram, the binding appears to remain low in recovered depressed patients, suggesting that it is not a marker
of the acute depressive state [17]. In addition 5-HT1A-receptor binding also is lowered in patients with panic
disorder [18]; however, these patients usually have high
rates of co-morbid depression.
Other 5-HT receptors, for example, the 5-HT2A receptor
and the 5-HT transporter, also have been studied widely in
imaging investigations in depressed patients. However,
the data from this work are contradictory [19,20], showing
that even the use of state-of-the-art imaging to measure 5HT receptors in unmedicated patients, classified with
reliable diagnostic systems, is insufficient to overcome
clinical heterogeneity or technical variability between
different laboratories. In schizophrenia the ability to use
PET and SPET to measure in vivo dopamine release has
produced important advances in understanding pathophysiology [21]. It has been difficult to apply the same methodology to measure 5-HT release in vivo, but a recent study
shows that it might be possible to use 5-HT1A ligand PET to
measure the release of 5-HT onto 5-HT1A autoreceptors in
the raphe nuclei [22]. This could be an important methodological development.
The data reviewed so far show unequivocally that unmedicated depressed patients have abnormalities in aspects
of brain 5-HT activity. Although this is a useful start, there
are many important caveats before we can link these
changes with depression. For example, with a few exceptions, we do not know how specific any of these changes are
for depression as opposed to other psychiatric disorders, for
example, anxiety states and schizophrenia. Also many of
the reported abnormalities apparently are present even
when patients are recovered from depression. Although the
latter observation suggests that the changes are not
merely epiphenomena of the acute depressed state, it does
raise questions about how these 5-HT abnormalities are
involved in the experience of being depressed. They might
be markers of vulnerability, indicating that people with
abnormal brain 5-HT function are more likely to become
depressed, perhaps when exposed to psychosocial stress. It
also could be, however, that these changes are scars of
previous acute depressive episodes and their treatment.
Studies in high-risk groups before the onset of clinical
illness will be needed to resolve this issue.
An interesting genetic vulnerability factor is allelic
variation in the promoter region of the gene encoding
the 5-HT transporter (5HTT), whereby carriers of the short
allele of 5HTT (the low-activity allele) have a greater risk
of developing depression in response to stressful life events
[23]. Intriguingly, subjects with a short allele of 5HTT also

Opinion
show increased neural responses to fearful facial expression in functional imaging studies [24], suggesting that the
short allele might facilitate processing of negative
emotional information, perhaps thereby worsening the
psychological impact of adverse life experiences. However,
this account has been criticized as implausible because of
the small effect size of polymorphic variation on the phenotypic expression of complex traits [25]. In addition
carriers of the short allele of the 5HTT do not seem much
more likely than carriers of the long allele to experience
depressive illness, which would be expected if they were
more susceptible to stress [26].
Tryptophan depletion, 5-HT and depression
Another approach to elucidating the role of 5-HT in depression is to assess the effect of lowering brain 5-HT function
and examining its clinical and neuropsychological consequences. The technique employed, tryptophan depletion
(TRD), takes advantage of the fact that is possible to lower
plasma and brain tryptophan by administering an amino
acid mixture free of the 5-HT precursor, tryptophan.
Because the conversion of tryptophan to 5-hydroxytryptophan by tryptophan hydroxylase is the rate-limiting step in
5-HT synthesis, TRD is able to produce a transient lowering of brain 5-HT activity [27].
In healthy volunteers with no vulnerability factors for
the development of depression, TRD does not produce
consistent changes in mood [28]. This suggests that lowering brain 5-HT is not in itself sufficient to cause clinical
depressive symptomatology. However, in recovered
depressed patients withdrawn from medication, TRD
causes, in a high proportion of participants, a transient
but striking return of depressive symptomatology [28,29].
This suggests that, in those with risk factors for depression, lowering brain 5-HT function can, indeed, cause
clinical depressive symptomatology. Interestingly the
same is found when catecholamines are depleted with amethyl-para-tyrosine [30], showing that recovered depressed patients are psychologically vulnerable to both 5-HT
and catecholamine deficiency.
Why should recovered depressed patients show a
depressive response to TRD when healthy controls do
not? It is worth noting at this point that people at high
genetic risk of depression who have not suffered a personal
episode of illness do not experience clinical symptomatology after TRD, although they might be more likely than
controls to experience transient dysphoria, measurable on
adjective check lists [28]. This suggests that most of the
overt clinical relapse experienced by recovered depressed
patients after TRD is related to the experience of having
been depressed before. This could be a consequence of
previous treatment, for example, with SSRIs, but in our
own TRD study [29] we also saw clinical relapses after TRD
in patients who had recovered spontaneously or after
psychotherapy. Therefore, it seems more likely that
depression itself might cause persistent neurobiological
changes that produce a depressogenic response when
affected individuals are exposed to low brain 5-HT activity.
What could these neurobiological changes be? We
already have noted that recovered depressed patients have
evidence of abnormal brain 5-HT activity, so perhaps the

Trends in Pharmacological Sciences

Vol.29 No.9

unusual psychological effect of TRD in such individuals is

due to an acute reduction in 5-HT availability interacting
with an already compromised 5-HT system. Another
possibility is that recovered depressed patients have
abnormalities in emotional processing circuitry, which
are exaggerated in a low brain 5-HT environment. We
have some evidence for the latter possibility in that
TRD, given in a dose insufficient to produce subjective
mood change, caused negative biases in the recognition
of facial expressions in recovered depressed patients but
not in healthy controls [31].
How could impaired 5-HT function cause the subjective
experience of depression?
The observation that lowering 5-HT levels can produce
changes in emotional perception suggests a way of conceptualizing how low 5-HT might cause vulnerability to
depression and depression itself in certain circumstances.
Generally it seems unlikely that 5-HT pathways influence
mood directly because drugs like SSRIs do not produce
mood elevation in healthy volunteers. In fact studies summarized by Harmer on this issue suggest that in depressed
patients 5-HT acts on mood indirectly by changing biases
in the processing of emotional information [32]. From this
we might expect that the persistent abnormalities in 5-HT
activity associated with the recovered state could be linked
to negative biases in emotional processing, which predispose individuals to experience depression [33]. Why in this
situation further lowering of 5-HT activity by TRD should
cause acute subjective symptomatic relapse is not clear.
However, it might be that further exaggeration of negative
biases in this low 5-HT state leads to individuals directly
accessing previously experienced negative memories and
beliefs.
Conclusions
Patients at risk of depression from having experienced
previous episodes have persistent abnormalities in brain
5-HT mechanisms. 5-HT pathways provide a prominent
innervation to emotional processing circuitry, including
the amygdala and prefrontal cortex. It seems plausible
that impaired 5-HT innervation of this circuitry is associated with negative biases in the processing of emotional
information, thereby placing recovered patients at
increased risk of clinical mood disorder particularly in
adversity. Lowering brain 5-HT activity through TRD in
recovered patients produces acute symptomatic relapse
possibly by bringing into conscious awareness previously
established associative networks of negative appraisals
and memories.
Future work will need to establish whether abnormalities in 5-HT activity in recovered depressed patients are
present in those at risk of illness but who have not suffered
personal episodes of depression. It also will be important to
understand in more detail the neuropsychological mechanisms by which TRD causes relapse in recovered depressed patients. In this way it should be possible not only to
establish that 5-HT dysfunction is linked to depression but
also to understand the way in which low 5-HT function can
lead to the subjective experiences characteristic of clinical
illness.
435

Opinion
Conflict-of-interest statement
The author has acted as a paid advisor to the following
companies involved in the development of antidepressant
drugs: Eli Lilly, GlaxoSmithKline, Lundbeck, Servier and
Wyeth.
Acknowledgements
The studies of the author were supported by the Medical Research
Council.

References
1 Coppen, A. (1967) The biochemistry of affective disorders. Br. J.
Psychiatry 113, 12371264
2 Cowen, P.J. (1990) A role for 5-HT in the action of antidepressant
drugs. Pharmacol. Ther. 46, 4351
3 (2005) Eli Lilly: correction and apology. BMJ 330, 211
4 Kirsch, I. et al. (2008) Initial severity and antidepressant benefits: a
meta-analysis of data submitted to the food and drug administration.
PLoS Med. 5, e45 DOI: 10.1371/journal.pmed.0050045 (http://
medicine.plosjournals.org)
5 Whittington, C.J. et al. (2004) Selective serotonin reuptake inhibitors
in childhood depression: systematic review of published versus
unpublished data. Lancet 363, 13411345
6 Lacasse, J.R. and Leo, J. (2005) Serotonin and depression: A disconnect
between the advertisements and the scientific literature. PLoS
Med. 2, e392 DOI: 10.1371/journal.pmed.0020392 (http://medicine.
plosjournals.org)
7 Healy, D. (2002) The Creation of Psychopharmacology, Harvard
University Press
8 Healy, D. et al. (1985) Peripheral adrenoceptors and serotonin
receptors in depression: changes associated with response to
treatment with trazodone or amitriptyline. J. Affect. Disord. 9, 285296
9 Yeo, I.B. (1888) An address on the therapeutics of the uric acid
diathesis. BMJ 1, 1619
10 Coppen, A. et al. (1978) Platelet 5 hydroxytryptamine accumulation in
depressive illness. Clin. Chim. Acta 87, 165168
11 Cowen, P.J. et al. (1989) Decreased plasma tryptophan levels in major
depression. J. Affect. Disord. 16, 2731
12 Cowen, P.J. (2001) Neuroendocrine markers of depression and
antidepressant drug action. In Antidepressants (Leonard, B.E. et al.,
eds), pp. 95107, Birkhauser Verlag
13 Bhagwagar, Z. et al. (2002) State and trait abnormalities in serotonin
in major depression. Br. J. Psychiatry 180, 2428
14 Grasby, P.M. (2002) Imaging the neurochemical brain in health and
disease. Clin. Med. 2, 6773
15 Sargent, P.A. et al. (2000) Brain serotonin1A receptor binding
measured by positron emission tomography with [11C]WAY-100635:
effects of depression and antidepressant treatment. Arch. Gen.
Psychiatry 57, 174180
16 Drevets, W.C. et al. (2007) Serotonin-1A receptor imaging in recurrent
depression: replication and literature review. Nucl. Med. Biol. 34, 865
877
17 Bhagwagar, Z. et al. (2004) Persistent reduction in brain serotonin1A
receptor binding in recovered depressed men measured by positron

436

Trends in Pharmacological Sciences Vol.29 No.9

18
19

20

21

22

23
24

25
26

27

28

29
30

31

32

33

34

35
36

emission tomography with [11C]WAY-100635. Mol. Psychiatry 9, 386

392
Neumeister, A. et al. (2004) Reduced serotonin type 1A receptor in panic
disorder. J. Neurosci. 24, 589591
Bhagwagar, Z. et al. (2006) Increased 5-HT2A receptor binding in
euthymic, medication-free recovered depressed patients: a positron
emission study with [11C]MDL 100, 907. Am. J. Psychiatry 163,
15801587
Bhagwagar, Z. et al. (2007) 5-HTT binding in recovered depressed
patients and healthy volunteers: a positron emission tomography
study with [11C] DASB. Am. J. Psychiatry 164, 18581865
Laruelle, M. and Abi-Dargham, A. (1999) Dopamine as the wind of
the psychotic fire: new evidence from brain imaging studies.
J. Psychopharmacol. 13, 358371
Sibon, I. et al. (2008) Decreased [18F]MPPF binding potential in the
dorsal raphe nucleus after a single oral dose of fluoxetine: a positronemission tomography study in healthy volunteers. Biol. Psychiatry 63,
11351140 (www.sciencedirect.com)
Caspi, A. et al. (2003) Influence of life stress on depression: moderation
by a polymorphism in the 5-HTT gene. Science 301, 386389
Hariri, A.R. et al. (2005) A susceptibility gene for affective disorders
and the response of the human amygdala. Arch. Gen. Psychiatry 62,
146152
Flint, J. and Muafo, M.R. (2007) The endophenotype concept in
psychiatric genetics. Psychol. Med. 37, 163180
Willis-Owen, S.A. et al. (2005) The serotonin transporter length
polymorphism, neuroticism, and depression: a comprehensive
assessment of association. Biol. Psychiatry 58, 451456
Young, S.N. et al. (1985) Tryptophan depletion causes a rapid
lowering of mood in normal males. Psychopharmacology (Berl.)
87, 173177
Ruhe, H.G. et al. (2007) Mood is indirectly related to serotonin,
norepinephrine and dopamine levels in humans: a meta-analysis of
monoamine depletion studies. Mol. Psychiatry 12, 331359
Smith, K.A. et al. (1997) Relapse of depression after rapid depletion of
tryptophan. Lancet 349, 915919
Berman, R.M. et al. (1999) Transient depressive relapse induced by
catecholamine depletion: potential phenotypic vulnerability marker?
Arch. Gen. Psychiatry 56, 395403
Hayward, G. et al. (2005) Low-dose tryptophan depletion in recovered
depressed patients induces changes in cognitive processing without
depressive symptoms. Biol. Psychiatry 57, 517524
Harmer, C.J. et al. (2004) Increased positive versus negative affective
perception and memory in healthy volunteers following selective
serotonin and norepinephrine reuptake inhibition. Am. J. Psychiatry
161, 12561263
Bhagwagar, Z. et al. (2004) Normalization of enhanced fear recognition
by acute SSRI treatment in subjects with a previous history of
depression. Am. J. Psychiatry 161, 166168
Mossner, R. et al. (2007) Consensus paper of the WFSBP task force on
biological markers: biological markers in depression. World J. Biol.
Psychiatry 8, 141174
Mendelson, S.D. (2000) The current status of the platelet 5-HT(2A)
receptor in depression. J. Affect. Disord. 57, 1324
Reddy, P.L. (1992) CSF amine metabolites in depression. Biol.
Psychiatry 31, 112118