Bipolar Disorders 2000: 2: 7792Printed in Ireland. All rights reser6ed
I-Shin Shiaha,b and LakshmiN Yathamaa Division of Mood Disorders, Departmentof Psychiatry, University of BritishColumbia, Vancouver, British Columbia,Canada, b Department of Psychiatry,Tri-Service General Hospital, NationalDefense Medical Center, Taipei, Taiwan,ROC
Key words: anticonvulsants bipolardisorder 5-HT lithium mania mood stabilizers serotonin
Received 7 May 1999, revised andaccepted for publication 9 July 1999
Corresponding author: Lakshmi N Yatham,MBBS, FRCPC, MRC Psych (UK), Associ-ate Professor in Psychiatry and Director,Mood Disorders Clinical Research Unit, De-partment of Psychiatry, The University ofBritish Columbia, 2255 Wesbrook Mall,Vancouver, BC, V6T 2A1, Canada. Fax:1 604 8227922; e-mail:firstname.lastname@example.org
Serotonin in mania and in the mechanism ofaction of mood stabilizers: a review ofclinical studies
Shiah I-S, Yatham LN. Serotonin in mania and in the mechanism ofaction of mood stabilizers: a review of clinical studies.Bipolar Disord 2000: 2: 7792. Munksgaard, 2000
Objecti7es: Serotonin (5-hydroxytryptamine, 5-HT) was implicated inthe pathophysiology of manic-depressive illness as early as 1958. Al-though extensive evidence has accumulated since then to support 5-HTs role in depression, relatively fewer studies examined its role inmania. The purpose of this paper was to review and summarize thecurrent state of knowledge on the role of 5-HT in mania and its treat-ment.
Methods: We systemically reviewed clinical studies of 1) 5-HT functionin mania and 2) 5-HT in the mechanism of action of mood stabilizers,including lithium and anticonvulsants.
Results: Review showed that cerebrospinal fluid, postmortem, platelet,neuroendocrine challenge, and tryptophan depletion studies providedsome evidence to support the hypothesis that a 5-HT deficit is involvedin mania and that enhancement of 5-HT neurotransmission exerts amood-stabilizing effect.
Conclusions: There is some evidence from clinical studies for the contri-bution of 5-HT in mania and in the mechanism of action of moodstabilizers. However, it is very likely that other neurotransmitters alsoplay important roles. Future directions for research include 1) in 6i6ostudy of 5-HT receptor subtypes using positron emission tomography,2) investigation of the interaction between 5-HT and other neurotrans-mitter systems, and 3) determination of the relationships between diag-nostic subtypes of mania and 5-HT function and otherneurotransmitter systems.
Serotonin (5-hydroxytryptamine, 5-HT) is an im-portant neurotransmitter within the central ner-vous system (CNS). The amount of 5-HT in theCNS is about 12% of that in the body (1). Be-cause this indoleamine transmitter cannot cross thebloodbrain barrier, all the neuronal 5-HT is syn-thesized locally (1, 2). The synthesis of 5-HT in thebrain involves 1) uptake of the essential amino acidL-tryptophan, which is a primary substrate for thesynthesis, from plasma into serotonergic neurons,2) hydroxylation of L-tryptophan to 5-hydroxy-tryptophan (5-HTP) via the enzyme tryptophanhydroxylase, and 3) decarboxylation of 5-HTP to5-HT through the enzyme aromatic amino acid
decarboxylase (2). It appears that 5-HT can besynthesized in both cell bodies and nerve terminals(2). 5-HT synthesized in the cell body is trans-ported into the terminals of dendrites and axonsfor release. Following its release, 5-HT is eitherinactivated by re-uptake into serotonergic nerveterminals, or interacts with a complex system ofreceptors. Once back inside the serotonergic neu-ron, this neurotransmitter is either re-stored in thevesicles or metabolized by the enzyme monoamineoxidase (MAO) and then converted into an inac-tive metabolite, 5-hydroxyindoleacetic acid (5-HIAA). To date, based on pharmacological ormolecular properties, at least 14 types of 5-HT
Shiah and Yatham
receptors have been identified in the mammalianbrain (3, 4). They are 5-HT1A, 5-HT1B, 5-HT1D,5-HT1E, 5-HT1F; 5-HT2A, 5-HT2B, 5-HT2C; 5-HT3,5-HT4; 5-HT5A, 5-HT5B, 5-HT6, and 5-HT7 recep-tors. Except for the 5-HT3 receptors, which areligand-gated ion channels, all 5-HT receptor sub-types belong to the superfamily of G-protein-cou-pled receptors (1, 5).
With regard to neuronal innervation, ascendingserotonergic fibers arise principally from two mid-brain nuclei, the dorsal raphe and the medianraphe, and project to virtually all forebrain re-gions, including thalamus, hypothalamus, caudate-putamen, hippocampus, and neocortex (6, 7). Thewidespread distribution of the serotonergic projec-tions may partially explain why alterations in 5-HTfunction in the brain would lead to modification ofa variety of physiological and behavioral re-sponses, such as sleep and wakefulness, food in-take, sexual activity, impulse control, learning andmemory (8, 9).
Given that 5-HT is involved in a variety ofphysiological and behavioral responses, it is notsurprising that disturbance in 5-HT function hasbeen implicated in the pathophysiology of manypsychiatric disorders, including mood disorders.5-HT was implicated in the pathophysiology ofmanic-depressive illness by Strom-Oslen and Weil-Malherbe as early as 1958 (10). However, Prangeet al. (11) were the first to formulate the permissivehypothesis of 5-HT function for bipolar disorder.They postulated that a deficit in central serotoner-gic neurotransmission permits the expression ofbipolar disorder, and that both the manic and thedepressive phases of bipolar disorder are character-ized by low central 5-HT function, but differ inhigh versus low central catecholaminergic (i.e.,norepinephrine (NE) and dopamine (DA)) neuro-transmission. This hypothesis was primarily basedon 1) the clinical observation that L-tryptophan,the precursor of 5-HT, was efficacious in the treat-ment of mania (11), 2) some earlier studies reveal-ing decreased levels of the 5-HT metabolite,5-HIAA, in cerebrospinal fluid (CSF) in both ma-nia and depression (1214), and 3) the animalstudies showing that lithium enhanced brain 5-HTfunction (1517). Although extensive evidence hasaccumulated since then to support a role for 5-HTin depression [(1822) for review], relatively fewerstudies examined its role in mania. This undoubt-edly relates to the difficulty in recruiting drug-freemanic patients for research studies.
The purpose of this paper was to review andsummarize the current state of knowledge on therole of 5-HT in mania and its treatment. We sys-temically reviewed clinical studies of 1) 5-HT func-
tion in mania and 2) 5-HT in the mechanism ofaction of mood stabilizers, including lithium andanticonvulsants. We also discussed the limitationsof these studies and indicated future directions forresearch.
5-HT function in mania
Several approaches have been used to study 5-HTactivity in mania: 1) CSF studies, 2) postmortemstudies, 3) platelet studies, and 4) neuroendocrinechallenge studies.
Cerebrospinal fluid studies
CSF studies measured the levels of 5-HIAA, themajor metabolite of 5-HT, with the assumptionthat 5-HIAA levels provide an index of central5-HT activity (23, 24). Although half the concen-tration of lumbar 5-HIAA comes from the spinalcord (25, 26), the 5-HIAA levels in lumbar CSFmay still reflect brain functioning, since many 5-HT axons and nerve terminals in cord originate infunctionally important brain stem nuclei. Indeed ithas been shown that lumbar CSF 5-HIAA concen-trations do correlate with brain 5-HIAA in hu-mans (27).
Studies of CSF 5-HIAA in manic patients haveproduced variable and inconsistent results. Forexample, baseline CSF 5-HIAA levels in manicpatients, compared to non-depressed controls,have been reported as decreased in four studies(1214, 28), unchanged in nine studies (2937),and increased in three studies (3840). However,when baseline CSF 5-HIAA levels were comparedbetween manic and depressed patients, the resultsof previous studies were much more consistent.With the exception of one study (29), all otherstudies found that CSF 5-HIAA levels in manicswere not different from depressed patients (13, 14,28, 3033, 3638, 40, 41).
Administration of probenecid blocks the activetransport of 5-HIAA out of the CSF, leading toaccumulation of 5-HIAA levels in CSF (42, 43).This provides a more dynamic measure of centralserotonergic activity than the basal level of CSF5-HIAA (4446). Of the four studies that exam-ined CSF 5-HIAA accumulation following admin-istration of probenecid in manics, depressives, andcontrols (30, 32, 38, 47), two reported that bothmanic and depressed patients have diminished CSF5-HIAA formation compared to controls (38, 47),and one reported that manic patients have signifi-cantly lower CSF 5-HIAA accumulation than de-pressives and controls (30). Furthermore, Goodwinet al. (32) found that manic and depressed patients
Serotonin, mania, and mood stabilizers
have similar CSF 5-HIAA levels 18 h after admin-istration of probenecid. Although there were nocontrol group data available for their 18-h study,comparison with values obtained in the controlsover a 9-h period suggested that the rate of accu-mulation of 5-HIAA in both the manic and de-pressed patients may have been reduced (46).Taken together, the results of the probenecid-in-duced accumulation studies appeared to supportthat both mania and depression are associated witha reduction in central 5-HT function.
The only study that assessed 5-HT and 5-HIAAlevels in postmortem brains of subjects with awell-documented history of bipolar disorder re-ported significantly lower 5-HIAA levels and lower5-HT:5-HIAA ratios in frontal and parietal cortex,compared to postmortem brains from controls(48). The results of this study, therefore, provide asupport for regional decreases in central 5-HTactivity.
Platelets share many properties with central 5-HTneurons, such as uptake, storage, and release of5-HT, 5-HT receptors and imipramine-bindingsites (49, 50). They are easily accessible for study,and therefore have been used as a peripheral modelfor 5-HT neurons to study 5-HT function in psy-chiatric patients.
Platelet 5-HT uptake. Meltzer et al. (51) examined5-HT uptake in 14 manic patients in comparison to20 healthy controls. They noted a tendency for adecrease in 5-HT uptake in four manic patientsand an increase in seven manic patients, but manicpatients as a group did not differ significantly fromhealthy controls. Similarly, Scott et al. (52) re-ported no difference in 5-HT uptake in eight manicpatients compared to 26 healthy controls. Meagheret al. (53) reported increased 5-HT uptake in 15manic patients compared to 19 healthy controls.However, manic patients as a group had a largevariation in their 5-HT uptake compared to thecontrol group in this study that could very well bedue to the effects of medication. Indeed, when fivedrug-free manic patients in this study were com-pared with controls, the 5-HT uptake was notdifferent between the two groups. Marazziti et al.(54), on the other hand, reported decreased 5-HTuptake in seven manics compared to 12 healthycontrols. Of the seven manics, only three weredrug-free, which confounds the interpretation ofresults.
Platelet imipramine-binding site. The imipramine-binding site that is present in both 5-HT neuronsand platelets appears to be closely linked to the5-HT active transport site. The binding affinity ofthis site to imipramine in platelets has been exam-ined to gain information about 5-HT neuron func-tion in manic patients. Of the four studies thatexamined imipramine binding in manic patients,two showed an increase (55, 56), while the othertwo showed no difference (57, 58) in binding inmanics compared to depressed patients. However,all four studies reported that binding was not dif-ferent in manics when compared to healthycontrols.
Taken together, the results of platelet studiesthat assessed 5-HT uptake or tritiated imipraminebinding did not show any consistent differencesbetween manic patients and controls. The dis-crepancy in findings between studies could be dueto methodological differences or problems, such aspatient heterogeneity, confounding effects of psy-chotropic medication, and sample sizes.
Other platelet 5-HT measures. In addition toplatelet 5-HT uptake and [3H]-imipramine binding,5-HT-induced platelet calcium mobilization, a pu-tative measure of 5-HT2 receptor sensitivity (59,60), has also been examined in patients with ma-nia. Okamoto et al. (61) reported that there is anincrease in the rapid peak and prolonged plateauphase in platelet intracellular Ca2 mobilization inten untreated acute manic patients compared to 14matched healthy subjects and ten euthymic bipolarpatients treated mainly with lithium. As for therapid peak, a similar result was also obtained inplatelets of patients with bipolar depression (62).This may suggest an increase in the sensitivity of5-HT2 receptors in both mania and bipolar depres-sion. In support of this hypothesis, a recent ex 6i6ostudy (63) showed that administration of 5-HT ledto a significant elevation in platelet intracellularcalcium levels in patients in both the manic anddepressive phases of bipolar disorder, compared tonormal controls and bipolar euthymic patients. Incontrast, a recent study, using [125I]-ketanserin asthe radioligand, reported that platelet 5-HT2 recep-tor-binding sites in 29 drug-free male manic pa-tients were not different from those in 29 malehealthy controls (64). Furthermore, another probefor assessing 5-HT2 receptor sensitivity, theadenosine diphosphate (ADP)-augmented 5-HT-induced platelet aggregation response, was re-ported to be significantly lower in both manic anddepressed patients compared to normal controls,and it was reversed following clinical recovery (65).The discrepancy between the studies that examined
Shiah and Yatham
platelet 5-HT2 receptor-mediated activity couldbe due to the differences in patient populationsand:or to the nature of the means of measuringplatelet 5-HT2 receptor activity.
Neuroendocrine challenge studies
The neuroendocrine challenge paradigm offers auseful means of assessing central (hypothalamic)5-HT function. This approach is based on theobservation that 5-HT exhibits an excitatory in-fluence on the release of cortisol, adrenocorti-cotropic hormone (ACTH), prolactin, and possi-bly growth hormone (GH) [(66, 67) for review],and the assumption that the extent of release ofhormones following a challenge with the 5-HTagonist provides an index of the responsivity ofthe 5-HT system (68). Depending upon the probeemployed, it is possible to assess the net central5-HT activity or the sensitivity of various 5-HTreceptor subtypes (68).
This paradigm has been used extensively to as-sess 5-HT function in depression [(6971) for re-view] and other psychiatric disorders [(7274) forreview]. There are a total of six neuroendocrinechallenge studies to date that examined the roleof 5-HT in mania by using different serotonergicprobes. Studies that examined pr...