Additional related information may be found at:
Neuropsychopharmacology: The Fifth Generation of Progress

Back to Psychopharmacology - The Fourth Generation of Progress

The Role of Serotonin in Schizophrenia

Bryan L. Roth and Herbert Y. Meltzer



Schizophrenia is generally considered to be a syndrome, probably a group of disorders, which produces diverse disturbances in cognition, reality testing, mood, interpersonal relations, social and work function. The disease is associated with some quantitative abnormalities in brain structure, e.g., enlarged ventricles and decreased temporal lobe volume (seeThe Neurobiology of Treatment-Resistant Mood Disorders, this volume), but these abnormalities are not specific to schizophrenia as they are also present in patients with mood disorders (20). Until recently, the hypothesis concerning the biological basis of schizophrenia and the mechanism of action of antipsychotic drugs concentrated primarily on the role of one neurotransmitter: dopamine (DA; see Brain Imaging in Mood Disorders, this volume). In recent years, the possibility that additional neurotransmitters, acting in concert with DA, contribute to the etiology of schizophrenia and the action of antipsychotic drugs has received more attention than DA hypothesis as originally formulated by Carlson and others (see Physiological Indicators of the Schizophrenia Phenotype, this volume).

This chapter focuses on the evidence for the role of serotonin (5-hydroxytryptamine; 5-HT) in the etiology and pathophysiology of schizophrenia and, to a limited extent, the mechanism of action of antipsychotic drugs. The potential of serotonergic agents to treat or exacerbate schizophrenia will be discussed. Next, the evidence for 5-HT/DA interactions in vivo as an important link between the DA hypothesis and a more modern approach to the pathophysiology of schizophrenia will be examined.



The first hypotheses concerning the involvement of 5-HT in schizophrenia was advanced by Wooley and Shaw (131) and Gaddum (25) based on the attribution of the psychotomimetic effects of lysergic acid diethylamide (LSD, which is structurally related to 5-HT),and its antagonists at brain 5-HT receptors. These investigators proposed that serotonergic activity might be decreased in schizophrenia. One of the major problems with this hypothesis was the recognition that the primary effect of LSD was to produce visual hallucinations, which are relatively rare in schizophrenia, not auditory hallucinations, which are the most common perceptual disturbance in schizophrenia (3, 118). Additionally, paranoid delusions, conceptual disorganization, and the wide range of cognitive impairments characteristic of schizophrenia: e.g., disturbances in working memory, semantic memory, and executive function, are generally absent during LSD intoxication (118). Another problem with the LSD hypothesis was that LSD is a full or partial agonist rather than an antagonists at many 5-HT receptors (30). As will be discussed, 5-HT antagonistss such as methysergide, cyproheptadine, ritanserin, and MDL 100907 are not psychotomimetic and may possibly have beneficial effects on specific aspects of psychosis. Furthermore, antipsychotics such as clozapine, risperidone, olanzapine, sertindole, and ziprasidone are antagonistss at multiple 5-HT receptors and have some advantages over selective DA receptor antagonistss.

The 5-HT deficiency hypothesis (113) was superseded by the proposal that the production of endogenous methylated indoleamines with psychotomimetic properties (e.g., N,N-dimethytryptamines) might be important to the etiopathology of schizophrenia (discussed in ref. 9). However, no consistent differences were found in the amounts of the N-methyl-transferase enzyme which synthesizes this type of compound in the brain, the level of methylated indoleamines, or their metabolites in plasma or urine of schizophrenic patients and normal controls (reviewed in ref. 9).


The decline of interest in the 5-HT deficiency hypothesis was followed by a nearly 25-year period during which consideration of the role of 5-HT in schizophrenia was limited. A relatively small group of biochemical and psychopharmacologic studies constituted the main effort in this regard. Biochemical studies of the density of specific types of 5-HT receptors in brain generally produced intriguing findings that supported the hypothesis of some type of serotonergic dysfunction in schizophrenia. Clinical trails of serotonergic precursors, agonists, or non-specific antagonistss were, however, generally disappointing, with no evidence comparable to the observation of DA agonists and antagonistss causing exacerbations or decreasing psychosis, respectively. The major breakthrough restoring interest in the role of 5-HT in schizophrenia was the identification of numerous 5-HT receptor subtypes and their extensive impact on multiple neurotransmitters and behaviors (41). In addition, it was thought that at least some of the extraordinary ability of clozapine to improve schizophrenic symptoms more effectively and with fewer side effects than typical neuroleptics could be attributed to its ability to block 5-HT2A receptors (79).

Before presenting this evidence, we will briefly discuss the influence of 5-HT on behavior. This topic is discussed in greater detail in other parts of this volume (see Serotonin and Behavior: A General Hypothesis, Indoleamines: The Role of Serotonin in Clinical Disorders, and Recent Studies on Norepinephrine Systems in Mood Disorders). Serotonin has been implicated in a variety of behaviors and somatic functions that are disturbed in schizophrenia: cognition, including memory; perception and attention; sensory gating; mood; aggression; sexual drive; appetite; energy level; pain sensitivity; endocrine function; and sleep. Many of these functions are clearly relevant to the etiology of positive and negative symptoms and the cognitive impairments which constitute the core abnormalities of schizophrenia. The biochemical and anatomical complexity and diversity of the serotonergic system (see Anatomy, Cell Biology, and Maturation of the Serotonergic System: Neurotrophic Implications for the Actions of Psychotropic Drugs, this volume) and its extensive interactions with multiple neurotransmitters (see Molecular Biology of Serotonin Receptors: ABasis for Understanding and Addressing Brain Function, this volume) provide the physiological substrate for the ability of 5-HT to influence all these behaviors. No clear statement of a revised 5-HT hypothesis of schizophrenia emerged until it was suggested, as part of a "serotonin-dopamine" hypothesis of schizophrenia, that there might be enhanced dopaminergic and serotonergic neurotransmission in subcortical areas in schizophrenia, leading to positive symptoms, and decreased dopaminergic and serotonergic activity, perhaps in the prefrontal cortex, which led to negative symptoms (1, 11, 72, 74, and 126). The major basis for this hypothesis was evidence supporting the hypothesis that the effects of clozapine and related antagonistss of both 5-HT2a D2 receptors (a group of drugs which now includes risperidone, olanzapine, sertindole, ziprasidone and quetiapine) upon schizophrenic psychopathology and cognition might be due to normalization of the combined influences of 5-HT and DA in a variety of brain regions, especially the from post-mortem studies that is consistent with regional abnormalities in 5-HT and DA activity (see below). A more general statement of the role of 5-HT and schizophrenia is that functional alterations in the serotonergic system ( including both pre-and postsynaptic function) affect multiple neurotransmitter systems (e.g., glutamate, GABA, norepinephrine [NE], acetylcholine, and DA) and cause the various behavioral disturbances in schizophrenia. Pharmacologic manipulation of the serotonergic system can either reduce or exacerbate positive, negative or disorganization symptoms and cognitive function, as well as modulate extrapyramidal function (e.g., extrapyramidal symptoms [EPS] and tardive dyskinesia or dystonia [TD]). Brier (11), Kapur and Remington (52) and Schmidt et al. (103) have recently reviewed the clinical and preclinical evidence, respectively, which supports this hypothesis.


The original DA hypothesis of schizophrenia has been supported by three types of findings: first, that all neuroleptics were effective D2 receptor antagonistss; second, that functional alteration in brain dopaminergic function were present in schizophrenia; and third, that drugs were increase brain dopaminergic activity can either induce psychosis in normal individuals or exacerbate symptoms in schizophrenes (see Brain Imaging in Mood Disorders , this volume). The extent to which comparable data is available with regard to 5-HT and schizophrenia will now be reviewed.

Fundamental to testing the 5-HT hypothesis of schizophrenia is the determination of whether serotonergic drugs can affect the psychopathology and cognitive dysfunction of schizophrenia (either to exacerbate or to ameliorate these components of the illness) as do the drugs which act on the dopaminergic system (e.g., amphetamine and neuroleptics). However, drug effects per se cannot provide the basis for concluding that 5-HT is involved in the etiology of schizophrenia. Alterations in serotonergic neurotransmission must also be demonstrable (e.g., receptor changes not due to drug to drug treatment). A large number of potential antipsychotic agents which are 5-HT2Areceptor antagonistss have been evaluated in patients with schizophrenia, providing evidence for the importance of this receptor for antipsychotic drug action (52,74). Additionally, a number of direct and indirect 5-HT agonists (fenfluramine, 5-hydroxytryptophan [ 5-HTP], mCPP, and tryptophan) sometimes exacerbate symptoms of schizophrenia (50,63,67,69). Finally, many studies of schizophrenia have demonstrated alterations in: 1) serotonergic neurotransmission as measured by the concentration of 5-HT and its metabolite 5-hydroxyindoleacetic acid (5-HIAA) and the density of 5-HT1A or 5-HT2A receptors in post-mortem brain specimens; 2) hormonal and behavioral responses to serotonergic drugs; and 3) abnormalities of 5-HT and its metabolites in blood or CSF.

The study of the role of 5-HT in schizophrenia has been advanced by the cloning of multiple 5-HT receptors. To date, 14 distinct 5-HT receptors have been identified via molecular biology and pharmacologic techniques (41; see Molecular Biology of Seotonin Rerceptors: A Basis for Understanding and Addressing Brain Function, Serotonin Receptor Subtypes and Ligands and Sertonin Receptors: Signal Transduction Pathways in this volume). Clozapine, the prototype of an atypical antipsychotic drug, was more effective than chlorpromazine or haloperidol in treatment-resistant schizophrenia (51) and has a very high affinity (,10nM) for the 5-HT2A, 5-HT2C, 5-HT6, and 5-HT7 receptors (79,98,99,104).Clozapine is also a moderately potent partial agonist at the 5-HT1A receptor and a 5-HT3 antagonists (see ref. 74 for review). The classical definition of an atypical antipsychotic drug is one that shows a marked dissociation between doses that block amphetamine-induced stereotype and those that cause catalepsy. There may also be a dissociation between the antipsychotic effect and extrapyramidal symptoms (EPS). A number of other atypical antipsychotic agents have also been found to have greater affinity for 5-HT2A than D2 receptors in vitro. The ratio of 5-HT2A / D2 affinities distinguishes between typical and atypical antipsychotics better than any other known pharmacologic feature (79,104). Risperidone and olanzapine, both of which have high 5-HT2A receptor blocking activity relative to their D2 affinity (78,104), have greater efficacy for treating negative symptoms of schizophrenia and produce fewer EPS than haloperidol (8,14). These observations are consistent with the hypothesis that 5-HT2A receptor antagonism is of value in an antipsychotic drug, since these agents do not share other properties that might produce this profile.

The fact that many atypical antipsychotic agents are potent 5-HT2A antagonistss is interesting in light of the evidence that stimulation of 5-HT2A or possibly 5-HT2C receptors is the basis for the hallucinogenic action of indoleamines such as LSD or psilocybin (30). Clozapine is a potent antagonistss at both of these receptor types. Clozapine and other potent 5-HT2A antagonistss such as olanzapine and mianserin ameliorate the psychosis due to L-DOPA or direct-acting DA agonists such as bromocriptine and pergolide in patients with Parkinson’s disease. The dose and plasma levels of clozapine that are effective in treating DA-induced psychosis in these patients is in a range in which 5-HT2A receptor blockade is likely to be the basis for its efficacy (75), rather than D4 receptors blockade, as suggested by Seeman et al. (106).This conclusion is supported by the evidence that two potent D4 antagonistss were ineffective antipsychotic agents in controlled clinical trails of schizophrenics. A comparison of the effect of the D2/D3 antagonists raclopride, the D1 blocker SCH 23390, the 5-HT1A agonist, 8-OH-DPAT, and the 5-HT2A / 2C antagonists ritanserin on catalepsy and conditioned avoidance (CAR) behavior also suggested actions on 5-HT1A as well as 5-HT2A receptors, but not the D receptor, might be the basis for at least some of clozapine’s advantages, since the two 5-HT agents inhibited raclopride-induced catalepsy and potentiated the effect of raclopride on the CAR (125). Many but not all atypical antipsychotic drugs are also potent antagonistss at 5-HT6 or 5-HT7 receptors; some typical neuroleptics are also potent 5-HT6 or 5-HT7 antagonistsi (99). 5-HT6 and 5-HT7 receptors are present in striatal and cortico-limbic areas, respectively (41,83). Blockade of these receptors could also contribute to some of the effects of atypical antipsychotic drugs, but consideration of the group of atypicals indicates that potent 5-HT6 or 5-HT7 receptor antagonism of high 5-HT6 / D2 or 5-HT7 / D2 ratios are not essential features of typical antipsychotic drugs (99). These receptors may play a modulatory role, however. Not until specific 5-HT6 or 5-HT7 agonists are available and have been adequately tested in the clinic will it be possible to determine the importance of these receptors in schizophrenia.



Cerebrospinal Fluid (CSF)

Many studies measuring the concentration of 5-HT metabolite, (5-HIAA), in the (CSF) of schizophrenic patients. CSF 5-HIAA levels reflect, in part, neuronal releases and / or metabolism of 5-HT and provide a partial estimate of 5-HT turnover in the CNS. Studies reviewed by Bleich et al. (9) suggest that CSF 5-HIAA levels correlate positively with frontal cortex and hindbrain levels of 5-HT.

Most investigators found no major differences in CSF levels of 5-HIAA in schizophrenic patients and controls, although several studies reported decreased CSF 5-HIAA concentrations in schizophrenia (see ref. 9 for review). A recent meta- analysis of CSF 5-HIAA studies in schizophrenia found 11 studies with adequate data for inclusion. Seven studies included patients in an acute exacerbation and six included chronic, stable patients. The effect sizes for either type of patient were small and non-significant (122). Some studies found a significant inverse relationship between CSF 5-HIAA concentrations and brain atrophy as determined by computerized tomography (CT) in schizophrenia (64,94). Low levels of CSF 5-HIAA (124) have been found in suicidal schizophrenic and depressed patients, while other studies have not replicated this result (see ref. 16fro review).

More recent studies have attempted to correlate CSF 5-HIAA levels with specific biological and psychological measures. Thus, Csernansky et al. (18) reported a positive correlation between CSF 5-HIAA concentration and BPRS Negative Symptom scores, work history, and a measure of cognitive impairment. Weinberger et al. (127) found CSF levels of 5-HIAA and homovanillic acid HVA correlated positively with prefrontal regional cerebral blood flow (rCBF) the Wisconsin Card Sorting test. The authors suggested that the primarily relationship might be with HVA, not 5-HIAA-i.e., with DA rather than 5-HT. Interestingly, CSF 5-HIAA concentrations did not correlate with rCBF under two other conditions in which frontal lobe activity was not predominant (resting state and a number matching test), suggesting specific relationship between CSF 5-HIAA levels and prefrontal cortical activity.

Lindström et al. (65) found a significant negative relationship between CSF 5-HIAA concentrations and abnormal brainstem-auditory evoked potentials in never treated first episode schizophrenics. These authors suggested brainstem dysfunction might be related to decreased serotonergic activity in schizophrenia. In a subsequent study, these authors noted that schizophrenic patients who showed reduced skin conductance responsiveness had low CSF 5-HIAA and HVA concentrations (89).

In light of the importance of both D2 and 5-HT2A receptor antagonism for the action of atypical antipsychotic drugs, the evidence that the ratio of dopaminergic to serotonergic activity as measured by the CSF HVA/5-HIAA may be more informative for schizophrenia than either measure alone is particularly interesting. Lewine et al. (64) reported that, while neither CSF HVA or 5-HIAA alone correlated with ventricular brain ratio (VBR) in 45 schizophrenia patients the ratio of HVA to 5-HIAA was negatively correlated with VBR (r = 0.45, p < 0.005). This suggests that structural brain abnormalities in schizophrenia might be related to a relative increased serotonergic compared to dopaminergic activity. Hsiao et al. (42) reported that typical neuroleptics produced a greater increase in CSF HVA / 5-HIAA than did clozapine, but the results were unrelated to chemical response. These authors also noted an increased correlation between CSF HVA and 5-HIAA after neuroleptic treatment and proposed that one effect of neuroleptic drug treatment may be to link more tightly the activities of 5-HT and DA neurons. Pickar et al. (92) found that low CSF ratios of HVA to 5-HIAA predicted a positive response to clozapine in 21 neuroleptic-resistant patients; neither metabolite alone predicted a good response to clozapine. They suggested that these findings were consistent both with the hypothesis that the efficacy of clozapine depended on a combined effect on 5-HT and DA neurotransmission and with the notion of a combined 5-HT/DA malfunction in schizophrenia. These results would imply that clozapine may be most effective in increasing dopaminergic and decreasing serotonergic activity, results consistent with the effect of clozapine to enhance prefrontal dopaminergic activity.

Taken together, these findings support, in a general way, the role of brain 5-HT dysfunction in schizophrenia. However, CSF 5-HIAA levels, at best, provide an integrated measure of serotonergic activity in multiple brain regions. They cannot distinguish between selective changes in different regions or provide any index of th necessary integration between serotonergic activity and that of other neurotransmitters. Additionally, because of the correlational nature of these studies, cause and effect relationships cannot be established. CSF studies are, at least for the time being, in declines a means of studying central serotonergic activity.


The rationale and background for the use of psychotropic challenge studies with endocrine and behavioral end points to clarify pathophysiology and study psychotropic drug action in psychiatric disorders has been previously summarized (84). Briefly, it has proposed that alterations in brain 5-HT receptor stimulation may be inferred from altered neuroendocrine responses to serotonergic agonists. Thus, for instance, a blunted prolactin (PRL), cortisol or temperature response to MK-212 (a direct-acting 5-HT2A / 5-HT2C agonist) could be attributed to a diminished number of 5-HT2A or 5-HT2C receptors and/or deficits in coupling to second messenger systems. Limitations to these types of studies often include: 1) the lack of specificity of the challenge agents; 2) the lack of suitable receptor subtype-specific antagonistss with which to define the actions of agonists, 3) the paucity of full dose-response studies; and (4) inability to assess the influence of other 5-HT and non-5-HT receptors that may influence the response to specific 5-HT agonists.

The 5-HT hypothesis of schizophrenia predicts that patients with schizophrenia should have abnormal responses to serotonergic challenge studies, if the hypothalamic neurons that transduce 5-HT responses are also abnormal. Among the serotonergic drugs that have been studied in schizophrenia are: 1) fenfluramine (which induces the release of 5-HT), 2)m-chlorophenylpiperazine (mCPP a full or partial agonist at multiple 5-HT receptors;ref. 44);3) L-tryptophan (a precursor of 5-HT, and thus, a potential agonist at all 5-HT receptors);and 4) MK-212 (a potent 5-HT2A and 5-HT2C agonist, and a weak agonist at 5-HT1A receptors). Previous investigators have suggested that mCPP was a selective 5-HT2c agonist (44). However, more recent data obtained with cloned 50HT receptors have shown that mCPP has high affinity (and efficacy) for at least seven 5-HT receptors subtypes: 5-HT1Da, 5-HT1Db, 5-HT2A, 5-HT2B, 5-HT2C, 5-HT6, 5-HT7 (Table 2). Additionally, mCPP is a partial agonist at cloned 5-HT2A receptors stably expressed in NIH3T3 cells (Roth, unpublished observation), Table 1 and Table 2.

Kolakowska et al. (54) reported a reduced growth hormone (GH) response and an enhanced PRL response to L-TRP in medicated schizophrenic patients. Price et al. (95) found that the PRL response to L-TRP was normal in non-medicated schizophrenic patients.

Studies with direct-acting agonists such as MK-212 and mCPP are somewhat easier to interpret than studies utilizing precursors. Lee et al. (59) reported that the temperature response to MK-212 was blunted in unmedicated schizophrenic patients. The PRL and cortisol responses to MK-212 were also blunted in unmedicated schizophrenic patients, compared with normal controls (Meltzer HY, unpublished data). These results are consistent with the hypothesis that 5-HT2A / 2C responsivity is diminished in schizophrenia and are consistent with post-mortem data of a decreased 5-HT2A receptor density in schizophrenia. The mCPP-induced elevations in serum PRL and Cortisol levels were also blunted in schizophrenic patients (67). Lerer et al. (63) found a blunted PRL response to fenfluramine in schizophrenia. Fenfluramine also exacerbated positive symptoms. On the other hand, Abel et al. (2) recently reported that schizophrenics had an enhanced PRL response to fenfluramine, and that this response was positively correlated with BPRS scores for depression, anxiety and guilt. This study was done in drugs-naïve, first-episode schizophrenic patients, but it was, however, not placebo-controlled, so the possibility of an enhanced stress response cannot be ruled out. The authors attributed the increased response to enhanced serotonergic "tone," consistent with the hypothesis of excessive serotonergic activity in schizophrenia (72). This relationship to affective symptomatology is consistent with our evidence of the antidepressant and anti-suicidal effects of clozapine. Meltzer et al. (77) recently found that the cimetidine-induced increase in plasma PRL levels was blunted in male but not female schizophrenics. This finding is of particular interest because of the numerous difference between male and female schizophrenics in age of onset and treatment response. Based on previous studies, suggesting that the cimetidine-induced increase in PRL was indirectly mediated via 5-HT2A receptors (perhaps via inhibition of 5-HT uptake which is promoted by histamine [56], these findings provided further evidence of decreased 5-HT2A receptor function in schizophrenia. 5-HT1A antagonistss might be clinical interest in female schizophrenics. Alternatively, increased stimulation of 5-HT1A receptors might be of clinical value.

The ability of 5-HT antagonistss such as clozapine and other 5-HT2 / D2 antagonistss to functionally block 5-HT receptors in vivo can be monitored by neuroendocrine means. Lemus et al. (62) reported that the fenfluramine-induced increase in PRL secretion was blunted by clozpine. Meltzer et al. (73) reported blockade by clozapine of the PRL, cortisol and temperature responses to MK-212. Clozapine, but not haloperidol, blocked the mCPP-induced ACTH and PRL responses in patients with schizophrenia (50). Treatment with clozapine also totally abolished the cimetidine-induced PRL response in schizophrenia (77). Based on previous studies that suggested that the cimetidine-induced increases in PRL were indirectly mediated via 5-HT2A receptors (perhaps via inhibition of 5-HT uptake, which is promoted by histamine [56]), these findings provide further evidence of decreased 5-HT2A receptor function in schizophrenia.

Thus, the results of these pharmacologic challenge studies also support the 5-HT hypothesis of schizophrenia, since: 1) most investigators have found blunted responses to indirect-acting 5-HT agonists (e.g., fenfluramine) and cimetidine or 5-HT2A / 2C agonists; and 2) atypical antipsychotic agents block the neuroendocrine responses to serotonergic agonists. Future studies with more selective agonists and antagonistss should help to clarify the 5-HT receptor subtype that contribute to the abnormal response to 5-HT agonists. Further study of 5-HT1A receptor function is of particular interest because of the post-mortem evidence for 5-HT1A receptor up regulation in some regions of the brain in schizophrenia.


Rodent studies indicated that clozapine, olanzapine, risperidone, quetiapine, ziprasidone and sertindole achieve 50% occupancy of cortical 5-HT2A receptors at lower doses than sertindole D2 receptors (104,115). The hypothesis was that these agents owe at least some of their special clinical features to high levels of 5-HT2A receptor blockade and minimal levels of D2 receptor blockade. PET or single photon emission computed tomography (SPECT) studies of atypical antipsychotic drugs can provide data that may be helpful in establishing the appropriate dose range for testing (34,52). Since clozapine is the prototype of this group of compounds, the goal of these studies is to identify the dose range which most closely approximates the ratio of 5-HT2A / D2 occupancy produced by clozapine. Were it to be found that this occupancy ratio did not produce optimal results with all antipsychotic drug of this type, the results would suggest that factors other than actions at these two receptors may be able to negate any advantages of this profile.

PET or SPECT studies may also be used to assess 5-HT and DA turnover with the atypical antipsychotic drugs. Methods are under development to measure 5-HT and DA synthesis uses 11C-labeled precursors. An indirect method has been proposed. Endogenous 5-HT and DA should displace loosely bound ligands in a quantitative manner. This has been done with SPECT for DA (57) in the striatum. More sensitive methods are needed to study the cortex and limbic areas. Such measurements are important or understanding the presynaptic effects of these agents.

Positron emission tomography studies permit measurements of 5-HT receptor density in the brains of schizophrenic patients, both on and off medication. Farde et al. (23) found that D2 receptor occupancy was only 20-67% during successful clozapine treatment, vs. 80-90% for typical neuroleptics, indicating D2 receptor antagonism alone does not explain the efficacy of clozapine. As expected, even at low doses (125-172 mg/day), clozapine occupied more than 80% of 5-HT2A receptors (23). Using [11C]N-methylspiperone (NMSP) as a ligand, our laboratory compared occupancy of cortical 5-HT2A and striatal D2 receptors in five controls (34). The relative occupancy of the 5-HT2A and D2 receptors in the clozapine-treated patients was 72 ± 10% and 19 ± 10%, compared with 23 ± 22% and 48 ± 5% in the typical neuroleptic patients. The 5-HT2A and 5-HT2A / D2 but not the D2 occupancies we4re significantly different between the two groups. Farde et al. (23) found that with risperidone, at 4 mg/day, the 5-HT2A and D2 occupancies were 78-88% and 15-80%, respectively. This ratio of 5-HT2A / D2 occupancy is significantly greater than that found with clozapine and suggests that lower doses, e.g.., 2 mg/day, would be more likely to produce an EPS profile closer to clozapine. Although this dose is lower than that reported to be optimal in multicenter trials (14), it is clear that many patients do respond at this dose, especially during maintenance treatment. It may be that the higher doses (6-8 mg/day) that were more effective in the trails of risperidone in patients with an acute exacerbation are necessary because of higher DA turnover in this phase of the illness. Perhaps for maintenance treatment, lower levels of D2 occupancy are sufficient. This would still require factors other than 5-HT2A and D2 receptor blockade to account for the efficacy of clozapine. Sertindole, another 5-HT2A / D2 drug, has recently been reported to occupy 80-90% of both 5-HT2A and D2 receptors (93). These results must be confirmed by other investigators.

In accordance with the concept of the importance of the relative activity of both the 5-HT and DA systems, Iyo et al. (46) used [11C]NMSP binding to compare the cortical 5-HT2A and striatal D2 binding in six men who had a history f methamphetamine abuse and susceptibility to recurrent psychosis, with 10 normal controls. The absolute binding in either brain region was not significantly different between the two groups, but the ratio of the binding in the two regions did differ: the ratio of striatal-to-cortical bindings was lower in the methamphetamine users than in the controls. This suggests that those patients vulnerable to relapse may have increased DA turnover relative to 5-HT turnover.


A number of authors have measured 5-HT receptor density in post-mortem brain tissue from patients with schizophrenia. Bennett et al. (9) found a decrease in [3H]LSD binding in Brodmann areas 6, 8, 11, 44, and 47, while Whitaker er al. found increased [3H]LSD binding in unmediated schizophrenics in cortical regions 4, 10, and 11. [3H]LSD binds to at least six distinct 5-HT receptors with high affinity (5-HT2C, 5-HT1E, 5-HT2A, 5-HT6, and 5-HT7; refs. 9, 128), making it difficult to determine which type of 5-HT2A receptor is responsible for the drug's effects. Using the 5-HT antagonists [3H]ketanserin as a ligand, Mita et al. (81) also found a decrease in binding in cortical area 9. [3H]ketanserin has high affinity for both a2-adrenergic receptors and tetrabenazine-sensitive sites (discussed in refs.99, 100). Arora and Meltzer (5) and Laruelle et al. (58), also using [3H]spiperone, found a decrease in binding in cortical areas 8 and 9. However, Joyce et al. (49), using quantitative receptor autoradiography, recently reported an increase of [125]LSD and [3H]ketanserin-labeled "5-HT2" receptors in temporal and posterior cingulate, frontal, and parietal cortices, the ventral putamen, nucleus accumbens and hippocampus, but not in the caudate nucleus or motor, prefrontal, entorhinal, or anterior cingulate cortices. These conflicting results may be related to the much greater age of the subjects in the Joyce et al. study, compared with earlier studies. Burner et al. reported 5-HT2A binding sites (as determined by [3H]ketanserin binding) were decreased in the dorsolateral prefrontal cortex (-27%) and parahippocampal gyrus (-38%) of schizophrenics. A similar trend was found in the cingulate gyrus, but not in superior temporal gyrus or striate cortex. Decreased 5-HT2A receptor mRNA as found in the dorsolateral prefrontal, superior temporal, anterior cingulate and striate cortices of the schizophrenics, but not in parahippocampal gyrus. Thus, there is strong evidence for a down regulation of 5-HT2A receptors in the cortex of schizophrenics. Since 5-HT2A receptor density is decreased by 5-HT2A receptor stimulation, this may be the result of increased 5-HT2A receptor activity. Blockade of this activity by drugs such as clozapine and other atypical antipsychotic agents that are 5-HT2A receptor antagonistss may contribute to their clinical profile.

Bennett at al. (see ref. 9) found no change in "5-HT1" receptor numbers in prefrontal cortex using [3H]5-HT. Hashimoto et al. (38) using [3H]-8-OH-DPAT as the ligand found an increase in 5-HT1A receptors in cortical areas 10 and 22 and in the "prefrontal" cortex. The difference in these two studies could result from the binding of 5-HT itself to many distinct 5-HT receptors (5-HT1A, 5-HT1Da/b, 5-HT1Ea/b, 5-TC-T,5H7. Suoyoshi et al. (116) reported an increase in [3H]-8-OH-DPAT binding in prefrontal cortex of 12 schizophrenics, compared with matched controls. The density of the high affinity site was increased by an average of 79% in the schizophrenic group. Joyce et al. (49), using quantitative autoradiographic techniques, also reported that 5-HT1A receptors were increased in several brain regions of schizophrenic patients (hippocampus, posterior cingulate cortex, motor and temporal cortex). Simpson et al. (111) used an autoradiographic method to study 3H-8-OH-DPAT binding to 5-HT1A receptors in orbital frontal cortex of 12 schizophrenic post-mortem brains and 18 controls. Increased 5-HT1A binding was present in three orbital frontal cortex regions. The increase was most characteristic of males. Since some 5-HT1A receptors are located on pre- or post-synaptic elements of glutamatergic synapses (reviewed in refs. 41,111) and some increased glutamate uptake sites and kainate and N-methyl-D aspartate (NMDA) receptor subtypes were also found in the same patients by Simpson et al. (111), it was suggested that 5-HT1A modulation of glutamatergic activity might be abnormal in schizophrenia. Burnet et al. (13) found a significant (23%) in 5-HT1A binding site densities in the dorsolateral prefrontal cortex of schizophrenics, with a similar trend in anterior cingulate gyrus. 5-HT1A receptor mRNA was not increased. There were no differences in either measure in the superior temporal gyrus, striate cortex, or hippocampus. Thus, there is some evidence in support of 5-HT1A receptor density increase in specific cortical areas in schizophrenia. 5HT1A receptor stimulation has potent effects on clozapine-induced increase in prefrontal DA release (Ichikawa et al., in preparation). Because of the synergistic effect of 5-HT2A antagonism and 5-HT1A agonism (78), the evidence for upregulation of 5-HT1A and down regulation of 5-HT2A receptors is most intriguing. This could signify a greater increase in 5-HT1A receptor stimulation than would be achieved by upregulation of only 5-HT1A receptors.

It is noteworthy that the dorsolateral prefrontal cortex had both decreased 5-HT2A and increased 5-HT1A receptor binding sites. This is the area of the brain that has been most implicated in schizophrenia. The resulting imbalance of the 5-HT1A to 5-HT2A receptor ratio could contribute to abnormalities in the function of cortical association pathways. Since these two 5-HT receptors often have opposite effects on the function of various neurons, the change in the ratio of their activities may have marked effects on specific types of neuronal activity.

Joyce et al. (49) reported a large increase in the number of 5-HT uptake sites in the dorsal and ventral striatum. They suggested that this was evidence for hyper-serotonergic innervation of the striatum in schizophrenia. A marked reduction in 5-HT uptake sites in the external and middle laminae of the anterior cingulate, frontal cortex, and posterior cingulate was also found (49). No differences in the number of uptake sites in the motor or temporal cortices were found. On the other hand, Dean et al. (19), using [3H]paroxetine to label the 5-HT transporter, found no difference in the number of uptake sites in the striatum, frontal cortex of hippocampus only. These studies differed significantly in methodology and subject characteristics. It is not possible at this time to draw any firm conclusion about the transporter for 5-HT in schizophrenia.


The platelet has been proposed as a model of 5-HT neurons, because there are a number of similarities between 5-HT physiology in platelets and neurons. Bleich et al. (9) concluded that platelet or whole blood concentrations of 5-HT did not differ in schizophrenic patients versus controls.

Studies of platelet–5-HT uptake in schizophrenia performed prior to 1988 yielded inconsistent results (9). The binding characteristics of the platelet 5-HT2A receptor are similar to those of the brain 5-HT2A receptor. Arora and Meltzer (6) reported that platelet 5-HT2 receptor density (Bmax) was increased in nine schizophrenic patients with a recent history of suicidal ideation or attempts, compared to 24 non-suicidal schizophrenic patients and 42 normal controls. The binding characteristics of the central and platelet 5-HT2A receptor are similar (90). The Bmax of platelet 5-HT2A receptor density and blood 5-HT levels were negatively correlated (Pho=-0.77, p = 0.0002, n = 18), suggesting that blood 5-HT levels may regulate this receptor (90). Grahame-Smith et al. (35) found that chronic treatment with phenothiazine or thioxanthenes (which have a high affinity for 5-HT2 receptors) caused an upregulation of platelet 5-HT2A receptors. Arora and Meltzer (6) found that chronic clozapine treatment also increased the numbers of platelet 5-HT2 receptors. It should be noted that these effects of antipsychotic drugs on the platelet 5-HT2A receptors are opposite to those found for brain 5-HT2A receptors (74). The fact that the effects of antipsychotic drugs on peripheral and central 5-HT2 receptors differ suggests that the platelet is not an ideal model for brain 5-HT2 receptor regulation in some circumstances.



High doses of precursors of 5-HT such as 5-HTP and tryptophan, together with a monoamine oxidase inhibitor given chronically, have generally been found to increase positive symptoms in schizophrenia (9). It is likely that high doses of 5-HT formed from precursors can displace endogenous DA, which causes the increase in positive symptoms.


Fenfluramine acutely induces 5-HT release and chronically depletes 5-HT, at least in rodents. Three recent studies have suggested that chronic fenfluramine administration may worsen psychotic symptoms in schizophrenics. Soper et al. (113) found that fenfluramine worsened performance in several cognitive test batteries. Marshall et al. (69) reported that fenfluramine worsened psychotic symptoms, and that this effect was inversely correlated with the fenfluramine-induced decline in blood 5-HT levels. Kolakowska et al. (54) reported exacerbation of psychosis by fenfluramine in some patients but improvement in others. Fenfluramine, in addition to enhancing the release of 5-HT, may also increase the release of DA. Either or both effects may contribute to the worsening of psychosis.


Chronic administration of a selective serotonin reuptake inhibitor would be expected to increase serotonergic activity by inhibiting 5-HT uptake and also by desensitizing the 5-HT autoreceptor, leading to increased 5-HT release. Chronic fluoxetine may also alter postsynaptic receptor responsivity (e.g., increasing responsivity to 5-HT1A agonists and enhancing cortical release of DA). Chronic fluozetine also has been reported to enhance the sensitivity of mesolimbic but not striatal D2 receptors in rat brains (15). Goff et al. (32) reported that the addition of fluoxetine to typical neuroleptics in nine treatment-resistant patients improved positive and negative symptoms. Similar results were reported with fluvoxamine (40, 119)m which increases plasma levels of clozapine by inhibiting its metabolism (40). In a randomized, double blind trial, the addition of fluvoxamine produced a greater decrease in negative but not positive symptoms in a group of 30 chronically hospitalized, neuroleptic-treated schizophrenics without worsening EPS in a six-week trial (110). Whether this was due to pharmacodynamic as well as pharmakinetic factors remains to be determined.

In another placebo-controlled, randomized trial, fluoxetine produced similar improvements in negative symptoms only when added to depot neuroleptic (32). The improvement in negative symptoms may be due to either enhanced serotonergic or mesocortical and mesolimbic activity, or both. The absence of a worsening of EPS after the addition of an SSRI to a neuroleptic may result from the fact that the activity of nigrostriatal DA neurons is already maximally diminished as a result of treatment with neuroleptic agent.


Buspirone, a partial 5-HT1A agonist, has been reported to either exacerbate, improve, or have no effect on positive and negative symptoms in schizophrenic patients receiving typical neuroleptics. It also has a beneficial effect on EPS and akathisia (12, 105). Further study of specific 5-HT1A agonists and antagonists as adjunctive treatments along with typical and atypical antipsychotics are needed. 5-HT1A receptor agonists may have antipsychotic-like properties (4) and can decrease neuroleptic-induced EPS (125). Ziprasidone, one of the new atypical antipsychotic drugs, is a potent 5-HT1A agonist, whereas quetiapine may be expected to be a 5-HT1A antagonists in vivo. Clozapine is a partial agonist at the 5-HT1A receptor. The clinical importance of these differing actions at the 5-HT1A receptor in patients remains to be determined. The increase in density of 5-HT1A receptors in some brain regions in schizophrenia may be due to inadequate stimulation of these receptors which would suggest further stimulation may be clinically useful.


The complex pharmacology of mCPP has been described above. In addition to stimulating a variety of 5-HT receptor subtypes (5-HT1D, 5-HT1C, 5-HT2, 5-HT2F, and 5-HT6; Table B), it may also have a-adrenergic and 5-HT uptake-inhibiting effects. Two studies have recently reported that iv. infusions of mCPP induced an exacerbation of positive symptoms of schizophrenia (44, 55). Krystal et al. (55) showed that pretreatment with ritanserin (a 5-HT2A/5-HT2C receptor antagonists) decreased the mCPP-induced increase in psychotic symptoms in four out of six patients. Clozapine, but not haloperidol, blocked the behavioral and neuroendocrine effects of mCCP (50). However, oral mCPP, at a dose of 0.5 mg/kg, only rarely caused significant changes in positive symptoms (67). Taken together, these results lend some support to the hypothesis of Meltzer that stimulation of 5-HT2A/2C receptors may contribute to the positive symptoms of schizophrenia (73)


Cyproheptadine, which can block the effect of 5-HT at multiple 5-HT receptors (including 5-HT2A and 5-HT2C receptors) had no effect on positive or negative symptoms in two placebo-controlled, double-blind studies (60, 109). In an uncontrolled study, the addition of cyproheptadine to perphenazine in four patients who relapsed after withdrawal of clozapine, was found to enhance the ability of perphenazine to treat positive symptoms (76). These data on the unique effect of the withdrawal of clozapine and the possible effectiveness of cyproheptadine in this syndrome provided further evidence for the role of 5-HT in clozapine treatment of psychosis.


There is now considerable evidence form studies of schizophrenia and other forms of psychoses, e.g., senile psychoses and L-DOPA psychosis, that the atypical antipsychotic drugs such as clozapine, olanzapine, quetiapine, risperidone, sertindone and ziprasidone are more effective with fewer EPS. These drugs have higher potencies as 5-HT2A antagonists than as D2 antagonistss or at almost any other receptor. At clinically effective doses, they all produce fewer EPS than typical neuroleptic drugs. However, marked EPS can occur with risperidone if the dose is increased. Clozapine (12.5-7.5 mg/day) as well as mianserin, a 5-HT2A/2C antagonists without D2 antagonists properties, can block L-DOPA psychosis without worsening EPS, raising the possibility that blockade of 5-HT2A or 5-HT2C receptors may contribute to the basis for the effects of these agents on positive symptoms. Seeman et al. have suggested that D2 receptor blockade may be the basis for the effect of clozapine L-DOPA psychosis (106), but this would not explain the effectiveness of mianserin in this condition. It has recently been reported that MDL 100907, a highly selective 5-HT2A antagonists, is more effective than placebo in affecting positive symptoms in schizophrenia (J. Shipley, personal communications). If confirmed in other studies, this would provide major support for the importance of 5-HT2A receptor blockade.

Some of the 5-HT2A/D2 receptor antagonistss (e.g., clozapine, risperidone, sertindole, and olanzapine) have been reported to be superior to typical neuroleptics for control of negative symptoms. No difference in this regard has yet been reported for quetiapine and ziprasidone. It may be that other pharmacologic effects of this group of drugs account for this action, or that some of the other pharmacologic properties of quetiapine and ziprasidone negate the effects of 5-HT2A receptor blockade.

An emerging area of great importance is the effect of this class of agents on cognition. Clozapine has been shown in several studies to improve attention and semantic memory. Risperidone has been reported to improve working memory. The role, if any, of 5-HT in these effects remains to be determined.

Because of the importance of 5-HT2A receptor blockade suggested by these results, it has been proposed that the addition of a selective 5-HT2A to a selective D2 antagonists, e.g., haloperidol or amisulpride, may provide some of the benefits of clozapine and related compounds. There is some evidence that this is the case with ritanserin (28). Trails of this type should utilize low doses of the selective D2 antagonistss in order to keep the occupancy of D2 receptors in the same range as that produced by clozapine ¾ e.g., 20 ¾ 50% with [11C]raclopride or C-NMSP as ligand (22,34). The evidence that risperidone loses some of its benefits with doses £ 10-12 mg/day is consistent with the hypothesis that high D2 receptor blockade can negate the beneficial effects of high 5-HT2A receptor blockade. All of this is quite speculative. Only careful dose-response studies with determination of 5-HT2A and D2 receptor occupancies and very specific drugs can clarify these important issues. Recently, sertindole at usual doses was found to produce high occupancy of both 5-HT2A and D2 receptors in five schizophrenic patients without causing EPS (93). This indicates that, under some circumstances, high 5-HT2A receptor occupancy can overcome the effect of D2 occupancy.


The 5-HT6 serotonin receptor has a high affinity for clozapine and olanzapine, while loxapine and amoxapine, two typical neuroleptics of a chemical class related to clozapine, have relatively weak affinities for this receptor (99). The 5-HT6 receptor is highly enriched in striatum, while the 5-HT7 receptor is concentrated in the hypothalamus (83,108). Schotte et al. (104) reported the affinities of a group of atypical antipsychotic drugs and haloperidol for cloned rat and human 5-HT1A, 5-HT1D, 5-HT2A, 5-HT2C, 5-HT3, 5-HT6, 5-HT7 receptors, as well as D2, D4, and a variety of other receptors. The novel agents studied included risperidone, olanzapine, ziprasidone, clozapine, sertindole, ORG-5222, seroquel and zotepine. The affinity of these drugs for the 5-HT2A receptor is the highest among all the 5-HT receptors and higher than that for the D2 receptor and virtually every other receptor measured. If the receptor profile of these agents is compared to the 5-HT2A receptor affinity, then marked differences between them and clozapine are evident. Clozapine has high affinity for the 5-HT2C, 5-HT6, 5-HT7, and 5-HT2A receptors. Risperidone and olanzapine have high affinities for the 5-HT2A, 5-HT2C, and 5-HT7 receptors. Olanzapine has high affinity for the 5-HT6 but not the 5-HT7 receptor. The reverse is true for risperidone. Sertindole has high affinity for the 5-HT2A, ziprasidone has high affinity for the 5-HT1A, 5-HT1D, 5-HT2A, and 5-HT7 receptors, similar to risperidone. Quetiapine has relatively high affinity for the 5-HT2A, 5-HT2C, and 5-HT7 receptors, compared with the D2 receptor. These conclusions are based upon in vitro affinities and may not occur in vivo (104). When one considers the marked variations in affinities for other key receptors in the dopaminergic, cholinergic, and noradrenergic systems, the great diversity among these compounds becomes quite apparent. This is the probable basis for the already apparent clinical differences among these compounds.


Evidence for linkage or association of at least one of the genes for a specific 5-HT receptor or one of the enzymes involved in the metabolism of 5-HT would be compelling evidence in favor of the hypothesis that 5-HT is involved in schizophrenia. The 5-HT2A receptor has been mapped to chromosome 13. Hallmayer et al. (37) found no evidence for linkage between a silent 5-HT2A receptor-gene polymorphism, the T102C variant of the 5-HT2A receptor, and schizophrenia in a Swedish kindred study. Ishigaki et al. (45) also found no association between this polymorphism and schizophrenia in a study involving 10 schizophrenics and 10 controls. However, Inayama and colleagues (43) reported a significant association between this mutation and schizophrenia in a sample of Japanese patients and healthy controls. Schizophrenic patients had a higher than expected proportion of allele 2 and genotype 2/2 of the T102C polymorphism. These findings have now been replicated (21,130). This suggests that the gene for the 5-HT receptor or a locus in linkage disequilibrium with it confers susceptibility to schizophrenia. Allele 2 is common in the general population. The T102C variant, or a nearby polymorphism, may affect a significant proportion of schizophrenic patients. In the large-scale study of Williams et al. (130). the relative risk of schizophrenia was 1.7 in individuals with the 2/2 or 1/2 genotype, compared with the 1/1 genotype. The attributable fraction ¾ i.e., the proportion of cases in which allele 2 or other genes that contribute to schizophrenia along with it ¾ that is etiologically necessary is 0.35, which is relatively high. This is because allele 2 (or other risk factors) is common in the population. Allele 2 itself is not a sufficient cause. The T102C polymorphism may act at the translation stage through effects on secondary structure and stability of the mRNA. Two other polymorphisms in the coding region of the 5-HT2A receptor gene, the substitution of threonine by asparagine at position 25 (Thr24Asn) and a substitution of histidine by tyrosine at position 452 (His452Tyr), were not associated with schizophrenia in a population that showed association with T102C (21).

The T102C polymorphism has also been found to be a risk factor for poor response to clozapine (7). Researchers found that the genotype C102/C102 was more common in patients who did not respond to clozapine than in those who did. However, these findings were not replicated by three independent groups (70,88,68). In addition, Nöthen (88) and Malhotra (68) reported a lack of allelic association between the His452Tyr polymorphism and response to clozapine. However, Masellis (71) found strong evidence that the His 452Tyr polymorphism was associated with a poor response to clozapine. Patients who failed to respond had a Tyr452 allele frequency of 0.16, compared with 0.06 in those who did respond. Poor responders to clozapine were more likely than those who did respond to be His/Try452 heterozygotes of Tyr/Tyr homozygotes. Differences in the time and method of ascertainment of who were responders to clozapine may have been the major difference between the studies which did not find that 452 polymorphism contributed to the clozapine response. Further study alone these lines is indicated.



Increased activity of mesolimbic DA neurons is delivered to be an essential factor in the etiology of positive symptoms (107). Negative symptoms and EPS have been postulated to be related to deficits in dopaminergic activity in the mesocortical and nigro-striatal systems, respectively (126,107). There is strong evidence, some of which will be reviewed below, indicating that the 5-HT system modulates dopaminergic activity and vice versa. This interaction occurs at the level of the cell bodies in the ventral tegmentum, substantia nigra and medial and dorsal raphe, as well as at various terminal areas of these three nuclei. Multiple types of 5-HT and DA receptors may be involved. To date, the 5-HT1A, 5-HT2A and D2 receptors appear to be the most important for these interactions. Evidence is strong for the model in which 5-HT activity has an overall inhibitory effect upon dopaminergic function, but examples of a facilitatory role exist as well (74).

As reviewed by Abi-Dargham et al. (1), decreasing serotonergic activity by lesioning the median raphe or inhibiting 5-HT synthesis increases spontaneous locomotor activity, an effect which may be mediated by projections from the midbrain raphe to the hippocampus (48). Lesioning of the raphe increases startle (29), which is analogous to decreased sensory gating in schizophrenia. Amphetamine-induced locomotor activity, an effect mediated by increased dopaminergic activity in the nucleus accumbens, is thought to provide one of the better models of positive symptoms in schizophrenia, since it is blocked by all known antipsychotic agents. Decreasing 5-HT function by a variety of means, including lesions of the median raphe and a tryptophan-free diet, enhances amphetamine-induced locomotor activity. Since the increase in locomotor activity produced by the direct-acting DA agonist apomorphine is also enhanced by inhibition of 5-HT synthesis, it is likely that the locus of the effect of 5-HT to enhance dopaminergic activity is post-synaptic.

Costall (17) demonstrated that the nucleus accumbens is one site of the inhibitory effect of 5-HT on DA-induced locomotor activity. Median raphe 5-HT neurons are the source of the projection to the nucleus accumbens. Decreased 5-HT activity from median raphe neurons in the accumbens might lead to disinhibition of mesolimbic DA neurons, one effect of which might be in increase in positive symptoms (1).

Turing behavior in the rodent has been used to assess the effect of both the median and dorsal raphe on striatal DA function. As reviewed by Abi-Dargham et al. (1), stimulation of median raphe 5-HT neurons inhibited the activity of nigro-striatal DA neurons, whereas stimulation of dorsal raphe 5-HT neurons enhanced the effect of DA release in the striatum.

There is also evidence that increased 5-HT activity leads to enhanced catalepsy (39), while 5-HT2A receptor blockade leads to decreased catalepsy. 5-HT1A receptor stimulation, by shutting down 5-HT release in the striatum, may also decrease catalepsy (39).

Serotonin neurons from the dorsal raphe project to the substantia nigra, the ventral tegmentum, the dorsal and ventral striatum, and to the frontal cortex. These terminal areas contain 5-HT1B/1D,5-HT2A, and 5-HT3 receptors. Serotonin may induce local release of dendritic nigral DA (131). Serotonin stimulation of 5-HT2A receptors inhibits the activity of DA neurons in the substantia nigra. 5-HT2A antagonistss increase activity of DA neurons in the ventral tegmental area (VTA) and substantia nigra DA neurons, and thus increase DA release in terminal areas (24,53,87). VTA neurons respond to lower concentrations of 5-HT2 antagonistss (33).5-HT2A antagonistss may also increase DA release in the prefrontal cortex by a direct action at 5-HT2A receptors on DA nerve terminals (82,93,117). 5-HT2A. antagonistss may also prevent the effect of decreased prefrontal glutamatergic activity that inhibits burst firing of VTA DA neurons. This might be a mechanism by which 5-HT2A antagonistss decrease negative symptoms (36,117).

Tonic release of DA in the mesolimbic system is dependent upon stimulation of 5-HT receptors (1). Thus, the 5-HT agonist DOI potentiates DA-induced DA synthesis and release. Blockage of this effect by 5-HT antagonistss may contribute to their ability to decrease negative symptoms (123).

Serotonergic dorsal raphe neurons may also regulate dopaminergic neurons. DA neurons from the substantia nigra and ventral tegmentum project to the dorsal raphe nucleus, which has a high concentration of DAD2-like but not DAD1-like receptors. Electrical stimulation of the substantia nigra leads to inhibition of the dorsal raphe neurons. Release of DA in the dorsal raphe increases the release of 5-HT in the dorsal raphe and decreases it in the striatum (24). This effect is blocked by pretreatment with the selective 5-HT1A antagonists WAY100135. Thus, inhibition of D2 receptor stimulation by neuroleptics should lead to increased 5-HT release in the striatum. This may contribute to their EPS liability via stimulation of 5-HT2A receptors in the striatum. Whether similar mechanisms exist for modulating serotonergic projections to the mesolimbic and mesocortical systems is not yet known, but it is likely that they do. Recent anatomical studies of 5-HT2 receptor protein suggests that DA 5-HT interactions could occur on postsynaptic elements as well. Garlow et al. (26) have demonstrated that 5-HT2A receptors are found in interneurons in caudate nucleus and certain cortical areas. Since at least some D receptors are found on similar neuronal populations, it is conceivable that direct 5-HT2A/D2 interactions occur on interneurons in caudate nucleus and other regions.

Given the probable heterogeneity of schizophrenia and the complexity of the serotonergic system, Meltzer and Fatemi (74) have suggested that no single type of abnormality of the serotonergic or other neurotransmitter system is likely to emerge as characteristic of all patients who meet current criteria for this syndrome. It seems more likely that specific signs and symptoms of schizophrenia, ranging from cognitive dysfunction, disorganization, negative and positive symptoms, motor abnormalities, decreased sexual development, insomnia and compulsive behaviors may, in part, result from specific individual or combinatorial serotonergic abnormalities. A comprehensive examination of pre-and postsynaptic serotonergic function in relation to dopaminergic activity in a wide range of schizophrenic patients at various stages in the evolution of the syndrome will be needed to assess this hypothesis.


Many investigators have speculated that schizophrenia may be a neurodevelopmental disease, perhaps involving the dopaminergic neuronal system. In this regard, it is interesting to note that serotonergic terminals have been shown to sprout after neonatal lesions of dopaminergic terminals (reviewed in ref. 47). Most recently, Jackson and Abercrombie (47) demonstrated that rats depleted of DA at birth showed enhanced release of 5-HT when evaluated as adults, and that ketanserin blocked the hypermotility of neonatally DA-lesioned rats. These results suggest that a functional deficit of DA release during the perinatal period can lead to functional supersensitivity of the 5-HT2A/5-HT2C receptors and enhanced release of 5-HT at adulthood.

As reviewed elsewhere in this volume (see Azmatia and Whitaker-Azmatia), the 5-HT1A receptor on astrocytes and glial cells regulates the release of the neurotrophic factor S-100B, which plays a crucial role in regulating the development of cortical neurons and astrocytes. Decreasing 5-HT1A release of these factors will have profound effects on cortical development, leading to stunted development of the cortical mantle. The role of the 5-HT1A receptor in maintaining cortical function in later life is not known and requires intensive study.

Roth et al. (98, 99) previously demonstrated that 5-HT2A and 5-HT2C receptors developed in a precisely regulated fashion. The principal increases in 5-HT2A and 5-HT2C receptors occurred postnatally, between postnatal days 7 and 14, with a subsequent decrease in 5-HT2A receptor and mRNA levels after this time. After postnatal day 21, little change in 5-HT2A or 5-HT2C receptors or mRNA levels were found. 5-HT2A receptor gene expression and receptor levels appeared to be unregulated by brain 5-HT levels; neonatal 5,7-dihydroxytryptamine lesions did not alter adult levels of 5-HT2 mRNA or receptors.

Taken together, these results suggest that the immediate perinatal period of the rat (which corresponds to the third trimester in humans) might represent a stage of vulnerability of the 5-HT2A/5-HT2C receptor system, and that dopaminergic, but not serotonergic, inputs could be important for regulating 5-HT2A receptor activity. Both DA and 5-HT may be important for the development of the cortical mantle. It is conceivable, then, that abnormalities in cortical anatomy and function and alterations in 5-HT1A and 5-HT2A receptors and receptor activity measured in schizophrenic patients could reflect disturbances during early brain development.


It is clear that role for 5-HT in the pathogenesis and treatment of schizophrenia is compelling. Quantitative differences in several 5-HT receptors have been found in schizophrenia by many (but not all) investigators. Neuroendocrine challenge studies are consistent with an altered sensitivity of 5-HT2A receptors. In addition, many typical and atypical antipsychotic agents bind with high affinity to 5-HT (particularly 5-HT2A) receptors. Alterations in serotonergic systems have been correlated with specific symptoms of schizophrenia, and novel antipsychotic agents which function as 5-HT2A antagonistss appear to be superior to neuroleptics for treating negative symptoms and treatment-resistant schizophrenia. Functional blockade of 5-HT2A receptors occurs in vivo in patients treated with atypical antipsychotic agents, and animal pharmacological studies are consistent with the notion that the 5-HT system may serve as one of the regulators of dopaminergic tone in vivo. Clearly, then, there is a large amount of internally consistent information to suggest a role for 5-HT in the etiology and/or treatment of schizophrenia. Future studies examining the role of multiple 5-HT receptors in the etiopathology and/or treatment of schizophrenia are likely to yield productive results.

published 2000