Blockade of the serotonin (5-hydroxytryptamine; 5-HT) transporter is the common pharmacological action shared by selective serotonin reuptake inhibitors (SSRIs; eg fluoxetine, sertraline, fluvoxamine, paroxetine, citalopram). Enhancement of the synaptic availability of 5-HT appears to be a critical component of the mechanism underlying the action of SSRIs in the treatment of depression, as demonstrated by the tryptophan depletion paradigm (Delgado et al, 1990,1999). While long-term adaptations to blockade of monoamine reuptake must be present to explain the delay in antidepressant efficacy (Duman et al, 1997), the relatively rapid (<7 h) return of depressive symptoms in recently remitted patients during the monoamine depletion studies suggests that the long-term adaptations, like enhanced synaptic availability of monoamines, are necessary, but not sufficient, for a therapeutic response. Both enhanced synaptic availability of monoamines and long-term adaptations appear to be necessary for the therapeutic effect in most depressed patients during the initial month(s) of treatment. However, there are little clinical data regarding which of the known 15 5-HT receptors are involved in mediating the actions of SSRIs in the diverse range of neuro-qpsychiatric syndromes in which they have shown efficacy (eg major depression, obsessive–compulsive disorder (OCD), pervasive developmental disorders (PDDs) such as autism, panic disorder, and post-traumatic stress disorder (PTSD)).

Recently, addition of the 5-HT2A/dopamine D2 receptor antagonist risperidone at low doses to ongoing SSRI treatment has been shown in an open-label study to enhance therapeutic responses in patients with major depression (Ostroff and Nelson, 1999). Patients with treatment-refractory OCD have been shown to benefit from a similar strategy in a double-blind placebo-controlled study (McDougle et al, 2000)). It is not known whether monotherapy with low-dose risperidone might also possess therapeutic effects in these disorders. Like the SSRIs, risperidone has demonstrated some efficacy as a monotherapy in the treatment of PDDs (McDougle et al, 1998; Research Units on Pediatric Psychopharmacology Autism Network, 2002).

The only receptor subtype that risperidone is likely or known to saturate at these lower doses (0.5–1 mg/day) is the 5-HT2A receptor (Schotte et al, 1996; Nyberg et al, 1999; Talvik-Lofti et al, 2000). Such findings suggest the hypothesis discussed in detail below, that activation of 5-HT2A receptors secondary to the blockade of 5-HT transporters by SSRIs functionally opposes the therapeutic effects that are achieved by the activation of non-5-HT2A receptors. This review will discuss preclinical evidence for opposing effects of 5-HT2A and non-5-HT2A receptors that may contribute to a synergistic clinical action between blockade of 5-HT2A receptors and activation of non-5-HT2A receptors. Then, evidence for the clinical efficacy of 5-HT2A antagonists, both in monotherapy and in combination therapy with SSRIs, will be examined.


At a cellular level, activation of 5-HT2A and 5-HT1A receptors exerts depolarizing and hyperpolarizing effects, respectively, on cortical pyramidal cells (Araneda and Andrade, 1991; Tanaka and North, 1993; Ashby et al, 1994; Aghajanian and Marek, 1997). These interactions have been observed both under in vitro conditions and during in vivo recordings from the rodent medial prefrontal cortex. Similar interactions have also been observed at a behavioral level. For example, stimulation of 5-HT1A receptors suppresses head shakes in rats induced by hallucinogenic drugs, which activate 5-HT2A receptors (Arnt and Hyttel, 1989; Schreiber et al, 1995). Previous studies have demonstrated that systemic administration of 5-HT1A agonists can block head shakes induced by direct infusion of the hallucinogen 1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane (DOI) into the medial prefrontal cortex (Granhoff et al, 1992; Willins and Meltzer, 1997).

Evidence for opposing effects of activation of 5-HT2A and 5-HT1A receptors has also been obtained from a behavioral screen for antidepressant drugs, rats performing on a differential reinforcement of low rate 72-s (DRL 72-s) schedule (Marek et al, 1989a,1989b; Marek and Seiden, 1994; Jolly et al, 1999). This behavioral screen measures a cardinal feature of prefrontal cortical ‘executive function,’ that is, withholding inappropriate responses (Fuster, 1997). At an operational level, water-deprived animals in this paradigm must wait at least 72 s following the previous response in order to receive a reinforcer (water) for making a response. The efficacy of 5-HT antagonists in exerting an antidepressant-like response on this screen (increased reinforcement rate, decreased response rate, and a cohesive rightward shift in the inter-response time (IRT) histogram) is related to the selectivity of the antagonists for 5-HT2A relative to 5-HT1A receptors (Marek et al, 1989a; Marek and Seiden, 1994).

Interactions have also been observed between 5-HT2A and 5-HT2C receptors at a behavioral level. Opposing effects exist for 5-HT2A and 5-HT2C receptors in modulating hyperlocomotion induced in mice by noncompetitive N-methyl-D-aspartate (NMDA) antagonists (Martin et al, 1997). A recent study has found clear evidence supporting opposing effects of 5-HT2A and 5-HT2C receptors in modulating head shakes in rats (Vickers et al, 2001). Activation of 5-HT2A and 5-HT2C receptors also leads to opposing effects on sexual function in rodents (Berendsen et al, 1990), in a manner suggesting that activation of 5-HT2A receptors could be responsible for the sexual side effects of SSRIs. DRL 72-s behavior also appears to involve reciprocal interactions between 5-HT2A and 5-HT2C receptors, since 5-HT2C agonists, similar to 5-HT2A antagonists, exert antidepressant-like actions in rats performing under this operant paradigm (Marek and Seiden, 1988; Martin et al, 1998; Marek et al, 2001a). In a similar vein, blockade of 5-HT2C receptors may attenuate the antidepressant-like effects observed on the DRL 72-s schedule following administration of drugs that block 5-HT2A receptors (Marek and Seiden, 1994).

Opposing interactions may also occur between other 5-HT receptors and the 5-HT2A receptor. One prominent effect of activating the 5-HT2A receptor in the cortex appears to be the induction of glutamate release from thalamocortical afferents (Marek et al, 2001b). This anatomical substrate may be of interest with respect to major neuropsychiatric disorders because the likely source of these thalamocortical afferents for the prefrontal cortex, the intralaminar and midline thalamic nuclei project to other important limbic-related forebrain regions such as the ventral striatum (nucleus accumbens) and the amygdala, as well as the prefrontal cortex (Berendse and Groenewegen, 1990,1991; Su and Bentivoglio, 1990). 5-HT itself is able to suppress late excitatory postsynaptic potentials (EPSPs) induced by the combination of the 5-HT2 agonist DOI and electrical stimulation of the white matter (Aghajanian and Marek, 1999). Preliminary experiments have suggested that activation of 5-HT1-like, 5-HT5, or 5-HT7 receptors may mediate this effect, which opposes 5-HT2A receptor activation (G Marek, unpublished observations).


The actions of risperidone are considered before those of other 5-HT2 antagonists because this drug is the most potent and selective (with respect to other 5-HT receptors) 5-HT2A antagonist available to clinicians. Blockade of 5-HT2A receptors is the most potent action of risperidone as assessed by in vitro receptor-binding studies (Schotte et al, 1996; Richelson and Souder, 2000). This is hardly surprising given that risperidone was developed using blockade of the behavioral effects of hallucinogenic drugs and potency in binding to 5-HT2A receptors as two of the principal drug targets. Risperidone is approximately 20 to 50-fold more potent at binding to the 5-HT2A receptor than to the α1-adrenergic, dopamine D2, histamine H1, and α2-adrenergic receptors ( Table 1). It is one of the most selective compounds clinically available with respect to selectivity at the 5-HT2A vs the 5-HT2C receptor. The 1000-fold selectivity of risperidone for the 5-HT2A vs the 5-HT1A receptor may also be relevant to the drug's therapeutic actions in mood disorders, OCD, and PDDs.

Table 1 In Vitro Receptor Potency of Risperidone, Olanzapine, and Mirtazapine (KI, nM)

The dose of risperidone reported to be effective for depression (0.5–1 mg/day) is in contrast to the doses of 3–6 mg/day used in schizophrenia. Risperidone at 4 mg is thought to produce 70–80% occupancy of striatal dopamine D2 receptors (Nyberg et al, 1999). Higher doses increase the likelihood of extrapyramidal side effects. In contrast, the 0.5–1 mg dose of risperidone probably saturates the 5-HT2A receptors. The previous lower estimates for cortical 5-HT2A receptor occupancy using N-methyl-spiperone as the radiotracer are now thought to be an underestimate of the true occupancy (Talvik-Lofti et al, 2000). Animal studies using ex vivo autoradiographic techniques have supported the increased potency of risperidone for displacement of binding to 5-HT2A receptors relative to histamine H1, dopamine D2, α1-adrenergic, and α2-adrenergic receptors (Schotte et al, 1996). Extrapolation of these results to humans would suggest that only the 5-HT2A receptor would show appreciable occupation at risperidone doses of 0.5–1 mg.


Several open-label reports with a total of 12 patients have suggested that addition of low-dose risperidone (0.5–1 mg/day) to ongoing treatment with SSRIs may have clinically significant antidepressant action in patients with major depression (O'Connor and Silver, 1998; Ostroff and Nelson, 1999). The report by Ostroff and Nelson is notable for the rapid (1–2 days) antidepressant effect of risperidone addition in the majority of patients. Another open series of five patients suggested that addition of risperidone (0.5–1.5 mg/day) to the monoamine oxidase inhibitor (MAOI) tranylcypromine resulted in a rapid antidepressant effect in patients lacking a satisfactory response to the MAOI alone (Stoll and Haura, 2000). Our own experience with depressed patients has been that the addition of risperidone to ongoing treatment with SSRIs does rapidly induce a therapeutic response, but that this response is not generally sustained (LL Carpenter, LH Price, unpublished observations). These open-label reports will require confirmation from double-blind placebo-controlled studies with respect to response efficacy, speed of response onset, and maintenance of therapeutic efficacy.

As in the depression literature, a number of open-label case series have suggested that addition of risperidone results in a therapeutic response in patients with OCD who have failed to obtain a satisfactory response to monotherapy with SSRIs after an appropriate duration of treatment (Jacobsen, 1995; McDougle et al, 1995b; Ravizza et al, 1996; Saxena et al, 1996; Stein et al, 1997). Recently, these earlier reports have been confirmed by a double-blind placebo-controlled study in which the addition of risperidone in a mean dose of 2.2 mg/day to ongoing SSRI treatment resulted in clinically significant improvement in refractory OCD patients (McDougle et al, 2000). The beneficial effects of risperidone did not appear to be limited to patients with chronic motor tics or schizotypal personality disorder, which had appeared to be the case with typical neuroleptic drugs (McDougle et al, 1994). Analyses of drug levels in a subset of patients did not reveal any evidence that risperidone might be enhancing the effects of the SSRIs via a pharmacokinetic, rather than pharmacodynamic, interaction.

In all, 14 anecdotal reports and open-label studies of risperidone in children, adolescents, and adults through 1998 have suggested that this atypical neuroleptic is effective in ameliorating repetitive behavior and aggression towards oneself, others, and property in patients with autism and other PDDs (McDougle et al, 1998). A double-blind placebo-controlled study in adults with autism and other PDDs has confirmed these earlier reports (McDougle et al, 1998). Monotherapy with risperidone in a mean dose of 2.9 mg/day decreased aggression, anxiety, depression, irritability, and repetitive behaviors, although core symptoms involving aberrant social behavior and language were not affected. While this dose of risperidone approaches doses currently recommended for the treatment of schizophrenia and would result in a significant occupation of dopamine D2 receptors, it is notable that acute dietary tryptophan depletion (McDougle et al, 1996) worsened symptoms in adult patients with autistic disorder. Specifically, an increase in repetitive behavior, depression, and anxiety was observed, without any change in social behavior or language. This and other evidence of the involvement of 5-HT in the pathophysiology of PDDs increases the likelihood that the therapeutic action of risperidone in these disorders might also involve 5-HT. In a recently completed multisite controlled trial, risperidone was significantly better than placebo for reducing self-injury, aggression, and agitation in autistic children and adolescents (Research Units on Pediatric Psychopharmacology Autism Network, 2002). The effect size for risperidone in this study was greater than what has been reported with the D2 antagonist haloperidol in this patient group.


Olanzapine has a more complicated profile of G-protein-coupled receptor targets than risperidone. In essence, the ‘dirty’ pharmacological profile of clozapine served as a template for designing an effective atypical neuroleptic medication free of major adverse effects, such as agranulocytosis and seizures. Olanzapine potently blocks 5-HT2A, 5-HT2C, and histamine H1receptors with a low nanomolar affinity ( Table 1). Olanzapine is four-to-ten-fold less potent at dopamine D2, α1-adrenergic, and muscarinic receptors than at the 5-HT2A receptor (Bymaster et al, 1996; Schotte et al, 1996). This mild selectivity for 5-HT2A receptors compared to dopamine D2 receptors is observed in human positron emission tomography (PET) neuroimaging studies at low olanzapine doses (eg 5 mg), but is lost at higher doses (20 mg) (Kapur et al, 1999). Unlike mirtazapine and mianserin (see below, Table 1), olanzapine is approximately 189-fold less potent at displacing binding to α2-adrenergic receptors than to 5-HT2A receptors. Olanzapine, similar to the other 5-HT2 antagonists used to augment SSRIs, is also about 412-fold less potent at displacing binding to 5-HT1A vs 5-HT2A receptors.


The atypical neuroleptic olanzapine has been found to augment the effects of the SSRI fluoxetine in depressed patients who were refractory to SSRI monotherapy (Shelton et al, 2001). The most striking finding from this double-blind, placebo-controlled study was that the greatest degree of improvement in the olanzapine+fluoxetine group occurred within the first week of treatment, while patients were receiving only 5 mg/day of olanzapine. This suggests an acceleration in the onset of action in addition to enhancement of efficacy. Notably, this dose of olanzapine would be expected to saturate about 98% of 5-HT2A receptors while saturating only about 55% of dopamine D2 receptors (Kapur et al, 1999).

Several open-label studies have also found that a relatively low dose of olanzapine (5 mg) improves the symptoms of patients with OCD who were refractory to SSRIs (Weiss et al, 1999; Bogetto et al, 2000). However, all of the patients from Bogetto et al and 5/8 positive responders from Weiss et al were treated with fluvoxamine, which could have elevated olanzapine levels by interactions with cytochrome P450 microsomal isoenzymes. Thus, pharmacokinetic data and experience with different SSRIs will be necessary to conclude that this effect reflects a pharmacodynamic interaction. Furthermore, the utility of olanzapine as an augmenting strategy in OCD may be compromised by sedation in this patient group (Weiss et al, 1999). In contrast to studies of olanzapine in the treatment of depression and OCD, there are no published double-blind studies concerning this drug's efficacy in the treatment of autism and other PDDs. However, an open-label case series described positive responses to olanzapine (7.8 mg/day) in patients with autistic disorder and other PDDs (Potenza et al, 1999). In a more recently published study employing a parallel-groups design, 12 children with autistic disorder were randomized to 6 weeks of open-label olanzapine or haloperidol (Malone et al, 2001). Mean final dosages were 7.9 mg/day for olanzapine and 1.4 mg/day for haloperidol. Five of six subjects in the olanzapine group and three of six in the haloperidol group were responders.


Unlike risperidone, the most potent pharmacological action of mirtazapine and mianserin is blockade of histamine H1 receptors, which is thought to play a role in the sedative side effects of these drugs (Table 1). Their second most potent pharmacological action is blockade of the 5-HT2 family of receptors. Mirtazapine and mianserin appear to block the 5-HT2A, 5-HT2C, and α2-adrenergic receptors with approximately equal affinity. Both drugs are potent antagonists, although at slightly lower affinities, of 5-HT3 receptors. Like risperidone, mirtazapine and mianserin are at least 300-fold more selective for blocking 5-HT2A vs 5-HT1A receptors. Both agents lack appreciable affinity for dopamine D2 receptors. Blockade of α2-adrenergic receptors by mirtazapine has attracted significant interest as a potentially important mechanism of therapeutic action by increasing synaptic concentrations of 5-HT in the hippocampus (De Boer, 1996; Haddjeri et al, 1997). Specifically, blockade of α2-adrenergic heteroceptors on serotonergic terminals in the prefrontal cortex is thought to play a role in alterations of hippocampal 5-HT. However, it should be noted that mirtazapine does not increase extracellular 5-HT levels in the prefrontal cortex, whereas norepinephrine and dopamine levels are increased (Millan et al, 2000). Furthermore, mianserin does not appear to possess appreciable affinity for the same adrenergic autoreceptor as does mirtazapine (De Boer et al, 1996). Hjorth and colleagues (Bengtsson et al, 2000) failed to observe increased extracellular 5-HT levels in the prefrontal cortex, the ventral hippocampus, or the dorsal hippocampus in response to either mirtazapine or idazoxan, at doses that did clearly block α2-adrenergic responses. Finally, there are no double-blind placebo-controlled studies demonstrating clinical antidepressant action for an α2-adrenergic antagonist lacking biologically significant 5-HT2A receptor affinity. Evidence reported thus far for the selective α2-adrenergic antagonist idazoxan suggests that the suspected beneficial effects of this drug for mood disorders may be mostly restricted to bipolar depression (Osman et al, 1989; Potter et al, 1994; Grossman et al, 1999). Taken together, it is far from clear that the beneficial effects from combining mirtazapine or mianserin with SSRIs in depressed patients (discussed below) is due to α2-adrenergic blocking activity of these compounds. Blockade of 5-HT2 receptors is an alternative hypothesis to explain the clinical reports.


Augmentation of SSRIs with 5-HT2 antagonists lacking appreciable dopamine D2 blocking properties has been investigated. An initial open-label trial reported significant improvement in nearly half of 20 depressed patients in whom mirtazapine was added to ongoing treatment with SSRIs, venlafaxine, desipramine+SSRI, or desipramine+venlafaxine (Carpenter et al, 1999). A follow-up double-blind placebo-controlled study corroborated the earlier open-label trial (Carpenter et al, 2002). In the controlled study, 84% of the moderate to severely depressed patients randomized either to placebo or mirtazapine (15–30 mg/day) were on SSRIs for a minimum of 4 weeks but did not have a satisfactory response. Using a 50% decrease in the Hamilton Depression Rating Scale (HDRS) as the primary outcome measure, 64% of the mirtazapine-treated patients were considered responders, compared with only 20% of the placebo-treated patients.

Efficacy in enhancing the action of SSRIs in depressed patients has also been examined for mianserin, a compound with a similar chemical structure as mirtazapine. Maes et al (1999) studied the combination of the SSRI fluoxetine (20 mg/day) with mianserin (30 mg/day) compared to fluoxetine alone in an in-patient setting using a double-blind controlled design. Using a 50% decrease in HDRS ratings as the outcome criterion, they observed a 60% response rate in patients given the fluoxetine+mianserin combination from day 1 compared to a 9% response rate in patients treated with fluoxetine monotherapy. A significant separation between groups was observed as early as the first week. A double-blind placebo-controlled comparison of mianserin (30 mg/day)+fluoxetine (20 mg/day) vs placebo+fluoxetine found a significant advantage for the combined treatment group only when excluding the drop-outs from the first 2 weeks of the 6-week trial (Dam et al, 1998).

Mianserin may also enhance the efficacy of SSRIs in refractory patients as well. A double-blind, placebo-controlled study comparing addition of mianserin (60 mg/day) or placebo to fluoxetine (20 mg/day) or placebo in patients with ≥25 score on the 17-item Hamilton Depression Rating (Ham-D) Score observed a significant 5-point advantage for the mianserin+fluoxetine group compared to the placebo+fluoxetine group. The mianserin+placebo group was intermediate between the other two groups. The remission rate was also over twice as great in the mianserin+fluoxetine group vs the placebo+fluoxetine group (Ferreri et al, 2001). A recent double-blind study (mianserin+sertraline vs placebo+sertraline) failed to observe beneficial effects of adding mianserin (30 mg/day) to sertraline (100 mg/day) in patients failing to respond to an initial open 6-week treatment with sertraline (50 mg/day×4 weeks; 100 mg/day×2 weeks; Licht and Qvitzau, 2002). Differences in depression severity may have contributed to these differences. The 17-item HAM-D scores required for study entry was ≥25 for the fluoxetine augmentation study after initial fluoxetine treatment (Ferreri et al, 2001), whereas the entry criterion was ≥18 for the sertraline augmentation study before the initial sertraline treatment (Licht and Qvitzau, 2002).

Literature regarding the efficacy of mirtazapine in OCD is limited to a single open-label trial. Only two of 10 patients (or 2/6 completers) improved during a 10-week course of treatment with mirtazapine (Koran et al, 2001). Four patients in the study had not had a previous unsuccessful trial with SSRIs, and both of the patients who improved were in this subgroup. To date, there have been no published studies in which mirtazapine was used to augment SSRI monotherapy in treatment-refractory OCD.

A recent open-label study of mirtazapine (mean dose, 32 mg/day) in children, adolescents, and young adults with PDDs (aged 4–24 years) found a 35% response rate in 26 patients, similar to the response rate generally observed in this population with SSRIs (Posey et al, 2001). This was a naturalistic study in which the nine mirtazapine responders were on either no concomitant medications (four patients); methylphenidate (two patients); clonidine and pimozide (one patient); paroxetine and risperidone (one patient); or clonazepam and guanfacine (one patient). As with risperidone (see above), clinical improvement was noted for aggression, self-injury, irritability, hyperactivity, anxiety, depression, and insomnia, with no improvement in core symptoms of socialization or communication impairment. Confirmation of these preliminary results under double-blind placebo-controlled conditions would support the hypothesis that blockade of 5-HT2A receptors plays an important role in the therapeutic effects of risperidone for autistic patients. The efficacy of 5-HT2 antagonists in combination with SSRIs for the treatment of PDDs remains to be explored.


While SSRIs have become the mainstay for the pharmacological treatment of major depression, a number of drugs are effective in monotherapy minus the ability to appreciably block either monoamine uptake or monoamine oxidase. These ‘atypical’ drugs (mirtazapine, mianserin, nefazodone, and trazodone) have been shown to have antidepressant efficacy in multiple double-blind, placebo-controlled studies (Pinder and Fink, 1982; Marek et al, 1992; Davis et al, 1997; Fawcett and Barkin, 1998). The common potent pharmacological target for these agents is blockade of the 5-HT2 family of receptors. Interestingly, there have been no controlled studies published suggesting that nefazodone or trazodone might enhance in a synergistic fashion the therapeutic effects of SSRIs. If the 5-HT1A and 5-HT2C receptors are important targets that SSRIs indirectly activate via blockade of the 5-HT transporter, then the lack of selectivity of nefazodone and trazodone at the 5-HT1A or the 5-HT2C vs the 5-HT2A receptors may limit the usefulness of these drugs as augmenting agents for patients refractory to SSRIs (Wander et al, 1986; Eison et al, 1990; Richelson, 2001). In addition to the well-known ‘atypical’ antidepressants listed above, there are double-blind studies demonstrating antidepressant activity of other potent 5-HT2 antagonists as well, such as ritanserin, pipamperone, pizotifen, danitracen, and etoperidone (Matussek et al, 1976; Ansoms et al, 1977; Standal, 1977; Kern and Richter, 1985; Bersani et al, 1990).

In contrast to the preliminary evidence that blockade of 5-HT2 receptors may augment the therapeutic effects of SSRIs in OCD, evidence from monotherapy with potent 5-HT2 antagonists in this disorder is generally negative. For example, the atypical neuroleptic clozapine, at doses (400–500 mg/day) that would provide significant blockade of 5-HT2A receptors (Nordstrom et al, 1993), did not improve any of 10 treatment-refractory patients participating in a 10-week open-label trial (McDougle et al, 1995a). Clinical experience with other available 5-HT2 antagonists in the treatment of OCD has been similarly unrewarding.

In contrast to the apparent benefit of atypical neuroleptic drugs for SSRI-refractory patients with OCD, a number of case reports (generally involving schizophrenic patients) have suggested that addition of atypical neuroleptic drugs can worsen obsessive–compulsive symptoms while simultaneously improving psychotic symptoms (Lykouras et al, 2000; Khullar et al, 2001; De Haan et al, 2002). The magnitude of this clinical problem remains to be fully explored as clozapine may be implicated more frequently than other antipsychotic drugs (De Haan et al, 1999). Obsessive–compulsive symptoms have been reported to be improved in schizophrenic patients treated with olanzapine (Poyurovsky et al, 2000). Addressing both the issue of clozapine and the primary psychiatric diagnosis, the open-label trial of clozapine in treatment-refractory OCD patients failed to observe worsening of obsessive–compulsive symptoms for any of the 10 patients treated for 10 weeks (McDougle et al, 1995a). The frequency for which exacerbation of obsessive–compulsive symptoms is observed in schizophrenic patients remains to be quantified. Thus, primary OCD vs obsessive–compulsive symptoms in patients with other diagnoses may be an important factor in whether addition of 5-HT2A antagonists improves patients' refractory to SSRI treatment.

The improvement of aggression in autism by both risperidone and mirtazapine is suggestive that blockade of 5-HT2A receptors may play a role in this response. This is interesting in light of a double-blind placebo-controlled crossover trial in which the selective 5-HT2A antagonist pipamperone decreased episodes of anger, aggressiveness, and impulsivity, and increased alertness and responsiveness in adult females with mental retardation (van Hemert, 1975). While this report did not specify the clinical diagnoses of the patients, PDDs are known to be over-represented in the population of mentally retarded individuals with prominent behavioral disturbances. This efficacy in treating aggression is consistent with previous open-label trials of pipamperone in diverse groups of patients (DeBerdt, 1976; Van Renynghe de Voxvrie and De Bie, 1976; Noordhuizen, 1977; Haegeman and Duyck, 1978). Pipamperone has since been shown to have significant in vitro selectivity as an antagonist at the 5-HT2A receptor compared to other receptors (dopamine D2, 20-fold; dopamine D3, 46-fold; 5-HT1A, 512-fold; 5-HT1Dα, 30-fold; 5-HT2C, 100-fold; histamine H1, 444-fold; α2A-adrenergic, 159-fold; α2B-adrenergic, 6.5-fold; α2C-adrenergic, 54-fold) (Schotte et al, 1996).


The central thesis argued above is that 5-HT2A receptors actually oppose the therapeutic effects of activating non-5-HT2A receptors in diverse neuropsychiatric syndromes such as depression, OCD, and PDDs. This broad range of clinical effects may be caused by localization of 5-HT2A receptors at critical sites in the prefrontal cortex regulating cortico-striatal-pallidal-thalamic loops. The target receptor(s) that actually mediate the therapeutic effects of increased synaptic 5-HT caused by SSRIs is far from clear and may be different for these and other psychiatric syndromes.

5-HT1A receptors may be involved in both the treatment (Pineyro and Blier, 1999) and the pathophysiology of major depression (Arango et al, 1995; Stockmeier et al, 1998). There have been some positive double-blind placebo-controlled studies with the azapirone class of drugs, such as buspirone and gepirone, which are partial 5-HT1A agonists that are not as effective as 5-HT itself in maximally activating postsynaptic 5-HT1A receptors (Robinson et al, 1990; Jenkins et al, 1990). However, these agents are not used as front-line monotherapeutic agents for patients with mood disorders. Furthermore, no positive double-blind placebo-controlled reports of antidepressant activity in depressed patients have been reported for the 5-HT1A partial agonists tandospirone and flesinoxan. A large multicenter, double-blind, placebo-controlled trial for ipsapirone failed to observe a significant antidepressant response, despite a placebo response rate of only 35% (Lapierre et al, 1998). Finally, it should be noted that no 5-HT1A partial agonists have been approved by the FDA for the treatment of major depression. It is not clear if the relative lack of success with 5-HT1A agonists in depression and OCD is caused by the following: (1) lack of compounds with sufficient agonist efficacy for postsynaptic receptors, (2) unfavorable pharmacokinetics of presently available drugs, or (3) inadequate therapeutic benefit from stimulation of 5-HT1A receptors alone.

The clinical overlap between migraine headache and major depression also raises the question of whether 5-HT1B/1D agonists might be useful in depression. Again, a central problem with 5-HT1B/1D receptors, as with 5-HT1A receptors, is that optimal clinical action might require blockade of presynaptic autoreceptors on the axon terminals of serotonergic neurons and activation of heteroceptors on the axon terminals of nonserotonergic neurons. Regarding presynaptic autoreceptor antagonists, the lack of efficacy of the 5-HT1A somatodendritic autoreceptor antagonist pindolol in augmenting the effects of SSRIs in SSRI-resistant depression (Perez et al, 1999) may simply mean that 5-HT1B terminal autoreceptors must also be targeted when attempting to increase serotonergic neurotransmission by blocking the 5-HT1A autoreceptors. Moreover, it has been argued that pindolol itself is a suboptimal tool for blocking 5-HT1A somatodendritic autoreceptors on serotonergic neurons (Kinney et al, 2000). Significant potential remains for potent and selective blockade of somatodendritic (5-HT1A) and terminal (5-HT1B/1D) autoreceptors on serotonergic neurons as a means to enhance the efficacy of SSRIs.

Activation of 5-HT2C receptors appears to cause effects that functionally oppose the effects resulting from activation of 5-HT2A receptors (see above). These two receptors are differentially regulated by antidepressants in a manner that could have clinical relevance (Berendsen and Broekkamp, 1991). Other preclinical work in rodents has found that 5-HT2C agonists have antidepressant-like action in the forced swim test, the olfactory bulbectomy model, and the DRL 72-s schedule (Martin et al, 1998; Cryan and Lucki, 2000). Moreover, a polymorphism for the 5-HT2C receptor has been linked to mood disorders (Lerer et al, 2001). This might result in an alteration of 5-HT2C function independent of receptor abundance. Alterations in RNA editing of 5-HT2C receptors has been reported in suicide victims (Niswender et al, 2001). Thus, simultaneous blockade of 5-HT2A receptors and activation of 5-HT2C receptors could result in an improved therapeutic benefit over either of these actions in isolation.

Blockade of 5-HT2C receptors could contribute to the enhanced efficacy of augmentation or combination therapy with mirtazapine, mianserin, or olanzapine. If blockade of 5-HT2C receptors mediates the antidepressant augmenting effects of 5-HT2 antagonists, then one might expect an SSRI with potent 5-HT2C receptor antagonist potency to have a faster onset of action. Fluoxetine has been suggested from extrapolations of preclinical data to block 5-HT2C receptors at doses used in the clinic (Bymaster et al, 2002). However, meta-analysis suggests that fluoxetine may have a slower onset of antidepressant action than other SSRIs (Edwards and Anderson, 1999). It remains to be determined if this is because of pharmacodynamic and/or pharmacokinetic considerations. Alternatively, suggestions that risperidone or pipamperone may augment the action of antidepressant drugs at doses with a measure of selectivity for blockade of 5-HT2A receptors might reflect the complex actions of 5-HT in mediating the therapeutic actions of antidepressant drugs (Ostroff and Nelson, 1999; Ansoms et al, 1977). In fact, the failure of the pharmaceutical industry to produce an effective antidepressant that selectively activates a single 5-HT receptor is consistent with the likelihood that indirect activation of multiple 5-HT receptor subtypes mediates the therapeutic actions of SSRIs.

The 5-HT5A/5B receptor is still poorly understood; indeed, its in situ postreceptor transduction pathway is not known. This receptor is most closely related to some of the 5-HT1 receptor subtypes on a pharmacological basis. Furthermore, in vitro studies have demonstrated that the 5-HT5A receptor, like the 5-HT1 family of receptors, can couple to Gi/Go proteins (Francken et al, 1998). The 5-HT5A/5B receptors are not known to be a target of antidepressant or neuroleptic drugs (Matthes et al, 1993). There is presently little data to indicate whether 5-HT5A/5B receptors mediate any of the clinical effects of increased synaptic 5-HT subsequent to blockade of 5-HT transporters. Interestingly, a recent report has implicated a polymorphism for this receptor in altered susceptibility to schizophrenia and mood disorders (Birkett et al, 2000).

There is some evidence suggesting that activation of 5-HT6 receptors might oppose the effects of 5-HT2A receptor activation in the frontal cortex. Activation of 5-HT2A receptors increases glutamate release from thalamocortical afferents, as discussed above. In contrast, blockade of 5-HT6 receptors has recently been shown to increase extracellular levels of glutamate measured using in vivo dialysis in the prefrontal cortex, hippocampus, and striatum (Dawson et al, 2000,2001). Activation of 5-HT6 receptors might, therefore, be expected to suppress extracellular levels of glutamate in the frontal cortex. Consistent with the hypothesis that functional relationships exist between 5-HT6 receptors and glutamate, blockade of ionotropic glutamate receptors decreases the expression of 5-HT6 receptor mRNA (Healy and Meador-Woodruff, 1999).

5-HT7 agonists might also be in functional opposition to the effects of 5-HT2A receptor activation in the prefrontal cortex. For example, high concentrations of 8-OH-DPAT (10 μM) suppressed delayed late EPSPs induced by concurrent activation of 5-HT2A receptors and electrical stimulation of the white matter in a delayed manner with only partial efficacy followed by complete recovery (G Marek, unpublished observations). However, effects of 8-OH-DPAT at 5-HT1A receptors in slice preparations are usually brisk in onset and only slowly fade away. Thus, these 8-OH-DPAT responses do not appear to be mediated by activation of 5-HT1A receptors. This action of 8-OH-DPAT was probably the result of activating a 5-HT1D/5A/5B/7 receptor. As discussed for the 5-HT6 receptor, blockade of ionotropic glutamate receptors also alters 5-HT7 receptor mRNA exprression (Healy and Meador-Woodruff, 1999). Further support for the hypothesis that activation of a Gqs-coupled receptor (positively coupled to adenylyl cyclase) might actually suppress glutamate release induced by 5-HT2A receptor activation comes from experiments in which the suppressant action of epinephrine or norepinephrine on DOI-induced late EPSPs is blocked by β2-adrenergic antagonists (B Ramos and G Marek, unpublished observations). β2-Adrenergic, like 5-HT4/6/7 receptors are well-known to be coupled to Gqs proteins. Activation of members of the 5-HT or adrenergic receptor superfamily that increase cAMP formation (eg 5-HT4/6/7), with subsequent upregulation of neurotrophins and transcription factors, has also been proposed to play an important role in the long-term effects of antidepressant drugs (Duman et al, 1997). Thus, activation of 5-HT6/7 receptors might have both acute and long-term effects that could contribute to the therapeutic actions of SSRIs.


SSRIs have been found to exert clinical efficacy in a wide variety of psychiatric syndromes, including major depression, OCD, autism and other PDDs, panic disorder, and PTSD. Drugs that block the 5-HT2 family of receptors, such as risperidone, olanzapine, mirtazapine, and mianserin, have been found to either augment the action of SSRIs in treatment-refractory patients or to demonstrate therapeutic efficacy as stand-alone agents. The most parsimonious explanation for these findings is that blockade of 5-HT2A receptors and activation of non-5-HT2A receptors may have similar effects. Further evidence relevant to this hypothesis will come from the systematic examination of other ‘atypical’ neuroleptics with respect to their clinical profile of action across diverse psychiatric syndromes. For example, ziprasidone is even more selective for the 5-HT2A receptor vs dopamine D2 or 5-HT2C receptors than is risperidone or olanzapine (Schotte et al, 1996; Tandon et al, 1997). Unlike most other neuroleptics, ziprasidone is a 5-HT1A partial agonist (Sprouse et al, 1999) that also blocks monoamine transport with sub-μM potency. This drug is relatively weak at displacing binding to histamine H1 receptors compared to other ‘atypical’ neuroleptics. A further difference from other ‘atypical’ compounds is the high potency that ziprasidone possesses for the 5-HT1D receptor, and blockade of the 5-HT1D autoreceptor enhances the effects of SSRIs on extracellular 5-HT levels (Rollema et al, 1996).

Support for the hypothesis that blockade of 5-HT2A receptors coincident with activation of non-5-HT2A receptors results in more robust clinical activity than either drug alone could come from clinical testing of highly selective 5-HT2A antagonists, such as M100907 (Kehne et al, 1996), in combination with SSRIs. Evaluation of novel compounds that combine equal and potent blockade of both 5-HT2A receptors and 5-HT transporters would also provide critical data (Puller et al, 2000; Schmidt et al, 2001). An appreciation of the opposing effects of different 5-HT receptor subtypes in mediating the therapeutic effects of drugs will be important in better defining the neurocircuitry involved in the pathogenesis of these disorders, and in developing treatments with more rapid onset and greater efficacy.