The role of serotonin receptors in the action of atypical antipsychotic drugs

The main class of atypical antipsychotic drugs (APDs) in current use includes the protypical atypical APD, clozapine, as well as aripiprazole, asenapine, iloperidone, lurasidone, olanzapine, quetiapine, risperidone, and ziprasidone. At clinically effective doses, these agents produce extensive blockade of serotonin (5-HT)_2A receptors, direct or indirect stimulation of 5-HT_1A receptors, and to a lesser extent, reduction in dopamine (DA) D₂ receptor-mediated neurotransmission. This contrasts with typical APDs, for example haloperidol and perphenazine, which are mainly DA D_2/D₃ receptor antagonists and have weaker, if any, potency as 5-HT_2A receptor antagonists. Some, but not all, atypical APDs are also effective 5-HT_2C receptor inverse agonists or neutral antagonists, 5-HT₆ or 5-HT₇ receptor antagonists. This diverse action on 5-HT receptors may contribute to significant differences in efficacy and tolerability among the atypical APDs. There is considerable preclinical and some clinical evidence that effects on 5-HT receptors contribute to the low risk of producing extrapyramidal side effects, which is the defining characteristic of an atypical APD, the lack of elevation in plasma prolactin levels (with risperidone and 9-hydroxyrisperidone being exceptions), antipsychotic action, and ability to improve some domains of cognition in patients with schizophrenia. The serotonergic actions of the atypical APDs, especially 5-HT_2A receptor antagonism, are particularly important to the differential effects of typical and atypical APDs to overcome the effects of acute or subchronic administration of N-methyl-d-aspartate (NMDA) receptor antagonists, such as phencyclidine, ketamine, and dizocipline (MK-801). 5-HT_1A receptor stimulation and 5-HT₆ and 5-HT₇ receptor antagonism may contribute to beneficial effects of these agents on cognition. In particular, 5-HT₇ receptor antagonism may be the basis for the pro-cognitive effects of the atypical APD, amisulpride, a D₂/D₃ receptor antagonist, which has no effect on other 5-HT receptor. 5-HT_2C receptor antagonism appears to contribute to the weight gain produced by some atypical APDs and may also affect cognition and psychosis via its influence on cortical and limbic dopaminergic activity.

Introduction

Antipsychotic drugs (APDs) are generally classified as either typical or atypical [1]. This classification is based on the increased liability of typical APDs to produce extrapyramidal side effects (EPS), including tardive dyskinesia (TD), at clinically relevant doses [2, 3]. Typical APDs are relatively selective dopamine (DA) D_2/D₃ receptor antagonists [4]. Antagonism of these DA receptors is the basis of both their efficacy and side effects [2, 4, 5]. Once the most widely used treatment for schizophrenia, this class of drugs is now seldom used in the US and other developed countries, having been supplanted by the atypical APDs, because of the latter's reduced risk of EPS and TD, leading to greater safety and tolerability [6]. It has been suggested that the atypical APDs are superior to typical APDs for improving some domains of cognition in schizophrenia [7, 8]. However, several recent studies have challenged these conclusions about the relative merits of these classes of drugs [9, 10, 11]. Clozapine, the prototypical atypical APD, has special efficacy over other atypical and typical APDs, which will not be discussed here. This article provides a brief discussion of the importance of serotonin (5-HT) to the mechanism of action of two classes of atypical APDs, the drugs that are characterized by high potency as 5-HT_2A receptor antagonists relative to their effect on D₂ DA receptors [12^•], and amisulpride, a substituted benzamide, which is now known to be a potent 5-HT₇ receptor antagonist [13^•] as well as a D_2/D₃ antagonist [14].

Medicinal chemistry efforts have yielded diverse series of compounds that conform to the pattern of receptor affinities noted above for the primary group of atypical APDs [15]. All three antipsychotic drugs approved in the US in the past two years: asenapine [16], iloperidone [17] and lurasidone [18], have greater potency as 5-HT_2A than D₂ receptor antagonists, consistent with the hypothesis that this profile is sufficient for an atypical APD [12^•]. Noteworthy, all three have high affinities for D₂ receptors, precluding the possibility of rapid D₂ receptor dissociation and inconsistent with the alternative ‘fast-off’ theory of atypical APDs proferred by Kapur and Seeman [19]. We suggest that this class of atypical APDs be referred to as serotonin spectrum dopamine modulators (SSDMs), because of the diversity of actions of various members of the group on 5-HT receptors, especially 5-HT_2A receptor antagonism and direct or indirect 5-HT_1A receptor agonism, but also 5-HT_2C, 5-HT₆ and 5-HT₇ receptor antagonism [20•, 21], and utilization of D₂/D₃ receptor antagonism or D₂/D₃ partial agonism, for example aripiprazole [14], to diminish DA D₂ receptor stimulation. The logs of the D₂(5-HT_2A/5-HT_1A) ratios and D₂(5-HT_2C/5-HT_1A) affinity ratios of eight atypical APDs were correlated with their effective clinical effective doses, consistent with the concept that actions at multiple 5-HT receptors and diminished D₂ DA receptor stimulation is important to the action of atypical APDs [22].

The importance of serotonergic mechanisms for the antipsychotic and pro-cognitive effects of atypical APDs is supported by many lines of evidence, particularly those which are based upon the hypoglutamate hypothesis of schizophrenia, now widely thought to better reflect the pathophysiology of schizophrenia than hyperdopaminergic models [23]. Disruption of corticolimbic glutamatergic neurotransmission by acute administration of N-methyl-D-aspartate (NMDA) receptor (NMDA-R) antagonists, for example phencyclidine (PCP), MK-801, or ketamine leads to intense, disorganized hyperlocomotor activity that is thought to model the positive symptoms of schizophrenia [23, 24•]. Subchronic administration of NMDA-R antagonists to rodents produces prolonged deficits in working and long term memory, attention, and social interactions, and has been suggested to model the cognitive impairment in schizophrenia [25, 26••]. It is now established that the SSDMs and selective 5-HT_2A receptor antagonists, for example M100907 and ACP-103 (pimavanserin), are more effective than typical APDs in blocking the NMDA-R antagonist model of psychosis and cognitive impairment [27•, 28, 29]. In contrast, typical APDs that are selective D₂ receptor antagonists, for example haloperidol, are much more potent at blocking dopaminergic stimulant-induced locomotor activity than the effects of NMDA-R antagonists [28]. The differential sensitivity of NMDA-R antagonism-induced hyperlocomotion to attenuation by atypical versus typical APDs is highly correlated with their relative degree of antagonism at D₂ and 5-HT_2A receptors [28]. This relationship is shown in Figure 1 [adapted from data in Millan et al. 1999; 28]. Plotting the ratio of the relative affinities of a number of antipsychotics, both typical and atypical, for D₂ and 5-HT_2A receptors against their relative potencies for inhibition of amphetamine-induced and PCP-induced hyperlocomotion, demonstrates a strong correlation between these actions (R = 0.92, P < 0.0001, Figure 1; Table 1). This analysis shows that the greater the ratio of affinities for 5-HT_2A over D₂ receptors for proven or putative antipsychotic drugs, the greater the relative potency for inhibition of PCP-induced versus amphetamine-induced hyperlocomotion. The more selective 5-HT_2A antagonists such as M100907 and SR46349B are the most potent antagonists for PCP-induced hyperlocomotion relative to amphetamine, while the reverse is true for more selective D₂ drugs such as haloperidol and amisulpride.

Selective 5-HT_2A receptor inverse agonists exhibit antipsychotic activity and potentiate both typical and atypical APDs in some preclinical models relevant to schizophrenia. Thus, the selective 5-HT_2A receptor inverse agonists M100907 and ACP-103, when administered alone, block the locomotor stimulating effects of NMDA-R antagonists [27•, 28, 29]. Raclopride, a D₂ receptor antagonist, in combination with M100907, but not M100907 alone, was effective in blocking the conditioned avoidance response, a classic model of antipsychotic activity [30]. ACP-103 potentiates the effect of haloperidol and risperidone to enhance cortical DA efflux in rats [31] and M100907 attenuated the increased DA efflux in the nucleus accumbens produced by haloperidol [32] or S-sulpiride [33]. The atypical APDs, lurasidone, melperone, N-desmethylclozapine, and risperidone, but not the typical APDs haloperidol or perphenazine, significantly reversed the impairment in novel object recognition (NOR), considered to be a model of declarative memory, produced by subchronic PCP treatment [34]. The selective 5-HT_2A receptor inverse agonists, ACP-103 and M100907, alone, at doses that fully occupy 5-HT_2A receptors [29, 35] did not reverse the effects of subchronic PCP on NOR [34]. However, the combination of ACP-103 or M100907, with subeffective doses of lurasidone, melperone, or risperidone, but not haloperidol, restored NOR performance in subchronic PCP-treated rats, suggesting that blockade of 5-HT_2A receptors contributes to the ability of atypical APDs to reverse the PCP-induced NOR deficit [34]. M100907 was ineffective when combined with haloperidol [34], indicating actions at receptors other than D₂ and 5-HT_2A receptors are required for reversal of the PCP-induced NOR deficit by the SSDMs, possibly other 5-HT receptor subtypes, as discussed later in this review [20^•]. Whether the importance of 5-HT_2A receptor blockade for the reversal of the effects of NOR in PCP-treated rats will generalize to other cognitive deficits remains to be studied. At best, these results in rodents are only partially mirrored by clinical data of the effect of the atypical APDs on declarative memory [7, 8, 11], the human cognitive function most similar to NOR.

Both subchronic PCP and clozapine treatment decrease 5-HT_2A receptor binding in frontal cortex. Morevoer, clozapine enhances this effect of PCP [35]. Decreased 5-HT_2A receptor binding has also been reported in cortex of post-mortem specimens from patients with schizophrenia [36]. A recent PET study found that antipsychotic drug-naive first epidsode schizophrenia patients have decreased binding potential for 5-HT_2A receptors in frontal cortex [37]. The binding potential was negatively correlated with severity of positive symptoms, but not cognitive function. Further, 60–70% occupancy of cortical 5-HT_2A receptors by quetiapine, an atypical APD, predicted improvement in positive symptoms [38^•]. Taken together, these findings are most consistent with the hypothesis that downregulation of 5-HT_2A receptors in the cortex of patients with schizophrenia is an adaptive mechanism [37, 39]. More evidence for the special advantages of SSDRs to amleiorate PCP-induced cognitive impairment is the report that clozapine, but not haloperidol, normalized DA turnover in dorsolateral prefrontal cortex, prelimbic cortex, and cingulate cortex, and reversed cognitive dysfunction due to PCP in vervet monkeys, seven days after 14 days treatment with PCP [40].

Clinical evidence also supports the importance of 5-HT_2A receptor blockade in the treatment of positive symptoms. Low doses of atypical APDs effectively treat l-DOPA psychosis without exacerbating motor symptoms, which typical APDs are unable to achieve, even at the lowest doses tried [41]. The efficacy of atypical APDs in l-DOPA psychosis has been suggested to be due to 5-HT_2A receptor blockade, which is supported by the ability of the selective 5-HT_2A receptor inverse agonist ACP-103 to accomplish the same results as clozapine and quetiapine [42]. Furthermore, in a double-blind, randomized trial of acutely psychotic patients, ACP-103 potentiated a subeffective dose of risperidone, 2 mg but not haloperidol, 2 mg (Meltzer et al. in preparation). SR43469B, another selective 5-HT_2A receptor antagonist, was more effective than placebo, but less effective than haloperidol, as monotherapy in acutely psychotic patients with schizophrenia [43].