Since the introduction of chlorpromazine and throughout the development of the new-generation antipsychotic drugs (APDs) beginning with clozapine, the D2 receptor has been the target for the development of APDs. Pharmacologic actions to reduce neurotransmission through the D2 receptor have been the only proven therapeutic mechanism for psychoses. A number of novel non-D2 mechanisms of action of APDs have been explored over the past 40 years but none has definitively been proven effective. At the same time, the effectiveness of treatments and range of outcomes for patients are far from satisfactory. The relative success of antipsychotics in treating positive symptoms is limited by the fact that a substantial number of patients are refractory to current medications and by their lack of efficacy for negative and cognitive symptoms, which often determine the level of functional impairment. In addition, while the newer antipsychotics produce fewer motor side effects, safety and tolerability concerns about weight gain and endocrinopathies have emerged. Consequently, there is an urgent need for more effective and better-tolerated antipsychotic agents, and to identify new molecular targets and develop mechanistically novel compounds that can address the various symptom dimensions of schizophrenia. In recent years, a variety of new experimental pharmacological approaches have emerged, including compounds acting on targets other than the dopamine D2 receptor. However, there is still an ongoing debate as to whether drugs selective for singe molecular targets (that is, ‘magic bullets’) or drugs selectively non-selective for several molecular targets (that is, ‘magic shotguns’, ‘multifunctional drugs’ or ‘intramolecular polypharmacy’) will lead to more effective new medications for schizophrenia. In this context, current and future drug development strategies can be seen to fall into three categories: (1) refinement of precedented mechanisms of action to provide drugs of comparable or superior efficacy and side-effect profiles to existing APDs; (2) development of novel (and presumably non-D2) mechanism APDs; (3) development of compounds to be used as adjuncts to APDs to augment efficacy by targeting specific symptom dimensions of schizophrenia and particularly those not responsive to traditional APD treatment. In addition, efforts are being made to determine if the products of susceptibility genes in schizophrenia, identified by genetic linkage and association studies, may be viable targets for drug development. Finally, a focus on early detection and early intervention aimed at halting or reversing progressive pathophysiological processes in schizophrenia has gained great influence. This has encouraged future drug development and therapeutic strategies that are neuroprotective. This article provides an update and critical review of the pharmacology and clinical profiles of current APDs and drugs acting on novel targets with potential to be therapeutic agents in the future.
Schizophrenia is a chronic and debilitating disorder with a lifetime prevalence near 1%.1 It is characterized by positive, negative, cognitive and affective symptoms, and may arise from neurodevelopmental and neurodegenerative pathophysiologic processes.2, 3 The serendipitious discovery of chlorpromazine in the early 1950s and development of clozapine in the late 1960s (with its reintroduction in the United States in 1989) represented two major milestones in the pharmacotherapy of schizophrenia.4 During the past half century, numerous first- (FGAs), second- and third-generation antipsychotics (SGAs, TGAs) were developed and dramatic growth of research in the area of pharmacological treatment of schizophrenia has advanced our understanding of the neurobiology and neuropharmacology of the illness.5, 6 However, the precise etiology of schizophrenia, including genetic and environmental diatheses, remains poorly understood.1
Although existing antipsychotic medications are often effective for treating positive symptoms, they have little impact on negative symptoms and cognitive deficits.7, 8 These symptom domains are recognized as core features of schizophrenia and their lack of responsiveness to treatment contribute to poor functional outcome.9 Furthermore, in a substantial number of patients, positive symptoms are resistant to currently available medications.10 Consequently, there is an urgent need for more effective and better-tolerated antipsychotic agents, and to develop mechanistically novel compounds that possess pharmacological activity for novel targets which address the various symptom dimensions of schizophrenia.
In recent years, a variety of new experimental pharmacological approaches have emerged and compounds acting on targets other than the dopamine D2 receptor for which clinical testing has begun.11, 12 Moreover, there has been a rapidly growing body of research on approaches for enhancing cognition and for improving negative symptoms in schizophrenia, either as monotherapies or as adjunctive treatments added to currently available antipsychotics.13, 14 To assess progress in the development of treatments for schizophrenia, this article will provide a critical review of the pharmacology and clinical profiles of current antipsychotic drugs (APDs) and on novel targets for future therapeutic agents.
Theories of mechanisms of action of APDs
Dopamine receptor modulation
The mechanism of action of APDs is based on the hypothesis that schizophrenia involves a dysregulation of neurotransmission in brain dopaminergic circuits with excess dopaminergic activity in the mesolimbic pathway and reduced dopaminergic signaling in the mescortical pathway.15 Based on this model, the antagonism of D2 receptors in the mesolimbic pathway will produce reductions in dopamine activity and psychotic symptoms.16 The dopamine D2 receptor is regarded as the primary target associated with therapeutic antipsychotic effects, as well as with the induction of extrapyramidal side effects (EPS) and prolactin elevation. All clinically approved and currently used APDs have nanomolar affinity for the D2 receptor and fully or partially block the actions of dopamine.11 D2 receptors mediate their physiological actions through both G-protein-dependent and independent (the scaffolding protein β-arrestin 2-dependent) signaling.17 Masri et al.18 have recently suggested that both FGAs and SGAs may share a common molecular mechanism of blocking dopamine-mediated interaction of the D2 long isoform with β-arrestin 2.
PET (positron emission tomography) and single photon emission computed tomography (SPECT) studies support the importance of in vivo D2 receptor occupancy as a predictor of antipsychotic response and side effects.19 Prospective studies demonstrate that, for at least some FGAs, antipsychotic effect requires a striatal D2 receptor occupancy of 65–70%, and D2 occupancy >80% significantly increases the risk of EPS.19 However, brain imaging studies found that there was a wide variation in D2 receptor occupancy among patients on the same dose of antipsychotic and within the same individual in different stages (first episode vs chronic) and phases (relapse vs remission) of the illness. Thus, the thresholds, although generally accurate, are not invariably reliable in predicting clinical events.20 Moreover, clozapine and quetiapine exhibit <60% striatal D2 receptor occupancy at therapeutically effective doses,21, 22 indicating that striatal D2 receptor blockade alone cannot explain therapeutic efficacy. However, the low occupancy of striatal D2 receptors by clozapine and quetiapine could account for their low EPS liability.
SGAs were termed ‘atypical antipsychotics’ because of their low incidence of EPS at therapeutically effective doses. Although there is debate as to what constitutes ‘atypicality’, the defining feature of this class medications is the separation of the dose that results in a therapeutic effect from that which is associated with an increasing risk of EPS.20 The concept of ‘atypicality’ is thought to be due to two properties: (1) their lower affinity for D2 receptors and (2) their high affinity for serotonin (5-HT)2A receptors. Meltzer et al.23 hypothesized that a high ratio of affinities for 5-HT2A receptors and D2 receptors was the critical feature of atypical drugs, while Kapur and Seeman24 proposed that low affinity for and fast dissociation from D2 receptors may be the critical property for ‘atypicality’. FGAs such as chlorpromazine and haloperidol bind more tightly than dopamine itself to the D2 receptor, and dissociate from it slowly.25 In contrast, SGAs such as clozapine and quetiapine bind more loosely than dopamine to the D2 receptor, with dissociation constants (expressed as ‘koff’) that are higher than those for dopamine. Preliminary evidence suggests that other SGAs, including amisulpride, aripiprazole and paliperidone, also have higher koff values, (that is, faster dissociation rates) than do FGAs, while asenapine, olanzapine, risperidone, sertindole and ziprasidone dissociate relatively slowly from the D2 receptor.26 Thus, this ‘fast-off-D2’ theory may not represent a general mechanism of ‘atypicality’.27 Moreover, it remains uncertain how long an SGA needs to bind to D2 receptors to provide maximum therapeutic efficacy with minimal EPS.11 Although more research is required to identify optimum D2 receptor occupancy levels for SGAs, it appears that continuous high levels of occupancy are not critical for maximum antipsychotic efficacy.28
Regionally selective binding of SGAs to dopamine tracts projecting to cortico-limbic areas has been proposed as another contributor to ‘atypicality’. Many animal studies have found preferential effects of SGAs on ventral tegmental area (A10) as compared with substantia nigra pars compacta (A9) dopamine neurons. In humans, PET and SPECT studies have demonstrated this cortico-limbic D2 receptor specificity for some SGAs.29, 30, 31, 32, 33 However, the finding cannot be replicated with different ligands for several SGAs. A recent meta-analysis of PET and SPECT studies supports the hypothesis that cortical D2/D3 receptors are likely to be an important site of antipsychotic action.34 However, this does not exclude the involvement of other brain regions including striatum. Further research is needed to clarify the neurochemical basis of ‘regional specificity’, as it raises the possibility of spatial targeting of APDs in the future.4
Partial D2 agonists represent another strategy of attempting to normalize dopaminergic imbalance in schizophrenia, without the side effects associated with full D2 antagonists. These agents have lower intrinsic activity at D2 receptors than full agonists, allowing them to act as either functional agonists or antagonists, depending on synaptic dopamine levels.35 Partial D2 agonist activity as reported for aripiprazole appears to inhibit endogenous dopamine activity where it is high and activate D2 receptors where it is low. In addition, such an agent should ideally maintain dopaminergic tone in the nigrostriatal and tuberoinfundibular pathways, thereby avoiding EPS and hyperprolactinemia normally associated with D2 antagonism. However, aripiprazole is not a simple partial D2 agonist; it can have partial agonist properties at D3, D4, 5-HT1A, 5-HT2C and, to a much lesser extent, 5-HT2A receptors.36, 37 At present, it is unclear to what extent binding to receptors other than D2 receptors contributes to the actions of aripiprazole.
Serotonin receptor modulation
The ‘dopamine–serotonin antagonism theory’ conceived by Janssen et al.38 and popularized by Meltzer et al.23 assumes that a high ratio of serotonin 5-HT2A receptor to D2 receptor blockade confers antipsychotic ‘atypicality’. 5-HT2A antagonism can increase dopaminergic transmission in the nigrostriatal pathway, thus reducing the risk for EPS, and could theoretically improve negative symptoms and cognitive impairment in schizophrenia by increasing release of dopamine, acetylcholine (Ach) or both in the prefrontal cortex (PFC). This hypothesis appears to apply, to some degree, to most SGAs, including asenapine, clozapine, iloperidone, olanzapine, paliperidone, perospirone, quetiapine, risperidone, sertindole, ziprasidone and zotepine. There are, however, critical limitations to this concept.20 For example, amisulpride has no meaningful affinity for the 5-HT2A receptor, and aripiprazole and blonanserin have higher D2 than 5-HT2A affinity, and yet clinically they have atypical profiles.35, 39 In addition, chlorpromazine and loxapine have relatively higher 5-HT2A than D2 affinity, but they do not have an atypical profile.23 Furthermore, risperidone and olanzapine exhibit high 5-HT2A receptor occupancy at doses that are not antipsychotic, and as doses of these drugs are increased beyond their usual therapeutic ranges, the risk for EPS increases despite maximal 5-HT2A receptor blockade.40 Thus, high 5-HT2A affinity may contribute to the modulation of dopamine in the striatum and PFC, but high 5-HT2A occupancy does not protect against the risk of EPS if D2 receptor occupancy is greater than the EPS threshold.20 The 5-HT2A/D2 hypothesis, therefore, does not satisfactorily explain ‘atypicality’.41 Furthermore, the apparent lack of efficacy of monotherapy with the selective 5-HT2A receptor antagonist M-100907 indicates that 5-HT2A antagonism alone does not account for the efficacy of SGAs.42, 43
Some but not all SGAs are 5-HT2C, 5-HT6 and 5-HT7 receptor antagonists as well as direct or indirect 5-HT1A receptor agonists.44 It has been suggested that partial agonism of 5-HT1A receptors, resulting in activation and blockade of pre- and postsynaptic receptors, respectively, contributes to the mechanism of action of some SGAs and TGAs, including aripiprazole, clozapine, perospirone, quetiapine and ziprasidone.45 5-HT1A receptors are located presynaptically in the raphe nuclei, where they act as cell body autoreceptors to inhibit the firing rate of 5-HT neurons, and are located postsynaptically in limbic and cortical regions, where they also attenuate firing activity. 5-HT1A partial agonist activity is thought to improve negative symptoms and cognitive impairment by enhancing dopamine release in the PFC. In addition, 5-HT1A receptor agonists have been reported to possess antidepressant and anxiolytic properties, and to attenuate the EPS liability of D2 antagonists.11 However, this remains to be proven in clinical studies.
NMDA receptor modulation
The ability of non-competitive N-methyl-D-aspartate receptor (NMDA-R) antagonists, such as phencyclidine (PCP) and ketamine, to induce a spectrum of positive, negative and cognitive schizophrenia-like symptoms has led to the hypothesis that NMDA-R hypofunction can contribute to the pathophysiology of schizophrenia.46, 47, 48
A wide range of preclinical studies have demonstrated that acute treatment with some SGAs, but not FGAs, selectively antagonizes the consequences of experimentally induced NMDA-R hypofunction at cellular and behavioral levels.49, 50 For example, clozapine and olanzapine, but not haloperidol or raclopride, inhibit electrophysiological effects of PCP in sectioned brain tissue, and attenuate NMDA-R antagonist-induced deficits in PPI (prepulse inhibition), social behavior and neurotoxicity. In addition, ketamine-induced brain metabolic activation is blocked by acute administration of clozapine and olanzapine, but not haloperidol.51 In contrast, chronic administration of haloperidol blocks PCP-induced deficits in PPI and ketamine-induced brain metabolic activation.52 Thus, adaptive changes elicited by both FGAs and SGAs appear to attenuate the effects of NMDA-R antagonists.11
The well-documented effects of SGAs on responses to NMDA-R antagonists raise the possibility that the therapeutic action of these agents may involve a correction of NMDA-R hypofunction. However, since none of the SGAs demonstrate direct affinity for NMDA-R, the mechanism by which these effects might be mediated is poorly understood. Animal studies have reported variable effects (increases, decreases, no change) on NMDA-R density in various brain regions after chronic antipsychotic administration.11 Inconsistent and often conflicting findings have been reported for NMDA-R subunit gene expression following long-term treatment with FGAs or SGAs. Differences in treatment regimens, brain regions examined and methods of assessment likely contribute to many of these discrepancies. A preclinical study showed that clozapine inhibits synaptosomal glycine transport through system A amino-acid transporters.53 Regulation of synaptic glycine levels represents a potential indirect mechanism for clozapine to potentiate NMDA-R function (Figure 1).
The development of [123I]CNS-1261 as the first usable SPECT ligand for the NMDA-R intra-channel PCP-binding site permitted the direct estimation of NMDA-R activity in living humans.54 Bressan et al.55 found that schizophrenia patients treated with clozapine had significantly reduced NMDA-R binding in all the brain regions. A trend in the same direction for FGAs was also observed. Further studies are needed to determine whether inhibition of the effects of NMDA-R antagonists by SGAs involves molecular modifications in glutamate receptors and/or other neurotransmitter–glutamate interactions.
Other receptor modulation
Most APDs interact with adrenergic, histaminergic and muscarinic neurotransmitter systems as well as with monoamine transporters. Interactions with these receptor systems contribute to many common antipsychotic-induced side effects but emerging data also indicate the potential for previously unrecognized therapeutic benefits.11 For example, the blockade of H1 receptors by APDs is probably related to weight gain and sedation, and α1-adrenergic receptor (AR) blockade is believed to contribute to orthostatic hypotension and sedation.4 M1 antagonist actions may cause central (for example, cognitive impairment) and peripheral (for example, constipation and dry mouth) anticholinergic adverse effects. In contrast, as will be mentioned later, M1 receptor agonism might be beneficial in treating the cognitive dysfunction as well as the psychotic symptoms in schizophrenia.56 Muscarinic M1 receptor agonism is relatively unique to clozapine and has been suggested as one mechanism that might underlie its superior efficacy.57 However, there is no definitive evidence that any neuroreceptors other than the D2 receptor play a significant role in their therapeutic efficacy.
Structural neuroimaging studies have demonstrated that pathomorphological brain changes, such as ventricular enlargement and reduction in gray- and white-matter volumes, may be present at the first psychotic episode of schizophrenia and possibly in the prodromal and premorbid phase.58, 59 Longitudinal studies have also found that cortical gray-matter loss may be progressive—especially early in the course of illness—and is associated with functional decline.60, 61 These findings raise the question of whether APDs could mitigate such pathophysiological progression in the early stages of illness. A number of investigators have found that APDs produce neuroplastic changes at structural and molecular levels in the brain.62, 63, 64, 65 For example, Lieberman et al.66 reported that long-term treatment with olanzapine, but not haloperidol, prevented progressive gray-matter volume reductions in patients with first-episode psychosis. Studies with similar results have been reported by the Utrecht group, indicating the specific neuroprotective value of APD treatment and particularly with clozapine and olanzapine.67, 68 Although the precise mechanisms by which APDs may arrest or delay the pathomorphological process remain unknown, there is growing evidence that some SGAs, both in vitro and in vivo, may have neuroprotective effects, including the production of neurotrophic factors, the attenuation of glutamate excitotoxicity, oxidative stress and apoptosis and the enhancement of neurogenesis and connectivity, all of which could provide a rationale for pharmacological intervention with SGAs during the initial stages of schizophrenia.65, 66 However, a recent large-scale longitudinal observational study of first-episode schizophrenia reported that APDs had a slight but significant association with brain tissue volume loss over time.69 This finding is further supported by a controlled study in macaque monkeys in which chronic administration of both olanzapine and haloperidol induced frontoparietal volume reductions associated with decreased astrocyte numbers.70, 71, 72 Thus, further well-designed clinical studies in first-episode patients are required to determine whether specific APDs may produce neuroprotective or possibly neurotoxic effects. Also, the clinical relevance of subtle changes in white and gray matter remains to be determined.
Pharmacological and clinical profiles of APDs
FGAs, also known as typical APDs, possess high affinity for and act as full antagonists at D2 receptors,73 and are associated with a high incidence of EPS and hyperprolactinemia.4 FGAs can be classified as either high- or low-potency medications with the former having a greater affinity for D2 receptors than the latter.73 Based on their chemical structure, FGAs may be divided into three groups: butyrophenones (for example, haloperidol), phenothiazines (for example, chlorpromazine) and a heterogeneous third group.74
The ability of FGAs to reduce positive symptoms and risk for relapse markedly improved clinical outcomes for many patients with schizophrenia. However, ∼30% of patients with acutely exacerbated symptoms have little or no response to FGAs, and up to 50% have only a partial response.75, 76 Moreover, FGAs offer little benefit for negative symptoms or cognitive impairment.
FGAs produce a variety of side effects including acute EPS, hyperprolactinemia, as well as TD (tardive dyskinesia) associated with long-term exposure, which are caused by the blockade of dopaminergic nigrostrial pathways.4 Low-potency FGAs possess high affinities for muscarinic M1 Ach, histaminergic H1 and α1 norepinephrine receptors, which can result in partially distinctive and overlapping side-effect profiles (for example, cognitive deficits and sedation).77 Largely because of the side-effect risks of EPS, some have argued that the only patients for whom FGAs are clearly preferable are those with a history of good response and tolerable side effects during treatment with an FGA.78
Second- and third-generation APDs
Clozapine, the prototypical SGA or atypical antipsychotic, was found to be an effective antipsychotic and did not cause EPS. Moreover, clozapine proved to be superior to chlorpromazine in treatment-resistant schizophrenia. However, clozapine also was associated with an elevated risk of potentially lethal hematotoxicity (agranulocytosis).79 Consequently, additional SGAs were introduced including risperidone, olanzapine, quetiapine and ziprasidone in an effort to provide the therapeutic benefits of clozapine without the associated risk of blood dyscrasias.11, 80 Aripiprazole is another pharmacologically different agent that is sometimes termed a TGA.81 Lower incidence of EPS is the major advantage of NGAs (new-generation antipsychotics) compared with most FGAs.8 However, most SGAs have an increased risk of causing weight gain and disturbances in glucose and lipid metabolism.
During the past decade, additional SGAs, including paliperidone, asenapine, iloperidone and lurasidone, have been approved by the United States (US) Food and Drug Administration (FDA).82 The pharmacology of these drugs is similar to other SGAs except that lurasidone is notable for its high affinity for the 5-HT7 receptors.83 However, the comparative effectiveness of these agents to existing FGAs and NGAs has yet to be determined.
Efficacy on psychotic symptoms
In a recent meta-analysis of randomized controlled trials (RCTs) comparing SGAs with FGAs, four SGAs (amisulpiride, clozapine, olanzapine and risperidone) were more efficacious than FGAs for treatment of positive symptoms, with small-to-medium effect sizes.8 However, the other SGAs (aripiprazole, quetiapine, sertindole, ziprasidone and zotepine) were only as efficacious as FGAs. Thus, there may be a modest advantage for some, but not all SGAs compared with FGAs.
There have been three large long-term pragmatic studies, which compared the effectiveness of SGAs with FGAs. The Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study compared the effectiveness of the FGA perphenazine with the SGAs olanzapine, quetiapine, risperidone and ziprasidone in 1493 chronic schizophrenic outpatients for up to 18 months of treatment.80 Results from the first phase of this study demonstrated that olanzapine had a significant advantage regarding all cause discontinuation, the primary study outcome measure. The time to the discontinuation of treatment for lack of efficacy was significantly longer in the olanzapine group than in the FGA perphenazine group (hazard ratio, 0.47), the quetiapine group (hazard ratio, 0.41), the risperidone group (hazard ratio, 0.45) or the ziprasidone group (hazard ratio, 0.59). However, the other SGA drugs did not differ from perphenazine on the primary outcome measure and all five medications displayed comparable changes in Positive and Negative Syndrome Scale (PANSS) positive scores.
In the CUtLASS (Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study), in which FGAs and SGAs were compared as classes (most patients in the first group were treated with sulpiride and most in the second group with olanzapine), no significant differences in positive symptoms were observed between patients treated with FGAs and those with SGAs.84
In a third large pragmatic trial, the EUFEST (European First-Episode Schizophrenia Trial)85, 86 close to 500 first-episode patients were randomly allocated to open-label treatment with either low-dose haloperidol or one of four SGAs (amisulpride, olanzapine, quetiapine or ziprasidone) for 1 year. Compared with haloperidol, all SGAs showed a significantly lower risk for early study discontinuation, the primary outcome variable but there were no group differences in PANSS total scores. In summary, CATIE, CUtLASS and EUFEST did not provide unequivocal evidence to support the superiority of SGAs over FGAs or of any individual non-clozapine antipsychotic over another on positive or overall symptoms.87
Efficacy on negative symptoms
Negative symptoms of schizophrenia include affective blunting, emotional withdrawal, poverty of speech, anhedonia and apathy. They may be divided into three subtypes that are often difficult to distinguish: (1) primary enduring (or deficit), (2) primary non-enduring and (3) negative symptoms that are secondary to other causes, such as depression, positive symptoms, EPS, substance abuse or iatrogenic effects such as understimulation during long-term hospitalization.88, 89 Approximately 70% of schizophrenic patients develop primary negative symptoms before the onset of positive symptoms.90 To date, there are no effective treatments for primary negative symptoms.
In a meta-analysis by Leucht et al.,8 four SGAs (amisulpiride, clozapine, olanzapine and risperidone) were more efficacious than FGAs for treatment of negative symptoms. However, the other five SGAs (aripiprazole, quetiapine, sertindole, ziprasidone and zotepine) were only as efficacious as FGAs. It should be noted that few of the studies included have specifically examined the effect of SGAs on primary negative symptoms. Thus, a direct beneficial effect of SGAs on primary negative symptoms has not been established with the possible exception of low doses of amisulpride.91 Some SGAs may be effective for secondary negative symptoms, since they have been proposed to show greater effects on symptoms of depression and anxiety and have lower risk of EPS.
Phase 2 of the CATIE trial provides a notable opportunity to compare the effectiveness of multiple SGAs in the treatment of negative symptoms. Patients who had discontinued their phase 1 APD due to lack of efficacy were randomized either to clozapine or to another SGA. There were no significant differences in PANSS negative symptom subscores either between or within treatment groups at baseline, 3 and 6 months.92 At 6 months, however, clozapine showed a trend toward greater reduction in negative symptom scores, as compared with other SGAs. Patients who had discontinued their phase 1 APD due to intolerable side effects were randomized to olanzapine, quetiapine, risperidone or ziprasidone. There were similarly no significant differences in PANSS negative symptom subscores among treatment groups at 12 months.93 Thus, the results of the CATIE study suggest only slight efficacy for SGAs in the treatment of negative symptoms of chronic schizophrenia, with a probable small advantage for clozapine.
Efficacy on cognition
Patients with schizophrenia typically perform one to two standard deviations below healthy volunteers on a variety of neuropsychological measures, particularly those that assess attention, verbal skills, processing speed and executive function.94 Given that cognitive deficits are among the strongest predictors of functional outcome in schizophrenia,95 treatments for these symptoms are currently considered to be the most urgently needed.
In a meta-analysis of 14 controlled studies with random assignment to treatment with either FGA or SGA (clozapine, olanzapine, quetiapine or risperidone), SGAs were superior to FGAs, haloperidol in particular, for ameliorating overall cognitive function, and specific improvements were noted in learning and processing speed.94 In another meta-analysis focusing on long-term memory measures, SGAs showed benefits over FGAs, with an effect size for differential improvement of <0.20.96 It is unclear, however, whether the improvements observed with SGAs in these studies represent true cognitive enhancement or only a relative reduction in EPS- and anticholinergic-related cognitive effects, as compared with FGAs.97, 98 Also, it is possible that the improvements may, at least in part, be due to practice effects.99
There has been a debate as to whether lower doses of FGAs might show comparable efficacy to SGAs for cognitive symptoms of schizophrenia.100 Several studies with improved methodology suggest that low doses of haloperidol can produce cognitive outcomes that approximate the SGA comparator.101, 102, 103, 104, 105 Other studies have also found mild benefits in cognitive performance with certain FGAs.106 The most compelling of these is the CATIE trial, in which perphenazine yielded more improvement in cognition after 18 months than any of the phase 1 SGAs.107 Notably, the magnitude of cognitive improvement was small and probably not clinically meaningful for all APDs assessed in this analysis.
Taken together, FGAs and SGAs, when dosed properly, may yield at most modest improvements in cognitive deficits of schizophrenia. Neither class appears to be clearly superior to the other. As described later, the use of adjunctive pharmacological agents might offer a viable approach, because they can be used to modulate specific neurotransmitter systems (for example, glutamate, Ach) hypothesized to be associated with particular cognitive functions.
Efficacy on mood symptoms
Mood symptoms, depression in particular but also mania, can occur in schizophrenia. Depressive symptoms are particularly common in all phases of the illness, and are associated with poorer outcomes, impaired social and vocational functioning, lower quality of life and an increased risk of relapse and suicide.108 The modal rate of comorbid depression in schizophrenia has been reported to be 25%.109
A meta-analysis demonstrated that five SGAs (amisulpride, aripiprazole, clozapine, olanzapine and quetiapine) were more efficacious than FGAs for treatment of depressive symptoms, whereas risperidone was not.8 However, it needs to be noted that most of the studies included were not primarily designed to evaluate depressive symptoms. Therefore, prospective trials with change in depression as the primary outcome are needed.
Efficacy for treatment-resistant schizophrenia
One-fifth to one-third of patients with schizophrenia who receive an adequate trial fail to respond to prescribed APDs, and are considered ‘treatment-resistant’.110, 111 Additional subsets of patients may be treatment-intolerant, treatment-non-compliant, slow to respond or only partial responders.111
Clozapine has been most extensively studied in treatment-resistant schizophrenia patients, and it is the only APD approved by the US FDA for this indication.79, 111 The drug's singular efficacy in treatment resistance was first conclusively demonstrated in a pivotal study in which 30% of patients receiving clozapine responded at 6 weeks, compared with 7% treated with chlorpromazine.79 A meta-analysis of seven controlled trials comparing clozapine to an FGA in treatment-resistant schizophrenia found clozapine to be superior to FGAs in terms of overall psychopathology reduction, EPS and compliance, although the magnitude of the drug's therapeutic effect was not consistently robust.112
In the past decade, a number of double-blind RCTs have reported the short-term (4–18 weeks) efficacy of clozapine to be comparable or greater than the other SGAs in treatment-resistant schizophrenia.113 The second phase of the CATIE study found that switching to open-label treatment with clozapine was more effective than to another SGA drugs in patients with schizophrenia who failed to improve after initial treatment with an SGA.92 In the CUtLASS 2 trial comparing the effectiveness of clozapine with that of other SGAs (amisulpride, olanzapine, quetiapine and risperidone) in patients with schizophrenia with clinician-defined poor response to 2 or more APDs, clozapine showed a statistically significant advantage with respect to total PANSS score, but not quality of life, over 1 year.114, 115
Because of clozapine's superior efficacy relative to FGAs and probably other SGAs, clozapine monotherapy represents the ‘gold standard’ for treatment of patients with refractory schizophrenia. Unfortunately, the risk of serious side effects and the need for regular blood monitoring render it unsuitable as a first-line medication. Even with clozapine treatment, a considerable number of patients do not respond or are only partially responsive. Thus, one strategy commonly employed in the management of such patients is to augment clozapine with other psychotropic medications. However, the evidence supporting this practice is very limited, and no drug has consistently demonstrated efficacy as an adjuvant to clozapine in treatment-resistant schizophrenia.4
Efficacy on disease-related complications and behavioral disturbances (suicide, violence, substance abuse)
Suicidal behavior is common in schizophrenia. Approximately 50% of patients with schizophrenia or schizoaffective disorder have suicidal behavior,116 and about 5% ‘succeed’ in taking their own lives.117 The InterSePT (International Suicide Prevention Trial), a double-blind RCT, compared suicidal behavior in 980 patients at relatively high risk for suicide, who were randomized to clozapine or olanzapine treatment and followed for up to 2 years.116 Suicidal behavior was significantly less common in patients treated with clozapine. Although the InterSePT study has been criticized for certain methodological limitations, such as lack of blinding among patients and clinicians, the results were compelling enough to garner the US FDA's approval of an indication unique to clozapine: reduction of the risk of recurrent suicidal behavior in patients with schizophrenia or schizoaffective disorder. In a secondary analysis of a large-scale safely study, the authors describe a lower suicide attempt risk for patients treated with sertindole compared with risperidone.118
Violent behavior is observed in a minority of patients with schizophrenia. It is etiologically heterogeneous (for example, positive symptoms, impaired impulse control, comorbid personality disorders).119 Current treatment of persistent violence in schizophrenia relies on APDs and mood stabilizers. The evidence of efficacy is relatively strong for clozapine120, 121 and olanzapine,122 but inconsistent for other SGAs.119 In the CATIE trial, SGAs did not reduce violence more than perphenazine.123 The antihostility and/or antiaggressive effects of clozapine appear to be independent of general antipsychotic effects.
Nearly 50% of patients with schizophrenia have a lifetime history of substance use disorder and substance abuse is a potent determinant for poor outcome in schizophrenia.124 To date, preliminary research has been conducted in patients with schizophrenia and co-occurring substance use disorders, and has suggested possible beneficial effects for aripiprazole, clozapine, olanzapine, risperidone and quetiapine.125 However, all of these drugs require further research with RCTs to fully assess their impact on such population in schizophrenia.
In general, SGAs offer the advantage of fewer acute EPS and reduced likelihood of TD, but produce greater metabolic side effects than FGAs.126 A meta-analysis demonstrated that all SGAs were associated with fewer EPS than haloperidol (even at low doses).8 However, with the exception of clozapine, olanzapine and risperidone, SGAs were not more effective than low-potency FGAs.
Among SGAs, amisulpride, paliperidone and risperidone have the potential to increase serum prolactin levels to an extent comparable to FGAs.4 In contrast, quetiapine, and clozapine, do not elevate serum prolactin levels, while aripiprazole actually suppresses prolactin. Olanzapine causes hyperprolactinemia only at high doses.127
Marked differences in liability for metabolic side effects are seen between the different SGAs. A meta-analysis by Leucht et al.8 found that with the exception of aripiprazole and ziprasidone, SGAs differentially induced more weight gain than haloperidol but not the low-potency FGAs. In another recent meta-analysis of 48 studies, clozapine and olanzapine produced the greatest elevation in weight, cholesterol and glucose followed by quetiapine, risperidone and sertindole with intermediate elevations. Amisulpride and aripiprazole demonstrated lower elevations and ziprasidone the lowest.128 These safety concerns are associated with potential long-term health risks for patients as well as decreased adherence to treatment regimens that may lead to relapse.
While modest, dose-dependent QTc prolongation has been reported with many antipsychotics; several FGAs (for example, haloperidol, mesoridazine, pimozide, thioridazine) and SGAs (sertindole, ziprasidone) can increase the QTc interval considerably, and all of these have been associated with fatal or potentially fatal arrhythmias.129, 130 In 2005, the US FDA issued a black box warning regarding an increased risk of mortality associated with the use of SGAs in elderly patients with dementia. On the basis of further observational research,131, 132 the FDA extended this warning to FGAs in 2008. A large retrospective cohort study demonstrated that current users of FGAs and SGAs had a similar, dose-related increased risk of sudden cardiac death.133 The evidence for other clinically relevant adverse events, such as sedation, sexual disturbances, hepatotoxic effects and orthostatic hypotension is limited and variable, especially with regard to differential risks between the different APDs.4
Current and future drug development strategies
D1 receptor antagonist or agonist
Earlier preclinical studies demonstrated that selective D1-like antagonists are active in most conventional rodent models predictive of antipsychotic-like activity.134 However, clinical trials of the selective D1-like antagonist SCH39166135, 136 and NNC 01-0687137 failed to demonstrate antipsychotic properties (Table 1).
Dopamine D1 receptors play an important role in cognitive function such as working memory. It is postulated that either insufficient or excessive D1-like receptor stimulation is deleterious to cognitive function of the PFC, thus an ‘optimal’ level of D1-like receptor activation is necessary for normal cognitive function.138, 139 Low doses of selective full D1-like receptor agonists, such as dihydrexidine (DAR-0100), A77636 and SKF81297, have been reported to have cognitive-enhancing actions in rodents and non-human primates.140, 141, 142, 143 In a pilot study, a single subcutaneous dose of dihydrexidine was well tolerated in patients with schizophrenia, but did not produce delayed clinical or neuropsychological improvements.144 However, it induced a significant increase in prefrontal brain activity (perfusion) compared with placebo, suggesting that dihydrexidine and other full D1-like receptor agonists may be able to modulate prefrontal dopaminergic function in schizophrenia.145 It should be noted, however, that chronic administration of full D1-like receptor agonists may lead to down-regulation of D1 receptors, which could exacerbate cognitive dysfunction in schizophrenia.56 Consequently, the dosing strategy of low and intermittent administration has been proposed as an alternative to conventional dosing.146
D2 receptor partial agonist
Several clinical studies with selective D2-like receptor partial agonists, including talipexole, preclamol, roxindole and pramipexole, failed to demonstrate a clear therapeutic effect on positive symptoms of schizophrenia, although the results suggested possible beneficial activity against negative symptoms.147, 148 In addition, the use of higher doses of partial D2 agonists is associated with worsening of positive symptoms, probably related to the activation of postsynaptic D2-like receptors.
Cariprazine (RGH-188), a novel putative APD, is currently in phase III clinical trials. It exhibits partial agonism at D2/D3 receptors, with preferential binding to D3 receptors, and partial agonism at 5-HT1A receptors.149, 150 It showed lower affinity for D2 or higher affinity for D3 receptors compared with aripiprazole.150 Cariprazine displayed partial agonist activity at D2/D3 receptors, with similar intrinsic activity but higher potency than those of aripiprazole.150 Preclinical studies have suggested that its propensity for EPS is low and that it may have pro-cognitive properties.
With the exception of aripiprazole and cariprazine, development of D2 partial agonists for schizophrenia has been de-emphasized due to their inferior therapeutic profile as compared with other marketed APDs. The intrinsic activity of D2-like receptor agonism may be critical in determining the efficacy and tolerability of such agents, and on which symptom dimensions.151, 152
SSR-181507 is a new partial D2 agonist and displays antagonist activity at D2 receptors as well as agonist activity at 5-HT1A receptors.153 Preclinical studies suggest that SSR-181507 has antipsychotic-like activity in the absence of extrapyramidal signs and cognitive deficits, with the additional benefit of apparent antidepressant and anxiolytic activities.154 OPC-34712 is a novel D2 partial agonist and exhibits enhanced affinity for 5-HT1A, 5-HT2A, 5-HT7 serotonin receptors.155 It is in phase III testing for schizophrenia.
D3 receptor antagonist
The D3 receptor has cortico-limbic distribution and is of interest because most APDs have high affinity for this receptor and it represents a promising target for enhancing cognition.156 Selective D3 antagonists have been developed (S33084, S33138, SB-277011-A, AVE5997),157, 158, 159 but there are limited animal behavioral data at this time. As aforementioned, cariprazine has ∼10-fold higher affinity for D3 vs D2 receptors.150 In a preliminary study, (+)-UH232, a D3 antagonist, actually worsened psychotic symptoms in drug-free patients with schizophrenia.160 In a recent 6-week double-blind RCT, a selective D3 antagonist ABT-925, used as monotherapy, failed to demonstrate any significant benefits for psychopathological symptoms in patients with acute exacerbation of schizophrenia, although the doses of ABT-925 used might have been too low to produce a measurable antipsychotic effect.161 Further controlled trials of D3 antagonists will help clarify the differential contributions of the D2 and D3 receptor in the mediation of antipsychotic effect (Table 1).
D4 receptor antagonist
D4 receptor antagonism has been proposed as a novel target and possibly contributes to the superior efficacy and atypical profile of clozapine. However, negative studies for the highly selective D4 antagonist L-745,870,162 the 5-HT2A/D4 antagonist finanserin163 and sonepiprazole164 suggest that the selective D4 receptor antagonist mechanism is ineffective, at least as monotherapy.11
5-HT1A receptor agonist
Preclinical studies suggest that 5-HT1A agonists may potentiate the antipsychotic activity of dopamine antagonists.165 Further, 5-HT1A receptors are up-regulated in postmortem PFC in schizophrenia, suggesting a deficit in 5-HT1A function.166, 167 Interestingly, the 5-HT1A partial agonists tandospirone and buspirone can enhance certain domains of cognition in patients receiving FGAs or SGAs.168, 169 Based on these data, compounds that possess 5-HT1A agonism combined with D2-like receptor antagonism, including SLV-313, SSR-181507, F-15063, S-16924, BSF 190555 (BTS 79018) and RGH-188, are being developed as potential APDs.148, 170 It is suggested that the balance between D2-like receptor antagonism and 5-HT1A receptor agonism may be critical in determining the efficacy of these compounds.170
5-HT2A receptor antagonist
As already described, 5-HT2A antagonists may contribute to ‘normalizing’ levels of dopamine release171 and theoretically possess antipsychotic activity, and also improve negative symptoms of schizophrenia.172 Three RCTs have demonstrated that addition of ritanserin, a relatively selective 5-HT2A/2C antagonist, to an FGA, produced significant reductions in negative symptoms and depressed mood in chronic schizophrenia.173, 174, 175 Moreover, ritanserin potentiated the clinical efficacy of risperidone on negative symptoms.176 However, the development of M-100907, a selective 5-HT2A antagonist, was discontinued after two phase III clinical trials demonstrated lower antipsychotic efficacy compared with haloperidol.42, 43 A phase II trial of the 5-HT2A/2C antagonist SR-46349B also demonstrated efficacy superior to placebo but inferior to haloperidol.177 Current data are insufficient to adequately judge the efficacy of 5-HT2A antagonists, but monotherapy with such agents may not be a viable treatment strategy (Table 1).
5-HT2A receptor inverse agonist
It has been recently recognized that many SGAs are inverse agonists at the 5-HT2A receptor rather than neutral antagonists.178 In contrast to antagonists, inverse agonists lack negative intrinsic efficacy and can attenuate basal constitutive signaling activity and block only agonist-induced responses.179 Pimavanserin tartrate (ACP-103) is the first 5-HT2A inverse agonist to enter clinical trials as a treatment for schizophrenia. In phase II clinical trials, pimavanserin seemed to be safe, well tolerated and potentiated the therapeutic effects of low-dose risperidone, and reduced haloperidol-induced akathisia.180
5-HT2C receptor agonist
The 5-HT2C receptor possesses a unique ability to tonically regulate dopamine release181 and play a critical role in mediating the interaction between serotonergic and dopaminergic systems.182 5-HT2C receptor agonists inhibit dopamine release in the mesolimbic dopamine pathways.181 Animal studies of the 5-HT2C receptor agonists WAY-163909 and CP-809,101 showed that they can attenuate psychostimulant- or NMDA-R antagonist-induced behaviors.183, 184 Furthermore, CP-809,101 was active in novel object recognition, an animal model of cognitive function.184 Thus, 5-HT2C receptor agonists appear to have a pharmacological profile similar to that of the SGAs and may be a novel approach in the treatment of schizophrenia. However, there is a concern that 5-HT2C receptor agonists may worsen cognition and cause EPS by reducing dopaminergic transmission in the mesocortical and nigrostriatal pathways.27 A 6-week placebo and olanzapine controlled phase II study with the full agonist vabicaserin showed a significant improvement over placebo for the 200 mg dose on PANSS total and positive scores as well as on the Clinical Global Impression Scale (CGI), but 400 mg per day did not separate from placebo.185
5-HT3 receptor antagonist
The 5-HT3 receptors tonically inhibit the release of Ach and also inhibit γ-aminobutyric acid (GABA) inhibitory interneurons.186 In preclinical models, 5-HT3 antagonists have been shown to have a broad spectrum of psychotropic effects, including correcting psychotic-like behavior, improving cognitive deficits and antagonizing locomotor hyperactivity induced by dopamine stimulants.187 Adjunctive treatment with ondansetron, a highly selective 5-HT3 antagonist, was shown to be effective for auditory P50 deficits,188 negative symptoms and cognitive impairments (visual memory) in patients with chronic schizophrenia on stable antipsychotic therapy.189, 190, 191 However, ondansetron did not improve global cognitive functioning or positive symptoms.190 Replication with larger sample sizes will be necessary before endorsing a role for adjunctive ondansetron in schizophrenia.
5-HT4 receptor agonist
5-HT4 receptors, expressed in nigrostriatal and mesolimbic systems, can modulate the release of Ach, dopamine, GABA and 5-HT.192 5-HT4 receptor agonists (for example, BIMU1, RS67333, RS17017) were shown to enhance memory in several animal models.193 The combined administration of partial 5-HT4 receptor agonists (for example, RS67333, SL65.0155) with cholinesterase inhibitors may enhance cognition to a greater extent than either treatment alone.194, 195 Although there have been no clinical trials of 5-HT4 receptor agonists as add-on therapy in patients with schizophrenia, 5-HT4 receptors may be an attractive target for improving cognition in schizophrenia, since currently available SGAs generally lack significant affinity for 5-HT4 receptors.193, 196
5-HT6 receptor antagonist
5-HT6 receptors are almost exclusively expressed in the central nervous system (CNS), particularly in areas associated with learning and memory.197, 198 While 5-HT6 receptor function has not been fully elucidated, its distribution and high affinity for certain SGAs (for example, clozapine, olanzapine) support the hypothesis that 5-HT6 ligands may have a therapeutic role in schizophrenia.199 In fact, there is an increasing body of evidence, suggesting that 5-HT6 receptor blockade improves memory processes.198 Preclinical studies indicate that 5-HT6 receptor antagonists, including SB-271046, SB-258510A, SB-399885, can enhance the release of cortical dopamine and cortical and hippocampal Ach and glutamate, and may also have longer-term neurotrophic actions in normal rats.197, 198 Moreover, 5-HT6 receptor antagonists can restore cognitive impairments in several animal models of schizophrenia.200, 201, 202 Interestingly, SB-399885 can potentiate haloperidol and risperidone-induced dopamine efflux in medial PFC and hippocampus.203 These data suggest a possible therapeutic role for 5-HT6 antagonists, as an adjunct to antipsychotics, to enhance cognitive function and/or to treat negative symptoms in schizophrenia (Table 1). The 5-HT6 antagonist GSK (SB)-742457 is currently in phase II clinical trials.202
5-HT7 receptor antagonist
The 5-HT7 receptor exerts important roles in circadian rhythms, mood and sleep.199 5-HT7 receptors are found in relatively high concentrations in hippocampus, thalamus and hypothalamus, with generally lower levels in cortex and amygdala.204 As with 5-HT6 receptors, 5-HT7 receptors bind certain SGAs (for example, amisulpride, clozapine, lurasidone, risperidone) with high affinity and may have important roles in learning and memory as well as antidepressant actions.199, 205 Moreover, the specific 5-HT7 receptor antagonist SB-258741 produced an unequivocal positive result in one animal model for positive symptoms of schizophrenia.206 Evaluation of its therapeutic potential is eagerly awaited in patients with schizophrenia.
Glycine site allosteric modulator
The NMDA-R hypofunction hypothesis of schizophrenia predicts a therapeutic effect for compounds that increase NMDA-R transmission.11 Glycine is a positive allosteric modulator (PAM) and obligatory co-agonist at the NMDA-R (Figure 1).207 The glycine allosteric regulatory site agonists, including glycine, D-cycloserine, D-serine and D-alanine, have been studied as potential drugs to augment NMDA-R-mediated neurotransmission and treat the symptoms of schizophrenia. In a recent meta-analysis of 26 double-blind, placebo-controlled studies, adjuvant glycine therapy was found to be effective in reducing total psychopathology, which was assessed by PANSS total scale or BPRS (Brief Psychiatric Rating Scale), positive symptoms and depressive symptoms.208 D-cycloserine, a partial glycine site agonist, was not effective in any symptom domains. D-serine, a full glycine site agonist, was found effective in total psychopathology, negative symptoms and cognitive symptoms. D-serine is more permeable than glycine at the blood–brain barrier, thus requiring a lower dosage.209 A recent trial, however, failed to show beneficial effects of D-serine on any clinical symptoms in patients with chronic schizophrenia.210 D-alanine, also acting as a full agonist on the glycine site of the NMDA-R, has been reported to improve both positive and negative symptoms.211 Furthermore, these glycine site agonists added onto SGA (excluding clozapine) were significantly better than FGA or clozapine, in improving negative symptoms and total psychopathology.208
However, neither glycine nor D-cycloserine have shown consistent benefits on any aspects of cognitive function in schizophrenia.13 Indeed, the results of the largest 16-week RCT of adjunctive glycine and D-cycloserine, the CONSIST (Cognitive and Negative Symptoms in Schizophrenia Trial), suggest that neither glycine nor D-cycloserine are effective for treating negative symptoms or cognitive impairments.212
Taken together, these results suggest that while adjunctive glycine and maybe D-alanine appear to convey some benefit against positive symptoms, this could not be demonstrated for any other of the glycine site modulators. Moreover, no consistent benefit was seen on negative and cognitive symptoms. D-Cycloserine was not found effective and the evidence for D-serine is not clear (Table 2).
Glycine transporter inhibitor
Glycine transporters, GlyT1 and GlyT2, are expressed on both neuronal and glial cells in the CNS (Figure 1). GlyT1 is thought to regulate the concentration of extracellular glycine at excitatory synapses containing NMDA-Rs.213 Thus, blockade of the GlyT1 transporter would increase NMDA-R-mediated transmission. To date, a number of pharmaceutical companies have been developing novel GlyT1 inhibitors.
Sarcosine (N-methylglycine), a potent and prototype GlyT1 inhibitor, when added to an existing regimen of APDs, showed efficacy for positive and negative symptoms in both chronically stable210, 214 and acutely ill patients with schizophrenia.215 However, sarcosine has not shown clinical efficacy when added to clozapine.216 Lane et al.210, 215 suggested that sarcosine is more efficacious than D-serine as adjuvant therapy for schizophrenia. However, optimal dosing for sarcosine and D-serine may be different.210 Sarcosine has also been studied as monotherapy in acutely symptomatic drug-free schizophrenic patients.217 To fully assess sarcosine's effects further placebo- or active-controlled, larger-sized studies are necessary.
In a recent large-scale double-blind phase II study of the potent and non-competitive GlyT1 inhibitor RG1678 (bitopertin), the compound significantly improved negative symptoms in the per protocol analysis and demonstrated a trend towards functional improvement in schizophrenia patients with predominant negative symptoms.218 In contrast, adding another GlyT1 inhibitor, Org-25935 to an SGA in a 12-week placebo-controlled study in patients with persistent negative symptoms in schizophrenia did not lead to a significant improvement.219 It will be interesting to learn whether these GlyT1 inhibitors are also effective in ameliorating cognitive deficits in schizophrenia.
Metabotropic glutamate receptor agonist
Metabotropic glutamate receptors (mGluRs), of which there are eight subtypes, are categorized into three groups: Group I (mGluR1/5), Group II (mGluR2/3) and Group III (mGluR4/6/7/8).220 Group II mGluRs (mGluR2/3) are located presynaptically on glutamate terminals where they may act as autoreceptors inhibiting glutamate release (Figure 1).220, 221 Extensive preclinical data demonstrate that orthosteric mGluR2/3 agonists, including LY354740, LY379268 and LY404039, exhibit an antipsychotic-like behavioral and neurochemical profile in animal models of schizophrenia.222 However, there is a possible concern that group II mGluRs may desensitize during chronic treatment with mGluR2/3 agonists.27
A double-blind proof-of-concept RCT reported significantly greater improvement on positive and negative symptoms after treatment with LY2140023 (Methionine-LY404039) compared with placebo, with few side effects, suggesting that mGluR2/3 agonists have antipsychotic properties without direct D2 antagonism and may provide a new alternative monotherapy for the treatment of schizophrenia.223 However, a follow-up phase II trial with LY2140023 and olanzapine in acute schizophrenia patients was inconclusive as both agents failed to show significant clinical improvement due to an unusually high placebo response rate.224 In a more recent 6-month open-label study, LY2140023 was inferior to the atypical antipsychotic standard of care in improvement on PANSS total score at the 6-month end point.225 Thus, the therapeutic effects of LY2140023 as a sole antipsychotic may not be large enough to claim equivalence to the existing APDs. Moreover, the impact of LY2140023 on cognitive dysfunction remains to be tested (Table 2). Despite the inconclusive results of phase II trials, phase III trials with LY2140023 have been initiated.
Metabotropic glutamate receptor modulator
PAMs of mGluR offer an attractive alternative to the direct activation of mGluR by orthosteric competitive agonists. PAMs have been discovered for mGluR1, mGluR2 and mGluR5.226 These molecules offer the potential to increase the efficiency of normal glutamate transmission without the risk of inappropriate stimulation.227
mGluR5 are located in cerebral cortex and nucleus accumbens and are expressed postsynaptically on glial cells (Figure 1). mGluR5 have been shown to enhance NMDA-R function in vitro, and activation of mGluR5 can improve learning and memory in normal animals as well as reverse cognitive impairments induced by NMDA-R blockade.226 PAMs of mGluR5 include 3-cyano-N-(1,3-diphenyl-1H-pyrazol-5-yl) benzamide (CDPPB)228 and ADX47273.227 Several preclinical studies have demonstrated that these compounds exhibit beneficial effects in preclinical behavioral models that predict efficacy in the treatment of both positive symptoms and cognitive deficits in patients with schizophrenia. Further results from clinical populations are awaited.
Ampakines, a class of compounds that allosterically enhance AMPA (α-amino-3-hydroxy-5-methy-isoxazole-4-propionic acid) receptor function, represent another potential class of adjunctive treatments for schizophrenia (Figure 1). Ampakines enhance excitatory glutamatergic transmission and facilitate long-term potentiation, learning and memory in rodents.229, 230 In a small double-blind RCT of schizophrenic patients who were partially refractory to treatment with FGAs, the ampakine CX-516 as a sole agent did not produce significant effects on positive symptoms or cognitive impairment.231 Moreover, in a placebo-controlled RCT of CX-516 added to clozapine, olanzapine or risperidone, no benefits were observed on cognition or symptoms of schizophrenia.232 Farampator (CX-691, ORG-24448), an AMPA potentiator, is currently in a phase II trial.233 In the case of AMPA ligands, it remains unclear whether agonists or partial agonists/modulators will have benefits in schizophrenia.
Glutathione (GSH) is the primary endogenous antioxidant and it plays a critical role in protecting cells from damage by reactive oxygen and other radical species.234 GSH also potentiates the NMDA-R response to glutamate.235 In drug-naive patients with schizophrenia, a decrease of GSH levels was observed in cerebrospinal fluid and medial PFC.236 As such, GSH supplementation could be of clinical benefit in the treatment of schizophrenia by preventing oxidative stress and enhancing neurotransmission at NMDA-Rs. However, oral administration of GSH has been shown to have little effect on brain GSH levels.234
N-acetylcysteine (NAC), a precursor of GSH, penetrates the blood–brain barrier and raises brain GSH levels in animal models (Figure 1).234 In a large-scale, double-blind, 6-month RCT of NAC as adjunctive therapy, NAC significantly improved psychopathological symptoms of schizophrenia, particularly negative symptoms in patients with chronic schizophrenia. Interestingly, NAC also improved akathisia, suggesting a beneficial effect of NAC for EPS. In an ancillary component of the primary study, NAC improved auditory cortical functioning as indexed by the mismatch negativity, a marker of glutamatergic function.237 Taken together, NAC supplementation appears to be a promising augmentation strategy in the treatment of schizophrenia that warrants further investigation.
α2 AR agonist or antagonist
Norepinephrine plays an important role in cognitive function of the PFC, particularly in working memory, by its actions at α2 ARs (Table 2).238, 239 Indeed, the α2 agonists (clonidine and a more selective α2A agonist guanfacine) have been shown to improve cognition in various preclinical studies.240, 241, 242 In an early clinical trial, clonidine improved PFC-mediated cognitive dysfunction in schizophrenia.243 A 4-week, double-blind RCT demonstrated a cognitive-enhancing effect of adjunctive guanfacine in patients with schizophrenia taking risperidone.244 The potential ability of α2 agonists to improve cognitive performance on tasks dependent on PFC function should be investigated in further large-scale RCTs.
On the contrary, clozapine and risperidone have potent antagonist properties at α2 ARs, which may contribute to the ‘atypicality’ of the SGAs by enhancing frontocortical dopaminergic transmission.245, 246 Several clinical and preclinical studies provide evidence that the α2 AR antagonist idazoxan enhances the antipsychotic efficacy of FGAs, olanzapine and risperidone.247, 248, 249 There are no current clinical trials investigating α2 AR antagonists in schizophrenia. In the case of α2 ARs ligands, it remains unclear whether agonists or antagonists will have greater potential benefits in schizophrenia.
Diverse evidence suggests that catechol-O-methyl transferase (COMT), a postsynaptic methylation enzyme that metabolizes released dopamine, is primarily responsible for synaptic dopamine inactivation in PFC, and that variation in COMT activity may affect prefrontal cortical activity, especially during working memory and executive function tasks.250, 251 Abnormalities of prefrontal dopamine function associated with working memory appear to be prominent features of schizophrenia, and specific alleles of the COMT gene run in families with a high incidence of the illness.252
Tolcapone, a CNS penetrant reversible selective inhibitor of COMT, has been reported to enhance prefrontal cognitive function in rodents253 and in normal human subjects.254 Tolcapone and a peripherally acting COMT inhibitor entacapone are currently in phase II clinical trials to investigate whether these COMT inhibitors offer a new adjunctive treatment for cognitive impairment in schizophrenia with and without the high-risk combination of COMT alleles.255 However, tolcapone was withdrawn from the market in Europe and Canada due to an increased risk of serious hepatotoxicity,256 and in the United States, restrictive liver enzyme monitoring measures are mandated that severely limits its use.257
α-7 nicotinic receptor agonist
Nicotinic Ach receptors (nAChRs) have been implicated in cognitive function and sensory processing. Among the nAChRs, the α7 nAChR subtype is a possible therapeutic target for schizophrenia, since preclinical, genetic and postmortem studies have demonstrated altered levels and function associated with the illness.258 Selective α7 nAChR agonists such as 3-2,4-dimethoxybenzylidene anabaseine (DMXB-A), R3487/MEM3454, PHA-709829, PH-399733, EVP-6124 and TC-5619 have been developed as potential candidates for adjunctive treatments of cognitive impairment and negative symptoms in schizophrenia (Table 2).259, 260, 261, 262, 263 Despite early concerns that rapid desensitization of the α7 nAChR would limit their therapeutic potential, several α7 nAChR agonists have already advanced to clinical trials. A phase II trial of DMXB-A found improvement in cognition that did not separate from placebo, while negative symptoms improved significantly.264 DMXB-A reduced the activity of the hippocampus during pursuit eye movement task as measured with fMRI (functional magnetic resonance imaging), suggesting its therapeutic effect on hippocampal inhibitory interneurons.265 It also improved default network function in schizophrenia.266 Continued expectations for DMXB-A as a pro-cognitive agent are reflected in ongoing clinical trials. In a phase Ib double-blind, 3-week RCT, EVP-6124 demonstrated significant improvement in measures of cognition and electrophysiological biomarkers of sensory processing when compared with placebo in schizophrenia patients treated with SGAs.259
There is additional evidence from a study with tropisetron, a high-affinity partial agonist for α7 nAChR and a potent 5-HT3 receptor antagonist. In a recent 8-week double-blind RCT, adjunctive administration of tropisetron improved auditory sensory gating P50 deficits and sustained visual attention compared with baseline in non-smoking patients with schizophrenia.267 However, the sample size of this trial was small and tropisetron was not superior to placebo on cognitive performance.
A phase IIa study of TC-5619, a selective α7 nAChR partial agonist, found that patients treated with TC-5619 performed significantly better on selective cognitive tests on the Cog State Schizophrenia Battery and showed improvement on the Scale for the Assessment of Negative Symptoms, CGI-Improvement (CGI-I) and Subject Global Impression Scale of Cognition (SGI-Cog). However, this effect was predominantly in patients who were tobacco smokers.268 Additional studies are required to determine the therapeutic potential of α7 nAChR agonists in schizophrenia.
α4–β2 nicotinic receptor agonist
The α4 β2 nAChRs are also known to be involved in cognition, and their agonists such as AZD3480 (TC-1734), varenicline, SIB-1553A and RJR2403 have been developed as adjunctive agents to APDs for the treatment of cognitive deficits in schizophrenia.56, 269 However, in a phase IIb trial of AZD3480, various domains of cognition were not improved in schizophrenia.270 A small open-label study of varenicline, which is also a full agonist at α7 nAChR, showed that the agent has some beneficial effects on cognition (verbal learning and memory) and reduces smoking behavior in schizophrenia patients.271 Another more recent larger study examined the effects of adjunctive varenicline in an 8-week randomized placebo-controlled trial. Antipsychotic doses remained constant through the study in which varenicline was titrated up to 1 mg daily. While the primary analyses of neurocognitive function showed no drug–placebo differences, some secondary analyses found advantages over placebo. Treatment effects differed between smokers and non-smokers.272 While not reported in patients with schizophrenia, there have been case reports and an US FDA advisory about possible psychiatric side effects of varenicline, including depression, suicidal ideation or attempts, and activation of psychotic or manic symptoms (Table 2).273, 274 Further study of varenicline in patients with schizophrenia is needed to understand whether a risk of behavioral toxicity also exists in clinically stable patients receiving APD.
Muscarinic receptor agonist
There is a large body of anatomical and pharmacological evidence demonstrating the potential for modulation of dopamine and glutamatergic neurons by cholinergic muscarinic receptors.275 Moreover, drugs enhancing muscarinic receptor function would be expected to lead to enhancement of GABAergic interneuron function.196 There is growing evidence that partial agonists of muscarinic receptors are active in animal models predictive of antipsychotic activity, and the SGAs clozapine and olanzapine are partial agonists for muscarinic M1, M2 and M4 receptors. In addition, muscarinic agonists have activity in animal models of negative symptoms, cognitive dysfunction and affective disorders, suggesting potential usefulness of muscarinic agonists either alone or in combination with APDs in the treatment of schizophrenia.220, 270, 276 Examples of these agents are the muscarinic M1/M4 agonist xanomeline, the M2/M4 agonists PTAC and BuTAC and the M1/M3 agonists CDD-0102, CI-1017 and YM-706.179, 275
In a pilot double-blind, 4-week RCT, xanomeline demonstrated efficacy on measures of cognition and total BPRS and PANSS scores when compared with placebo in unmedicated patients with schizophrenia.277 This was a small trial, and further studies are necessary to determine whether xanomeline is effective for cognitive deficits and general psychopathology in schizophrenia, either as monotherapy or as adjunctive treatment.13
Acetylcholinesterase (AchE) is an enzyme responsible for catabolizing Ach in the synaptic cleft. AchE inhibitors can enhance cholinergic function by increasing Ach concentrations and potential activity at both nicotinic and muscarinic receptors.13 A number of studies have examined AchE inhibitors, including donepezil, rivastigmine and galantamine, as potential cognitive enhancers added to APDs in schizophrenia. The results of double-blind RCTs of donepezil and rivastigmine suggest that these agents are not particularly effective for improving cognitive function in schizophrenia.13
Galantamine is an AchE inhibitor that also acts as an allosteric modulator of the α4 β2 and α7 nAChRs. Of the seven double-blind, placebo-controlled trials in patients with schizophrenia, four have provided some evidence of improvements in cognitive function with galantamine compared with placebo on total cognitive battery scores and on measures of memory and processing speed.278, 279, 280, 281 However, results from three recent studies found no cognitive benefits in schizophrenia.282, 283, 284 Moreover, one double-blind RCT of galantamine failed to show improvements in negative symptoms.285 These results provide no global support for galantamine augmentation for cognitive and negative symptoms in schizophrenia, although specific cognitive deficits may be responsive to adjunctive galantamine in some patients.
Histamine H3 receptor antagonist/inverse agonist
The histamine H3 receptor (H3R) is a presynaptic autoreceptor that inhibits histamine release in the brain and also regulates the release of numerous other neurotransmitters (for example, dopamine, ACh, noradrenaline and GABA) via a parallel role as a heteroreceptor.286, 287 A number of H3R antagonists/inverse agonists have been synthesized and evaluated for their potential utility in the treatment of schizophrenia.286, 287 Preclinical studies suggest that these compounds can enhance the release of such neurotransmitters simultaneously, which play important roles in cognitive processes. Moreover, pitolisant (BF2.649), a high affinity and selective non-imidazole H3R antagonist/inverse agonist, showed significant inhibitory activity in several mouse models of schizophrenia.288 As such, H3R antagonists/inverse agonists deserve attention as a possibly novel class of drugs endowed with pro-cognitive properties (Table 2). Several H3R antagonists/inverse agonists, including ABT-288, pitolisant (BF2.649), GSK-239512 and MK-0249, are currently under clinical investigation in phase II trials for the treatment of schizophrenia.286
Cannabinoid-1 receptor antagonist
Acute cannabis intoxication and long-term cannabis use can induce schizophrenia-like symptoms. In addition, cannabis is thought to be a risk factor for developing schizophrenia.289 Accumulating evidence suggests that cannabinoid-1 (CB1) receptors and their accompanying system of endogenous activators may be dysregulated in schizophrenia,290 and CB1 receptors also play an important role in glucose and lipid metabolism.291 Consequently, the potential utility of CB1 receptor antagonists for the treatment of psychiatric symptoms and obesity in schizophrenic patients has received considerable attention. In a double-blind RCT, a selective CB1 receptor antagonist, SR-141716, used as monotherapy, failed to demonstrate benefits for psychopathological symptoms in schizophrenia, although only a single dose of the experimental agent was studied.177 Rimonabant, another CB1 receptor antagonist that had approval for weight loss in Europe, did not reduce body weight or improve metabolic parameters in a small study of overweight patients with schizophrenia292 and another small-scale placebo-controlled study employing the Repeatable Battery for the Assessment of Neuropsychological Status total score as the primary outcome measure found no improvement of global cognitive functioning.293 The agent was withdrawn from the European market in 2009 due to increased risk of psychiatric adverse effects. This also led to the early termination this RCT. Cannabidiol, a cannabis constituent that does not appear to act on CB1 receptors, may have some potential for antipsychotic effects.294
GABA is the major CNS inhibitory neurotransmitter and GABAergic interneurons inhibit neuronal output of glutamatergic pyramidal cells in the PFC through activation of GABA type A (GABAA) receptors containing α2 subunits.295 Agents that increase GABAergic inhibition of cortical pyramidal cells have been hypothesized to improve working memory and other cognitive impairments in schizophrenia (Table 2).
In a small double-blind RCT, adjunctive administration of MK-0777, a GABAA α2/α3 partial agonist, improved delayed memory performance and decreased reaction time on selected measures of prefrontal cortical function in patients with chronic schizophrenia.296 However, a recent double-blind RCT of MK-0777 failed to show any significant benefits for cognitive deficits or psychopathological symptoms in schizophrenia.297 MK-0777 is a relatively weak GABAA α2 partial agonist, with 10–20% of the potency of a full GABAA α2 agonist, thus a more potent partial agonist with greater intrinsic activity might be necessary.
Antiinflammatory agents (Celecoxib)
In line with the findings indicating immunological alterations in schizophrenic patients, non-steroidal antiphlogistic agents have been explored. In a placebo-controlled, double-blind and randomized study, Akhondzadeh et al.298 added 400 mg per day celecoxib, a cyclooxygenase-2 inhibitor, to risperidone in 60 acutely ill schizophrenic in-patients. The cyclooxygenase-2 group outperformed the placebo group in terms of positive and general psychophathology symptoms as well as total PANSS scores although improvement was found in both groups. In another placebo-controlled, randomized, double-blind study, Müller et al.299 confirmed an earlier study of theirs and found an improvement of PANSS and CGI scores in first-episode schizophrenic patients after co-administration of celecoxib to a therapeutic dose of amisulpride.
The therapeutic potential of minocycline, a second-generation tetracycline, was shown in several studies using animal models of schizophrenia. For example, minocycline attenuated cognitive impairments and PPI deficits induced by NMDA-R antagonists.300, 301, 302 In addition, minocycline attenuated the increase of dopamine levels in frontal cortex and striatum induced by a NMDA-R antagonist,302 and ameliorated neurotoxicity caused by methamphetamine.303 The mechanism by which minocycline acts is not fully understood, but it has demonstrated neuroprotective effects in animal models of ischemic brain injury, inflammatory-mediated neurotoxicity and classic neurodegenerative disorders, possibly via its inhibition of nitric oxide synthase, inhibition of, microglial activation, and antiapoptotic properties.304
In a recent 24-week double-blind RCT, minocycline added on to SGAs showed a beneficial effect on negative symptoms and cognitive functioning, including executive function, spatial working memory and spatial recognition memory, in patients with early-phase schizophrenia.305 Additional work is needed to more fully delineate the potential of this agent.
Neurokinin-3 receptor antagonist
Neurokinin-3 (NK3) tachykinin receptors are located in the brain regions implicated in schizophrenia and appear to regulate midbrain dopamine neuronal activity.306 Preclinical studies have demonstrated that a potent and selective non-peptide NK3 receptor antagonist, osanetant (SR-142801) selectively inhibits dopamine release in certain brain regions.307 In a 6-week RCT, osanetant monotherapy was superior to placebo on global assessment of efficacy and measures of positive symptoms in schizophrenia.177 A double-blind phase II RCT of another selective NK3 receptor antagonist, talnetant (SB-223412), has also shown significant improvement in positive symptoms and cognitive impairment with no major side effects in schizophrenia.308 However, both osanetant and talnetant have significant pharmacokinetic limitations309 and development of these compounds for schizophrenia has been stopped.310 Nevertheless, NK3 antagonists have demonstrated some potential for antipsychotic and pro-cognitive effects and this mechanism clearly merits further study. AZD2624 is a NK3 receptor antagonist. It has reached phase II clinical trials.
Epidemiological studies have shown that female patients have a considerably later onset of schizophrenia, better therapeutic responses and outcomes than males, and women are more likely to develop late-onset schizophrenia after menopause.311 Clinical studies have demonstrated the inverse relationship between estradiol levels and positive symptoms over the menstrual cycle in premenopausal women with schizophrenia.312 In preclinical studies, estrogen has been shown to modulate the sensitivity of D2 receptors in the brain and it can reduce behavioral abnormalities induced by dopamine agonists.313 These findings have been interpreted to provide a theoretical rationale for the antipsychotic-like effects of estrogen.314 In three double-blind RCTs, adjunctive estrogen showed a beneficial effect on positive symptoms in women with schizophrenia.311, 315, 316 Although estrogen may be a useful treatment for women with schizophrenia, its therapeutic potential in men with schizophrenia remains unclear.
Accumulating evidence suggests that neuroactive steroids, including pregnenolone, pregnenolone sulfate, allopregnanolone, dehydroepiandrosterone (DHEA) and DHEA sulfate (DHEAS), act at inhibitory GABAA and excitatory NMDA-Rs, and have neuroprotective and neurotrophic properties.317, 318 Thus, neurosteroids may be candidate modulators of the pathophysiology and therapeutics of schizophrenia.319
In a small double-blind RCT, pregnenolone as adjunct to SGAs significantly reduced negative symptoms, but not cognitive impairments, in patients with schizophrenia or schizoaffective disorder compared with placebo.320 A small double-blind RCT of adjunctive DHEA in patients with chronic schizophrenia and prominent negative symptoms found that adjunctive DHEA was superior to placebo in treating negative, depressive and anxiety symptoms, especially in women.321 In another study in olanzapine-treated patients, DHEA augmentation was associated with an improvement in negative symptoms, but no effect on positive symptoms.322 Interestingly, some improvement in EPS was also seen in patients receiving DHEA, consistent with a previous study in which DHEA significantly decreased antipsychotic-induced Parkinsonism.323 Finally, in another study, DHEA augmentation produced modest pro-cognitive effects while showing no effects on positive, negative or Parkinsonian symptoms.324 Although the findings with neurosteroid augmentation in schizophrenia are somewhat inconsistent and the mechanism of action of pregnenolone and DHEA requires further characterization, the possibility that these compounds have efficacy in schizophrenia should be further investigated in longer and larger studies with broader dose ranges.
Omega-3 fatty acid
The two main omega-3 fatty acids in fish oil, eicosapentaenoic acid and docosahexaenoic acid have important biological functions in the CNS. Eicosapentaenoic acid can affect neuronal activity and docosahexaenoic acid is a major structural component of neuronal membranes.325 Moreover, omega-3 fatty acids have antiapoptotic, antioxidant and antiinflammatory effects in the brain.326, 327 A number of studies suggest that deficient uptake or excessive breakdown of membrane phospholipids may be associated with schizophrenia.328 These have led to research interest in the possible therapeutic benefits of adjunctive use of omega-3 fatty acids in schizophrenia, especially given their favorable side-effect profiles and their potential to restore disturbances in the metabolism of synaptic membrane phospholipids that may be responsible for dysfunction of dopaminergic and serotonergic receptors.
Results of several double-blind RCTs of omega-3 fatty acids in schizophrenia remain inconclusive.329, 330 However, a recent double-blind RCT demonstrated that 12-week administration of omega-3 fatty acids can reduce the risk of conversion from a prodromal state into first-episode psychosis.331 Omega-3 fatty acids were well tolerated, thus they could potentially represent a viable and safe preventative monotherapy in people at ultra-high risk of psychosis.
The posterior pituitary hormone oxytocin is a neuropeptide that plays a key role in the regulation of a number of diverse behavioral and cognitive processes, including social attachment, affiliation and memory.332, 333 In addition, preclinical studies have shown that oxytocin has antipsychotic-like efficacy in several animal models of schizophrenia.334, 335 In human studies, acute intranasal oxytocin administration increased interpersonal trust, eye contact, performance on tests of face emotion recognition and theory of mind as well as social reciprocity in healthy336, 337 and autistic subjects.338 Feifel et al.339 recently reported that 3-week administration of intranasal oxytocin, given adjunctive to APDs, produced significantly greater reductions in schizophrenic symptoms compared with placebo. Moreover, in a preliminary double-blind RCT, 2-week daily intranasal administration of oxytocin improved measures of social cognition and psychotic symptoms, especially paranoia in schizophrenia.340 In order to characterize the antipsychotic potential of oxytocin, further trials are needed with larger sample sizes, a broader dose range, and longer durations of treatment.
Phosphodiesterase 10A (PDE10A) is a dual specificity enzyme hydrolyzing the cyclic nucleotide messengers, cAMP and cGMP.341 PDE10A is abundant only in brain tissue, with a predominant distribution in the putamen and caudate nucleus in the medium spiny neurons of the striatal complex.342 PDE10A inhibitors, including papaverine, TP-10 and MP-10 (PF-2545920), demonstrate positive effects in several animal models, spanning positive, negative and cognitive symptoms of schizophrenia.341, 343 However, a phase II trial of MP-10 was terminated before completion for undisclosed reasons.341
Secretin is a gastrointestinal peptide increasingly recognized as an important neuropeptide with extensive CNS receptor expression and activity. Preclinical studies have demonstrated that systemically administered secretin is capable of modulating conditioned fear344 and can partially reverse impairment in PPI induced by PCP.345 In a small pilot double-blind RCT, a single intravenous dose of secretin produced transient improvement in the CGI-I scale in patients with treatment-refractory schizophrenia, although no significant differences in psychiatric symptoms were found between secretin- and placebo-treated patients.346 A recent double-blind RCT showed that subcutaneous administration of secretin significantly increased eye-blink conditioning in schizophrenia, suggesting a possible role for secretin in modulating cerebellar-mediated learning by classical conditioning.347 Further studies evaluating the efficacy and safety of secretin are warranted, especially among less severely ill patients with schizophrenia.
The hematopoietic growth factor erythropoietin (EPO) possesses multifaceted direct neuroprotective properties such as antiapoptotic, antiinflammatory, antioxidant, neurotrophic, angiogenic and synaptogenic activity.348, 349 EPO was generally found to be clinically safe and crosses the blood–brain barrier.348 In a double-blind RCT, weekly intravenous administration of high-dose rhEPO (recombinant human EPO) over 12 weeks, given adjunctive to APDs, produced significantly greater improvement in a set of cognitive functions as compared with placebo in patients with chronic schizophrenia.350 Moreover, rhEPO was able to delay the progressive atrophy in the brain areas typically affected in schizophrenia.351 Importantly, the gray matter protection by rhEPO was highly associated with improvement in attention and memory functions.351 Thus, EPO may have the potential to prevent progressive neuropathological changes and to improve cognitive function in schizophrenia. Further clinical studies are warranted with larger sample sizes, a broader dose range and longer durations of treatment.
Ginkgo, a traditional Chinese medicine, has antioxidant and immunostimulatory properties and it can improve brain circulation at the microvascular level.352 In a recent meta-analysis of six trials, adjunctive ginkgo to APDs demonstrated moderate improvement in total symptomatology and in negative symptoms in patients with chronic schizophrenia.352 However, ginkgo appears to be effective only when added to FGAs. Moreover, it remains unclear as to what properties play a therapeutic role in schizophrenia.352
Future strategies for drug development in schizophrenia
Since the introduction of chlorpromazine and throughout the development of the new-generation APDs beginning with clozapine, the D2 receptor has been the ‘Holy Grail’ for the development of APDs. Pharmacologic actions to reduce neurotransmission through the D2 receptor represent the only proven therapeutic mechanism for psychoses. A number of novel non-D2 mechanisms of action of APDs have been explored over the past 40 years but none has definitively been proven effective. Consequently, it remains unclear as to which pharmacologic mechanism will provide the highest level of efficacy while avoiding serious side effects in the treatment of schizophrenia. Moreover, there is still an ongoing debate as to whether drugs selective for singe molecular targets (that is, ‘magic bullets’) or drugs selectively non-selective for several molecular targets (that is, ‘magic shotguns’, ‘multifunctional drugs’ or ‘intramolecular polypharmacy’) will lead to more effective new medications for schizophrenia.3, 27, 152, 179, 353, 354 Given the complexity of schizophrenia as a disease, and the heterogeneity of the patient population, it would seem logical to develop compounds with at least a modicum of D2 receptor affinity that also bind one or more favored targets such as 5-HT1A, 5-HT2A, 5-HT2C, 5-HT6, 5-HT7, glutamatergic and/or nicotinergic receptors, while avoiding engagement of problematic targets such as α1-adrenergic, H1, M1 and M3 receptor activity.355 Another option is to develop single-target agents that can be used to augment multi-target agents. The conundrum of single-target vs multi-target agents will likely remain at the forefront of drug development until the pathophysiology of schizophrenia is fully elucidated.
In this context, current and future drug development strategies can be seen to fall into three categories: (1) refinement of precedented mechanisms of action to provide drugs of comparable or superior efficacy and side-effect profiles to existing APDs; (2) development of novel (and presumably non-D2) mechanism APDs; (3) development of compounds to be used as adjuncts to APDs to augment efficacy by targeting specific symptom dimensions of schizophrenia and particularly those not responsive to traditional APD treatment.
Given that cognitive deficits in schizophrenia are widely prevalent and appear to be correlated with functional outcome, much effort has been focused on developing cognitive-enhancing drugs. As described above, a variety of molecular targets with potential pro-cognitive effects in schizophrenia have been identified. However, as yet there is no proven mechanism for developing such drugs and many barriers still exist in the translation from basic science to drug discovery.356 As such, developments in translational neuroscience are needed to bridge from preclinical to proof-of-concept studies so that they can, to the greatest extent possible, use similar outcome measures.77 Although hypotheses of cognitive deficits in schizophrenia separately implicate dopaminergic, cholinergic, noradrenergic, serotonergic, glutamatergic and/or GABAergic deficits, it is very likely that the actual circuit dysfunction involves interactive changes across some or most of these systems.242 Accordingly, such complexity may limit the potential effectiveness of agents targeting a single mechanism, and more creative approaches may be necessary.13
From a mechanistic perspective, a more rational approach to the treatment of cognitive deficits in schizophrenia would be to develop compounds that specifically target the mesocortical system, with particular selectivity for the PFC and medial temporal lobe, while avoiding other dopaminergic pathways (for example, D1 mediated). In order to develop such agents, it is necessary to learn more about how individual compounds affect specific brain regions in both preclinical and clinical studies.77 Several neurophysiological and non-invasive neuroimaging biomarkers, including oculomotor neurophysiology assessments, fMRI, evoked response potentials, MR spectroscopy and magnetoencepholography, hold great promise to enhance translational research. Reliable and sensitive measures of brain function will likely prove essential for enhancing our ability to characterize the biological effects of novel therapeutic agents in the early stages of drug development.77
In recent years, significant progress has been made in identifying various susceptibility genes in schizophrenia, including dysbindin, neuregulin 1, COMT, disrupted in schizophrenia 1 (DISC1) and others, by genetic linkage and association studies.56 Many of these genes appear to regulate synaptic plasticity and chemical neurotransmission particularly by glutamate. Research on the products of these implicated genes will promote rational drug development. Indeed, investigators have pursued a number of genetic animal models for schizophrenia based on single genes as etiologic agents of the illness.357 However, these animal models cannot be expected to recapitulate the breadth of the clinical phenotype of schizophrenia, because schizophrenia is a complex multi-factorial, polygenic brain disorder involving environmental interactions. More sophisticated bioinformatic animal models based on etiological considerations could produce new opportunities to identify the molecular pathophysiology of the disease phenotype.357 However, at present, we have no molecular target for any of the dimensions of schizophrenia.
There is a great need for the development of novel methods to identify optimal individualized treatment plans. In particular, the efficacy and tolerability of antipsychotics could be directly influenced by genetic variations in cytochrome P450 (CYP) enzymes that affect drug metabolism. The activity of drugs may also be influenced by genetic alterations affecting the drug target molecule, such as monoaminergic receptors, neurotransmitter transporters and metabolizing enzymes. Further development of genetic tests for the prediction of drug response and side effects, and pharmacogenetic research into genetically determined drug metabolic polymorphisms, as well as pharmacogenomic strategies to the identification of novel factors influencing response, would lead to a better understanding of the rational basis for the personalization of antipsychotic treatment.358 In addition, APDs may also be targeted to specific patient subgroups based on profiling and the identification of endophenotypes of schizophrenia. Clinical implementation of such practices could have a strong impact in reducing adverse effects and improving treatment adherence and efficacy.358
In recent decades, a focus on early detection and early intervention in psychosis has emerged.359 As already described, various lines of evidence indicate that pathophysiological changes occur in cortical and subcortical brain regions in patients with schizophrenia, which may be associated with disease progression, clinical deterioration and functional decline.65 If that is the case, early intervention aimed at halting or reversing progressive pathophysiological processes in schizophrenia could yield substantial improvements in outcomes in the future. To do so, more sensitive and specific diagnostic tools as well as safe and effective interventions including novel therapeutic agents are required.1
Pharmacotherapy for schizophrenia is highly effective but at the same time there is tremendous unmet clinical need. Over the last half century, there has been only limited progress in innovating mechanisms of action and developing novel therapeutic agents for the treatment of schizophrenia. APDs have mainly been developed in the procrustean mold of D2 receptor antagonists. Despite refinements in this mechanism of action, the therapeutic effects of the NGAs are not yet sufficiently robust to claim superiority over earlier treatments, with the possible exception of clozapine. Moreover, novel drugs in the development for schizophrenia have not yet been proven effective or validated alternative mechanisms of action. Other novel approaches to enhance efficacy and change the ultimate course and prognosis of schizophrenia need to be explored. The breadth of potential targets and compounds under investigation clearly demonstrate the interest and importance in pursuing innovative drug development. The discovery of effective novel therapeutic agents for the treatment of schizophrenia will require continued research efforts and collaboration by both academic and industrial laboratories to further our understanding of the molecular and functional pathophysiological mechanisms operative in schizophrenia.