Emerging Approaches for Treatment of Schizophrenia: Modulation of Cholinergic Signaling

Introduction

Schizophrenia is a devastating and complex psychiatric illness that affects approximately 1% of the population worldwide. Schizophrenia usually manifests with an onset of symptoms in late adolescence/young adulthood and these symptoms can be grouped into three main categories. Positive symptoms include disordered thoughts/speech, delusions, and hallucinations. Negative symptoms are comprised of blunted affect, anhedonia, and social withdrawal; while cognitive impairments include disturbances in attention, working memory, and executive functions (Lewis and Lieberman, 2000). There is strong evidence that genetics plays a key role in schizophrenia; however, it is becoming increasingly evident that this disorder is polygenic in nature and likely involves the interaction between hundreds of genes and environmental factors (Sullivan et al., 2012). Similar to the genetic nature of schizophrenia, the neurobiology underlying this disease is complex. Post-mortem and neuro-imaging studies have demonstrated altered connectivity and function in numerous brain regions including (but not limited to) the prefrontal cortex, hippocampus, striatum, and thalamus (Ross et al., 2006). Until recently the prevailing hypothesis was that the primary pathology of schizophrenia was a dysregulation of signaling by monoamines such as dopamine and serotonin. However, it is becoming increasingly clear that the disease etiology cannot be explained solely as a result of changes in monoamine signaling. Recent research suggests that altered signaling via the N-methyl-D-aspartate (NMDA) receptor subtype of glutamate receptors, GABA_A receptors in the prefrontal cortex, and cholinergic muscarinic (mAChR) and nicotinic (nAChR) receptors may contribute to all three symptom clusters of schizophrenia (Lisman et al., 2008; Marin, 2012; Noetzel et al., 2012; Sarter et al., 2012).

Current therapies for schizophrenia include the use of both typical (e.g., haloperidol) and atypical antipsychotics (e.g., clozapine and risperidone) which both exert their primary effects through D₂ dopamine receptor antagonism as well as activity at a host of other monoamine receptors (Sawa and Snyder, 2003). These treatments are at least partially effective in treating the positive symptoms of schizophrenia. However, studies have shown that up to 74% of schizophrenia patients discontinue the use of atypical antipsychotics within 18 months due to adverse effects such as parkinsonian-like motor effects and metabolic syndrome which are tied to their mechanism of action (Lieberman et al., 2005). Even for those patients that do not experience pronounced side-effects, the treatments available do not reduce the cognitive and negative symptoms. Given the shortcomings of current therapies it is imperative that novel approaches are developed that can be used to treat all three symptom clusters associated with schizophrenia with fewer adverse effects.

Targeting Muscarinic Receptors for Treatment of Schizophrenia

Multiple lines of evidence suggest that increasing cholinergic signaling via both muscarinic (mAChR) and nicotinic receptors (nAChR) could be beneficial in schizophrenia. Post-mortem studies have demonstrated a decrease in mAChR expression levels in several brain regions of schizophrenic patients including the pre-frontal cortex, striatum, and hippocampus (Crook et al., 2000; Dean et al., 1996; Zavitsanou et al., 2004). Consistent with this finding, administration of nonselective mAChR antagonists can induce psychotic-like symptoms in healthy human volunteers (Osterholm and Camoriano, 1982), and exacerbate existing symptoms when administered to schizophrenic patients (Tandon et al., 1991). Conversely, increasing cholinergic signaling via administration of acetylcholinesterase (AChE) inhibitors can improve cognitive performance (Feldman et al., 2007; Forette et al., 1999), making cholinergic signaling an attractive target to treat a multitude of disorders associated with cognitive impairments. Unfortunately, most agents that target cholinergic signaling have failed in clinical trials due to dose-limiting effects caused by activation of peripheral receptors.

Most mAChR agonists that have entered clinical testing were originally developed for treatment of Alzheimer’s disease (AD). Of the mAChR agonists, the M₁/M₄ preferring agonist xanomeline progressed the furthest in the clinical development for treatment of cognitive deficits associated with Alzheimer’s disease. While xanomeline showed a trend towards improving cognitive function in AD patients, the cognition-enhancing effects observed in a phase III trial failed to reach statistical significance. However, a surprising finding from this study was that this M₁/M₄ preferring agonist had robust therapeutic effects on psychotic symptoms and behavioral disturbances associated with AD. In particular, xanomeline produced robust, dose-dependent reductions in delusions, hallucinations, vocal outbursts, suspiciousness, and agitation while improving blunted affect and other behavioral disturbances (Bodick et al., 1997a; 1997b). More recently, a four-week, double-blind, placebo-controlled outcome trial in subjects with schizophrenia (n=20) was performed to further evaluate the antipsychotic efficacy of xanomeline. This study found robust improvements in both the positive and negative symptoms of schizophrenic patients compared to the placebo group. Both the magnitude and the time-course of xanomeline efficacy were superior to those observed with typical antipsychotics. Therapeutic effects of xanomeline were statistically significant after one week of treatment as opposed to the multi-week delay observed with typical antipsychotics. In terms of treating cognitive impairments, xanomeline produced statistically significant improvements in some measures of cognitive function including verbal learning and short-term memory (Shekhar et al., 2008). Unfortunately, adverse effects, particularly gastrointestinal-related, were observed with xanomeline and dose-limitations have removed it from consideration for long-term clinical use. However, these clinical studies provide strong validation for mAChR agonists as potential therapeutics to treat psychosis and behavioral disturbances in a wide spectrum of diseases including schizophrenia and neurodegenerative diseases.

Allosteric Modulators of Muscarinic Receptors

While results of clinical studies with mAChR agonists such as xanomeline are encouraging, the clinical utility of nonselective muscarinic agonists and AChE inhibitors is limited, primarily due to the dose-limiting side effects including brachycardia, GI distress, excessive salivation, and sweating. The mAChRs are members of the family A class of G-protein coupled receptors and include five subtypes named M₁-M₅. Evidence suggests that the most prominent adverse effects of AChE inhibitors and other nonselective cholinergic agents are mediated by peripheral M₂ and M₃ receptors (Bymaster et al., 2003). Accordingly, efforts were made to specifically target M₁, M₄, or M₅ receptors with the goal of retaining the therapeutic efficacy of non-selective cholinergic agents without the adverse side-effects. However, the acetylcholine (ACh) binding site (also called the orthosteric site) is highly conserved across the five subtypes, making it difficult to develop subtype-selective ACh-site ligands. To circumvent this problem, an approach of targeting binding sites that are separate from the ACh-site and are referred to as allosteric sites has been adopted. These allosteric, non-ACh binding pockets are less well conserved between receptor subtypes, and targeting these sites has been highly successful for multiple G-protein coupled receptors including muscarinic receptors (Conn et al., 2009a; May et al., 2007). Allosteric activators can function mechanistically as either allosteric agonists, which in the course of binding to the allosteric site directly induce receptor activation, or as positive allosteric modulators (PAMs) which do not activate the receptor directly but potentiate the ability of the endogenous agonist ACh to induce receptor activation. These mechanisms are not exclusive in nature and in some cases a single molecule can have both allosteric potentiator and allosteric agonist activity (Conn et al., 2009b). As discussed below, the discovery of subtype-specific allosteric modulators has greatly advanced our understanding of the physiological role of various muscarinic receptor subtypes in the brain and have emphasized the potential utility of mAChR subtypes as targets for the development of novel antipsychotic treatments.

M₄ Muscarinic Receptor

While the receptor subtypes that mediate the clinical antipsychotic efficacy of xanomeline have not been definitively established, several lines of evidence suggest that M₁ and/or M₄ are likely responsible. The M₄ receptor is expressed in many brain regions including the striatum, cortex, and hippocampus (Hersch et al., 1994; Levey et al., 1995; 1991). This receptor canonically signals via activation of the Ga_i/Ga_o G-proteins with concomitant inhibition of adenylyl cyclase and cAMP production. M₄ knock-out mice show an overall hyper-dopaminergic phenotype with increased locomotor activity, and increased basal- and psychostimulant-induced accumbal dopamine levels (Gomeza et al., 1999; Tzavara et al., 2004). M₄ knock-out mice, unlike wild-type mice, are not susceptible to muscarinic agonist-induced reductions in striatal and accumbal dopamine (Threlfell et al., 2010), suggesting that M₄ negatively regulates dopaminergic tone in these areas. M₄ receptors have been found to co-localize with D₁ dopamine receptors in the striatum (Ince et al., 1997), and selective deletion of M₄ from D₁-expressing cells is sufficient to enhance hyperlocomotor activity and increase behavioral sensitization to psychostimulants (Jeon et al., 2010). Further evidence of the importance of this subpopulation of M₄ receptors to antipsychotic efficacy comes from the finding that these mice are no longer responsive to the antipsychotic effects of xanomeline as assessed by amphetamine-induced locomotion (Dencker et al., 2011).

The above studies suggest that M₄-specific modulators could represent a novel therapeutic approach for the treatment of schizophrenia. A major breakthrough in M₄ PAMs occurred in 2008 with the independent discoveries of two structurally distinct M₄ PAMs, VU0010010 and LY2033298. The M₄ modulator VU0010010 demonstrated robust and potent (EC₅₀ = 400nM) activity at M₄ and was void of activity at all the other mAChR subtypes. Extensive in vitro characterization demonstrated that VU0010010 bound to an allosteric site on M₄ and that binding did not cause activation of the receptor directly, but instead increased the affinity of the receptor for ACh. In brain slice preparations VU010010 potentiated M₄-mediated depression of glutamatergic, but not GABAergic, afferents onto CA1 pyramidal cells of the hippocampus indicating a key regulatory role for M₄ in hippocampal function (Shirey et al., 2008). Chemical optimization of VU0010010 led to the discovery of two related PAMs, VU0152100 and VU0152099, demonstrating M₄ subtype-specificity and possessing improved physiochemical properties making them suitable for in vivo administration. Importantly, both VU0152100 and VU0152099 have demonstrated efficacy in reversing amphetamine-induced hyperlocomotion in rats (Brady et al., 2008). LY2033298 represents another structurally-distinct M₄-selective PAM that has specificity for the M₄ receptor and can be administered to rodents in vivo. Systemic administration of LY2033298 produces robust effects in many preclinical models of psychosis including conditioned avoidance responding, and amphetamine-induced locomotion (Chan et al., 2008; Leach et al., 2010; Suratman et al., 2011). The collective data from these subtype-selective PAMs demonstrate that activation of M₄ can be beneficial in numerous preclinical models predictive of anti-psychotic-like activity. These findings support the hypothesis that M₄ is partially responsible for the clinical efficacy of xanomeline and highlight the potential of M₄ PAMs as novel therapeutics for schizophrenia.

M₁ Muscarinic Receptor

The M₁ receptor is the predominant mAChR subtype in the CNS and is found in many brain regions including high levels of expression in the striatum, cortex, and hippocampus (Levey et al., 1995; 1991). This receptor canonically signals via activation of the Ga_q G-protein with concomitant activation of phospholipase C leading to phosphoinositide hydrolysis and calcium mobilization. M₁ knock-out mice are deficient in tasks requiring the medial prefrontal cortex (mPFC), and exhibit enhanced amphetamine-induced hyperactivity and hyperactive dopamine transmission (Anagnostaras et al., 2003; Gerber et al., 2001; Miyakawa et al., 2001). The high levels of M₁ receptor expression in brain regions that are dysregulated in schizophrenia, combined with the clinical efficacy of the M₁/M₄ preferring agonist xanomeline, suggests that M₁-selective compounds could be useful therapeutic agents for treating schizophrenia.

Over the past several years, numerous M₁ allosteric agonists and PAMs have been discovered and have greatly furthered our understanding of M₁ function in the brain. A major breakthrough in the discovery of M₁ PAMs came with the discovery of BQCA, which is a highly M₁-selective compound that interacts via an allosteric region of the receptor and has favorable pharmacokinetics and CNS exposure making it useful for in vivo studies (Ma et al., 2009). In wild-type, but not M₁ knock-out mice, BQCA potentiates mAChR agonist-induced increases in inward currents, and the frequency of spontaneous EPSCs in mPFC pyramidal cells. In addition, BQCA increases the firing rates of these cells in vivo and enhances prefrontal cortical-dependent forms of cognitive function, providing strong collective evidence that M₁ is an important regulator of prefrontal cortical function (Shirey et al., 2009). BQCA is also efficacious in reversing deficits in the acquisition of contextual fear induced by the administration of a non-selective mAChR antagonist, suggesting an important role for M₁ in hippocampal function. Systemic administration of BQCA was also found to reverse amphetamine-induced hyperlocomotion (Ma et al., 2009) in a manner similar to that observed with M₄ PAMs or the atypical antipsychotic clozapine. However, further studies with other M₁-selective agents have not demonstrated robust effects on amphetamine-induced hyperlocomotor activity, suggesting that allosteric activators of M₁ may not all have efficacy in this model that is often used to predict efficacy in treating the positive symptoms of schizophrenia. Together, these findings suggest that M₁ PAMs may have greater efficacy in treating cognitive disturbances in schizophrenia patients, whereas M₄ PAMs may be preferable for treating positive symptoms. However, future clinical studies will be required to establish the potential efficacy of M₁- and M₄-selective PAMs in reducing different symptom clusters in schizophrenia patients.

In addition to M₁ PAMs, a number of systemically-active M₁ agonists have been reported that appear to act as allosteric agonists and demonstrate high selectivity for activation of M₁ relative to other mAChR subtypes. The mechanism of action for these allosteric agonists including 77-LH-28-1, VU0364572, and others, is complex and current data suggests that these molecules may bind in a bitopic manner involving interactions with both allosteric and orthosteric sites on the receptor (Avlani et al., 2010; Digby et al., 2012b). Several lines of evidence suggest that M₁ activation via 77-LH-28-1, or VU0364752 and related compounds, can regulate hippocampal function and plasticity at the CA1-Schaffer collateral synapse via modulation of NMDA receptor currents in a context-dependent manner (Buchanan et al., 2010; Digby et al., 2012a; Jo et al., 2010; Jones et al., 2008; Langmead et al., 2008; Lebois et al., 2011). Unfortunately, while these M₁ allosteric agonists can selectively activate M₁ relative to other mAChRs, this selectivity is based on functional selectivity rather than selective high affinity binding to M₁ (Digby et al., 2012b). In addition these compounds can differentially induce M₁-mediated signaling to various pathways including Ca²⁺ mobilization and β-arrestin activation and can have differential effects that depend on varying levels of receptor reserve in different neuronal populations (Digby et al., 2012a). These issues, along with the bitopic nature of binding of most M₁ allosteric agonists, complicate the activity of these compounds relative to the highly selective PAMs, such as BQCA. Based on this, M₁ PAMs may represent a more viable approach to treatment of cognitive disturbances in schizophrenia than would allosteric agonists.

M₅ Muscarinic Receptor

The M₅ subtype of muscarinic receptor accounts for <2% of the total mAChRs expressed in the CNS (Yasuda et al., 1993). However, the expression pattern is highly-restricted to the cerebrovasculature as well as midbrain dopaminergic neurons, where it is the only detectable mAChR subtype (Weiner et al., 1990; Yamada et al., 2001). Similar to M₁, the M₅ receptor canonically signals via activation of the Ga_q G-protein. Consistent with its localization to dopamine neurons, M₅ knock-out mice display altered striatal dopamine release and blunted responses to drugs that modulate the dopaminergic system such as cocaine and opiates (Basile et al., 2002; Bendor et al., 2010; Thomsen et al., 2005). These mice also display deficits in prepulse inhibition (Thomsen et al., 2007) and exhibit reductions in amphetamine-induced hyperlocomotion (Wang et al., 2004). In addition to the knock-out studies in mice, genetic studies in humans have found that the M₅ mAChR gene (CHRM5) along with the α7 nAChR gene (CHRNA7) are linked to schizophrenia susceptibility (De Luca et al., 2004).

The restricted pattern of M₅ expression in midbrain dopamine neurons, along with the genetic evidence detailed above, suggest that M₅ represents another compelling target for the development of novel antipsychotics. However, the lack of subtype selective ligands has limited our understanding of the function of this mAChR in circuitry relevant to schizophrenia. Recently, great strides have been made in developing M₅-selective compounds, ultimately leading to the discovery of the first M₅-selective PAM VU0238429. This compound interacts with the receptor via an allosteric binding site and robustly potentiates responses at M₅ without activity at other mAChR subtypes (Bridges et al., 2009). Studies in brain slices have demonstrated that VU0238429 can potentiate both inward currents and Ca²⁺ mobilization induced by a non-selective mAChR agonist in dopamine neurons of the substantia nigra (Foster et al., 2012), demonstrating that activation of M₅ can robustly modulate the dopaminergic system. Unfortunately, VU0238429 does not possess physiochemical properties amenable to in vivo dosing. Further optimization of M₅-selective modulators is in progress (Bridges et al., 2010a; 2010b), and future in vivo studies will be necessary to fully elucidate the potential utility of M₅- modulators as antipsychotics.

Targeting Nicotinic Receptors for Treatment of Schizophrenia

Numerous clinical and preclinical findings suggest that modulating nicotinic receptors (nAChRs) could provide therapeutic efficacy in treating schizophrenia. In healthy humans, administration of nAChR antagonists can induce schizophrenic symptoms including psychosis and amnesia. Interestingly, schizophrenics have a drastically higher incidence of cigarette smoking (>80%) compared to both the general population and to individuals with other psychiatric disorders (Miwa et al., 2011). While the cause of this is not completely understood, it is clear that acute nicotine exposure in schizophrenic patients can improve cognitive processes, including sensory information processing. Unfortunately, the effects observed with nicotine disappear quickly after administration (Adler et al., 1993). The transient efficacy and the associated cardiovascular and abuse liability limit the utility of nicotine as a therapeutic for patients with schizophrenia. However, novel subtype-selective agonists and PAMs of nicotinic receptors may provide positive cognitive outcomes that are sustained over time with fewer adverse effects.

Nicotinic receptors are pentameric ligand-gated ion channels that are comprised of β and/or a subunits and can be either homomeric or heteromeric in nature. Differences in subunit composition can alter receptor function by changing the receptors ligand-sensitivity, Ca²⁺ permeability, and desensitization kinetics. The α₄β₂ and α₇ nAChR subtypes are the most prominently expressed subtypes in the brain and are located in several circuits shown to be imbalanced in schizophrenia (Fabian-Fine et al., 2001; Perry et al., 2002). α₄β₂ agonists have been found to be effective in modulating attention and memory in rodents and non-human primates (Buccafusco et al., 1995; Hahn et al., 2003). While there are currently no clinical reports regarding the efficacy of these molecules on schizophrenic patients, it has been found that the α₄β₂ agonist TC-1734 was well-tolerated and robustly improved age-associated memory impairments (Dunbar et al., 2011). Further studies are needed to determine the utility of such α₄β₂ agonists in treating schizophrenia.

The expression level of the α₇ nAChR receptor is decreased in schizophrenic patients (Court et al., 1999) and the gene for this receptor has been linked to schizophrenia vulnerability (De Luca et al., 2004; Leonard et al., 2002). Among the compounds targeting nAChRs for the treatment of schizophrenia, the α₇ agonist DXMB-A has made it the farthest in clinical development showing pro-cognitive efficacy in both a phase I study (Kitagawa et al., 2003), and a phase II study examining the effects of single-day administration of DMXB-A to schizophrenic patients (Olincy et al., 2006). Unfortunately, a follow-up phase II study examining the effects of DXMB-A after 4 weeks of daily dosing found no significant changes in cognitive improvements compared to placebo (Freedman et al., 2008). However, improvements were detected in the Scale for the Assessment of Negative Symptoms (SANS), particularly in the anhedonia and alogia sub-scales. It is difficult to definitively conclude if the absence of pro-cognitive efficacy after 4 weeks of DXMB-A administration is due to drug-induced desensitization, or whether the efficacy is masked by a significant practice-effect that was observed across all groups during test repetition. Recently, numerous α₇ PAMs have been discovered, which due to their mechanism of action may be more refractory to causing desensitization. These PAMs (including PNU-120596 and A-867744) and several novel α₇ agonists have been demonstrated to be efficacious in several preclinical models of schizophrenia (see Jones et al., 2012 for a comprehensive review), and represent exciting new avenues for treating the cognitive symptoms of schizophrenia.

Summary

Schizophrenia is a complex disease that is epitomized by symptoms spanning numerous clusters including cognitive, behavioral, and emotional disturbances. Current therapies are only partially effective at treating the behavioral symptoms and do not alleviate the cognitive or emotional deficits. Several preclinical and clinical studies suggest that cholinergic receptors represent novel targets with the potential to provide comprehensive and efficacious treatment of schizophrenia. While broad-spectrum modulators of muscarinic and nicotinic receptors have demonstrated clinical efficacy they have ultimately failed due to peripheral side-effects and a loss of efficacy over time. However, the discovery of receptor subtype-specific modulators has raised hopes that these molecules can provide long-lasting clinical efficacy with fewer adverse-effects. The coming years will be instrumental in elucidating the feasibility of modulating the M₁, M₄, M₅, α₄β₂, and α₇ cholinergic receptors as therapeutic treatments for schizophrenia.