Schizophrenia Research Forum - A Catalyst for Creative Thinking

SOBP 2009—Basic Science and Drug Discovery Converge on Muscarinic Agonism for Schizophrenia

We are pleased to present several meeting reports from the recent annual meeting of the Society for Biological Psychiatry. This is the first summary from researcher C. Anthony Altar, NeuroDrug Consulting.

9 June 2009. New targets for schizophrenia and its cognitive deficits remain more in demand than ever, it seems. An outstanding symposium at the 2009 Society for Biological Psychiatry in Vancouver, British Columbia, Canada, showed with clinical, preclinical, and pharmacological data why muscarinic receptor agonism is a promising approach.

The first speaker, Jurgen Wess, (NIMH, Bethesda, Maryland), reviewed muscarinic receptor subtypes; M1, 3, and 5 are Gq11-linked to turn on PLC-beta, IP3, and DAG, while M2 and 4 are Gi/o-coupled and linked to inhibition of adenylyl cyclase. (See Wess et al., 2007 for an excellent review of therapeutic uses of muscarinic agonists and antagonists and other aspects of muscarinic function.)

M1 is found in most brain areas relevant to schizophrenia. However, other muscarinic receptors are abundant in the periphery where they slow the heart, regulate the force of smooth muscle contraction, and affect the digestive system, glands, and arteries. Efficacy-limiting side effects via these interactions remain problematic for muscarinic drug development.

Complicating matters further, most regions in brain or peripheral organs express multiple muscarinic subtypes. This diversity and the paucity of selective muscarinic ligands cloud the specific roles for each subtype. However, M1 through -5 knockout mice, and M1/3, M2/4, M1/4, and M2/3 double knockouts, have helped elucidate the roles of each subtype, said Wess.

Mice with deletions of M1 and, less convincingly, M4 receptors are promising schizophrenia-like models, suggesting the potential anti-schizophrenia role for M1 agonism. M1-/- mice show increased spontaneous locomotion and increased extracellular dopamine in the striatum (Miyakawa, 2007; Gerber et al., 2001). M4-/- mice show increased locomotion in response to D1 agonists, decreases in sensorimotor gating, and increased dopamine efflux in nucleus accumbens. M4 receptors are co-localized with D1 striatal medium spiny neurons of the direct striatal output pathway.

The Wess team made conditional deletions of the M4 coding sequence by first surrounding it with two LOX-CRE sites in one transgenic mouse. Offspring of these animals make functional M4 receptors, and when crossed with D1-CRE transgenic mice. their offspring, in turn, showed a CRE-mediated deletion of M4 receptors and only in D1 receptor expressing striatal neurons.

Amphetamine (2 mg/kg) hyperactivity was increased ~50 percent in the D1-M4-KO mice, as was hyperactivity following D1 agonist (SKF82958) or cocaine treatment. Dopamine release in nucleus accumbens was also increased in these mice following amphetamine, providing a pro-dopaminergic role for M4 antagonism.

Catalepsy induced by haloperidol (0.3 to 1 mg/kg) or risperidone (1 mg/kg) was 90 percent blocked at 30 minutes, and 50 percent at 90 minutes, in the D1-M4-KO mice, showing an anti-cataleptic effect of M4 antagonism, at least during D2 blockade. This suggests that selective M4 agonism may induce EPS in patients, particularly if they are co-treated with D2 blockers. The possible EPS liability of M4 agonism, and possible anti-dopaminergic role for M4 agonism, leave unresolved whether full or partial M4 agonists might be better indicated for treating schizophrenia. Thus, the M1 agonist approach was highlighted by this and the remaining presentations.

Symposium organizer Elizabeth Scarr (U. Melbourne, Victoria, Australia) showed that decreases in muscarinic receptors in schizophrenia have been replicated in postmortem studies conducted in at least four laboratories. An early study by Scarr and colleagues showed 25 percent decreased [3H]pirenzepine binding to muscarinic receptors in schizophrenia postmortem cortex, ~30 percent decreases in M1 protein, and ~60 percent decreases in M1 mRNA, whereas neither M4 protein nor mRNA changed (Dean et al., 2002). In a greatly expanded study, these researchers show two distinct populations of [3H]-pirenzepine binding in schizophrenia (Scarr et al., 2008). For about three-fourths of the patients, it is about equal to or below those of controls, but the remaining one-fourth of patients reveal a distinct 75 percent binding decrease from controls or the other patients. Termed "muscarinic receptor deficient schizophrenia" (MRDS), these patients show decreased M1 receptors labeled by [3H]pirenzepine in brain sections, and a large (up to 50-fold) right-shift of M1 activation by muscarinic agonists. In control postmortem Brodmann area 44, M1 receptors are dense in layers I-V, lower in VI, and much lower elsewhere in neocortex. Schizophrenia patients in the non-MRDS showed ~15 percent decreases in both sets of layers and ~75 percent decreases were found in both layers for MRDS patients. The team also saw about half these losses for both groups in [3H]AF-DX 384 ligand binding to M2/4, and[3H]4-DAMP binding to M3 receptors.

That raised questions from the audience: could these broad muscarinic decreases be due to anti-cholinergic drugs often used to quell motor side effects of antipsychotic drugs? The MRDS and non-MRDS schizophrenia groups did not differ in content of the Gq guanine nucleotide binding protein, suggesting that the deficit is proximal to the M1 receptor and not associated with anticholinergic drugs, which can alter this G protein. In addition, animal studies have shown that one month’s treatment with the anti-cholinergic benztropine does not affect levels of [3H]pirenzepine binding (Crook et al., 2001). Also, a study found that hippocampus from subjects with schizophrenia that had been prescribed anti-cholinergics had higher levels of [3H]pirenzepine binding than schizophrenia subjects not prescribed these drugs (Scarr et al., 2007).

Scarr and her colleagues will use heterozygous M1-/- mice to identify other factors that may create a deficit of M1 signaling in the brains of MRDS patients. The group will initiate a study to identify MRDS patients postmortem using [3H]pirenzepine binding, and to determine if the decreases in metabolic gene expression that characterize schizophrenia—and linked to deficiencies in muscarinic M1 receptor stimulation (Altar et al., 2008)—will be even more robust in the MRDS cases and less so in the remaining patients.

The next speaker, Thomas Raedler (University of Calgary) used [123I]QNB binding by in vivo PET analysis in controls and seriously ill schizophrenia inpatients to identify drugs that occupy M1 receptors. Cerebellum, devoid of specific binding, was a negative control. As expected, PET images showed extensive cortical binding, but in unmedicated patients, binding was ~25 percent lower in the caudate, putamen, thalamus, medial and lateral frontal cortex, and temporal and occipital cortex, with p values mostly <0.01. There was a strong, negative correlation between striatum or frontal cortex [123I]QNB binding and schizophrenia symptom score, again consistent with a muscarinic agonist approach for this disease. Raedler reviewed the findings by Shekhar and colleagues, in which the partial muscarinic agonist xanomeline was found to improve several domains of cognitive function and positive symptoms of schizophrenia (Shekhar et al., 2008).

Olanzapine was found, at 5 or 20 mg doses, to occupy ~30 percent to ~60 percent of [123I]QNB binding, respectively, in cortex, thalamus, pons, putamen, and the caudate, among other areas. Clozapine at 150-450 mg doses blocked 40-80 percent of [123I]QNB binding. In a [123I]IDEX SPECT binding study, Lavalaye and colleagues generated similar data but overall less occupancy by olanzapine than in the [123I]QNB binding studies (Lavalaye et al, 2001). Clozapine and olanzapine are antagonists at the M1 receptor, although the metabolite N-desmethylclozapine (NDMC) is a potent, partial M1 agonist. NDMC is also a partial agonist at D2 and D3 receptors, a 5HT2A binder, and increases ACh and DA release in FCx and HC. Phase 3 studies with NDMC may have failed due to a chance exclusion of MRDS patients. A biomarker is needed to identify MRDS patients if this subgroup is particularly responsive to muscarinic agonists. A selective M1 PET ligand would be a big help for detecting MRDS patients and developing compounds with M1 agonism.

In the final talk, Jeannette Watson, (GSK, Chelmsford, UK) began by describing a recent study by Chouinard and colleagues. They found clinical efficacy for some cognitive domains in patients treated with AChE inhibitors and antipsychotics (Chouinard et al., 2007). This is impressive, as some of the APDs were clozapine and olanzapine, which themselves posses anticholinergic properties. AChE inhibitors may overcome this by non-specifically activating muscarinic and nicotinic receptors in brain.

GSK is developing a CNS penetrant, selective M1 allosteric agonist, 77-LH-28-1. First presented at CINP last year, 77-LH-28-1 is about 1,000-fold less potent at binding to M2-5. It has about 85 percent intrinsic efficacy at M1 receptors, like oxotremorine-M, exceeding that of xanomeline or AC42 (~50 percent), determined by increases in GTP γ-S binding to human postmortem cortex. Surprisingly, n-desmethyl clozapine was inactive in increasing GTP binding in this human neocortex assay, yet could potently block all of the oxotremorine agonism! Could this be a second explanation for the lack of clinical efficacy of NDMC (ACP-104) in the Acadia clinical trials?

In partial support of the MRDS hypothesis Scarr presented earlier, Watson's team saw a threefold right shift of potency of oxotremorine in the cohort of schizophrenia patients with low pirenzepine binding.

In rats, 77-LH-28-1 increased electrical activity of cornu ammonis cell firing in rats and scopolamine blocked this effect. 77-LH-28-1 also doubled the increase in hippocampal firing produced by NMDA, reduced the threshold for LTP induction in hippocampal slices in vitro, and elevated the fEPSP slope by 70 percent. 77-LH-28-1 at 1 and 3 mg/kg was active in the rat novel object recognition model with 24-hour natural forgetting, whereas vehicle-treated rats forgot the old object, as evidenced by their 50 percent time divided equally between the novel and familiar objects. 77-LH-28-1 showed fewer muscarinic-related adverse events (altered body temperature, salivation, defecation) compared to the non-selective M agonist, milameline (3 mg/kg). With this impressive effort and basic science, could other selective M1 allosteric activators, or partial agonists, be far behind? We hope not, as questions about the mGluR2/3 approach and the cognition liability of D2 antagonists—both discussed at this meeting—increase demand for new antipsychotic drug targets.—C. Anthony Altar.

Comments on Related News

Related News: Muscarinic Receptor Agonists: Not Created Equal

Comment by:  Bryan Roth, SRF Advisor
Submitted 15 July 2012
Posted 16 July 2012
  I recommend the Primary Papers

This is an interesting and important paper which serves as a cautionary reminder regarding the potential activity of allosteric modulators to have "functionally selective" effects on GPCR signaling (Urban et al., 2007; Allen and Roth, 2011). Thus, it has been known for decades (see Urban et al., 2007, for review) that GPCR "agonists" and "antagonists" can activate distinct signaling pathways and that a drug can appear to be an "agonist" for one pathway and have "antagonist" activity at another. As we have recently suggested, such activity could have potentially therapeutic implications for schizophrenia and related disorders (Allen et al., 2011).

Although it has been recognized for some time that allosteric modulators of GPCR activity may also have functionally selective actions (Sheffler and Conn, 2008) in vitro, it was unknown if these effects of signaling bias were therapeutically relevant.

What is important here is that Digby et al. rigorously demonstrate signaling bias in vitro and in situ, and then follow these findings up by showing that the signaling "fingerprint" is relevant to the apparent electrophysiological effects of the drugs in vivo.

Although the drugs lacked apparent antipsychotic drug-like actions, they clearly had different effects on synaptic physiology.

Taken together with many other papers of this sort which have appeared over the past several years, the findings reinforce the principle that relying on a single cellular readout for demonstrating the "agonist," "antagonist," or "allosteric" actions of a small molecule could ultimately be problematic for therapeutics.


Allen JA, Roth BL. Strategies to discover unexpected targets for drugs active at G protein-coupled receptors. Annu Rev Pharmacol Toxicol. 2011 Feb 10;51:117-44. Abstract

Allen JA, Yost JM, Setola V, Chen X, Sassano MF, Chen M, Peterson S, Yadav PN, Huang XP, Feng B, Jensen NH, Che X, Bai X, Frye SV, Wetsel WC, Caron MG, Javitch JA, Roth BL, Jin J. Discovery of β-arrestin-biased dopamine D2 ligands for probing signal transduction pathways essential for antipsychotic efficacy. Proc Natl Acad Sci U S A. 2011 Nov 8;108(45):18488-93. Abstract

Sheffler DJ, Conn PJ. Allosteric potentiators of metabotropic glutamate receptor subtype 1a differentially modulate independent signaling pathways in baby hamster kidney cells. Neuropharmacology. 2008 Sep;55(4):419-27. Abstract

Urban JD, Clarke WP, von Zastrow M, Nichols DE, Kobilka B, Weinstein H, Javitch JA, Roth BL, Christopoulos A, Sexton PM, Miller KJ, Spedding M, Mailman RB. Functional selectivity and classical concepts of quantitative pharmacology. J Pharmacol Exp Ther. 2007 Jan;320(1):1-13. Abstract

View all comments by Bryan Roth

Related News: Muscarinic Receptor Agonists: Not Created Equal

Comment by:  Rick Neubig
Submitted 19 July 2012
Posted 20 July 2012

This is a very useful paper, but it is worth pointing out that there is really no evidence for "ligand bias" or "functional selectivity" of these compounds. The VU analogs are partial agonists but show no bias toward one signal pathway or another. The low ERK and α-arrestin signaling are due to the lower degree of receptor reserve for those two pathways and the requirement for more efficacious agonists to signal. The correlation of the in-vitro studies with the different in-vivo readouts does provide very important guidance for pharmacologists about what properties are needed for M1 agonists to produce different physiological effects. It doesn't, however, give any real information about "biased agonists."

View all comments by Rick Neubig

Related News: Muscarinic Receptor Agonists: Not Created Equal

Comment by:  Anantha ShekharAmanda Bolbecker
Submitted 27 July 2012
Posted 27 July 2012

The discovery of M1 allosteric agonists caused a great deal of excitement because they act on the M1 receptor fairly exclusively. The new research by Digby et al. suggests that this newer class of compounds can influence very specific molecular signaling pathways initiated by M1 activation, and that these pathways mediate specific aspects of cognition. These findings could have profound effects on the development of novel therapeutic agents for cognitive impairments seen in schizophrenia and other severe neuropsychiatric disorders.

View all comments by Anantha Shekhar
View all comments by Amanda Bolbecker

Related News: Muscarinic Receptor Agonists: Not Created Equal

Comment by:  Bryan Roth, SRF Advisor
Submitted 1 August 2012
Posted 1 August 2012

Regarding the comment by Rick Neubig: while it could be argued that the paper does not provide "gold standard" pharmacological data for bona fide functional selectivity/stimulus bias, in fairness to the authors I provide their conclusions directly from the paper:

"Thus, the differential effects of these M1 agonists on CNS responses may reflect a combination of partial agonist activity that is impacted by differences in receptor reserve and by an inherent stimulus bias at M1 so that these compounds are not capable of fully activating some responses, even in systems in which the receptor is highly expressed…. Thus, different levels of M1 expression are likely to contribute to the differential responses to VU0357017 and VU0364572 observed in these studies. However, it was interesting to find that VU0357017 never achieved full efficacy in activation of ERK1/2 phosphorylation, even in cell lines with strong induction of M1 expression to levels that induced high receptor reserve in the calcium mobilization assay. Also, VU0357017 did not induce robust β-arrestin responses in the original cell line or in the TREx hM1 cells. Thus, the differential effects of these M1 agonists on CNS responses may reflect a combination of partial agonist activity that is impacted by differences in receptor reserve and by an inherent stimulus bias at M1 so that these compounds are not capable of fully activating some responses, even in systems in which the receptor is highly expressed."

View all comments by Bryan Roth