Neuroscience 2008—Cholinergic Neurons in Schizophrenia: Nicotinic and Muscarinic Approaches
Editor’s Note: Scarcity of resources—money and time—keeps many researchers at home during scientific conferences. You can help them by sharing another critical resource—information. Meeting reports are an important way that you can provide a critical community service to many researchers toiling away in far-flung laboratories and offices, trying to piece together small patches of the schizophrenia puzzle. We encourage you to contact us if you would like to help spread the wealth that is information.
We are very fortunate to have C. Anthony (Tony) Altar as our guest correspondent for the Neuroscience 2008 meeting in Washington, DC. Tony is a unique researcher who keeps active collaborations going across academia, industry, and government.
3 December 2008. Presentations at the 2008 meeting of the Society for Neuroscience (SfN) provided support for the benefit of enhancing nicotinic and muscarinic cholinergic receptor signaling to improve schizophrenia, particularly the cognitive deficit and negative symptoms.
Nicotinic receptors and schizophrenia
We all know that nicotine elevates attention, memory, and cognitive performance, and that selective agonists for the α7 nicotinic receptor subunit are being tested for their ability to do this and more in schizophrenia and Alzheimer's disease. Unfortunately, the α7 nicotinic receptor partial agonist anabaseine (DMXB-A) failed to alter MATRICS measures or BPRS scores (Freedman et al., 2008; see SRF related news story). Before closing the doors on this approach, researchers continue working on several development programs in the nicotinic area. MEM 3454 is an α7 partial nicotinic receptor agonist and 5-HT3 antagonist. It improves cognitive functions in young adult and aged rats or mice. Mei Huang, Herb Meltzer, and colleagues at Vanderbilt University (abstract 874.23) worked with Memory Pharmaceuticals to evaluate MEM 3454 on cortical and hippocampal neurotransmitter release using microdialysis. Over a 0.1-10 mg/kg range of subcutaneous doses, MEM 3454 elevated dopamine (DA) and acetylcholine (ACh) release in the cortex and hippocampus, but not in the nucleus accumbens. MEM 3454 produced relatively small, ~40-60 percent overall increases in dopamine, and ~20-40 percent increases in acetylcholine, release, based on integrated area-under-the curve analysis, and an abrupt inverted U-shaped dose-response curve that peaked at the 0.5 mg/kg dose. Comparable increases in DA and ACh release produced by aripiprazole, risperidone, olanzapine, or quetiapine were augmented by a 0.45 mg/kg dose of MEM, but the effects of haloperidol were not. The MEM 3454 elevations of DA but not ACh release were blocked by the selective α7 receptor antagonist, MLA, whereas increases in ACh release but not DA release were blocked by the 5-HT3 antagonist, CPBG. Those straightforward studies show that MEM 3454 increases in DA and ACh release are mediated by α7 agonism and 5-HT3 antagonism, respectively, and are consistent with the improved cognitive function expected with this class of drug—potentially by itself, but as shown here, in combination with atypical antipsychotics. Drugs that possess nicotinic α7 agonism/5-HT3 antagonism may be useful adjuncts to existing antipsychotics, and represent receptor mechanisms that can be included in the design of drugs that block D2/5HT2A receptors. Anything like that out there?
It was also interesting to learn from Herb that his group has found comparable effects across many antipsychotic drugs on ACh and DA release in the hippocampus and frontal neocortex. Thus, you dialysists out there can study one or the other region, not both. They also mentioned that 0.3 to 0.5 mg/kg doses of Abilify produce the greatest increases in dopamine release in rat hippocampus or neocortex, with less effect at higher or lower doses. This translates well within the human dose of 2-20 mg per patient, which is pretty much the clinical dose range. Their findings support the use of hippocampal or cortical dopamine release, and not necessarily on both, for evaluating antipsychotic candidates and as one basis for selecting human doses.
Another α7 nicotinic agonist, WAY 317538, is a full agonist. As conducted in collaboration with researchers lucky enough to live in Siena, Italy (Siena Biotech), Wyeth researchers including Chad Beyer (abstract 657.16) showed that WAY 317538 more than doubled glutamate release in the rat medial frontal cortex and for at least three hours, without changing DA release. The lack of effect on DA may be due to its lack of H3 antagonism, which is a property of MEM 3454 (reviewed above). This glutamate-but-not-dopamine-release profile characterizes all α7 nicotinic agonists the Wyeth scientists discussed. Maybe it is for that reason that WAY 317538 was described as not going forward in development.
Amanda Williams and her colleagues at AstraZeneca showed a remarkable potency for the compound AZD0328. At 100 ng/kg, or only 20 ng per rat, AZD0328 improved performance in the cognitive memory task and novel object recognition test (abstracts 292.6, 906.23). Efficacy was lost at higher doses. When exposed to a long delay between the first and second test, most rats respond to a novel object with the same zeal and duration they would if they had never seen it. Even with the delay, AZD0328-treated rats spent only one-third of their time with the familiar object, and the rest of their time with the novel object. The report of a 67 percent decrease in α7 nicotinic receptors in frontal cortex and hippocampus after administration of AZD0328 is not surprising for nicotinic agonism, but warrants evaluations of its efficacy with chronic administration. This agonist is in clinical development at AstraZeneca in the areas of schizophrenia and Alzheimer's disease.
Another promising approach from AstraZeneca, in collaboration with Targacept, was reported by W.C. Moore and colleagues with AZD3480, an α2/β4/β2 nicotinic agonist (abstracts 329.4, 329.9). Moore described the use of this compound (formerly known as TC-1734) in Phase 2 trials for otherwise-unmedicated patients with schizophrenia. AZD3480 was studied 12 weeks after cholinergic deafferentation of the hippocampus by fimbria-fornix knife cuts. The degree of LTP, induced by stimulation of the entorhinal cortex in their in vitro hippocampal slice preparation, was measured by an enhanced EPSP response to subsequent stimulation. LTP was increased in tissues from intact rats by 60 nM AZD3480. This is very impressive, and the decrease in LTP produced by cholinergic lesion was restored to near-normal levels by AZD3480, and not 150 nM donepezil, which is not surprising since fornix lesion should leave little ACh for the donepezil to augment. These results may bode well for the use of the compound in illnesses like Alzheimer's disease, where cholinergic losses are well known, or for schizophrenia, where cholinergic inputs and signaling via muscarinic receptors may be diminished (as discussed in the even more exciting "Muscarinic" section of this report). Also promising was the efficacy of AZD3480 in the radial 8 arm maze and novel object recognition tests in intact animals. The behavioral efficacy of the compound in fimbria-fornix lesioned rats, or in those challenged with a low dose (like 0.3 mg/kg) of scopolamine, has not been evaluated, but it may be expected to work based on the LTP study.
Should investigators wish to confirm that their α7 nicotinic acid receptor-binding drugs actually get in the brain and bind to this target, [11C]CHIBA-1001 was reported at this meeting by Kenji Hashimoto of Chiba University (abstract 328.1) as the first and only PET ligand available for this purpose (Hashimoto et al., 2008). Some of the details of this method seemed less rosy. [11C]CHIBA-1001 shows only a 46 nM affinity for the α7 nicotinic acid receptor, defined by [125I]α-bungarotoxin, and was only tested at 28 other receptors for inactivity. Seemingly too much of the label (75 percent) persisted in rhesus macaque brain even when co-treated with 5 mg/kg of unlabeled SSR 180711, an α7 nicotinic receptor agonist. No [11C]CHIBA-1001 label was displaced with an α4β2 nicotinic agonist. The PET signal persisted for two hours post-intravenous infusion, but the high degree of non-displaceable labeling is a concern. Fortunately, in human brain, Masatomo Ishikawa, also at Chiba University, showed binding to be more (~50 percent) specific in healthy subjects, and visibly displaceable occupancy in the neocortex and cerebellum was obtained with a 10 mg dose of a "proprietary" α7 nicotinic agonist (abstract 328.2). Hopefully that agonist was something different from CHIBA-1001 itself.
Muscarinic receptors and schizophrenia
While clinical proof of concept for muscarinic agonism has been provided for treating cognitive deficits of schizophrenia (Shekhar et al., 2008) and the delusions and hallucinations of Alzheimer's disease patients (Bodick et al., 1997), a deficiency in muscarinic transmission that may subserve this approach has been more clearly provided at this year’s meeting. Studies by Holt and colleagues showed decreases in cholinergic innervation of the nucleus accumbens in schizophrenia (Holt et al., 2005; Holt et al., 1999). Modeling this loss in rats with bilateral accumbens infusions of the selective cholinergic toxin, Francois Laplante and colleagues from the University of Montreal used saporin-anti-ChAT IgG immunoglobulin to deplete 50 percent of ACh neurons in this region (abstract 761.18). Though selective, this procedure did not change basal startle but decreased prepulse inhibition of startle by up to 40 percent, and slightly increased basal locomotion, grooming, and "other" movements. Some of these changes were increased by the D2 agonist quinpirole. These findings support the old-as-haldol DA/ACh balance hypothesis by showing that increases in this ratio predispose to psychosis (or psychosis-related behaviors in animals), and, equally important, that decreases in the ratio—by blocking D2 receptors or increasing cholinergic transmission—can treat the disease.
More support for the benefit of augmenting muscarinic cholinergic tone in schizophrenia was obtained by Hasib Salah-Uddin, Jeannette Watson, Brian Dean, and colleagues at the University of Leicester, GlaxoSmithKline, and the Mental Health Research Institute, Australia (abstract 54.19). They provided compelling new support for altered muscarinic receptor signaling in schizophrenia. Prior reports of decreased muscarinic receptors in the cortex, caudate putamen, and entorhinal cortex in schizophrenia by Dean and colleagues (Dean et al., 2000; Crook et al., 2001) are impressive enough, but do those decreases translate to a decrease in muscarinic signaling? Using postmortem dorsolateral prefrontal cortex (DLPFC), from schizophrenia and control subjects, they have shown that about 25 percent of people with schizophrenia have a ~70 percent reduction in [3H]-pirenzepine binding. This group has been termed "muscarinic receptor-deficit schizophrenia" (MRDS) (Scarr et al., 2008). The MRDS subgroup in the new study showed not only the ~70 percent decrease in [3H]-pirenzepine binding, but a 0.5 log unit right-shift in the potency of the orthosteric muscarinic receptor agonist oxotremorine-M in stimulating [35S]GTPγS binding in membranes prepared from these brains. About a 20 percent greater degree of [35S]GTPγS binding, stimulated by the orthosteric agonist, also characterized the MRDS patients. The functional activity of the allosteric site agonist, AC-42, was not affected in the MRDS population. The potency and relative efficacy of M1 receptor-agonist coupling for oxotremorine-M and AC-42 was similar for controls and non-MRDS subgroups. These findings combine nicely with deficient expression of genes in hippocampal neurons of schizophrenia patients (Altar et al., 2005), which could be reversed by oxotremorine-M in human neuronal cell line (Altar et al., 2008). Many of those gene decreases in schizophrenia were replicated in a second cohort of patients, and were consistent with a metabolic deficit hypothesis of schizophrenia.
As long as we are on the topic of a metabolic deficit in schizophrenia, Rosalinda Roberts, newly ensconced at the University of Alabama, and her colleagues (abstract 656.6) found 30 percent less cytochrome oxidase (COX; p <0.001) in the putamen of schizophrenia cases versus normal controls, and these were lower regardless of whether the schizophrenia patients were on or off medications at the time of death. Also arguing against medication status in these effects was their finding in rats that COX activity was not decreased by haloperidol or clozapine compared to vehicle-treated controls. The decrease in COX activity in the putamen of schizophrenia cases agrees with prior reports by others including those of Prince and colleagues (see, e.g., Prince and Oreland, 1998), and may be due to mitochondrial abnormalities in the putamen, a possibility that is supported by the diminished number or size of mitochondria in schizophrenia cases also reported by this group. It might be of interest for Rosie to see if there is an MRDS-like segregation in these cases and if so, whether they co-segregate with deficient muscarinic receptor coupling to [35S]GTPγS binding.
While muscarinic receptors have been known to facilitate long-term potentiation (LTP) in the hippocampus, muscarinic receptor subtypes responsible for this were unknown. Katherine Buchanan of University College, London, used the drug 77-LH-28-1, a selective allosteric M1 muscarinic acetylcholine receptor agonist, to show that M1 activation facilitates LTP in the young rat hippocampus (abstract 335.11). When applied to hippocampal slice preparations maintained in vitro, 77-LH-28-1 depolarized and increased membrane resistance in CA1 neurons, but did so without changing glutamate release. In response to "theta burst stimulation,” 77-LH-28-1 increased LTP due to glutamate release, and a similar LTP increase by 77-LH-28-1 was produced in response to increased endogenous ACh release from stimulated cholinergic afferents, or by the non-selective, non-hydrolyzable cholinergic agonist, carbachol. Their work demonstrates a mechanism by which M1 acetylcholine receptors can amplify LTP induction in the hippocampus and thereby enhance cognitive function. This would represent another example of tonic amplification of cholinergic stimulation, whose attractiveness lies in the selectivity it imparts during heightened cholinergic transmission.
That the entorhinal cortex itself may contribute to a habituation impairment in schizophrenia was suggested by the findings of Segev Barak and Ina Weiner of Tel Aviv University (abstract 418.3). They noted that subjects with schizophrenia have a deficiency in latent inhibition (LI), as they are more likely to attend to irrelevant stimuli presented during preconditioning trials. As shown previously with systemic injection of the muscarinic antagonist scopolamine, bilateral infusion of 1-10 μg scopolamine into the rat entorhinal cortex disrupted LI. This showed that at least in this region, muscarinic receptor agonism supports the ability to inattend to irrelevant stimuli. That feature of inattention is one in which subjects with schizophrenia do poorly on. Barak and Weiner also showed that muscarinic receptor agonism in the amygdala promoted a re-attending to previously irrelevant stimuli that are now relevant. Thus, like the Buchanan study with LTP described in the previous paragraph, known roles for cholinergic function in LTP or attention mechanisms have now been shown to be mediated by muscarinic receptors. Both the LTP and LI models can be considered as tests for drugs designed to work through muscarinic receptor agonism, although the potentially counterproductive role of amygdala muscarinic agonism may be problematic for muscarinic agonist therapeutics.
While the desired preference of muscarinic agonists for M1/3/5 over M2/4, and for M1 over other receptors, is likely for CNS targeting and for avoiding peripheral cholinergic side effects, the design of selective compounds for these targets is challenging. Muscarinic receptors are highly conserved in the vicinity of the acetylcholine orthosteric binding site. John Ellis and Edward Stahl of Penn State University showed that muscarinic receptors possess allosteric sites, at which receptor activity may be positively or negatively modulated (abstract 532.5). They reported that amiodarone enhances the activity of ACh at the M5 receptor relative to M1, probably by binding within extracellular loop(s) of the receptor.
Ditte Dencker and Anders Fink-Jensen of the University of Copenhagen used mice produced by Jongrey Jeon (NIH), in which an elegant conditional inactivation of mouse M4 muscarinic cholinergic receptors is produced when those receptors reside on DAD1-receptor-containing striatonigral GABA neurons, and medium spiny striatal cholinergic interneurons (abstract 743.2). These mice showed a 75 percent attenuation of 0.3 mg/kg haloperidol catalepsy, and nearly as much loss at a whopping 1 mg/kg haloperidol dose. Even a 5 mg/kg dose of the non-selective muscarinic receptor blocker, scopolamine, was unable to block what little haldol catalepsy remained in the M4-deletion mice. Thus, the EPS side effects of typical antipsychotics like haloperidol appear to be promoted by endogenous cholinergic activation of M4 receptors on D1 receptor-containing neurons. Their findings are consistent with the use of anticholinergic drugs to block EPS, and refine this to suggest that a centrally selective M4 antagonist/M1 agonist might be a useful adjunct for treating Parkinson's disease or the EPS side effects associated with typical antipsychotics, while simultaneously improving cognition. Now that's a drug discovery challenge!—C. Anthony Altar.