INTRODUCTION

Schizophrenia is a complex syndrome defined by the presence of positive and negative symptoms, as well as cognitive dysfunctions (Pantelis et al, 1999; Raedler et al, 2007). Post mortem studies (Dean et al, 1996; Crook et al, 2000; Zavitsanou et al, 2004; Deng and Huang, 2005; Scarr et al, 2007), and a recent neuroimaging study (Raedler, 2007), have consistently shown that widespread decreases in the levels of muscarinic acetylcholine receptors (CHRM) occur in the central nervous systems (CNS) of patients with schizophrenia. Significantly, protein and mRNA levels for the M1 (CHRM1; Dean et al, 2002; Mancama et al, 2003), but not the M4 (CHRM4; Dean et al, 2002) or M2/M3 (CHRM2/CHRM3; Scarr et al, 2006) subtypes, have been shown to be decreased in the frontal cortex of patients with schizophrenia. These data support the hypothesis that decreases in the binding of the CHRM subtype-selective antagonist, [3H]-pirenzepine (PZP), reflect decreases in the CHRM1 in the frontal cortex of patients with the disorder.

In other syndromes, such as diabetes (Gale, 2001), the ability to sub-divide patients into biologically more homogenous groups using biological markers has often underpinned the beginnings of defining the pathologies of different diseases within the syndrome. Therefore, it is highly significant that it has recently been reported that a distinct sub-population of patients with schizophrenia can be defined that have a 60–80% reduction in cortical [3H]-PZP binding (Scarr et al 2008). This low [3H]-PZP binding subgroup has been defined as having ‘muscarinic receptor-deficit schizophrenia’ (MRDS) (Scarr et al, 2008). To date, it has not been possible to determine whether changes in CHRM1 density in the CNS of patients with schizophrenia is associated with a change in receptor function; something that might be predicted to occur in MRDS patients, given the marked downregulation of CHRM1 protein within this subgroup.

CHRM1 couples preferentially to heterotrimeric G proteins of the Gq/11 sub-family to exert the majority of its cellular actions (Caulfield and Birdsall, 1998). This involves the activated, ligand-bound receptor interacting with a Gq/11 protein to facilitate GTP-for-GDP exchange on the Gαq/11 subunit. In the presence of a radiolabeled, non-hydrolysable GTP analog [35S]guanosine-5′-O-(3-thio)triphosphate ([35S]-GTPγS), the receptor will facilitate the binding of [35S]-GTPγS to Gq/11 proteins and by immunoprecipitating Gαq/11 subunits using specific antibodies, it is possible to quantify the transduction of receptor activation to a proximal downstream step in the signal transduction pathway (DeLapp et al, 1999). By using a novel adaptation of this method, [35S]-GTPγS-Gαq/11 immunocapture (Salah-Uddin et al, 2008), it is now possible to measure G protein-coupled receptor-Gαq/11 coupling in membranes prepared from human post mortem tissue to assess agonist potency and efficacy. Here, this technique has been used to determine whether agonist-dependent CHRM1 signalling in human cortex (Brodmann area 9) is altered in the tissue from patients with MRDS compared with that in other forms of schizophrenia (non-MRDS), and patients with no known history of psychiatric or neurological disorders (controls). Our data indicate that although CHRM agonist potency is decreased in MRDS, there is an increase in the efficacy of CHRM1-Gq/11 coupling. These new data indicate that despite marked declines in CHRM1 expression in some patients with schizophrenia, the ability of this receptor to initiate signal transduction through Gq/11 proteins is undiminished.

MATERIALS AND METHODS

Materials

[35S]-GTPγS (1000–1200 Ci/mmol) and anti-rabbit-IgG-coated SPA beads (RPNQ0016) were obtained from GE Healthcare. Complete protease inhibitor cocktail was purchased from Roche Applied Science. All other chemicals and reagents were obtained from Sigma-Aldrich. The Gq/11α antiserum was generated (against the C-terminal sequence (C)LQLNLKEYNLV) as described earlier (Akam et al, 2001). AC-42 (4-n-butyl-1-[4-(2-methylphenyl)-4-oxo-1-butyl]-piperidine hydrogen chloride) was synthesized by GlaxoSmithKline, Harlow, UK.

Tissue Collection

Approval was granted for tissue acquisition and use in this study by the Ethics Committee of the Victorian Institute of Forensic Medicine and the North Western Mental Health Programme Behavioural and Psychiatric Research and Ethics Committee, and consent for the use of the samples for research purposes was approved by the donor's relatives. Blocks of tissue containing Brodmann area 9 (BA9) were excised from the left hemisphere of patients, who were retrospectively defined as having MRDS or non-MRDS, as well as age/sex-matched controls (see Table 1). BA9 was taken as the region of the CNS on the lateral surface of the frontal lobe and includes the middle frontal gyrus superior to the inferior frontal sulcus.

Table 1 Demographic, Post Mortem and Pharmacological Data for Subgroups of Patients with Schizophrenia and Controls

As part of the clinical assessment using the Diagnostic Instrument for Brain Studies (DIBS) (Hill et al, 1999), a number of parameters were calculated. Post mortem interval (PMI) was calculated as the time from death to autopsy when deaths were witnessed. In cases in which death was witnessed, the time between death and autopsy was taken as the PMI. In the latter case, the tissue was only taken from individuals who had been seen alive up to 5 h before being found dead, and PMI was taken as the interval halfway between the donor being found dead and last being seen alive. In all cases, the cadavers were refrigerated within 5 h of being found and the tissue was rapidly frozen to −70°C within 30 min of autopsy. The pH of the CNS tissue was measured as described earlier (Kingsbury et al, 1995). When available, post mortem toxicology was reviewed to exclude recent substance misuse and levels of anti-psychotic and anti-cholinergic drugs in the blood were recorded. Duration of illness (DOI) was calculated as the time from first contact with a psychiatric clinical service to death. The most recently prescribed anti-psychotic and anti-cholinergic drugs and their final recorded prescribed doses were recorded and converted to standardized drug doses (see Table 1).

Diagnostic Evaluation

For a patient to be included in this study, sufficient information was needed to be available from clinical case records to enable a psychologist and psychiatrist to reach a diagnostic consensus using the DIBS and thus be able to make a diagnosis according to the DSM-IV criteria (American Psychiatric Association, 2000). Patients with schizoaffective disorder were excluded.

[3H]-Pirenzepine Binding

In situ radioligand binding and autoradiography were performed using five 20 μm frozen tissue sections from each block of BA9. The binding of [3H]-PZP (15 nM) was measured in the presence (non-specific binding: two sections) or absence (total binding: three sections) of 1 μM quinuclidinyl xanthene-9-carboxylate hemioxalate (Dean et al, 1996) after incubation in 10 mM KH2PO4, 10 mM Na2HPO4; pH 7.4 (buffer A) at 25°C for 30 min. Sections were washed twice for 2 min in ice-cold buffer A, dipped in ice-cold water and thoroughly dried before being fixed overnight in paraformaldehyde fumes in a desiccator. The sections, and a set of [3H]-micro-scales™, were apposed against a BAS-TR2025 imaging plate until an image of appropriate intensity was obtained for scanning in the BAS 5000 phosphoimager. Exposure time related to both the density of binding sites and the specific activity of the radioligand were used. The intensity of the phosphoimages was then measured by comparison with the intensity of the blocks of radioactivity on the [3H]-microscales using AIS image analysis software, with the results being expressed as d.p.m. per mg of estimated wet weight tissue equivalents (ETE) and then converted to fmol per mg ETE. In this way, [3H]-PZP binding was measured using a single-point saturation analysis, which provides a good approximation of the density of radioligand binding sites in tissue sections (Dean et al, 2002).

Western Blotting

Tissue samples from Brodmann area 9 (BA9) were solubilized in sample buffer containing 1% SDS, 1 mM Na3VO4 and 10 mM Tris–HCl (pH 7.5). Proteins were loaded (25 μg/well) in duplicate on to 10% SDS polyacrylamide gels and resolved by gel electrophoresis for 1 h at 150 V. Gels were then equilibrated in Towbin's transfer buffer (Tris/glycine/methanol) for 15 min. Proteins were then transferred to Hybond nitrocellulose membranes (GE Healthcare) for 1 h at 100 V in Towbin's buffer. Membranes were blocked for 1 h at room temperature in Tris-buffered saline/0.1% Tween 20 (TBS-T) containing 5% non-fat milk powder and then incubated overnight at 4°C in 7.5 ml TBS-T containing rabbit anti-Gαq/11 (1 : 500 dilution). On the next day, the membranes were washed 3 × 5 min in TBS-T at room temperature and incubated for 2 h at room temperature in TBS-T containing Dako goat anti-rabbit IgG:HRP-conjugated secondary antibody diluted 1 : 2000. The membranes were washed for 3 × 5 min in TBS-T at room temperature and incubated for a further 5 min at room temperature in SuperSignal ECL solution (Pierce). Excess solution was drained and blotted and a single 5 min exposure was captured using a 440CF Kodak imaging station. Band (43 kDa) intensity is reported as a ratio to an internal control (IC). A representative western blot of relative Gαq/11 levels in a control patient and a patient with schizophrenia is shown in Figure 1b. Before measuring Gαq/11 in the cases, sufficient protein homogenate (IC) was prepared from the frontal cortex of a volunteer with no history of psychiatric or neurological illness. Aliquots of this homogenate were run in each of 12 wells on 2 gels (24 samples over all), and the OD of each immunopositive Gαq/11 band was measured as described above. These experiments showed that the anti-human Gαq/11 antibody bound to a CNS protein of appropriate molecular weight. Moreover, using data from the IC, both the inter- and intra-gel variation for the measurement of Gαq/11 was shown <15%. Subsequently, a sample of IC was included in two lanes of each subsequent gel on which protein from samples was separated and the OD of each sample was expressed as a ratio of the IC to control for gel-to-gel variation in our analyses (Dean et al, 2002).

Figure 1
figure 1

Representative concentration–response curves for oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding in membranes prepared from dorsolateral prefrontal cortex (Brodmann area 9 (BA9)) tissue of control and muscarinic receptor-deficit schizophrenia (MRDS) and non-MRDS groups (a). Basal and agonist-stimulated specific [35S]-GTPγS binding to Gαq/11 proteins was determined as described in the Materials and Methods section. Data for one individual from each subgroup are shown to illustrate the rightward curve shift, increase in maximal response and steepened slope in the MRDS group relative to non-MRDS and controls. (b) Shows a representative western blot of relative Gαq/11 levels in the dorsolateral prefrontal cortex (BA9) of a control and a patient with schizophrenia. The lanes show images of the band in the internal control (IC), a control (C) and a patient with schizophrenia (S).

Membrane Preparation

Brodmann area 9 of each individual donor was homogenized using a Polytron in 10 volumes of 10 mM HEPES, pH 7.4, containing 1 mM EGTA, 1 mM dithiothreitol (DTT), 10% sucrose and a complete protease inhibitor cocktail. The resultant homogenate was diluted 10-fold and centrifuged at 1000 × g for 10 min at 4°C, the supernatant saved and the pellet re-homogenized and centrifuged as above. The combined supernatants were then centrifuged at 11 000 × g for 20 min at 4°C. The resulting pellet was re-homogenized in 40 volumes of 10 mM HEPES, 1 mM EGTA, 1 mM DTT, 1 mM MgCl2, pH 7.4 and centrifuged at 27 000 × g for 20 min at 4°C. The resulting pellet was re-suspended in the same buffer at a protein concentration of 1 mg/ml, aliquots snap frozen in liquid nitrogen and stored at −80°C. To minimize membrane degradation, assays were conducted within 24 h of membrane preparation.

[35S]-GTPγS Binding/Immunocapture Assay

[35S]-GTPγS-Gαq/11 immuno-specific binding using a 96-well SPA-based method was performed using the method described by Salah-Uddin et al (2008). Assays were performed blind to patient diagnosis and extracts from each cohort of matched patients with schizophrenia and controls were run together in the same set of experiments. Membranes were pre-treated with 10 mM N-ethylmaleimide (NEM) for 60 min on ice and subsequently diluted in an assay buffer (20 mM HEPES, 100 mM NaCl, 10 mM MgCl2, pH 7.4) at a final protein concentration of 25 μg per assay point. GDP (0.1 μM) was added at 55 min into the NEM-containing incubation medium. Experimental reactions were performed in a final volume of 100 μl in 96-well Optiplates™. To ensure equilibrium [35S]-GTPγS binding, 60 μl of membranes were added to each well containing an agonist and incubated at 25°C for 20 min. Nucleotide exchange was initiated through the addition of 20 μl [35S]-GTPγS to each well to give a final concentration of 500 pM and membranes were incubated for 60 min at 25°C. Ice-cold 0.27% Igepal-CA630 was used to terminate the reaction. Rabbit anti-Gαq/11 antibody (1 : 300) was then added to each well and the plate agitated at 4°C for 60 min. Finally, anti-IgG-coated PVT-SPA beads were added to each well, the plate agitated once again for a further 30 min at 4°C after which it was centrifuged. Bound radioactivity was measured using a TopCount detector.

Data Analysis

Concentration–response curves were fitted by non-linear regression analysis with variable slope using GraphPad Prism 4 (GraphPad Prism Software Inc., San Diego, CA). The residuals of all the experimental responses were analyzed to identify the distribution. G protein and [3H]-PZP were normally distributed and signal window (c.p.m.), Hill slope and pEC50 were log-normally distributed. A one-way analysis of variance (ANOVA) was used to analyze the G protein, [3H]-PZP, Hill slope and pEC50, and analysis of covariance was used to analyze binding-over-basal with age being fitted as the covariate. Planned comparisons using Student's t-test were then investigated and adjusted for multiplicity (using Tukey's test) to identify whether there were significant differences in the responses among the three diagnostic groups.

Pearson product–moment correlations, assuming a straight-line best fit, were used to analyze relationships between experimental parameters, and relationships between the continuous clinical parameters and the experimental parameters. The clinical responses, age, PMI, brain weight and brain pH were analyzed using one-way ANOVA, and planned comparisons using Student's t-test were then investigated and adjusted for multiplicity to identify whether there were significant differences in the responses among the three diagnostic groups; DOI and last recorded drug dose were analyzed using Student's t-test, and sex and suicide were analyzed using Fisher's exact test. Any significant differences observed in the experimental data set were not attributable to demographics.

RESULTS

Demographic Data

Dorsolateral prefrontal cortex (BA9) tissue was obtained from 14 male and 6 female patients with schizophrenia, and 8 male and 2 female controls. There was no statistical difference in the mean ages and gender of the patients with schizophrenia and controls (F (2, 27)=0.008, P=0.992 and P=0.153, respectively). Similarly, there were no significant differences in CNS pH (F (2, 27)=0.929, P=0.407), brain weight (F (2, 20)=2.083, P=0.151), PMI (F (2, 27)=0.441, P=0.648), DOI (F(1,18)=0.302, P=0.589), or last recorded drug doses (F (1,18)=0.493, P=0.492). An increased suicide rate was observed in the combined group of patients with schizophrenia, but was not significantly different among the subgroups of patients with schizophrenia (P=0.076).

[3H]-Pirenzepine Binding

[3H]-pirenzepine binding was significantly decreased in tissue sections prepared from BA9 from patients with schizophrenia compared control brains (105±17 vs 183±12 fmol per mg ETE; (F (2, 27)=53.618, P<0.0001). Patients with schizophrenia were divided into two subgroups based on [3H]-PZP binding: ‘low’ [3H]-PZP (MRDS; <100 fmol per mg ETE) and ‘normal’ [3H]-PZP (non-MRDS; >100 fmol per mg ETE) binding groups; [3H]-PZP binding in these subgroups was 38±9 and 174±12 fmol per mg ETE, respectively (Figure 2). [3H]-PZP binding in the MRDS group was significantly different from both the control and non-MRDS groups (P<0.001), whereas there was no difference between the control and non-MRDS groups (P=0.806).

Figure 2
figure 2

[3H]-Pirenzepine (PZP) binding in dorsolateral prefrontal cortex (Brodmann area 9) sections from low PZP (muscarinic receptor-deficit schizophrenia (MRDS)) and normal PZP (non-MRDS) subgroups of patients with schizophrenia and controls. Specific [3H]-PZP binding was determined as described in the Materials and Methods section and expressed as d.p.m. per mg estimated wet weight equivalents (ETE). The horizontal line indicates the mean value for each group. Statistically significant differences between groups are indicated as ***P<0.001.

Western Blot Analysis

Immunoblotting for Gαq/11 proteins in BA9 of all patients showed no differences among controls, MRDS or non-MRDS groups (F (2,27)=0.164, P=0.850). The ratio of band intensity compared with the IC was 1.09±0.17 and 1.00±0.15 for the MRDS and non-MRDS groups, respectively, and 0.98±0.06 for the control group. No correlation was found between Gαq/11 immunoreactivity and [3H]-PZP binding (P=0.667).

[35S]-GTPγS Binding/Immunocapture Assay

To assess the CHRM1 function in membranes prepared from BA9, agonist-stimulated Gαq/11-[35S]-GTPγS immunospecific binding was assessed in controls, MRDS and non-MRDS groups. The CHRM full agonist oxotremorine-M and the CHRM1-selective allosteric partial agonist, AC-42 (Spalding et al, 2002; Langmead et al, 2006), were used to stimulate receptors in all membrane preparations. Representative concentration-dependent responses for oxotremorine-M- and AC-42-stimulated Gαq/11-[35S]-GTPγS binding in the different groups are shown in Figure 3. The potency of oxotremorine-M differed among groups (F (2,27)=4.234, P=0.025) and was significantly decreased in the MRDS group (pEC50 (M), 5.69±0.16) compared with the control group (pEC50=6.17±0.10; P=0.024). No difference in potency was observed between the control and the non-MRDS groups (pEC50=6.05±0.07; P=0.806). Scatter graphs of these data are shown in Figure 4. Similar trends were also seen with respect to AC-42-stimulated Gαq/11-[35S]-GTPγS binding in the different groups (pEC50 (M) values: control, 5.31±0.19; MRDS, 4.72±0.11; non-MRDS, 5.14±0.14; F (2, 27)=4.195, P=0.076) with a significantly decreased potency in the MRDS group (P=0.024).

Figure 3
figure 3

Representative concentration–response curves for oxotremorine-M- and AC-42-stimulated Gαq/11-[35S]-GTPγS binding in membranes prepared from dorsolateral prefrontal cortex (Brodmann area 9) tissue of control and low and normal [3H]-pirenzepine binding in subgroups of patients with schizophrenia. Basal and agonist-stimulated specific [35S]-GTPγS binding to Gαq/11 proteins was determined as described in the Methods section. Curves for 10 donors in each subgroup were constructed and analyzed to provide the data shown in Figures 1, 4 and 5.

Figure 4
figure 4

Potency estimates (pEC50) in dorsolateral prefrontal cortex (Brodmann area 9 (BA9)) sections from muscarinic receptor-deficit schizophrenia (MRDS) and non-MRDS groups and controls. Oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding concentration-effect curves were analyzed for membranes prepared from dorsolateral prefrontal cortex (BA9) tissue of control and low (MRDS) and normal (non-MRDS) [3H]-pirenzepine binding subgroups of patients with schizophrenia. A pEC50 (M) value was determined for each concentration–effect curve as described in the Methods section. A statistically significant difference between groups is indicated as *P<0.05; **P<0.01.

No differences were detected in the basal levels of Gαq/11-[35S]-GTPγS binding between the groups (basal values (c.p.m. per 25 μg membrane protein): control, 2147±96; MRDS, 2138±93; non-MRDS, 2384±152; F (2,27)=2.074, P=0.145). We therefore chose to analyze relative agonist efficacy by expressing agonist-stimulated increases in Gαq/11-[35S]-GTPγS binding as the magnitude of the signal window generated in c.p.m. (Figure 5). Maximal oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding in BA9 membranes (measured as concentration–response curve maxima−basal calculated by GraphPad Prism) was significantly different in the MRDS group (signal window, 4703±296 c.p.m.) compared with the controls (3240±190 c.p.m.; P=0.003). In contrast, no significant difference was observed between the control and non-MRDS groups (P=0.464). Similar, but not statistically significantly different, trends were seen with respect to AC-42 relative efficacy differences between the groups with respect to the increase in Gαq/11-[35S]-GTPγS binding-over-basal (c.p.m): control, 1964±173; MRDS, 2664±228; non-MRDS, 2234±176; F(2, 27)=2.183, P=0.062). The intrinsic activity of AC-42 relative to oxotremorine-M ((maximal AC-42 response/maximal oxotremorine-M response) × 100) remained unchanged in all subgroups (control, 53±3%; MRDS, 54±4%; non-MRDS, 50±3%).

Figure 5
figure 5

Relative efficacy estimates in dorsolateral prefrontal cortex (Brodmann area 9 (BA9)) sections from muscarinic receptor-deficit schizophrenia (MRDS) and non-MRDS groups and controls. Oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding concentration–effect curves were analyzed for membranes prepared from dorsolateral prefrontal cortex (BA9) tissue of control, MRDS and non-MRDS groups. The signal window (in c.p.m.) was determined as an increase in [35S]-GTPγS binding to Gαq/11 proteins over basal as described in the Methods section. A statistically significant difference between groups is indicated as **P<0.01.

Representative concentration–response curves from the control subgroups of patients with schizophrenia are shown in Figure 1. In addition to illustrating the dextral shift and increase in maximal response observed for the MRDS group, it can also be seen that there are apparent differences in the slope factors (Hill coefficients) for the three curves (F (2, 27)=14.835, P<0.001). We therefore determined Hill coefficients for oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding responses (Figure 6) in the controls and sub-group patients with schizophrenia. Hill coefficients for oxotremorine-M concentration–response curves for the MRDS (nH=0.82±0.02) were significantly greater than those determined for control (0.64±0.01; P<0.001) and non-MRDS (0.68±0.02; P=0.0013) groups. There was no significant difference between the latter two groups.

Figure 6
figure 6

Hill coefficients from analysis of oxotremorine-M concentration–response Gαq/11-[35S]-GTPγS binding curves in dorsolateral prefrontal cortex (Brodmann area 9) sections from MRDS, non-MRDS groups and controls. Statistically significant differences between groups are indicated as ***P<0.001.

When potencies, relative efficacies and Hill coefficients for oxotremorine-M-stimulated receptor-Gαq/11-[35S]-GTPγS binding were plotted against [3H]-PZP binding values for all patients, significant correlations were obtained for all parameters; pEC50 (P<0.001, r2=0.325; Figure 7a), maximal Gαq/11-[35S]-GTPγS binding-over-basal (P=0.046, r2=0.134; Figure 7b) and Hill coefficient (P<0.001, r2=0.551; Figure 7c). Although the observed correlation between receptor expression and pEC50 value was consistent with receptor theory (ie, a decrease in pEC50 as receptor expression increases), the increasing maximal responsiveness with decreasing receptor expression was not.

Figure 7
figure 7

Correlation analysis of pEC50 (a), oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding signal window (b) and Hill coefficients (c) with [3H]-pirenzepine binding in dorsolateral prefrontal cortex (Brodmann area 9) sections from muscarinic receptor-deficit schizophrenia (MRDS), non-MRDS groups and controls.

DISCUSSION

In this study we have shown that a subgroup of patients with schizophrenia defined as MRDS (Scarr et al, 2008), display distinct functional changes with respect to the proximal downstream signalling consequences of CHRM activation. In particular, despite a marked reduction in CHRM1 expression (down by 75–80% compared with both a control group and a subgroup of patients with schizophrenia displaying normal [3H]-PZP binding), agonist-stimulated [35S]-GTPγS binding to Gαq/11 proteins was increased in MRDS membranes prepared from dorsolateral cortex (BA9), relative to both the other non-MRDS and control groups.

Changes in CHRM expression in the brains of patients with schizophrenia have been widely reported, with the majority of studies reporting subtype-specific deficits (Dean et al, 1996, 2002; Crook et al, 2000, 2001; Raedler et al, 2003). On the basis of the criteria used recently to define a CHRM-deficit endophenotype (Scarr et al, 2008), we have performed a pharmacological analysis of the CHRM function in membranes prepared from BA9 of MRDS and non-MRDS patients, and controls. We have used a Gαq/11-[35S]-GTPγS binding immunocapture assay (Salah-Uddin et al, 2008), which allows us to assess productive receptor-G protein coupling in membrane preparations and hence report a proximal functional readout of relative efficacy, as well as expression level of CHRM, within human dorsolateral cortex for the different patient groups. In addition, we have shown earlier in human cortical (BA23/25) membranes that the oxotremorine-M-mediated signal is wholly attributable to CHRM1 stimulation, as agonist-stimulated [35S]-GTPγS binding to Gαq/11 proteins is completely prevented by pre-incubation with the selective CHRM1 toxin, MT-7 (Salah-Uddin et al, 2008).

Initial experiments showed that the full agonist oxotremorine-M and the CHRM1-selective allosteric partial agonist, AC-42 (Spalding et al, 2002; Langmead et al, 2006), each caused concentration-dependent increases in Gαq/11-[35S]-GTPγS binding in controls and both sub-group patients with schizophrenia. This is of particular relevance in the MRDS endophenotype as it might have been predicted that CHRM1 function would be diminished because of the observed decrease in the receptor number, and that therapeutic strategies targeting the CHRM1 in this group might not be effective. However, although the receptor number was decreased in the MRDS group, the residual CHRM1 population coupled with greater efficiency to this key functional readout such that maximal responses were undiminished. Data were generated using oxotremorine-M for the comparison of potency, efficacy and receptor/G protein cooperativity in healthy and diseased tissue. This is because full agonist, oxotremorine-M, provides a larger signal window compared with the partial agonist AC-42 that allows for subtle differences in the pharmacology to be measured. This is very important when investigating pathological mechanisms. Furthermore, the intrinsic activity of AC-42 relative to oxotremorine-M was similar in all groups. The selectivity and novel binding site of AC-42 at the CHRM1 affords it, and similarly acting compounds, potential therapeutic advantages (Langmead and Christopoulos, 2006) and therefore our observation is important in showing that AC-42 is effective in functionally activating the CHRM1 subtype with similar relative efficacy in brain tissue from both normal donors and patients with schizophrenia. The effects of AC-42 followed trends similar to oxotremorine-M, but were less pronounced in patient groups because of the smaller signal window.

Given the marked decrease in [3H]-PZP binding sites it is not surprising that the potency of oxotremorine-M to stimulate Gαq/11-[35S]-GTPγS binding was reduced two- to three-fold in the MRDS subgroup compared with controls and the non-MRDS subgroup. Classical studies in which receptor expression levels are incrementally decreased by irreversible alkylation have shown that the concentration–effect curves become increasingly right-shifted and then collapse once all ‘spare’ receptors have been eliminated (see Kenakin, 2006).

In contrast, the relative efficacy of oxotremorine-M was significantly greater in the MRDS subgroup compared with the control and non-MRDS groups, despite the marked decrease in [3H]-PZP binding observed in this subgroup. This difference cannot be explained or predicted simply by receptor number as efficacy is generally determined by multiple components in the signal transduction system measured. An alternative way of showing this between-sub-groups efficacy difference is presented in Figure 8. Here we have calculated the increase in [35S]-GTPγS bound to Gαq/11 proteins (as fmol per mg membrane protein) stimulated by a maximally effective oxotremorine-M concentration and compared this with the receptor density (assuming that 15 nM [3H]-PZP provides an estimate of the receptor density in each BA9 membrane preparation). In membranes prepared from control and non-MRDS groups, this yields a stoichiometry of <1; however, for the MRDS subgroup this value rises to 4, representing an eight- to ten-fold change in the CHRM1-to-Gαq/11-[35S]-GTPγS stoichiometry in this schizophrenia endophenotype, and suggesting a fundamentally altered receptor-G protein coupling in this group. Given that the intrinsic activity of AC-42 does not vary between subgroups we can conclude that a comparable stoichiometry change also occurs for this allosteric partial agonist in MRDS.

Figure 8
figure 8

Assessing the stoichiometry of agonist-stimulated Gαq/11-[35S]-GTPγS binding relative to CHRM expression in dorsolateral prefrontal cortex (Brodmann area 9). Individual values for the net increase in Gαq/11-[35S]-GTPγS binding stimulated by oxotremorine-M (100 μM) were calculated as fmol per mg protein. This allowed a direct comparison with CHRM1 expression, assessed by [3H]-pirenzepine binding (see Materials and Methods). The stoichiometry (the molar ratio of Gαq/11-[35S]-GTPγS binding to receptor level) is shown for controls, muscarinic receptor-deficit schizophrenia (MRDS) and non-MRDS groups (n=10 per group). Statistically significant differences between groups are indicated as ***P<0.001.

Another potentially important observation made in this study is that the slope factor (Hill coefficient) of the concentration–response curves for oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding varies among the groups. The Hill coefficient can be used to indicate cooperativity between receptor and G protein and, in turn, be used as a measure of receptor-G protein coupling efficiency. Hill coefficients of <1 were consistently observed in all patient groups; however, significantly greater values were observed in the MRDS group. There are a number of possible explanations of why Hill coefficients within this range are observed with respect to receptor-G protein coupling. These include the possibility of receptor populations existing in different affinity states for the agonist (eg, ‘free’ receptors vs pre-coupled receptor-G protein ternary complexes (De Lean et al, 1980)), or the compartmentalization of receptors and/or G proteins constraining productive coupling. Irrespective of the precise explanation, the observed increase in Hill coefficient for oxotremorine-M-stimulated Gαq/11-[35S]-GTPγS binding concentration–response curves in the MRDS group may be a further indication of an adaptation to the decreased CHRM1 number in cortical neuronal populations of patients with schizophrenia categorized within the MRDS endophenotype. This suggests an enhanced CHRM1/Gαq/11 coupling and supports the altered stoichiometry of receptor-facilitated Gαq/11-[35S]-GTPγS binding observed in this subgroup (Figure 8).

The stoichiometry of productive receptor-G protein coupling is likely to be influenced by a variety of factors. Here we have shown that levels of Gαq/11 protein expression within BA9 are not altered in schizophrenia. An earlier study, which undertook quantitative analysis of Gαq/11 protein expression in human cerebral cortex, indicated that this G protein subtype is expressed at a level of 20–30 pmol per mg protein (López de Jesús et al, 2006). This indicates that Gαq/11 proteins are likely to be expressed in considerable (>10-fold) excess compared with the CHRM1 in the membrane preparations used here, and thus Gαq/11 protein availability for productive coupling to agonist-occupied CHRM1 is unlikely to be rate-limiting. Therefore, it may be possible for each activated CHRM1 to recruit multiple Gq/11 proteins and in the MRDS endophenotype the data would suggest that an increased receptor/G protein ratio may manifest itself as a greater cooperativity between receptor and G protein (higher Hill coefficient) along with increased [35S]-GTPγS binding per CHRM1.

Irrespective of what underlies the relative efficacy difference, we have clearly shown in using a Gαq/11 protein-specific [35S]-GTPγS binding assay that although a sub-population of patients with schizophrenia show a marked decrease in dorsolateral prefrontal cortical CHRM1, this is (super-) compensated for by an adaptative change in receptor-G protein coupling efficiency. Whether this adaptation represents a manifestation of the underlying disease process, or is a secondary process mitigating (or militating) the decline in CHRM1 expression remains to be determined. The CHRM1 subtype is predominant among CHRMs in the cortex, striatum and hippocampus (Weiner et al, 1990; Levey et al, 1991), where it is expressed on the majority of neuronal post synaptic nerve terminals (Hersch and Levey, 1995). The CHRM1 subtype has been shown to be involved in cognitive processes, most recently through the use of knockout mice (Hamilton et al, 1997; Anagnostaras et al, 2003; Wess et al, 2007). An array of clinical and basic science evidence has implicated dysfunctional prefrontal cortical circuitry in the pathophysiology of schizophrenia (Perlstein et al, 2001; Ragland et al, 2007), with cholinergic deficits being specifically implicated (Hyde and Crook, 2001). To date, it has not been possible to separate the MRDS endophenotype from other patients with schizophrenia by CHRM1 sequence, gender, age, suicide, DOI or any particular drug treatment (Scarr et al, 2008). However, our observation of altered CHRM1 number and efficacy (assessed as receptor-stimulated [35S]-GTPγS-for-GDP exchange on Gαq/11 proteins) provides new insight into how the prefrontal cortical cholinergic circuitry may change and adapt in different endophenotypes of patients with schizophrenia. Interestingly, very recent data from a small clinical trial of patients with schizophrenia (Shekhar et al, 2008) showed that the CHRM agonist, xanomeline, which shows some degree of selectivity for CHRM1, separated from placebo for verbal learning and short-term memory indices. These data suggest that this therapeutic intervention may be effective in the treatment of cognitive deficits in schizophrenia and it will be important to establish the relative effectiveness of CHRM1 agonists in the MRDS and non-MRDS subgroups of patients with schizophrenia.