INTRODUCTION

Acetylcholine (ACh) plays an important role in motor function and various domains of cognition, for example attention, learning, and memory (Winkler et al, 1995; Perry et al, 1999). Cholinergic dysfunction has been shown to be central to the pathophysiology of Alzheimer's disease (Cummings and Benson, 1987) and has also been postulated to contribute to the cognitive deficits of various neuropsychiatric disorders, including schizophrenia (Tandon and Greden, 1989; Sarter and Bruno, 1998). There are five known (M1–M5) muscarinic acetylcholine receptors in the human genome (Kubo et al, 1986; Bonner et al, 1987; Brann et al, 1993). Of these, the M1 receptor has been most closely linked to schizophrenia. The M1 receptor subtype is the most abundant of the muscarinic receptors in the cortex and hippocampus (Levey et al, 1991; Wei et al, 1994), brain regions crucial to cognitive function. Decreased M1 receptor binding has been reported in postmortem studies of the prefrontal cortex, hippocampus, and striatum from patients with schizophrenia (Dean et al, 1996; Crook et al, 2000, 2001; Katerina et al, 2004); and decreased M1-receptor cDNA levels in the frontal cortex have also been reported (Mancama et al, 2003). This has contributed to the suggestion that enhancement of central cholinergic neurotransmission by M1 agonists might be useful to treat the cognitive impairments of schizophrenia (Sur et al, 2003; Weiner et al, 2004).

Clozapine, the prototypical atypical antipsychotic drug (APD), was the first APD shown to be effective in treating the cognitive dysfunction of schizophrenia (Hagger et al, 1993), a finding which has been replicated, and is shared by other compounds with a similar pharmacology, for example, olanzapine, quetiapine, risperidone, and ziprasidone (Woodward et al, 2005). Clozapine and olanzapine have been reported to have antimuscarinic properties (Herrling and Misbach-Lesenne, 1982; Bymaster et al, 1996). Clozapine has nanomolar affinity for all five cloned muscarinic receptors (Bolden et al, 1992). It has been reported to act as an antagonist at M1 receptor (Bolden et al, 1992; Zorn et al, 1994; Sur et al, 2003; Weiner et al, 2004) and M2/3/5 receptor (Bymaster et al, 1996; Michal et al, 1999). However, clozapine is also reported as M1/2/4 partial agonist (Zorn et al, 1994; Fritze and Tilmann, 1995; Zeng et al, 1997; Olianas et al, 1997). The discrepancy from these studies may be due to the methodology difference as these experiments involved CHO cells. To add to the complexity, N-desmethylclozapine (NDMC), the major active metabolite of clozapine in rodent and man (Aravagiri and Marder, 2001; Baldessarini et al, 1993; Weigmann et al, 1999), has its own unique muscarinic receptor pharmacology. Clozapine is rapidly metabolized to NDMC in rats and, thus, high serum levels of NDMC are seen after oral administration of clozapine, producing brain levels comparable to serum levels (Baldessarini et al, 1993; Weigmann et al, 1999). NDMC has been reported to be a potent M1 agonist in vivo (Sur et al, 2003; Weiner et al, 2004) and, like clozapine, to have high affinities for 5-HT2A and 5-HT2C, and weaker, but still significant affinities, for D2 receptors (Kuoppamaki et al, 1993, Weiner et al, 2004). This receptor-binding profile is similar to clozapine, suggesting that NDMC might have antipsychotic properties. NDMC also demonstrates a high affinity for M4 and M5 receptors, comparable to that observed for M1 receptors (Weiner et al, 2004). Acute administration of NDMC, like clozapine, significantly increases c-Fos expression in the medial prefrontal cortex (mPFC) and nucleus accumbens (NAC), consistent with its atypical APD pharmacologic profile (Young et al, 1998).

Recently, Weiner et al (2004), using a cell-based functional assay, compared the effects of NDMC and clozapine on muscarinic receptors, and observed that NDMC displayed high potency and significant agonist efficacy at multiple muscarinic receptor subtypes, most notably the M1 receptor. By contrast, clozapine behaved as an antagonist. Moreover, the M1 agonist activity of NDMC was blocked by both atropine and clozapine. Furthermore, NDMC, but not clozapine, increased the phosphorylation of mitogen-activated protein kinase (MAP kinase) in the CA1 regions of mouse HIP, a response consistent with M1 and not M2–M5-receptor activation (Berkeley et al, 2001). These results suggest that NDMC is a potent M1 agonist, whereas clozapine displays potent M1 antagonist actions in vivo. NDMC is the only commonly used antipsychotic agent that has been reported to have M1 agonist activity (Weiner et al, 2004).

The ability of APDs to improve some or all aspects of the cognitive deficit in schizophrenia (Meltzer and McGurk, 1999; Woodward et al, 2005) has been attributed, in part, to their ability to preferentially increase the release of dopamine (DA) (Imperato and Angelucci, 1989; Moghaddam and Bunney, 1990; Kuroki et al, 1999) and ACh in the cortex and HIP (Ichikawa et al, 2002a, 2002b; Shirazi-Southall et al, 2002; Chung et al, 2004), while the anticholinergic activity of clozapine, olanzapine, thioridazine, and mesoridazine has been suggested to interfere with memory (Eitan et al, 1992; Adler et al, 2002; McGurk et al, 2004). The increased DA release induced by the atypical APDs may be due, in part, to blockade of serotonin 5-HT2A and D2 receptors, and direct or indirect stimulation of 5-HT1A receptors (Ichikawa et al, 2001). The mechanism by which clozapine increases ACh release in the mPFC is distinct from the mechanism by which clozapine increases cortical DA release, since 5-HT1A receptor stimulation is not a factor in clozapine-induced ACh release (Ichikawa et al, 2002a).

In order to test the hypothesis that NDMC is an M1 agonist and that the M1-antagonist effect of clozapine may diminish the M1-agonist effect of NDMC, the present study examined the effect of NDMC alone, and following pretreatment with telenzepine, an M1-preferring antagonist (Schudt et al, 1988; Noronha-Blob et al, 1988), or low-dose clozapine, on DA and ACh release in the mPFC and, in some experiments, the NAC and HIP as well. We have previously found that telenzepine inhibited the ability of clozapine to increase DA and ACh release in rat mPFC (Ichikawa et al, 2004). We also examined the ability of WAY100635, a 5-HT1A-receptor antagonist reported to block the effects of clozapine on DA but not ACh release (Ichikawa et al, 2002a), to inhibit the effect of NDMC on mPFC DA and ACh release.

MATERIALS AND METHODS

Animals

Male Sprague–Dawley albino rats (Zivic-Miller Laboratories, Porterville, PA) weighing 250–350 g were housed two per cage and maintained in a controlled 12 : 12-h light/dark cycle and under constant temperature at 22°C, with free access to food and water. Animals used in this study were cared for in accordance with the guidelines of the Institutional Animal Care and Use Committee of Vanderbilt University. ‘Principles of laboratory animal care’ (NIH Publication No. 85-23, revised 1985) were followed.

Surgery and Microdialysis

Rats were anesthetized with the modified Equithesin mixture (810 mg pentobarbital, 4.3 g choral hydrate, 2.12 mg MgSO4, 14 ml ethanol, and 29 ml propylene glycol were dissolved in saline and the final volume was 100 ml), and mounted in a stereotaxic frame (Stoetling, Wood Dale, IL). Stainless guide cannula (21-gauge) with a dummy probe were placed and fixed by cranioplastic cement (Plastic One, Roanoke, VA) onto the cortex dorsal to both the mPFC and the NAC. Rats received dual probe implantation for the mPFC, NAC, or HIP (coordinates: A +3.2, L +0.8 (10°C inclination), V −5.5 mm; A +2.0, L +1.5 to +1.7, V −7.5 mm; and A +5.6, L +5.0, V −7.0 mm, respectively, relative to bregma). The incision bar level was 3.0 mm, according to the atlas of Paxinos and Watson (1998).

The microdialysis probes were constructed in our laboratory. A silica-glass capillary tube (150 μm o.d., 75 μm i.d., Polymicro Technologies, Phoenix, AZ) was inserted through the inner bore of a 25 G stainless tube. The stainless tube was inserted into a 28 G Teflon tubing and then the Teflon tubing was inserted into the inner bore of a 18 G stainless tube. The hollow fiber dialysis membrane (polyacrylonitrile/sodium methalylsulfonate polymer, 310 μm o.d., 220 μm i.d., 40 000 Da cutoff, AN69HF, Hospal; CGH Medical, Lakewood, CO) was fitted over the glass capillary and into the end of the 25 G stainless tube. This junction (0.5 mm) was glued with epoxy (5-Min Epoxy; Devkon, Danverse, MA, USA) after the length of the hollow dialysis fiber was cut to 3 mm and the tip of the membrane (0.5 mm) was plugged with epoxy. The length of exposed nonglued surface for dialyzing was 3 mm.

At 3–5 days after cannulation, a dialysis probe was implanted into the mPFC and NAC under slight anesthesia with isoflurane (Metofane, Pitman-Moore, Mundelein, IL). Rats were then housed individually overnight in a dialysis cage. After the overnight perfusion at 0.4 μl/min of the probe, the flow was increased to 1.5 μl/min. After 1 h, the dialysate samples were collected every 30 min. The perfusion medium was Dulbecco's phosphate-buffered saline solution (Sigma, St Louis, MO), including Ca2+ (138 mM NaCl, 8.1 mM Na2HPO4, 2.7 mM KCl, 1.5 mM KH2PO4, 0.5 mM MgCl, 1.2 mM CaCl2, pH 7.4). No AChesterase inhibitor in the dialysate is required with this procedure (Ichikawa et al, 2002b). After stable baseline values in the dialysates were obtained, each rat received two injections, vehicle/NDMC, WAY100635/NDMC, telenzepine/NDMC, or clozapine/NDMC. The locations of the dialysis probes were verified at the end of each experiment by brain dissection. The procedures applied in these experiments were approved by the Institutional Animal Care and Use Committee of Vanderbilt University in Nashville, TN, where the present studies were completed.

Biochemical Assays

Determination of DA

Dialysate samples were directly applied onto a high-performance liquid chromatography (HPLC) with electrochemical detection, and analyzed with a Millennium chromatogram manager (Waters, Milford, MA). DA was separated (BDS Hypersil 3 μm C18, 1.0 × 100 mm2; Keystone Scientific, Bellefonte, PA) at 35°C maintained by column heater (LC-22C Temperature Controller; BAS, West Lafayette, IN). The mobile phase consisted of 48 mM anhydrous citric acid and 24 mM sodium acetate trihydrate containing 0.5 mM EDTA-Na2, 10 mM NaCl, 2 mM dodecyl sulfate sodium salt, and 17 % (v/v) acetonitrile, adjusted to pH 4.8 with concentrated NaOH, and was pumped (0.05 ml/min) by LC-10AD (Shimadzu, Kyoto, Japan). A Unijet working electrode (MF-1003, BAS) was set at +0.58 V (LC-4C, BAS) vs an Ag/AgCl reference electrode. Reagents used were analytical or HPLC grade.

Determination of ACh

The method has been described previously (Ichikawa et al, 2002a). In brief, dialysate samples are directly injected onto the liquid chromatography/electrochemistry (LCEC) system assisted by a chromatography manager (Millennium; Waters, Milford, MA), and analyzed for ACh. ACh is separated on a coiled cation exchanger ACh column (analytical column) (Sepstik 10 nm ID 530°C 1.0 nm; BAS, West Lafayette, IN), followed by the post-IMER (immobilized enzyme reactor) (BAS), which consists of choline oxidase (ChO)/AChesterase. ACh is hydrolyzed by AChesterase to form acetate and choline in the post-IMER, and then choline is oxidized by ChO to produce betaine and hydrogen peroxide (H2O2). H2O2 is detected and reduced to H2O on a Unijet amperometric detector cell with a peroxidase-redox-coated glassy carbon electrode (MF-9080; BAS), set at +100 mV (LC-4C; BAS) vs Ag/AgCl reference electrode. This reduction is analyzed with the detector (LC-4C; BAS) as signal indicating ACh in the chromatogram.

Drugs

NDMC (ACADIA Pharmaceutical Inc.) and clozapine (Sandoz, East Hanover, NJ) was dissolved in a small amount of 0.1 M tartaric acid and the pH was adjusted to 6–7 with 0.1 N NaOH. WAY100635 (Wyeth Laboratories, Philadelphia, PA) and telenzepine (Research Chemical Inc.) were dissolved in deionized water. Vehicle or drugs in a volume of 1.0 ml/kg were administered subcutaneously to randomly assigned rats.

Data analysis

Mean predrug baseline levels (time −60, time −30, and time 0) were designated as 100%. Following a significant overall repeated measures ANOVA (treatment × time), Fisher's protected least significant difference post hoc pairwise comparison and one-way ANOVA (StatView® 4.5 for the Macintosh) were used to determine group differences. A probability p<0.05 was considered significant in this study. All results are given as mean±SEM.

RESULTS

Basal extracellular DA levels in the dialysates obtained from all the rats used in this study were 1.93±0.11 (mean±SEM fmol/10 μl; N=51) for the mPFC, 2.28±0.09 (mean±SEM fmol/10 μl; N=45) for the HIP, and 15.26±0.52 (mean±SEM fmol/20 μl; N=42) for the NAC, respectively. Basal extracellular ACh levels in the dialysates obtained from all the rats used in this study were 7.85±0.22 (mean±SEM fmol/10 μl; N=40) for the mPFC, 6.15±0.37 (mean±SEM fmol/10 μl; N=38) for the HIP, and 4.28±0.65 (mean±SEM fmol/20 μl; N=45) for the NAC, respectively. The ACh concentration in the mPFC or HIP was significantly higher than that in the NAC. There were no significant differences in basal extracellular DA or ACh levels between treatment groups within each region.

As shown in Figure 1, NDMC, at doses of 10 and 20 mg/kg, but not 5 mg/kg, dose-dependently increased extracellular DA concentrations in the mPFC (F(1,12)=14.77, p=0.0002; F(1,11)=32.49, p<0.0001, and F(1,10)=1.27, p=0.26, respectively). NDMC, at 10 and 20 mg/kg, but not 5 mg/kg, also significantly increased cortical ACh release, but in a nondose-dependent manner (F(1,10)=4.18, p=0.04; F(1,9)=6.8, p=0.01; and F(1,10)=2.02, p=0.16, respectively). High doses of NDMC and clozapine produced a similar effect on DA release (250% over the baseline) (Kuroki et al, 1999). However, at a low dose, 5 mg/kg, clozapine had a greater effect in cortical DA release than NDMC since at 5 mg/kg NDMC had no effect on DA release but clozapine produced a significant increase in DA release in the mPFC (Kuroki et al, 1999). Clozapine produced a much greater increase in ACh release than NDMC since both low (5 mg/kg) and high (20 mg/kg) doses of clozapine produced a great increase in ACh release in the mPFC (Ichikawa et al, 2002b).

Figure 1
figure 1

Time course effects of N-desmethylclozapine on extracellular dopamine (a) and acetylcholine (b) levels in the medial prefrontal cortex. The arrows indicate drug injection times. Data are means±SEM (N=4–7) of the dialysate dopamine or acetylcholine levels, expressed as a percentage of each predrug baseline dopamine or acetylcholine value.

In the HIP, NDMC, 10 and 20 mg/kg, significantly but nondose-dependently increased DA release (F(1,8)=13.54, p=0.0004 and F(1,10)=13.18, p=0.004, respectively) as well as ACh release (F(1,10)=19.48, p<0.0001 and F(1,9)=32.83, p<0.0001, respectively) (Figure 2). However, 5 mg/kg of NDMC did not increase either DA or ACh release in this region (F(1,7)=3.025, p=0.756 and F(1,7)=3.339, p=0.705, respectively) (Figure 2). In the NAC, neither 10 or 20 mg/kg of NDMC had any effect on DA (F(1,10)=0.64, p=0.43 and F(1,11)=0.6, p=0.44, respectively) or ACh release (F(1,10)=0.56, p=0.49) (Figure 3).

Figure 2
figure 2

Time course effects of N-desmethylclozapine on extracellular dopamine (a) and acetylcholine (b) levels in the hippocampus. The arrows indicate drug injection times. Data are means±SEM (N=5–6) of the dialysate dopamine or acetylcholine levels, expressed as a percentage of each predrug baseline dopamine or acetylcholine value.

Figure 3
figure 3

Time course effects of N-desmethylclozapine on extracellular dopamine and acetylcholine levels in the nucleus accumbens. The arrows indicate drug injection times. Data are means±SEM (N=4–7) of the dialysate dopamine levels, expressed as a percentage of each predrug baseline dopamine or acetylcholine value.

Telenzepine, 3 mg/kg, completely blocked 10 mg/kg NDMC-induced DA (Figure 4; F(1,12)=5.71, p=0.018) and ACh (Figure 4; F(1,9)=38.29, p<0.0001) release in the mPFC. Clozapine, 1.25 mg/kg, which itself had no effect on mPFC DA or ACh release (Figure 5), blocked NDMC (10 mg/kg)-induced ACh (Figure 5; F(1,9)=9.63, p=0.003) but not DA (Figure 5; F(1,10)=0.0003, p=0.99) release. WAY100635 partially and significantly blocked the increased mPFC DA release produced by NDMC, 20 mg/kg (Figure 6; F(1,8)=4.73, p=0.03), but had no effect on ACh release produced by the same dose of NDMC (Figure 6; F(1,8)=4.73, p=0.03).

Figure 4
figure 4

The effect of the M1-receptor antagonist telenzepine (3 mg/kg, s.c.) on extracellular dopamine (a) and acetylcholine (b) release induced by N-desmethylclozapine (10 mg/kg, s.c.) in the medial prefrontal cortex. Rats were pretreated with telenzepine 30 min prior to administration of N-desmethylclozapine. The arrows indicate drug injection times. Data are means±SEM (N=5–7) of the dialysate dopamine or acetylcholine levels, expressed as a percentage of each predrug baseline dopamine or acetylcholine value.

Figure 5
figure 5

The effect of clozapine (1.25 mg/kg, s.c.) on extracellular dopamine (a) and acetylcholine (b) release induced by N-desmethylclozapine (10 mg/kg, s.c.) in the medial prefrontal cortex. Rats were pretreated with clozapine 30 min prior to administration of N-desmethylclozapine. The arrows indicate drug injection times. Data are means±SEM (N=5–6) of the dialysate dopamine or acetylcholine levels, expressed as a percentage of each predrug baseline dopamine or acetylcholine value.

Figure 6
figure 6

The effect of the 5-HT1A-receptor antagonist WAY100635 (0.2 mg/kg, s.c.) on extracellular dopamine (a) and acetylcholine (b) release induced by N-desmethylclozapine (20 mg/kg, s.c.) in the medial prefrontal cortex. Rats were pretreated with WAY100635 30 min prior to administration of N-desmethylclozapine. The arrows indicate drug injection times. Data are means±SEM (N=4–8) of the dialysate dopamine or acetylcholine levels, expressed as a percentage of each predrug baseline dopamine or acetylcholine value.

DISCUSSION

The main findings of the present study are that (1) NDMC, the major active metabolite of clozapine, significantly increased DA and ACh release in the mPFC and HIP, but not the NAC; (2) the M1-preferring antagonist telenzepine completely blocked DA and ACh release in the mPFC produced by NDMC; (3) NDMC (10 mg/kg)-induced ACh release was completely blocked by clozapine (1.25 mg/kg), consistent with previous reports that NDMC is a potent M1 agonist, while clozapine has M1 antagonist properties in vivo; (4) clozapine pretreatment did not block NDMC-induced cortical DA release, indicating M1 agonism did not contribute to this effect of NDMC; and (5) the increases in DA, but not ACh, release in the mPFC produced by NDMC was partially blocked by the 5-HT1A antagonist WAY100635, indicating that cortical DA release is partially dependent upon 5-HT1A-receptor stimulation.

Effect of NDMC on DA Release

Like clozapine and other atypical APDs, NDMC preferentially increased DA release in the mPFC and HIP compared to the NAC. Low-dose NDMC (5 mg/kg) had no effect on DA release in the mPFC, whereas the same dose of clozapine significantly increased mPFC DA release (Kuroki et al, 1999). However, NDMC and clozapine, at a dose of 20 mg/kg, produced similar increases in DA release. This suggests that NDMC may contribute to the ability of clozapine to increase cortical DA release in the rodent.

The fact that the increased DA release induced by NDMC in the mPFC was completely blocked by the M1-receptor antagonist telenzepine indicates that the cortical DA release produced by NDMC is dependent upon activation of M1 receptors. However, telenzepine also partially or completely blocked the effect of clozapine and risperidone, respectively, to increase DA release in the mPFC (Ichikawa et al, 2004). Risperidone, which has very low affinity for any muscarinic receptor subtype, is not an effective agonist at M1 receptors (Schotte et al, 1996; Weiner et al, 2004). This suggests that both NDMC and risperidone increase cortical DA release by a mechanism that is not dependent upon direct stimulation of M1 receptors, but could involve indirect mechanism as well, and so does not prove that NDMC is acting through a direct M1 mechanism. The same may be true for ACh release. The M1 receptor is the primary muscarinic receptor in the human frontal, temporal, parietal, and occipital cortical areas (Flynn et al, 1995). Cortical M1 receptors are localized mainly on postsynaptic dendrites and spines associated with both glutamatergic and cholinegic transmission (Mrzljak et al, 1993). As the density of M1 receptors is much greater than M4 receptors in the cortex (Levey et al, 1991; Volpicelli and Levey, 2004), it seems more likely that the effect of atypical APDs to increase ACh release is more likely to be M1- rather than M4-mediated. We are currently investigating whether the effect of NDMC and clozapine is cortically mediated through local injection studies.

The ability of NDMC, like clozapine, to increase cortical DA release was also partially blocked by the 5-HT1A-receptor antagonist, WAY100635. Thus, both NDMC and clozapine increase cortical DA release, in part by a 5-HT1A-dependent mechanism. NDMC has a higher affinity for the 5-HT1A (111 nM) than for the D2 (265 nM) receptor (P Herrling and P Neumann, personal communication, 1989) and is most likely a 5-HT1A partial agonist, as is clozapine. However, WAY100635 also inhibits the increase in DA release produced by olanzapine and risperidone, neither of which are 5-HT1A partial agonists (Ichikawa et al, 2001), suggesting an indirect mechanism that includes 5-HT1A receptor stimulation.

NDMC also has higher affinities for the 5-HT2A and 5-HT2C receptors than the parent compound clozapine (Kuoppamaki et al, 1993; Weiner et al, 2004). The antagonism of 5-HT2A, 5-HT2C and D2 receptor by NDMC may contribute to its ability to increase DA release in the mPFC and HIP, as is the case for other atypical APDs (Kuroki et al, 1999; Liegeois et al, 2002; Meltzer et al, 2003).

Effect of NDMC on ACh Release

NDMC, like clozapine and other atypical APDs (Ichikawa et al, 2002a), significantly increased ACh release in the mPFC and HIP, but not the NAC. The NDMC-induced ACh release in the mPFC was blocked by telenzepine but not WAY100635, as has been previously reported for clozapine and risperidone (Ichikawa et al, 2002a, 2004). This suggests that NDMC-induced cortical ACh release may be mediated by direct or indirect stimulation of M1 but not 5-HT1A receptors. In the present study, low-dose clozapine attenuated NDMC-induced cortical Ach, but not DA release, suggesting the M1-receptor antagonism of clozapine blocked the M1 agonism of NDMC. Therefore, the net effect of clozapine to increase cortical ACh release in vivo may be due, in part, to its metabolite NDMC, which would be partially attenuated by the M1 antagonist actions of clozapine. NDMC displays high potency interactions with all five human muscarinic receptors, with marked agonist activity at the M1, M4, and M5 receptors (Weiner et al, 2004). The M1 receptors involved in DA and ACh release may be located on DA and ACh postsynaptic nerve terminals in the cortex, HIP, or elsewhere in the forebrain on circuits that regulate the release of these neurotransmitters by 5-HT1A receptors as well as glutamatergic and GABAergic mechanisms. Johnson et al (2005) recently reported that intra-hippocampal infusion of 10 μM clozapine and 100 μM olanzapine, but not intra-septal infusion, by reverse dialysis, increased HIP ACh efflux to an extent comparable to that of systemic administration. Cholinergic neurons from the mesopontine cholinergic nuclei (Ch5, Ch6) project to the DA cell bodies in the VTA (Bymaster et al, 2002). However, mainly M5, not M1, muscarinic receptors are localized on these neurons (Weiner et al, 1990). Further studies are required to determine if M1 receptors located elsewhere, for example, the ventral tegmentum, nucleus basalis Meynert, or the septum, are involved in the effect of clozapine or NDMC in enhancing cortical or HIP DA release.

Previous in vivo microdialysis studies suggest that the muscarinic autoreceptor modulating ACh efflux in the mammalian medial pontine reticular formation (Baghdoyan et al, 1998), striatum (Billard et al, 1995), and cortex (Iannazzo and Majewski, 2000; Douglas et al, 2001) is M2. Therefore, clozapine, NDMC, and olanzapine which are, to varying extents, M2 antagonists (Bymaster et al, 2002; Weiner et al, 2004), may also enhance cortical and HIP ACh release via blockade of M2 autoreceptors (Bymaster et al, 1996). Since completion of this study, Johnson et al (2005) reported that clozapine and olanzapine, 10 mg/kg, produced a marked increase in extracellular ACh in the HIP while ziprasidone produced a small increase. Based upon correlation of the ED400% and in vitro functional potencies at muscarinic M2 receptors, these authors concluded that the increase in ACh release produced by these compounds was due to M2 antagonism. It should be noted that the study of Johnson et al (2005) used neostigmine in the dialysate fluid. We have shown elsewhere that this may alter the effect of some but not all psychotropic drugs. As ziprasidone and risperidone, which lack significant M2 antagonism produce large increases in ACh release in the HIP, which are blocked by telenzepine, as is the case with clozapine (Chung et al, 2004; Ichikawa et al, 2004), we propose that M1 agonism, direct or indirect, rather than M2 antagonism, is primarily responsible for the release of ACh in the HIP.

Clinical Significance: NDMC, M1-Receptor Agonism and Cognition

As previously mentioned, the M1 receptor subtype is the most abundant of the muscarinic receptors in the cortex and hippocampus (Levey et al, 1991; Wei et al, 1994), brain regions crucial to normal cognitive function. M1 receptors in the hippocampus have been shown to activate extracellular signal-regulated kinases (ERK), which are crucial for many neural functions, including learning, memory, and synaptic plasticity (Berkeley et al, 2001). These authors concluded that M1 receptor-mediated ERK activation provides a mechanism by which M1 receptors could modulate learning and memory. M1 receptor agonists have been reported to improve working memory in animals (Aura et al, 1997; McDonald et al, 1998). Muscarinc antagonists with weak specificity for the M1 receptor may worsen working memory in patients with schizophrenia (Spohn and Strauss, 1989; King, 1990), while more specific M1 antagonists do so in laboratory animals (Bymaster et al, 1993; Roldan et al, 1997). Mice lacking M1 receptors exhibit deficits in measures of spatial learning and memory, indicative of impaired hippocampal and cortical function (Anagnostaras et al, 2003). Learning deficits in the radial arm maze and fear-conditioning paradigm have also been reported in M1-knockout mice (Miyakawa et al, 2001). Moreover, M1-deficient mice have significantly elevated DA neurotransmission in the striatum (Gerber et al, 2001), significantly increased locomotor activity and increased response to the stimulatory effects of amphetamine, evidence of an inhibitory effect of the M1 receptor on dopaminergic transmission, which suggests a possible basis for an antipsychotic effect of M1 agonists. As previously mentioned, the M1/4 agonist xanomeline has been reported to mimic the effect of D2 antagonists to produce an antipsychotic-like profile in rats (Stanhope et al, 2001). It has been reported that NDMC dose-dependently potentiated NMDA receptor currents in CA1 pyramidal cells by 53% (Sur et al, 2003). Decreased glutamatergic activity in pyramidal neurons has been hypothesized to be a major factor in the pathophysiology of schizophrenia (Moghaddam, 2004; Javitt, 2004). Thus, the M1 agonism of NDMC may, by stimulating glutamatergic activity, be of particular importance to the beneficial effects of NDMC and the parent compound, clozapine, on cortical function. Patients with schizophrenia who are heterozygous for the C267A polymorphism (267C/A) of the M1 receptor have been reported to produce more correct responses and less perseverative errors on the Wisconsin Card Sort test, which is dependent upon prefrontal cortical function (Morice, 1990; Berman et al, 1995), than those who were homozygous for 267 C/C, providing additional genetic evidence suggesting that M1 receptors have an important effect on prefrontal cortical function (Liao et al, 2003).

The effect of clozapine on DA or ACh release is most likely the result of the combined effect of clozapine and NDMC, the agonist/antagonist mixing. Thus, high NDMC levels, and particularly high NDMC/clozapine ratios, would increase M1 muscarinic receptor stimulation, as predicted by mass action and by agonist/antagonist mixing studies (Brauner-Osborne et al, 1996). Brain clozapine concentrations in the rat during chronic treatment have been reported to exceed those of NDMC during chronic treatment by three-fold (Weigmann et al, 1999). There is no information on what the relative levels are in man. High concentrations of NDMC are found in plasma samples in some patients treated with clozapine (Hasegawa et al, 1993). High NDMC levels, and a high NDMC/clozapine ratio even more so, would increase M1 muscarinic receptor stimulation. The present data on the blockade of NDMC-induced ACh release by clozapine are consistent with clinical data from our laboratory, which suggest that the NDMC/clozapine ratio is a better predictor of clinical response to clozapine than clozapine levels alone (Frazier et al, 2003; Mauri et al, 2003; Weiner et al, 2004).

In conclusion, NDMC preferentially increased DA and ACh release in the mPFC and HIP but not the NAC, similar to the effect of clozapine and other atypical APDs. The blockade of NDMC-induced ACh release by telezenpine and clozapine indicates that the stimulation of M1 receptors contributes to the ability of NDMC to increase cortical DA and ACh release, confirming that NDMC has significant M1 agonistic actions, whereas the parent compound, clozapine, is an antagonist.