Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Muscarinic receptors have been characterized according to their pharmacological sensitivity and high affinity for specific antagonists (Hulme et al., 1990). M₁ receptors show sensitivity and affinity for pirenzepine, whereas M₂ receptors show lower affinity for pirenzepine and higher sensitivity for compounds such as AF-DX 116 and methoctramine. Receptors of the M₃ subtype were found to be more sensitive to 4-DAMP (piperidinium, 4-[diphenylacetyloxy]-1,1-dimethyl-iodide) and HHSiD (Silanol, cyclohexylphenyl[3-(1-peperidinyl)- propyl]-monohydrochloride). An M₄ receptor has been identified, but there are fewer data on selective agonists/antagonists for this receptor, and pharmacological characterization of endogenous M₅ receptors in tissue preparations is minimal (Caulfield and Birdsall, 1998).

Five subtypes of muscarinic receptors (M₁-M₅) have been demonstrated to be distinct gene products on the basis of molecular characterization. These receptors were shown to share a high level of sequence homology and belong to the super family of G protein-coupled receptors possessing seven transmembrane-spanning domains (Kubo et al., 1986; Bonner et al., 1987). In most neurons, M₁, M₃, and M₅ receptors appear to be positively coupled to phospholipase C, resulting in the formation of diacylglycerol and the release of inositol trisphosphate, which in turn liberates free, intracellular Ca²⁺. In contrast, M₂ and M₄ receptors are negatively coupled to adenylyl cyclase.

There are several clinical disorders with impaired cognitive function that may benefit from drugs that improve memory and learning. The largest population would be individuals suffering from Alzheimer's disease (AD), an age-related neurodegenerative disorder presenting progressive loss of cognitive function and characterized by brain neurofibrillary tangles and amyloid-containing plaques. Several neurotransmitters have been shown to be altered during the course of the disease, the most consistent pathological findings being the loss of forebrain cholinergic neurons, as evidenced by a decrease in choline acetyltransferase activity (Whitehouse et al., 1981). Based on these findings, it has been proposed that precursor therapy (e.g., choline, lecithin), acetylcholinesterase inhibitors (AChEIs), muscarinic agonists, and acetylcholine (ACh)-releasing agents would be useful in the symptomatic treatment of AD. To date, only the AChEIs Cognex (Warner-Lambert Co., Morris Plains, NJ; Knapp et al., 1994) and Aricept (Pfizer Inc., Groton, CT; Doody, 1999) have met regulatory agency requirements necessary for marketing these compounds in the United States as AD treatments.

The relationship between changes in muscarinic receptors and AD remains controversial, with decreases, increases, and no changes observed (Pavia et al., 1998). During disease progression, it may be that presynaptic M₂ receptors are lost, whereas postsynaptic M₁ receptors remain largely intact (Mash et al., 1985). Thus, agonists acting directly at these sites should restore lost cholinergic activation even when the disease has progressed to the point that cholinergic function is no longer adequate to support the efficacy of an AChEI. M₁-selective agonists may be especially useful in AD, because they have the additional property of having fewer side effects.

Demonstrating clinical efficacy with first generation muscarinic agonists (e.g., arecoline, oxotremorine, RS-86) has been difficult. Although in some cases the clinical trials were flawed by use of too few patients or poor clinical outcome measures, all of these agents produced dose-limiting peripheral cholinergic side effects at doses equal to or below those achieving efficacy. In addition, they also suffered from limited oral activity, minimal central nervous system penetration, short duration of action, or poor biodistribution (Jaen and Schwarz, 1997).

Milameline is a novel muscarinic receptor agonist that overcomes many of the limitations of first-generation agonists (Fig. 1). The results described herein present the in vitro biochemical characterization of milameline and demonstrate that it possesses central activity in rats and rhesus monkeys at doses that do not routinely produce undesirable peripheral cholinergic side effects.

View larger version (8K): [in this window] [in a new window]   Fig. 1.   Chemical structure of milameline (CI-979, PD129409, RU35926).

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

[³H]CMD and [³H]QNB Receptor Binding in Rat Brain Tissue

The cerebral cortices from male Long-Evans rats (180-200 g) were homogenized in 10 mM Na⁺/K⁺ buffer (pH 7.4) and then aliquoted (~7 mg wet weight tissue) to tubes along with various concentrations of agonist and either 0.03 nM [³H]QNB (43.6 Ci/mmol) or 1 nM [³H]CMD (55.5 Ci/mmol) in a final 2-ml volume. Each concentration was run in triplicate tubes. Specific binding was defined by the presence of 1 µM atropine. Samples were incubated for 2 h at 25°C and then filtered through GF/B filters (Whatman Inc., Clifton, NJ) presoaked in 0.05% polyethylenimine with a Brandel cell harvester (Brandel Laboratories, Gaithersburg, MD). Filters were washed three times with ice-cold buffer and then counted with liquid scintillation techniques. IC₅₀ values were determined by nonlinear curve fitting.

Cell Growth Conditions

Chinese hamster ovary (CHO)-K1 cell lines expressing the five subtypes of muscarinic receptors were initially obtained from Dr. M. R. Brann (National Institute of Neurological Disorders and Stroke, University of Vermont, Burlington, VT) and grown at 37°C with 5% CO₂ in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 1% nonessential amino acids. The cells were passed on a weekly basis, and receptor levels appeared stable for up to 4 months in culture. The total number of receptors was controlled by the confluency of the cells, with the highest number of receptors observed just below confluency. For whole-cell binding studies, cells were harvested at 70 to 90% confluency with either light trypsinization (0.05% trypsin with EDTA) or a 0.02% EDTA solution in PBS. No significant difference in binding was observed with either of these two treatments. To harvest membranes, cells were washed with PBS, scraped into ice-cold 5 mM Tris-HCl with EDTA, homogenized with a Polytron disrupter, sonicated, and then pelleted at 20,000 rpm for 30 min. Membranes were resuspended in 10 mM Na⁺/K⁺ PO₄ buffer, aliquoted, and kept frozen at 80°C until the day of assay.

[³H]N-Methylscopolamine ([³H]NMS) Binding in Transfected CHO Cells

Whole-Cell Binding. Whole-cell binding with the transfected CHO cells was performed essentially according to the methods of Fisher (1988). Intact cells were harvested, washed with buffer A (142 mM NaCl, 5.6 mM KCl, 2.2 mM CaCl₂, 3.6 mM Na₂HCO₃, 1 mM MgCl₂, 5.6 mM glucose, and 30 mM HEPES; pH 7.4), and then aliquoted (~40-80 µg of protein) to 2 ml of buffer A containing [³H]NMS (78.9 Ci/mmol) and the appropriate muscarinic agonist. The concentration of the ligand was 0.04 nM for human (h)M₁, hM₂, hM₃, and hM₄ cells and 0.06 nM for hM₅ cells. Nonspecific binding was defined as that unaffected by the inclusion of 50 µM atropine. The reaction was initiated by the addition of the cells and allowed to proceed for 60 min at 37°C with oxygenation and shaking. Termination of the incubation was achieved by rapid vacuum filtration through GF/B Whatman filters with a Brandel cell harvester. Filters were washed three times with 5 ml ice-chilled 0.9% NaCl. Ten milliliters of Beckman Ready Gel scintillation cocktail (Beckman Instruments, Fullerton, CA) was added to the filters in vials, and the samples were allowed to set overnight before vortexing and counting in a Beckman 2800 scintillation counter. IC₅₀ values were obtained with a logit equation adapted from Parker and Waud (1971), with K_i values calculated from the equation of Chang and Prusoff (1973).

Cell Membrane Binding. Membrane aliquots of between 10 and 40 µg were added to 2-ml volumes of 10 mM Na⁺/K⁺ PO₄, pH 7.4, containing appropriate concentrations of test compound and 0.1 nM [³H]NMS. Atropine at 1 µM was used to determine nonspecific binding. Incubation proceeded for 120 min at 25°C with constant agitation and was terminated by rapid vacuum filtration through GF/B Whatman filters with a Brandel cell harvester. Filters were washed three times with 5 ml of ice-cold 10 mM Na⁺/K⁺ PO₄, placed in vials with 10 ml of Beckman Ready Gel, allowed to sit overnight, vortexed, and then counted. Results were calculated as described above.

Phosphatidylinositol (PI) Hydrolysis

Transfected CHO Cells. Cells were grown in 12-well plates under the conditions described above. Four to 6 days after seeding the plates, the nutrient medium was aspirated and hM₁, hM₃, or hM₅ CHO cells were labeled with 1 µCi/ml of [³H]myo-inositol (specific activity, 15.0-18.8 Ci/mmol) in 0.5 ml of media per well. After 48 h, the medium was aspirated, and the cells were washed twice with 1 ml of minimum essential medium (MEM) containing 10 mM LiCl (MEM/LiCl). A volume of 0.5 ml MEM/LiCl was then added to each well and allowed to incubate at 37°C for at least 15 min. The stimulation period was initiated by the addition of 10 µl of the appropriate agonist concentration and allowed to proceed for 15 min, at which time the reaction was terminated by the aspiration of medium and the addition of 0.5 ml ice-cold 5% trichloroacetic acid (TCA). After waiting at least 15 min, the TCA extract was applied to Dowex-formate columns (Bio-Rad AG 1-X8 resin, formate form, 100-200 mesh; Bio-Rad Labs., Richmond, CA). The wells were rinsed with 0.5 ml of distilled H₂O, which was also applied to the columns. The columns were washed four times with 3 ml of 5 mM myo-inositol, and then total [³H]inositol phosphates were eluted into scintillation vials with two 2-ml washes of 1 mM NH₄ formate containing 0.1 M formic acid. Beckman Ready Gel scintillation cocktail (10 ml) was added, and the samples were counted in a Beckman 2800 scintillation counter. As an index of efficacy, the percentage of maximal carbachol (1 mM) stimulation was calculated for each agonist tested, whereas, as a measure of potency, the concentration producing a half-maximal response (EC₅₀) was calculated.

Rat Cortical Slices. Male Long-Evans rats (300-400 g) were sacrificed, and the cerebral cortex (minus the olfactory bulb) was dissected from the rest of the brain. The tissue was cut into 0.35 × 0.35-mm slices on a McIlwain tissue chopper and dispersed in Krebs-Ringer HEPES-buffered medium, pH 7.4. Composition of the medium was: 119 mM NaCl, 4.75 mM KCl, 1.25 mM CaCl₂, 1.20 mM KH₂PO₄, 1.18 mM MgSO₄, 22 mM HEPES, and 10 mM glucose. After two washes with medium, the slices were incubated with shaking and under oxygenated conditions for 30 min at 37°C and then put into medium with 10 mM LiCl present. Aliquots of 10 to 15 mg of slices were then pipetted into 12 × 75-mm glass tubes to which [³H]myo-inositol (15-18 Ci/mmol; 0.5 µM final concentration) was added. The tubes were then incubated for 60 min, at which time agonists or vehicle was added. The final volume was 0.3 ml. The incubation was allowed to proceed for another 60 min, and the reaction was terminated by the addition of 0.94 ml of 1:2 chloroform/methanol. After 15 min, 0.31 ml of chloroform and 0.31 ml of H₂0 were added, and all samples were vortexed and then centrifuged for 15 min at 500g. From the upper aqueous phase, 0.75 ml was put on the Dowex columns, and samples were processed as above for the CHO cell to measure total [³H]inositol phosphates produced.

cAMP Accumulation in Transfected CHO Cells. Muscarinic receptor-mediated inhibition of forskolin-stimulated cAMP accumulation was measured in hM₂ and hM₄ CHO cells. The cells were grown as described above in six-well plates. Cells were washed twice with serum-free MEM and then incubated in medium containing 1 mM isobutyl-1-methylxanthine (IBMX) for 5 min at 37°C. The experiment was initiated by the addition of agonist alone for 2 min followed by the addition of 5 µM forskolin, with the incubation allowed to proceed for an additional 2.5 min. Aspirating the assay medium and extracting the cells with cold 5% TCA terminated the reaction. cAMP was assayed with Amersham's scintillation proximity assay (Arlington Heights, IL). The scintillation proximity assay reaction was carried out in 96-well trays, incubated at room temperature for 15 to 20 h with shaking, and counted in a Wallac 1205 betaplate reader. EC₅₀ values were then determined for each agonist.

Release of [³H]ACh from Slices of Rat Cerebral Cortex. Male Long-Evans rats (200-300 g) were sacrificed, and the cerebral cortex (minus the olfactory bulbs) was dissected out from the rest of the brain. Cuboidal slices (0.3 × 0.3 mm) were cut on a McIlwain tissue chopper and dispersed in ice-chilled Krebs-Ringer HEPES-buffered medium, pH 7.2. This was composed of 119 mM NaCl, 4.75 mM KCl, 1.25 mM CaCl₂, 1.20 mM KH₂PO₄, 1.18 mM MgSO₄, 22 mM HEPES, and 10 mM glucose. After a quick centrifugation at 500g, the supernatant was discarded, and the slices were incubated with [³H]choline (80 Ci/mmol; 0.01 µM final concentration) in 10 ml of normal medium for 15 min at 37°C with oxygenation and shaking. The slices were washed three times with 5 ml of medium and then resuspended in a volume of medium such that 0.2-ml aliquots of this suspension would be equivalent to at least 10 to 15 mg of tissue. The slices were further incubated for 15 min in a final volume of 3 ml in normal medium or medium with elevated K⁺ (and lowered Na⁺) in the presence or absence of muscarinic agonist. Placing the samples on ice followed by the separation of medium and tissue by rapid centrifugation terminated the reaction. Tissue samples were acidified with perchloric acid and then homogenized. The amount of radioactivity was measured in both fractions by liquid scintillation counting techniques. Under these conditions, >85% of the ³H counted resulting from Ca²⁺-dependent K⁺-stimulated release is [³H]ACh.

Inhibition of Rat Brain AChE Activity. The method was based on that of Ellman et al. (1961) with modifications by Marks et al. (1981). Male Long-Evans rats (200-300 g) were sacrificed, and the cerebral cortex (minus the olfactory bulbs) was dissected out from the brain and homogenized in 100 mM K⁺PO₄ buffer, pH 7.4. Aliquots of the tissue (~0.2 mg wet weight) were dispensed to 12 × 75-mm glass tubes with various concentrations of test agent and 0.5 mM acetylthiocholine for a final volume of 200 µl. Tubes were incubated at 37°C for 30 min, at which time 0.9 mM neostigmine was added, and the tubes were placed on ice to terminate the reaction. Color was developed with the addition of 300 µl of dithio-bis(2-nitrobenzoic) acid (final concentration of 425 µM). The absorbance of each tube's contents was measured on a spectrophotometer set at 412 nm. A blank tube used to null the absorbance to zero contained the appropriate concentration of the compound tested plus 5 µM the specific AChEI BW284C51 [1,5-bis-(4-allyldimethyl ammonium phenyl)-pentan-3-one dibromide]. The concentration of protein in the homogenate was measured by the Bradford method, with bovine -globulin as the standard (Bradford, 1976).

Spontaneous and Scopolamine-Induced Swimming Activity in Rats. Male Long-Evans rats (Blue Spruce Farms Inc., Altamont, NY) weighing 180 to 240 g were used in all testing. These animals were housed under a 12-h light-dark schedule in group cages (4-6 rats/cage) with food and water available ad libitum. Swimming activity (total distance) was recorded continuously during a 5-min test session in a 1.2 × 1.2 × 0.6-m white plastic enclosure, filled to a depth of 0.3 m with water (21-23°C) that was made opaque with dry milk. The total distance traveled was measured with a computerized videotracking system (Columbus Instruments, Columbus, OH). Spontaneous swimming activity was measured in rats given milameline (0.1, 0.32, 1.0, and 3.2 mg/kg; n = 5/dose) or vehicle (0.9% saline; n = 5) p.o. 30 min before testing. The effect of milameline on scopolamine-induced swimming activity was measured in rats injected i.p. with scopolamine (0.32 mg/kg; n = 5) or vehicle (0.9% saline; n = 5) 30 min before testing. Milameline (3.2, 10.0, and 32.0 mg/kg; n = 5/treatment) was administered s.c. 15 min after scopolamine, which was 15 min before testing. Data were analyzed with ANOVA followed by Newman-Keuls post hoc comparisons.

Effects on Spatial Memory in Basal Forebrain-Lesioned Rats. Male Long-Evans rats (Blue Spruce Farms) weighing 280 to 320 g were used. These animals were singly housed under a 12-h light-dark schedule with food and water available ad libitum. Rats were anesthetized with sodium pentobarbitol (55 mg/kg i.p.) and placed in a small animal stereotaxic apparatus (model 900, David Kopf Instruments, Tujunga, CA). Cannulas (27 gauge) were directed at four sites within the region ventral to the globus pallidus encompassing the basal forebrain cholinergic neurons, two sites in each hemisphere. Coordinates were (from stereotaxic zero): A 8.5 mm, L ±2.5 mm at the anterior two sites, and A 7.5 mm, L ±2.7 mm at the posterior two sites. Each cannula was lowered 8.0 mm below the dural surface. One microliter of ibotenic acid (3 µg/µl) in artificial cerebrospinal fluid was delivered at each site over 5 min via syringe pump (model 555; Sage Instruments, Cambridge, MA). The cannula was left in place for 30 s after completion of each infusion and then withdrawn. Sham-lesioned animals (artificial cerebrospinal fluid injections only) were similarly prepared. After 5 weeks recovery in their home cages, rats were tested for their ability to acquire a spatial water-maze task. The water-maze task consisted of a pool (1.2 × 1.2 × 0.6 m) filled with water (21-23°C) made opaque by the addition of powdered milk. The goal was a submerged platform (11.5 cm square) centered in one of the four quadrants of the pool with its surface 2.5 cm below the water level. Each rat was tested in a single trial on each of 5 consecutive days, followed by 2 days off, and then an additional single trial on day 8. Performance was assessed 30 min after a daily s.c. injection of milameline (0.1 and 0.32 mg/kg; n = 21/treatment) or vehicle (0.9% saline; n = 21). Sham lesioned rats (n = 21) received daily injections of saline only. Lesions were confirmed by histological evaluation at the completion of testing. Data were analyzed with ANOVA followed by Newman-Keuls post hoc comparisons

Rat Cortical Electroencephalogram (EEG) Activity. Male Long-Evans rats (Blue Spruce Farms) weighing 250 to 350 g were used. Animals were anesthetized with sodium pentobarbitol (55 mg/kg i.p.) and placed in a small-animal stereotaxic apparatus (David Kopf). These rats were singly housed under a 12-h light-dark schedule with food and water available ad libitum. Animals were surgically implanted with stainless steel electrodes screwed into the skull surface overlying the frontal and parietal cortex. Coordinates were (from bregma): A 2.0 mm, L ±3.0 mm (frontal); P 4.0 mm, L ± 2.0 mm (parietal). Electrodes were secured to a plastic connector (model MS-363, Plastic Products, Roanoke, VA) that was permanently attached to the skull surface with dental acrylic. Animals were given 1 week to recover from surgery before testing. After recovery, animals were placed into sound-attenuating chambers during the light cycle of the rats' normal diurnal cycle and allowed 2 h to habituate to these environments. Thereafter, EEG was recorded continuously for 10 min before and 2 h after drug administration from each conscious, freely moving rat. EEG samples were taken continuously and converted once each second from the time to the frequency domain (Mi² model M100 A-D converter, Modular Instruments Inc., Southeastern, PA) with fast Fourier transformation (FFT). The power spectra from these FFTs were summed across 15-min epochs to yield a mean power spectrum for the period. Mean power within the 0 to 4 (), 5 to 8 (), 9 to 15 (), 16 to 25 (1), and 26 to 35 (2)-Hz bands was calculated separately for each bandwidth. Total power within all bandwidths also was calculated. Milameline (1.0, 3.2 mg/kg, and 10.0) or vehicle (0.2% carboxymethylcellulose) was administered p.o., immediately after a 15-min collection of EEG data as reflected in the 0 time point. Each rat (n = 7) received all treatments in random order with no more than one treatment per week. Data for each bandwidth and total power were analyzed separately via nonparametric-ranked ANOVA (Kruskal-Wallis rank transformed analysis with repeated measures). Comparisons among means were made with Duncan's post hoc tests at inclusive intervals before treatment and at 15 to 30, 45 to 60, and 90 to 120 min after drug administration.

Rat Core Body Temperature. Male Long-Evans rats (Blue Spruce Farms) weighing 180 to 240 g were used. These animals were housed under a 12-h light-dark schedule in group cages (4-6 rats/cage) with food and water available ab libitum. Rectal measurement of core body temperature was made with a telethermometer (model 43TA, Yellow Springs Instruments Company Inc., Yellow Springs, OH). Rats were administered milameline (0.1, n = 5; 0.32, n = 5; 1.0, n = 10; and 3.2 mg/kg, n = 10) or vehicle (0.9% saline; n = 15) p.o. 30 min before body temperature was measured. Data were analyzed with ANOVA followed by Newman-Keuls post hoc comparisons.

Local Cortical Blood Flow in Rats. Male Long-Evans rats (Blue Spruce Farms) weighing 350 to 500 g were used. These rats were singly housed under a 12-h light-dark schedule with food and water available ab libitum. Animals were anesthetized with sodium pentobarbitol (55 mg/kg i.p.) and placed in a small-animal stereotaxic apparatus (David Kopf). Polarizable 4-mm-long platinum-iridium electrodes (Rhodes Medical Instruments, Woodland Hill, CA) insulated to a 1-mm bared tip were surgically implanted in the frontal cortex at coordinates (from stereotaxic zero): A 2.0 mm, L 2.0 mm, and 1.0 mm below the dural surface. A nonpolarizable reference electrode was placed anterior to the blood flow electrode. Electrodes were secured to a plastic connector (model MS-363, Plastic Products) that was permanently attached to the skull surface with dental acrylic. Animals were given 1 week to recover from surgery before testing. After recovery, local cortical blood flow was measured with the hydrogen clearance technique from platinum electrodes in the rat frontal cortex (Lou et al., 1987). Animals were placed in a chamber into which hydrogen gas (5-6%) in room air was infused. The gradient of bulk-tissue partial pressure of inhaled hydrogen gas to the electrode surface was detected (hydrogen molecules at the polarized electrode surface are oxidized to permit a measurable current), and its rate of change was recorded (Gould model 2800S, Gould Instruments, Valley View, OH) after cessation of hydrogen infusion. The rate of hydrogen clearance from the tissue served as an indirect measure of local blood flow. Rats were pretreated with methylscopolamine (0.32 mg/kg s.c.) to block peripheral effects, followed by two sequential blood flow measurements taken from each conscious, freely moving rat. After this, milameline (0.32, 1.0, 3.2, and 10.0 mg/kg; n = 6/dose) dissolved in 0.9% saline was given s.c., and measurements were made at 15-min intervals for 120 min. Data were analyzed via ANOVA followed by Newman-Keuls post hoc comparisons.

Gastrointestinal (GI) Motility in Rats. Male Long-Evans rats (Blue Spruce Farms) weighing 175 to 200 g were used. These rats were singly housed under a 12-h light-dark schedule with food and water available ab libitum and were fasted 36 to 48 h before testing with water available ab libitum. Fasted rats were administered 20 red Delrin pellets (Ball Delrin 1/16 inch in diameter, Small Parts, Inc., Miami, FL) by oral gavage 20 min before sacrifice. After sacrifice by CO₂ asphyxiation, their stomach and small intestines were removed and aligned on a calibrated light table. The pellets in the lumen of the intestines and stomach were counted (stomach emptying), and the distance traversed by the lead pellet (intestinal transit) was recorded. Other observable cholinergic side effects occurring between dosing and sacrifice were also noted. Milameline (0.1, 0.32, 1.0, and 3.2 mg/kg; n = 5/dose) or vehicle (0.2% carboxymethylcellulose; n = 5) was administered p.o. 30 min before sacrifice, which was 10 min before pellet injection. Data were analyzed via ANOVA followed by Newman-Keuls post hoc comparisons.

Reversal of Scopolamine-Induced Deficit in Monkeys. Test- and drug-experienced rhesus monkeys (Macaca mulatta, 12 years old, 5.7-11.3 kg, 4 females and 2 males) from in-house stock were used. These monkeys have been used in this task over several years to test the effects of various cholinomimetic compounds. Animals were fed 16 to 20 h before testing and were singly housed in a vivarium adjoining the test room. Water was available ad libitum. Monkeys were transported from the vivarium to the testing chambers (Industrial Acoustics Company, Inc., Bronx, NY) in specially designed cages, which they freely entered and exited. This permitted moving animals without physical restraint and direct human contact. Testing consisted of measuring the number of responses made by monkeys on a microcomputer-controlled continuous-performance task (Callahan et al., 1993). In this task, monkeys were presented a stimulus consisting of a yellow square randomly displayed on a 19-inch color television monitor. The animals were trained on the task by first being rewarded for orienting to the monitor, then for responding to the touch-screen, and finally for responding to the presentation of the stimulus object. Stimulus duration (1, 2, or 4 s) and intertrial interval (1, 2, or 3 s) were randomly determined in a complete block design. Delivery of a 190-mg banana-flavored food pellet and presentation of a tone (ascending series of tones, 500 ms) rewarded correct responses (touching the yellow square). Inaccurate responses were signaled by a tone (700 Hz, 500 ms) but were not rewarded. Test sessions consisted of 150 trials (50 at each stimulus duration). An individually titrated performance-impairing dose of scopolamine (0.003 or 0.006 mg/kg) was injected i.m. 90 min before testing. Doses of scopolamine were minimized to produce a significant decrease, but not abolition, of performance. Milameline (0.001, 0.003, and 0.010 mg/kg) was given i.m. 60 min after scopolamine, which was 30 min before testing. Data were analyzed with ANOVA followed by Newman-Keuls post hoc comparisons.

Monkey Cortical EEG. Two male rhesus monkeys (8 and 18 years old; 9.9 and 9.7 kg, respectively) were used. Monkeys were preanesthetized i.m. with ketamine (10 mg/kg), xylazine (0.6 mg/kg), and atropine (0.05 mg/kg) before intubation with an endotracheal tube. Anesthesia was maintained with halothane in oxygen (1-2%). Animals were placed in a primate stereotaxic apparatus (model 200-25A, David Kopf) and surgically implanted with stainless steel electrodes screwed into the skull surface overlying the frontal and parietal cortex. Coordinates were (from stereotaxic zero): A 22.0 mm, L ± 15.0 mm (frontal), P ± 5.0 mm, L ± 5.0 mm (parietal). All procedures were performed by sterile surgical technique. Electrodes were secured to a plastic connector (model MS-363; Plastic Products) that was permanently attached to the skull surface with dental acrylic. Animals were given 1 month to recover from surgery before testing. Monkeys were chaired in a macaque restrainer (model R001; Primate Products, Inc., Redwood City, CA) and placed into a sound attenuating chamber (Industrial Acoustics) at the end of the light cycle of the monkeys' normal diurnal cycle. The animals were allowed 30 min to habituate to the environment. EEG was recorded continuously for 15 min before and 2 h after drug administration. EEG samples were taken continuously and were converted once each second from the time to frequency domain (Mi² model M100 A-D converter; Modular Instruments) with FFT. The power spectra from these FFTs were summed across 15-min epochs to yield a mean power spectrum for the period. Mean power within the 1- to 4- (), 5- to 7- (), 8- to 13- (), 14- to 21- (1), and 22- to 32- (2) Hz bands was calculated separately for each bandwidth. Total power within all bandwidths also was calculated. Milameline (0.010, 0.017, and 0.032 mg/kg) or vehicle (0.9% saline) was administered i.m. in the right flank. Treatments were given in random order, with no more than two treatments per week and at least 2 days between treatments. Because only two monkeys were used, statistical analyses were not performed.

Side Effects in Monkeys. Rhesus monkeys (8-29 years old, 6.6-14.4 kg, 11 females and 3 males) were divided into two sex- and age-matched groups. The repeated-treatment group (n = 9) received daily i.m. injection of milameline (0.032 mg/kg) for 14 consecutive days. The acute group (n = 5) received no treatment during this period. On the day of testing, both groups received milameline in a rising dose paradigm (0.003-, 0.010-, 0.017-, 0.032-, and 0.056-mg/kg total cumulative dose), with 30 min between doses. Monkeys were observed continuously for salivation and emesis and evaluated subjectively (mild, moderate, severe, or no side effects) by an experimenter blind to the drug histories of the monkeys.

Materials. Milameline was synthesized within the Parke-Davis Department of Chemistry (Ann Arbor, MI). The ligands, [³H]NMS and [³H]myo-inositol were obtained from Amersham. [³H]CMD and [³H]QNB were obtained from New England Nuclear (Boston, MA). Carbachol, forskolin, IBMX, ibotenic acid, scopolamine hydrobromide (scopolamine), and scopolamine methylnitrate (methylscopolamine) were obtained from Sigma Chemical Co. (St. Louis, MO). All doses used in these studies refer to free base.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Receptor-Binding Studies

[³H]CMD and [³H]QNB Binding in Rat Brain Tissue. When using rat brain homogenates and a high-affinity muscarinic agonist ligand (e.g., [³H]CMD), most muscarinic agonists will inhibit binding at low nanomolar concentrations, whereas the use of muscarinic antagonist ligands (e.g., [³H]QNB or [³H]NMS) will yield inhibition constants generally in the micromolar range. This is in contrast to muscarinic antagonists, which inhibit binding in both types of assays in the nanomolar range. The results with milameline were consistent with an agonist profile, with K_i values of 20 ± 4 nM with [³H]CMD and 3060 ± 670 nM with [³H]QNB as the ligand.

[³H]NMS Binding in Transfected CHO Cells. In Table 1, it can be seen that milameline binds with approximately equal affinity to the five subtypes of human muscarinic receptors when using a whole-cell preparation from transfected CHO cells. Similar to the results obtained with [³H]QNB in rat brain tissue, K_i values were in the low micromolar range and ranged from 2.3 ± 0.1 µM (hM₁ receptors) to 4.3 ± 0.4 µM (hM₅ receptors). When using membranes from the CHO cells, the K_i values were about twofold higher than with whole cells, and milameline showed slightly higher affinity for hM₂ receptors compared with the other receptor subtypes. The reason for the affinity shift and higher M₂ affinity is unclear, but it may be the different assay conditions used.

Stimulation of PI Hydrolysis in hM₁, hM₃, and hM₅ CHO Cells and in Rat Cortical Slices. As can be seen in Fig. 2A, milameline was able to increase the production of total [³H]inositol phosphates in both hM₁ and hM₃ CHO cells but produced no measurable effect in hM₅ cells at concentrations of 0.1 to 1000 µM. The maximal effect produced in hM₁ cells was 44% that of carbachol, whereas it was 30% of the carbachol response in hM₃ cells, the response produced by 1 mM carbachol being set to 100%. Concentrations of milameline producing EC₅₀ values for hM₁ and hM₃ cells were 5.4 ± 0.6 and 9.5 ± 0.1 µM, respectively, whereas comparable values for carbachol were 4.7 ± 1.2 and 2.5 ± 0.4 µM.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Stimulation of PI hydrolysis in transfected CHO cells (A) and rat cortical slices (B) by milameline and carbachol. Cells and brain slices were prepared and labeled as described in Experimental Procedures. At concentrations of 0.1 to 1000 µM, both agonists were examined for the ability to increase total [3H]inositol phosphates produced during a 15-min incubation period in CHO cells transfected with hM1, hM3, and hM5 receptors or during a 60-min incubation with rat cortical slices. Results for the CHO cells are expressed in terms of the maximal response produced by carbachol at 1 mM concentration (set to 100%), whereas data for the slices is expressed as percent stimulation. Results shown are representative of one experiment, with each point the mean of triplicate (cells) or quadruplicate (brain slices) samples.

Inhibition of Forskolin-Stimulated cAMP Accumulation in CHO Cells. Forskolin stimulates the production of cAMP by acting directly on adenylyl cyclase, the enzyme responsible for cAMP synthesis. Both milameline and carbachol produced a concentration-dependent inhibition of forskolin-stimulated cAMP accumulation in hM₂ (Fig. 3A) and hM₄ (Fig. 3B) CHO cells. IC₅₀ values for milameline were 15.2 ± 1.2 and 4.2 ± 0.2 µM, whereas those for carbachol were 4.4 ± 0.3 and 6.3 ± 0.6 µM, respectively. The maximal inhibition was ~60% for both agonists in each cell line, and there was no differentiation between full and partial agonists, as was observed in PI hydrolysis.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Inhibition of forskolin-stimulated cAMP accumulation in hM2 (A) and hM4 (B) CHO cells by milameline and carbachol. The ability of milameline and carbachol to block the increase in accumulation of cAMP produced by 5 µM forskolin during a 2.5-min incubation was examined in transfected CHO cells. Cells were prepared as described in Experimental Procedures. Agonists were added 5 min before the addition of forskolin, and total amounts of cAMP accumulated in the presence of 1 mM IBMX were measured. Results shown are representative of one experiment, with each point the mean of three or four samples.

Decreased Release of [³H]ACh from Rat Cortical Slices. Muscarinic agonists have been shown to decrease K⁺-stimulated [³H]ACh release from rat brain tissue by activating presynaptic autoreceptors of the M₂ subtype (Raiteri et al., 1989). Milameline, like the reference muscarinic agonist arecoline, was able to decrease [³H]ACh release in a concentration-dependent manner, with a maximal inhibition of ~50% at 100 µM (Fig. 4A). These decreases in release were inhibited by the muscarinic antagonist scopolamine (10 µM), demonstrating that the effect of both agonists was mediated through activation of muscarinic receptors (Fig. 4B).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Inhibition of K+-stimulated [3H]ACh release from rat cortical slices by milameline and arecoline (A) and reversal by scopolamine (B). The ability of milameline and arecoline to decrease the release of [3H]ACh produced by 20 mM K+ during a 15-min incubation period was examined with slices produced from rat cerebral cortex (minus olfactory bulbs). The muscarinic antagonist scopolamine (10 µM) was used to reverse the inhibition of 20 mM K+-stimulated [3H]ACh produced by milameline and arecoline (both at 10 µM). Slices were prepared as described in Experimental Procedures. Each point represents the mean and S.E. of two or three separate experiments. *p < .01 versus control; **p < .01 versus drug alone; Dunnett's test.

Inhibition of Rat Brain AChE. Because it is possible that any cholinergic-mediated behavioral effects observed in animals could be the result of either direct receptor activation or inhibition of AChE, milameline was examined for the ability to inhibit AChE activity. Milameline showed no significant inhibition of rat brain AChE up to a concentration of 100 µM. For comparison, under the same experimental conditions, tacrine and physostigmine were shown to produce concentration-dependent effects, with IC₅₀ values of 0.20 and 0.13 µM, respectively (data not shown).

Spontaneous and Scopolamine-Induced Swimming Activity in Rats. Vehicle-treated rats swam 40 to 50 m during a 5-min test session. Milameline dose dependently decreased spontaneous swimming activity (Fig. 5). This decrease was statistically significant [F(4,20) = 13.03, p < .01] at an oral dose of 3.2 mg/kg. Scopolamine (0.32 mg/kg i.p.) consistently and significantly increased total distance traveled by >20 m, with rats displaying stereotyped, thigmotaxic (wall-hugging) swimming patterns with infrequent, spontaneous redirections of movement. This differed markedly from vehicle-treated animals. Excessive and stereotyped locomotor activity was similarly observed in scopolamine-treated rats in open-field testing (Sanberg et al., 1987). Milameline at doses of 3.2 mg/kg significantly [F(4,20) = 45.80, p < .01] reversed the increase in swimming activity produced by scopolamine (Fig. 5).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Spontaneous swimming activity was decreased in rats administered milameline p.o. 35 min before testing, with the distance traveled being recorded continuously for 5 min after drug treatment with a videotracking system. Scopolamine (0.32 mg/kg i.p.) administered 15 min before testing increased swimming. The scopolamine-induced increases were reduced by milameline administered s.c. 30 min before testing. Each bar represents the mean ± S.E. from groups of five drug and test naive animals. *p < .05 versus vehicle 0.0 dose; **p < .01 versus scopolamine; ##p < .01 versus vehicle 0.0 dose; ANOVA followed by Newman-Keuls test.

Effects on Spatial Memory in Basal Forebrain-Lesioned Rats. Lesions of the basal forebrain in rats produce cortical cholinergic hypofunction and impair behavioral performance (Araki et al., 1986). These lesions mimic the cholinergic cell loss and resultant cortical hypofunction accompanying AD. In our studies, basal forebrain-lesioned rats receiving daily injections of vehicle acquire a spatial water-maze task more slowly than similarly treated sham controls. Milameline (0.1 and 0.32 mg/kg) given daily improved the performance of lesioned rats in a water maze (Fig. 6). The improved performance at the 0.32-mg/kg dose was statistically significant on trials 3 [F(3,80) = 3.16, p < .05] and 4 [F(3,80) = 3.56, p < .05]. Sham-operated rats often swam directly to the platform, whereas lesioned rats often swam around the periphery of the pool, resembling the stereotypical swimming patterns of scopolamine-treated rats. Lesioned rats treated daily with milameline acquire this task in a manner similar to the sham-operated rats.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Basal forebrain-lesioned rats administered milameline s.c. 30 min before testing improved their ability to acquire a spatial water-maze task. Animals were tested in a single daily trial 5 weeks after bilateral injections of ibotenic acid or vehicle (sham lesion) at a region encompassing the basal forebrain cholinergic neurons. Lesioned animals given 0.32 mg/kg of milameline acquired this task at a rate equivalent to sham-lesioned rats. Each point represents the mean ± S.E. for 21 animals. *p < .05 versus vehicle 0.0-dosed lesioned rats; ANOVA followed by Newman-Keuls test.

Rat Cortical EEG Activity. Cholinomimetics produce characteristic changes in cortical EEG activity recorded from the skull surface of rodents (Yamamoto, 1988; Danober et al., 1993). These agents can shift the EEG record from one dominated by slow, sleeplike activity to a record dominated by low-voltage desynchronized activity. Milameline produced a similar pattern of EEG activity at doses of 1.0 mg/kg in rats consistent with the action of a central cholinomimetic. Qualitatively, the EEG records of milameline-treated rats were dominated by low-voltage desynchronized activity with little or no discernible synchronized slow activity. This pattern is reflected quantitatively as a dose- and time-dependent decrease [F(1,18)=36.08, p < .01] in total power in the EEG spectrum relative to the vehicle control (Fig. 7). Decreased total power can be attributed to a large decrease in power in the slower EEG bandwidths ( and ) and to a lesser extent in the faster bandwidths. However, the profile of the changes for individual bandwidths follows that shown for total power in Fig. 7. This EEG activity pattern has been associated with an alert or awake state.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Total power in the rat cortical EEG was decreased by p.o. doses of milameline (, 0.0 mg/kg; , 1.0 mg/kg; , 3.2 mg/kg; , 10.0 mg/kg). EEG activity was recorded continuously for 120 min, and total power was summed across 15-min intervals. Comparisons across the means were performed at 30-, 60-, 90-, and 120-min time points. Significant decreases in total power were evident at all three doses Activity persisted throughout the test session at the 10.0-mg/kg dose. Each point represents the mean ± S.E. of seven animals. *p < .05 versus vehicle 0.0 dose; nonparametric ranked ANOVA (Kruskal-Wallis rank transformed analysis with repeated measures) with comparisons among means by Duncan's post hoc tests.

Rat Core Body Temperature. Reduction in body temperature after cholinomimetic treatment is thought to arise from central activation of brain stem cholinergic neurons (Spencer, 1965). Milameline dose dependently decreased body temperature in rats, reaching statistical significance [F(4,40) = 4.82, p < .01] at 0.32 mg/kg (Fig. 8). This characteristic decrease in body temperature produced by milameline is similar to that for other muscarinic agonists in rodents (Palacios et al., 1986).

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Core body temperature was dose dependently decreased in rats administered milameline p.o. 30 min before measurement with a rectal probe. Each point represents the mean ± S.E. of 5 to 15 animals at each treatment. Local cortical blood flow, estimated with the hydrogen clearance technique from platinum electrodes, was increased in rats when measured 15 to 30 min after s.c. doses of milameline. Duration of action was dependent on the dose administered, persisting beyond the 2-h test session after the 10.0-mg/kg dose. Each point represents the mean increase in blood flow of six animals pretreated with methylscopolamine (0.32 mg/kg i.p.). *p < .05 versus vehicle 0.0 dose: ANOVA followed by Newman-Keuls test.

Local Cortical Blood Flow in Rats. Measurement of local cortical blood flow provides an indirect index of cortical neuronal activity (Lou et al., 1987). Milameline significantly increased cortical blood flow in the frontal cortex of rats pretreated with methylscopolamine to block peripheral cholinergic actions (Fig. 8). This increase was statistically significant [F(4,25) = 10.08, p < .01] at milameline doses of 1.0 mg/kg s.c. Duration of action was dependent on the dose administered, persisting beyond the 2-h test session after a 10.0-mg/kg dose.

GI Motility in Rats. Milameline dose dependently increased both stomach emptying [F(4,20) = 15.81, p < .01] and intestinal transit [F(4,20) = 8.88, p < .01], as shown in Fig. 9. This action appears to affect smooth muscle contractility in the rat GI tract. In addition, signs of excessive peripheral cholinergic stimulation (salivation) were noted in rats at the 3.2-mg/kg dose.

View larger version (18K): [in this window] [in a new window]   Fig. 9.   GI motility was increased in fasted rats administered milameline. Milameline was administered p.o. 30 min before sacrifice, which was 10 min before gavage injection of 20 Delrin plastic pellets. After sacrifice, stomachs and intestines were removed, and the distance pellets traveled through the intestines were measured. This provided two measures of GI motility: stomach emptying (percentage of pellets leaving the stomach) and intestinal transit (distance traveled by the lead pellet). Each point represents the mean ± S.E. of five animals. **p < .01 versus vehicle 0.0 dose; ANOVA followed by Newman-Keuls test.

Reversal of Scopolamine-Induced Deficit in Monkeys. In monkeys, scopolamine significantly decreased the number of correct responses recorded on a continuous-performance task, due to omissions and not inaccuracy of response. Inaccurate responses were rare and did not increase after any treatment. Milameline reversed this impaired performance and was statistically significant when the "best dose" (0.003 or 0.010 mg/kg) was retested for each monkey (Fig. 10). This best dose of milameline in each monkey correlated with the maximum tolerated dose free of side effects.

View larger version (15K): [in this window] [in a new window]   Fig. 10.   Scopolamine (0.003-0.006 mg/kg) administered i.m. 90 min before testing impaired the number of responses by rhesus monkeys on a continuous-performance task. This scopolamine-induced impairment was reversed by milameline administered i.m. 30 min before testing. The "best dose" of milameline (0.003 or 0.01 mg/kg) in each monkey correlated with the individual maximum-tolerated dose that was free of side effects. Each bar represents the mean ± S.E. from six animals. *p < .05 versus scop; ANOVA followed by Newman-Keuls test.

Monkey Cortical EEG. Qualitatively, the EEG records of vehicle-treated monkeys were characterized by high-voltage slow activity consistent with the monkeys being asleep or quiescent. Milameline administration produced EEG records dominated by low-voltage desynchronized activity that would be expected of monkeys awake and alert. This qualitative assessment was confirmed by the power spectral analyses. Milameline (0.001-0.010 mg/kg) dose dependently produced a decrease in total power that persisted throughout the 2-h test session at the highest doses (data not shown, similar to rat data in Fig. 7). This decrease in total power can be attributed to decreases in slow wave ( and  bandwidths) activity, without a concomitant decrease in fast activity.

Side Effects in Monkeys. Milameline in monkeys produced excessive salivation and emesis (Table 2). Other signs of excessive cholinergic stimulation were not observed at the doses tested. No relationship between sex and age of monkeys and the occurrence of cholinergic side effects were obvious in either the repeated- or acute-treatment group. At 0.01 mg/kg, milameline induced salivation in one monkey in each group and affected progressively more animals with increasing severity as the dose was increased. Emesis first appeared in monkeys at 0.017 mg/kg and also increased in severity and prevalence as the dose was increased. In the acute-treatment group, two monkeys salivated excessively and four showed moderate to severe emesis at 0.032 mg/kg. Because of the severity of the side effects, only two of the five monkeys from the acute-treatment group received the 0.056-mg/kg dose. Of the two, one showed moderate emesis, whereas the other completed the regimen side-effect free. In the repeated-treatment group, the side effects were moderate, with no severe emesis, and all of the monkeys received every dose of milameline. Three of these monkeys completed the regimen side-effect free.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Milameline is a novel muscarinic agonist that overcomes many of the major limitations of first-generation agonists (Schwarz et al., 1991). It is an orally active agent with potent central nervous system (CNS) penetration, good bioavailabilty, and a reasonable duration of action. In vitro, milameline binds to muscarinic receptors in rat brain tissue at micromolar concentrations when using an antagonist ligand and nanomolar concentrations with an agonist ligand. In transformed CHO cells expressing human muscarinic receptors, milameline displays equal affinity at the five subtypes of receptors. Using functional assays in the CHO cells, milameline shows partial agonist activity, with PI hydrolysis being stimulated to ~30 to 45% of the full agonist carbachol in hM₁ and hM₃ cells (no effect was observed in hM₅ cells). It also inhibits forskolin-stimulated cAMP accumulation in hM₂ and hM₄ cells. Like other muscarinic agonists, this compound decreases the release of [³H]ACh from rat brain slices, showing an interaction with putative M₂ presynaptic autoreceptors. All functional responses are blocked by muscarinic antagonists, such as scopolamine or atropine, showing that the responses are due to activation of muscarinic receptors. Additionally, milameline did not inhibit rat brain AChE. Thus, milameline has all the biochemical characteristics expected from a partial muscarinic agonist in the assays used.

Little or no binding was observed to several other CNS neurotransmitter receptors. With NovaScreen evaluation, no significant binding was shown to occur at the following sites: adenosine A₁ and A₂; ₁-, ₂-, and -adrenergic receptors; -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; benzodiazepine; calcium L-channel; dopamine D₁ and D_2; forskolin; -aminobutyric acid; glutamate; glycine; histamine H₁; kainate; leukotrienes B₄ and D₄; nicotinic receptors; N-methyl-D-aspartate; µ, , and  opiates; phorbol ester; 1-(1-phenylcyclohexyl)piperidine; 5-hydroxytryptamine 1 and 2; and thromboxane A₂. Thus, with regard to the sites examined, milameline appears to be selective for muscarinic receptors.

Milameline possesses in vivo central cholinomimetic activity at doses lower than those eliciting unwanted cholinergic side effects. In rats, milameline decreased body temperature and significantly improved behavioral performance of basal forebrain-lesioned rats at a dose of 0.32 mg/kg. It also decreased swimming activity and increased cortical EEG arousal and blood flow in rats at a dose of 1.0 mg/kg. Significant cholinergic side effects were not seen in rats until doses of 3.2 mg/kg, with the exception of enhanced GI motility, which first appeared at a dose of 1.0 mg/kg. In monkeys, milameline improved behavioral performance impaired by scopolamine at doses of 0.003 to 0.01 mg/kg and induced an EEG arousal reflecting increased central cholinergic tone at 0.01 mg/kg. Doses >0.01 mg/kg produced increased salivation and emesis in monkeys, but these side effects appeared to tolerate with repeated administration.

Numerous studies show positive effects of cholinomimetics on behavioral deficits in animals induced by the muscarinic receptor antagonist scopolamine (M'Harzi et al., 1995; Rupniak et al., 1989). Scopolamine, a peripheral- and central-acting anticholinergic (but not methylscopolamine, a mainly peripheral-acting anticholinergic) impairs performance of rats in a radial-arm maze (Cassel and Kelche, 1989). Similar effects were observed in swimming activity of rats in a water maze (Saucier et al., 1996). In monkeys, sustained attention was shown to be impaired through blockade of central muscarinic receptors by directly infusing scopolamine into the lateral ventricle (Callahan et al., 1993). Because the ability of a muscarinic agonist to reduce the effects of scopolamine is mediated centrally, it would appear that milameline is a centrally acting cholinomimetic in the whole animal.

Patients with AD show a significant loss of neurons (up to 90%) in the nucleus basalis of Meynert, a major source of cholinergic innervation of the cerebral cortex (Whitehouse et al., 1981). Lesions of the nucleus basalis in rats produce cortical cholinergic hypofunction and impair behavioral performance (Smith, 1988). Similar to the impaired performance in water maze, lesioned rats show severe deficits in retention on an operant-delayed matching task (Dunnett, 1985) and impaired acquisition of a conditioned avoidance response in a two-way shuttle box (Araki et al., 1986). In this study, milameline reversed the spatial memory deficit in lesioned rats. To the extent that deficits in cholinergic function impair behavioral performance, these data support the potential therapeutic utility of milameline for improving cognitive deficits arising from cholinergic hypofunction, as occurs in AD.

In AD, slowing of the EEG, disproportional to the changes seen during normal aging, has been observed (Soininen et al., 1982). EEG changes in AD are associated with diminished or absent , and increased  and  activities. Experimental evidence suggests that the cholinergic nucleus basalis plays a pivotal role in neocortical EEG activation (Riekkinen et al., 1991). The marked decrease in the number of nucleus basalis neurons and the slowing of EEG activity in patients with AD implies the cholinergic system is involved in the regulation of EEG activity (Bird et al., 1983). Furthermore, EEG changes have been correlated with Mini-Mental State scores (Primavera et al., 1990). Treatment of AD patients with cholinergic drugs that improve memory and attention shift the EEG spectrum into more normal patterns, whereas the anticholinergic drugs have the opposite effect (Agnoli et al., 1983). The effects of milameline on EEG activity in rats and monkeys are consistent with the pattern of activity seen with centrally acting cholinomimetics.

Several first-generation muscarinic agonists have been evaluated in AD trials, with mixed clinical results. Intravenous infusions of arecoline have been shown to produce positive cognitive effects in small groups of AD patients. However, its short half-life made it clinically unacceptable (Christie et al., 1981; Soncrant et al., 1993). Although preliminary observation of AD patients treated with intraventricular infusions of bethanechol suggested clinical improvement, the results from subsequent studies were more equivocal (Harbaugh et al., 1984; Gauthier et al., 1986). Out of seven AD patients given oral oxotremorine, five showed unexpected depressive reactions, and no conclusions on the cognitive performance of this agonist were reached (Davis et al., 1987). Oral doses of pilocarpine, with and without lecithin, produced no positive effects in four AD patients (Caine, 1980). Almost 500 AD patients have received RS-86, a spiro-piperidyl compound with cholinergic activity in animals. Although some individuals showed positive responses, the majority either showed no effect or were unable to continue dosing because of troublesome cholinergic side effects (Wettstein and Spiegel, 1984; Hollander et al., 1987). The ability to overcome the limitations observed with these agonists may allow newer-generation compounds, such as milameline, to show positive clinical results in AD.

More recently, it has been suggested that M₁ subtype-selective muscarinic receptor agonists should have a more desirable therapeutic profile for the treatment of AD than nonselective agonists (Jaen and Schwarz, 1997). Activation of M₁ receptors would target receptors in the cortex and hippocampus, which are intimately involved in memory and learning, while avoiding those receptors (presumably M₂ and M₃ receptors) responsible for peripheral cholinergic side effects (e.g., salivation, nausea etc.). Although milameline is not subtype selective, it does show modest separation between doses active in certain cognitive models versus those producing dose-limiting side effects. It remains unknown, however, whether these results are predictive for humans.

Nicotinic receptor agonists may also have potential for the symptomatic treatment of AD. Activation of nicotinic receptors has been shown to increase attention and improve learning and memory in both animals and humans. Because AChEIs have been found to be clinically active in AD, it may be that activation of nicotinic receptors by increased levels of ACh is an important component to the overall clinical effect. More recently, the identification of nicotinic receptor subtypes based on subunit composition has driven the identification of CNS selective agonists for the treatment of AD (Decker and Brioni, 1997).

In phase I trials conducted in normal healthy subjects, milameline was well tolerated at single and multiple (q 6 h) doses up to 1.0 mg. Peripheral cholinergic signs (e.g., sweating, hypersalivation, and increased urinary frequency) were observed at 0.5 and 1.0 mg. Dose-limiting GI signs (e.g., stomach pains, emesis) were found at 2-mg/q 6 h dose, and all adverse effects resolved when milameline was discontinued. In aged, healthy volunteers and patients with probable AD, gradual dose escalation extended the tolerated dose to 3.0 mg (Sedman et al., 1995).

Certainly, the goal for developing clinically meaningful treatments for AD is to halt or delay the progression of the disease. Cholinergic drugs, AChEIs, muscarinic agonists, and nicotinic agonists, appear to be palliative treatments at best, although there is some interesting in vitro evidence that they alter amyloid precursor protein processing (Roberson and Harrell, 1997). Whereas there is much hope that targeting amyloid processing and/or neurofibrillary tangle formation will achieve the targeted goal, it is doubtful that agents with these mechanisms will be approved and marketed in the next 2 to 5 years. Thus, obtaining preclinical and clinical results with novel muscarinic agonists will aid in creating more therapeutic choices for AD patients in the short term.