7 April 2008. The cholinergic hypothesis of Alzheimer disease tends to elicit a yawn from scientists who reserve their excitement for anti-amyloid treatments or more radically new biologic approaches such as immunotherapy. Still, it’s what has yielded most AD drugs the doctor can prescribe today, and researchers clearly feel there is more to be gained from it. Even as amyloid and tau pathologies are considered central to the pathogenic process, the cholinergic system is indispensable for learning, attention, and information processing. At the Keystone conference, held 24-29 March in Keystone, Colorado, two speakers—one from a small biotech, one from pharma giant Merck—presented their active approaches of propping up a failing cholinergic system in AD. One targets a serotonin, one a muscarinic acetylcholine receptor.
First, the more clinically advanced drug. J. Thomas Megerian of Epix Pharmaceuticals in Lexington, Massachusetts, updated the audience on PRX-03140, a 5-HT4 serotonin receptor agonist his company is developing together with GlaxoSmithKline. Epix’s claim to fame is its in-silico drug design technology, which has generated some potential drugs against G protein-coupled receptors (GPCRs). This large family of transmembrane signaling proteins is thought to contain targets for a range of diseases, but they have been difficult to exploit in part because no crystal structures are available to guide drug discovery. The target of PRX-03140 is one such GPCR, i.e., the 5-HT4 receptor.
According to Megerian, preclinical studies have suggested that this small-molecule agonist boosts cholinergic transmission by increasing levels of the acetylcholine receptor on demand, that it stimulates release of the growth factor BDNF, decreases Aβ levels, and promotes α-secretase cleavage of APP to generate the neurotrophic sAPPα fragment. Epix is attempting to develop PRX-03140 either alone or as a combination therapy with donepezil, which increases acetylcholine levels in the resting state by simply blocking its degradation.
Last December, Epix released top-line results of their recently concluded Phase 2a study but soon after issued a correction when errors in the initial data analysis emerged. At Keystone, Megerian presented these data: the trial recruited 80 people with mild AD at 17 U.S. sites and treated them for two weeks with PRX-03140. Some patients took donepezil for three months and then added one of five different doses (5 to 200 mg/day) of PRX-03140. Others took one of two doses (50 or 150 mg) of study drug but no donepezil. The trial measured safety, exposure to donepezil, and cognitive endpoints.
PRX-03140 caused no serious side effects in monotherapy, but the higher doses of combination therapy showed expected cholinergic side effects, mostly gastrointestinal. The drug did not alter drug exposure of donepezil. Patients on the high dose of monotherapy improved a statistically significant 3.6 points over baseline on the ADAS-cog battery of tests and 4.5 points over placebo. The dose-response effect between placebo, 50 mg, and 150 mg was also statistically significant. These effect sizes are within the range of modest improvement typically seen in cholinesterase inhibitor trials. The difference Megerian pointed to is that their drug achieves this effect after two weeks, whereas the cholinesterase inhibitors can take months to do so. The combination therapy showed no signal on ADAS-cog. On a commercial, computerized cognitive assessment (called Mindstreams, by NeuroTrax), patients on both types of therapy showed improvement on either the spatial or memory index scores, Megerian added. The company also uses EEG, measuring the ratio of alpha:theta waves, and reports seeing a signal on the high-dose monotherapy.
Based on these results, Epix is planning two larger Phase 2b trials. A three-month monotherapy trial will compare PRX-03140 to donepezil and offer a three-month extension to people who get randomized to placebo; a six-month combination trial will test PRX-03140 added onto a stable dose of donepezil.
Presented by William James Ray of Merck and Co. in West Point, Pennsylvania, a different way of tapping the cholinergic system for future therapies drew wide praise from other meeting attendees. It targets the M1 muscarinic receptor, one of five GPCRs that constitute the metabotropic acetylcholine receptors. Abundant in the hippocampus and neocortex, the M1 receptor mediates γ oscillations in hippocampal networks that are thought to underlie memory (Fisahn et al., 2002). Researchers have long known that stimulating M1 can be effective in AD, but previous trials had to be abandoned because the drugs were insufficiently selective for M1 and produced intolerable cholinergic side effects.
Ray said that his group became excited about M1 activation when Frank LaFerla at the University of California, Irvine, reported that AF267B, the M1 agonist developed by Abraham Fisher at the Israel Institute for Biological Research in Ness-Ziona, removed plaque pathology, selectively reduced Aβ42 levels, and shifted APP processing toward α-cleavage in triple transgenic mice (see ARF related news story; Caccamo et al., 2006). Ray said these data made his group wonder whether the cholinergic system in the brain normally functions to repress accumulation of Aβ42. If this were true, a drug activating the M1 receptor would be a “home run,” Ray said, because it would combine the known symptomatic boost of drugs such as donepezil with a new disease-modifying effect on amyloid. (See also McLaurin Keystone story on links between cholinergic system and Aβ.)
A challenge to making this work is to ensure the drug is truly selective for the M1 receptor, Ray said. This is difficult to achieve with an agonist that binds the acetylcholine site of the receptor, because that site is highly conserved among all types of muscarinic receptor. For this reason, the scientists instead tried to find a drug that binds the M1 receptor in some other pocket where this receptor is distinctive from the M2 to M5 receptors. Such allosteric binding could tweak the M1 receptor and stimulate it. A subsequently developed allosteric potentiation assay identified one compound, called benyzl quinolone carboxylic acid (BQCA). It left unaffected not only the other muscarinic receptors, but also 300 human receptors and enzymes that are routinely screened against new compounds of interest, Ray said.
BQCA binds to a pocket in the extracellular domain of M1, far away from its acetylcholine site. The compound lowers the energy required for the receptor to adopt its active conformation. This effectively reduces the concentration of acetylcholine needed for signaling, and in this way could sensitize the hippocampus to the transmitter’s dwindling supply in AD, Ray said. BQCA meets some requirements for a drug: it can be taken orally, crosses the blood-brain barrier, and is not metabolized too rapidly in the liver. BQCA caused no cholinergic side effects at therapeutic doses in rodents, ferrets, dogs, and monkeys, and it has some functional data in mouse behavioral tests under its belt. Even so, more preclinical work yet needs to be done, particularly on defining the relationship between M1 activation and Aβ42 levels.
In addition to these efforts, a number of biotech and pharmaceutical companies are developing agonists for the α7 nicotinic acetylcholine receptor, but none presented at this conference.—Gabrielle Strobel.