Seven months after failing a Phase IIa trial in bipolar disorder, Memory Pharmaceuticals Corp.'s MEM 1003 also missed in a Phase IIa trial in Alzheimer's disease.

Shares of the Montvale, N.J.-based company (NASDAQ:MEMY) fell 69 cents, or 39.4 percent, to close at $1.06 on Monday.

Memory discontinued its work with calcium channel modulator MEM 1003 in bipolar disorder after the previous Phase IIa trial missed its primary and secondary endpoints. Yet President and CEO James Sulat said the current Phase IIa data in Alzheimer's disease deserve a closer look before any decisions are made regarding MEM 1003's future in that indication. (See BioWorld Today, March 6, 2007.)

The double-blind, placebo-controlled Phase IIa study enrolled 183 patients with mild to moderate Alzheimer's disease. The trial was initially designed as a monotherapy study but, due to enrollment difficulties, an adjuvant arm was added to include patients on stable doses of the cholinesterase inhibitors Aricept (donepezil, Pfizer Inc.), Exelon (rivastigmine, Novartis AG) or Razadyne (galantamine, Ortho-McNeil Neurologics Inc.). Patients in both the monotherapy and adjuvant arms were randomized to receive twice-daily doses of MEM 1003 at 30 mg, MEM 1003 at 90 mg or placebo for 12 weeks.

Overall, the trial missed its primary endpoint of improving the Alzheimer's disease Assessment Scale - cognitive subscale (ADAS-cog) score after 12 weeks. During a conference call, Memory's Chief Medical Officer Stephen Murray said that may be at least partially attributable to a "striking and frankly quite unexpected" placebo effect.

In the overall trial population, patients receiving placebo demonstrated a 3.2 point improvement in ADAS-cog scores, while those receiving low-dose MEM 1003 showed a 2.3 point improvement and those receiving high-dose MEM 1003 showed a 1.8 point improvement. In comparison, the average ADAS-cog improvement with existing Alzheimer's drugs is about 2.7 points.

Breaking the placebo data down by treatment arm showed that placebo patients in the monotherapy group had a whopping 5 point improvement in ADAS-cog scores, while placebo patients in the adjuvant arm had a slight worsening in scores. The results seen in the adjuvant group were "more typical of the results seen in previous Alzheimer's disease trials," Sulat said.

Teasing out what caused such a difference between the monotherapy and adjuvant arms will be a "key question" in subsequent data analyses, Murray said. He added that the initial analysis did not reveal any obvious differences other than the fact that patients in the monotherapy arm were slightly younger, but he noted that enrollment difficulties in the monotherapy arm could have resulted in an "unusual" patient population.

Although the overall trial failed, MEM 1003 did show positive trends in the adjuvant arm, both in the primary endpoint and in all four secondary endpoints. In that subgroup, ADAS-cog scores improved by 2 points for patients receiving low-dose MEM 1003 and by 1 point for patients receiving high-dose MEM 1003. Those improvements were not statistically significant due to the small sample size, and Murray suggested not reading too much into the apparent inverse dose-response for the same reason.

MEM 1003 was generally well-tolerated, with similar adverse event rates occurring in all arms of the study. Five MEM 1003-treated patients had serious adverse events, including two deaths, but none were deemed to be treatment-related.

Moving forward, Memory intends to conduct a thorough analysis of the trial data before making any decisions regarding the future of MEM 1003.

Elsewhere in its pipeline, Memory expects data next month from another Phase IIa Alzheimer's trial, this time with nicotinic alpha-7 agonist MEM 3454. Murray said the company believes the Alzheimer's patients enrolled in the trial are sufficiently different from the MEM 1003 trial that "we will not see the same placebo response phenomenon."

MEM 3454 and the rest of Memory's nicotinic alpha-7 agonist program are partnered with F. Hoffmann-La Roche Ltd. The deal calls for Memory to lead development through Phase IIa, with Roche holding a co-development option. (See BioWorld Today, Sept. 19, 2003.)

Earlier in development, Memory has a PDE10 inhibitor program for neurological and psychiatric disorders partnered with Amgen Inc. and a 5HT6 antagonist program for various neurology indications. The company also recently regained all rights from Roche to a PDE4 inhibitor program for Alzheimer's and other central nervous system disorders. (See BioWorld Today, Oct. 18, 2005, and June 20, 2007.)