ARTICLE IN BRIEF
Myriad Pharmaceuticals is halting further development of tarenflurbil (flurizan) for Alzheimer disease after a large multicenter phase 3 study showed that it did not alter cognitive function in patients with mild disease.
Tarenflurbil (flurizan) — a nonsteroidal anti-inflammatory drug (NSAID) — was considered a hopeful therapy for Alzheimer disease (AD) based on lab and animal tests that found it modulated amyloid beta (Abeta). But the therapy will now be shelved for AD after a large multicenter phase 3 study, reported Dec. 16, 2009, in the Journal of the American Medical Association, showed that it did not alter cognitive function in patients with mild disease.
“We were sad that the drug didn't work but it was a valuable study,” said Robert C. Green, MD, MPH, an AD researcher and clinician at Boston University School of Medicine. Tarenflurbil inhibits gamma secretase, an enzyme involved in amyloid-beta (Abeta) synthesis. The hope was that it would modulate the production of Abeta and slow the progressive of disease. (For more about this mechanism, see “The Role of Gamma-Secretase in Amyloid-Beta Modulation.”)
The study made it through phase 2 of drug testing in mild to moderate AD patients but it didn't show a benefit. When the investigators separated mild patients from those with more advanced disease there was a slight signal that it may have benefited those in the earliest clinical stages. This study puts that hope to rest.
Sponsored by Myriad Pharmaceuticals of Salt Lake City, the trial was the largest study of an experimental drug for the treatment of AD with more than 1,600 patients and involving 133 sites from all corners of the country.
Dr. Green said that it was important to move forward with the phase 3 study to “see whether it had an effect on the disease process and if it did, to determine what would be considered a useful benefit.” Patients with mild AD in the phase 2 study had a dose-related slower rate of decline on cognitive tests. The study was much smaller, around 200 patients. A small subset had mild disease. “There is always a caveat that you will be finding something due to chance,” Dr. Green said.
STUDY PROTOCOLS, RESULTS
The 1,649 patients in the current study were allowed to continue taking the standard AD fare, donepezil (Aricept) or memantine (Namenda). Study patients received either 800 mg twice a day for 18-months or placebo pills. Every three months they would return to the clinic for neurocognitive testing. At the end of the study, the two groups were compared. The primary endpoint was a positive change on the Alzheimer's Disease Assessment Scale (ADAS-Cog) and an improvement in functional ability as measured by the Alzheimer's Disease Cooperative Study activities of daily living. There were also a number of secondary outcome measures.
But by the end of the 18months, patients on placebo or tarenflurbil lost an average of seven points on the ADAS-Cog test. And other cognitive measures continued to drop. The study investigators debated long and hard over what would constitute an improvement for a disease-modifying drug. “It could have been a little effect (a point or two on ADAS-Cog) if it lasted for a few years,” said Dr. Green.
It's not clear why the drug works to lower Abeta in the test tube and animal models but not in patients, he said. It could be a dose effect, or it might be that it is already too late to make a difference.
“I'm very disappointed that the glaring flaw in the tarenflurbil trial hasn't been more prominently reported,” said Samuel E. Gandy, MD, PhD, the Mount Sinai Professor of Alzheimer's Disease Research and associate director of Mount Sinai's Alzheimer's Disease Research Center in New York City. He noted that Douglas R. Galasko, MD, a professor in residence at the University of California-San Diego, showed in a 2007 study that the tarenflurbil dose used in the study didn't change CSF levels of Abeta 42.
The idea behind the growing number of anti-amyloid medicines in development is that they rid the brain of the plaques and thus the protein would increase in levels in the CSF as it makes its way out of the body.
“To say that the flurizan trial is inconclusive in an understatement,” Dr. Gandy said. “One cannot draw any conclusions about the role of amyloid in Alzheimer disease unless there are direct data to show whether amyloid burden increased, decreased, or remained unchanged during an anti-amyloid drug trial. The only data that we have, those of Galasko, prove that the drug probably never changed the brain or CSF Abeta 42 one iota.”
Other NSAIDs have worked in the lab but showed no benefit in patients. Tarenflurbil, at least in the lab, slows the production of Abeta-42 proteins, Dr. Gandy said. It is the mirror-image molecule of ibuprofen. Its structure seems to block the standard side effects of the NSAIDs.
It is too early to know whether anti-amyloid agents would have to be given years before even the first clinical signs. There is mounting evidence that the pathological disease begins much earlier than the first signs of faulty memory, Dr. Gandy said.
Myriad Pharmaceuticals is stopping its development of the drug. There are other gamma-secretase inhibitors being tested. It will take time, Dr. Gandy said, to figure out how these medicines will ultimately help in stalling or stopping
THE ROLE OF GAMMA SECRETASE IN AMYLOID-BETA MODULATION
The amyloid-beta (Abeta) peptide of Alzheimer disease (AD) is 40 to 42 amino acids long embedded inside the sequence of a precursor that is nearly 800 amino acids long. In order to release Abeta, the precursor must be cut in two places. The “molecular scissors” that make the “right hand” or carboxyl terminal cut is an enzyme or protease known as gamma secretase.
When gamma secretase cuts after amino acid 40, a short, fairly inert form of Abeta is generated (Abeta 40). However, if gamma secretase cuts after amino acid 42 (Abeta 42), the Abeta is very sticky and clumps together to begin the formation of toxic substances known as oligomers.
Gamma secretase inhibitors turn off both the 40 cut and the 42 cut so levels of all forms of Abeta go down. Gamma secretase modulators aim selectively for the 42 cut, so only levels of the sticky Abeta 42 go down. Either strategy will lower oligomer concentration and ultimately would be predicted to be beneficial, according to Samuel E. Gandy, MD, PhD, the Mount Sinai Professor of Alzheimer's Disease Research and associate director of Mount Sinai's Alzheimer's Disease Research Center.
• Green RC, Schneider LS, Zavitz KH, et al, for the Tarenflurbil Phase 3 Study Group. Effect of tarenflurbil on cognitive decline and activities of daily living in patients with mild Alzheimer disease: A randomized controlled trial. JAMA 2009;302(23):2557–2564.
• Galasko DR, Graff-Radford N, Koo EH, et al. Safety, tolerability, pharmacokinetics, and Abeta levels after short-term administration of R-flurbiprofen in healthy elderly individuals. Alzheimer Dis Assoc Disord 2007;21(4):292–299.