With another failure in Alzheimer’s Disease, is it finally time to admit we have held on to the amyloid hypothesis for too long?
It seems like every couple of months we sit down to write the same blog post. Just last September, we wrote one entitled “Another One Bites the Dust in Alzheimer’s Disease” after intepirdine’s demise. Maybe this one should be called “And Another One Down,” because even though it’s a new year and a completely different drug, the fundamental story hasn’t changed: another Phase III drug has failed in Alzheimer’s Disease (AD). Last year, intepirdine was the last man standing in the symptomatic 5HT class (Pfizer pulled one, Lundbeck’s failed), so the writing was more or less on the wall. In this case, Merck’s BACE inhibitor verubecestat was our first real look at this potentially disease-modifying mechanism of action. The drug’s early discontinuation in prodromal AD (Merck press release, February 13, 2018) makes us wonder whether it will take the whole drug class—including candidates from AstraZeneca, Eli Lilly, Biogen, and Eisai—down with it.
BACE1 is the rate-limiting step in Aβ production, thus BACE inhibitors are designed to stop amyloid plaque formation upstream rather than remove plaques that are already in the brain (like the many MAbs in development). As such, there was a sense that early-stage patients would stand to benefit most from these drugs. A year ago, Merck terminated their EPOCH study of verubecestat in mild to moderate AD for lack of efficacy (Merck press release February 14, 2017), joining the ranks of many other DMTs to fail in established dementia. This left the verubecestat development program reduced to only the APECS study, which was evaluating the drug in prodromal AD (i.e., amnestic MCI) patients and represented the first real test of the theory that amyloid-targeting drugs need to be administered before the onset of full-blown AD.
KOLs we interviewed in 2017 knew that verubecestat’s success or failure would have strong implications for other developers of BACE inhibitors. As one U.S. neurologist put it:
“Verubecestat certainly didn't make a difference in people with mild to moderate disease. If we peel out the mild patients in the EPOCH study, especially the very mild patients—people that score over 22 on the MMSE—and there is no signal, not even a trend toward a signal, a separation between drug and placebo, it will be worrisome. I think it will make you a little more skeptical about what the MCI study is going to show and what it predicts for those companies that combine MCI and very mild patients.”
Now with two failures for verubecestat, the most recent in a prodromal (amnestic MCI) study that did not combine MCI and very mild patients, we have to ask, what now? What does this mean for other companies developing BACE inhibitors and their mixed prodromal/early mild cohorts in Phase III trials? It certainly calls into question their trial designs but maybe more importantly, it calls into question if prodromal intervention or BACE inhibition can work at all and only casts further doubt on the amyloid hypothesis.
Adding fuel to the fire was Biogen’s acknowledgement at an investor conference one day later that a blinded look at their aducanumab trials, slated to enroll 2,700 patients, uncovered more variability than expected on the primary end point, and have increased Phase III enrollment by more than 500 patients. As we consider what that means for the fate of that drug, the chorus of voices suggesting that the field and the industry finally need to move beyond the amyloid hypothesis is getting louder.
The amyloid hypothesis will continue to have its defenders, despite the latest news, who may point to the intriguing preliminary data for aducanumab demonstrating a dose-and time-dependent decrease in amyloid plaque burden, and a potential slowing of cognitive decline—data that encourage us all to wait and see. They may also point to the many ongoing prevention trials in asymptomatic, at-risk patients as the only good test of the hypothesis—and potentially the ideal scenario for BACE inhibitors, which many have suggested will be most effective in a primary prevention role. We’ll have to wait for those to read out beginning in 2019/2020.
That aside, it feels like now, more than ever, the time has come (or, some argue, has long past) for the industry to put real investment into exploring other approaches. Some shifts are happening. Tau, which has long been an industry afterthought, is now, finally, starting to have its day: AbbVie and AC Immune have advanced antibodies targeting tau into the clinic and Biogen has partnered with IONIS on a tau antisense oligonucleotide. However, it could be years before we know if tau therapies work (the earliest trial readout is 2020 and that’s only a Phase II study). Other approaches are being explored, but that road hasn’t been any easier. In the past month and a half, we have seen Boehringer Ingelheim scrap their PDE9 inhibitor BI409306, Takeda and Zinfandel Pharmaceuticals terminate their pioglitazone program, and Pfizer shutter its discovery and early clinical development neuroscience division including four Phase I AD molecules. The one program we expect to read out in 2018 is vTv’s RAGE inhibitor, azeliragon, which also targets pathways upstream of amyloid production. But knowing that azeliragon is being assessed in mild AD patients, it’s hard to believe the drug will be able to succeed where so many others have failed. Moreover, if combination therapy—e.g., amyloid + tau or BACE+ MAb—is really the only way forward, then there could be a regulatory quagmire awaiting those efforts if both investigational products are individually untested or ineffective.
Despite huge risks and terrible odds, large pharma and small biotechs alike have maintained a valiant commitment to research in AD. But the latest string of failures suggest that it may be time to stop bringing new anti-amyloid drugs into the clinic, at least until the trials that are already ongoing (including the prevention trials) readout, and instead, to seize this opportunity to invest in new ideas. There is a mounting cost to continuing down the current path that extends beyond the financial; today’s late-stage clinical trials are enrolling on average about 1,300 patients per study, and most developers are running two trials. It’s an enormous commitment from patients and their caregivers and huge draw on the precious resource that these hard to recruit, difficult to identify, early stage patients represent. We must invest these resources as wisely as we can.