Yes indeed, Merck announced last night that the first Phase III trial of their beta-secretase (BACE) inhibitor verubecestat was stopped because of futility. The monitoring committee, after looking over the data so far (the trial’s been running since 2012) concluded that there was no real chance of seeing efficacy. Merck had been enrolling patients up until early last year (edit: mixed up my dates at first!) in this trial (EPOCH), which focuses, well, focused, on patients with mild-to-moderate Alzheimer’s. They have another one going on patients even earlier in the disease process, and that’s where any hopes left for the drug will reside. Others are still trying their own compounds, too.
Are there such hopes? You really have to wonder. The list of Alzheimer’s clinical failures is impressive, but the list of failures to clinically validate the amyloid hypothesis is even more so. Recall that this is the same Merck compound that got a lot of press back in November, when the company published results that showed that it really did seem to be affected various cleaved amyloid concentrations in both humans and primates. Even back in 2012, though, when the company announced the further clinical development whose results we see today, people were wondering if they’d have a chance in this patient population. (I’ll pause to note that back then I called a headline about an effective Alzheimer’s drug being perhaps three years away “insane”, and I’m sticking with that judgment).
Beta-secretase inhibitors have failed in the clinic. Gamma-secretase inhibitors have failed in the clinic. Anti-amyloid antibodies have failed in the clinic. Everything has failed in the clinic. You can make excuses and find reasons – wrong patients, wrong compound, wrong pharmacokinetics, wrong dose, but after a while, you wonder if perhaps there might not be something a bit off with our understanding of the disease. Remember, every time one of these therapies comes around, it builds on the failures before it. Better and better attempts are made – I mean, verubecestat seems to be a pretty good compound, from the preclinical drug discovery perspective. It’s surely the best swing anyone’s taken at beta-secretase (and there have been many). But it just flat out did not work.
And let me take a minute to congratulate the drug discovery folks there on the compound itself. The fact that it does not work in the clinic is not their fault; it’s because we don’t know what the hell we’re doing in Alzheimer’s. Once you’ve decided that “We’re going to go after beta-secretase, because it’s absolutely one of the most solid ideas in the field”, the die is cast. But getting a good inhibitor for that target has not been easy. This one goes back to the Schering-Plough organization, later taken over by Merck, who came up with some interesting BACE-binding motifs while doing fragment screening. Verubecestat’s unusual “head group” came out of this work, and it’s worth thinking about how if you drew that structure on the glass of your fume hood, you’d probably have several people stopping by to tell you that you’re a fool. (Edit: structure updated, since I left out an amide carbonyl while I was redrawing it yesterday. That analog apparently didn’t work!) They’d especially be interested in your belief that a sulfonylguanidine would be a CNS drug. It doesn’t look like your run-of-the-mill combination of functional groups, but it gets into the brain, hits its target, and was good enough to be put into a gigantic series of Phase III trials. No, the people who came up with this deserve a lot of credit – mind you, I know for a fact that a good number of people who are on those papers have since been laid off by Merck, which tells you something about our industry, too.
So no, I’m not expecting much out of the prodromal Alzheimer’s patient trial that the compound is in. I will be very happy to be wrong about that, but by now, I think we have to assume that the focus on amyloid as the cause of Alzheimer’s, or even just as a method of treatment for it, is in very serious doubt.