AAN 2025 | From mutation to meaning: revisiting the PSEN1 Alzheimer’s story

Hello, I’m Alberto Aspay from the University of Cincinnati. Today I’m interested in kind of looking at the story of the most common familial form of Alzheimer’s disease associated with presenilin 1 mutation and how the data that we have is different to what we have been trying to tell ourselves about how that particular most common form of familial Alzheimer’s works. These are my disclosures, and I trust that none of them are particularly relevant to what I’m about to present to you. The first is to go through what kind of story we have here. And essentially, as you see in this diagram, we start with this large protein called amyloid precursor protein, or APP, that undergoes sequential catalysis by two enzymes, beta-secretase and gamma-secretase. And the story is that these mutations, presenilin-1 and presenilin-2, and APP to a certain extent, increase the gamma-secretase activity, which is the enzyme APP codes. In doing so, it creates more of the A-beta-42, which we think is the protein that is associated with Alzheimer’s, as well as the plaques from which it comes. That’s the story. Now, this particular enzyme breaks down this large 99 amino acid peptide that comes from APP and essentially trips it into its 42 amino acid segment, giving rise to A-beta 42. And so the problem, though, is that the mutation affects that enzyme activity, affects the enzyme substrate, and therefore what happens, in fact, in presenilin-1 mutations is that gamma secretase activity is stalled and the production of A-beta-42 goes down. So there is here a genetic example of a loss of function of gamma secretase. In fact, most presenilin-1 mutations directly reduce the levels of A-beta 42. There are some exceptions, but here in this study, it’s been shown that over 90% of these mutations lead to a reduced production of A-beta 42 and A-beta 40. A-beta 40 is not as relevant to Alzheimer’s as A-beta 42 is because A-beta 42 is what transforms into plaques. and plaques are of course the diagnostic hallmark of Alzheimer’s. Presenilin-1 carriers have lower levels of A-beta-42 than even sporadic Alzheimer’s, as you see in this study that’s including 14 individuals with sporadic allele in Alzheimer’s compared to 8 individuals with presenilin-1. And what’s interesting is that very recently, there was a presenilin-2 mutation carrier shown to be exceptionally resilient, which essentially means instead of developing Alzheimer’s based on the mutation that this individual had, he was now in his 70s, and we suspected that there must have been a reason among them, those listed at the bottom of this slide, as the authors suggested. But in fact, as you see, one reason we could argue this individual was resilient is because the levels of A-beta 42 were higher than those of others in the cohort that were non-resilient. Now, here is the latest presenilin-1 analysis, and this is a Lancet of Neurology paper. As you see, 190 presenilin-1 carriers in this study. It’s looking at a variety of different factors associated with these carriers, including A-beta-42 levels and gamma secretase activity. And it turns out that gamma secretase activity is lowered. And the lower the gamma secretase activity in presenilin-1 mutations, the faster the increase in amyloid accrual by PET, the more rapid decreases in hippocampal volume, lower MMSE, higher CDRSPS scores, worse memory delay recall, so globally worse cognitive function, and lower levels of A-beta 42. So, presenilin-1 mutations, decrease this gamma secretase activity and reduce CSF A-beta-42 levels. And then, if you look at the tertiles of activity, lowest gamma secretase tertile compared to the highest gamma secretase tertile, you find that there is an earlier cognitive decline in the case in which there is lowest gamma secretase activity, and an earlier age of onset. So all of this indicating that lower gamma secretase activity is detrimental in individuals with presenilin-1 mutation. The conclusions of the Lancet Neurology paper, however, is that these findings support targeting gamma secretase. Targeting gamma secretase can be increasing or decreasing. You would think it’s increasing. But in fact, that is not what the authors are presenting in the discussion. They assume that we should continue on lowering gamma secretase activity through modulators. And the reason we speak of modulators rather than direct inhibitors is because when we have tested gamma secretase inhibitors, as the example of verubesistat, here what has happened is that there has been a dose-dependent reduction in A-beta-42 levels, but also a reduction in cognitive function or decline in cognition. And we’ve assumed that these data suggest that there must be some other explanation, such as notch signaling inhibition. And therefore, we’ve come with this concept that gamma-secretase modulators would still decrease A-beta-42 on the idea that A-beta-42 is toxic without affecting the notch signaling. So not sparing gamma secretase inhibitors. But this is what we have so far of trials that use gamma secretase inhibitors, beta secretase inhibitors, and the sole gamma secretase modulator that so far has been reported. In general, the picture isn’t attractive at all and should have given us pause to continue along the path of reducing gamma secretase function in any way. Now, so when we look back at this particular graph, the idea is to know that while we have said to ourselves that presenilin-1 increase the gamma secretase activity, in fact, it decreases the gamma secretase activity, lowering the levels of A-beta-42 in spinal fluid. So presenilin-1 is leading to a loss of function through a reduction in gamma secretase activity and lower levels of A-beta 42. Plaques do form, even though A-beta 42 production is low, and we don’t know why that is, but plaques still form, and it’s possibly the result of A-beta 42 becoming stuck in the membrane and therefore increasing the local concentration of A-beta 42 so that it undergoes transformation into amyloid plaques. Now, we have lecanemab, and lecanemab is in many ways thought of as the vindication of the amyloid hypothesis, but it is interesting that it was really a result published almost three decades after donepezil, making donepezil look much, much better. We’ve always thought that donepezil is modest, but compared to lecanemab, it’s actually almost twice as effective. And if we look at the key indices that clinicians must have to determine whether to use a therapy or not, you can see that donepezil is having a number needed to treat of 10 to 12, whereas we don’t know what that number is for lecanemab. The number needed to harm, the number of people that we would need to give this drug for one to be harmed is about 12 in donepezil, but three in lecanemab. And we know that lecanemab accelerates brain atrophy and ARIA, that we don’t know what the significance of that may be. So what makes lecanemab seems exceptional? Essentially, it is this idea that while it reduces amyloid, it also increases A-beta-42, but nearly 300 picograms per ml. So here you see these are the monoclonal antibodies. All of them have been associated with a reduction to different levels of amyloid, but all of them have also increased A-beta 42. Donanemab is the only one we don’t know the data because the donanemab trial did not include spinal fluid analysis. So what we’ve done with this data is to then determine the extent to which we can replicate what has been done before, which is that across trials, certainly not within them, there is a relationship whereby the lower the amyloid, the better the cognitive endpoint. And we’ve replicated that with what we’ve seen through the trials, but we used the other piece of data, the increases in A-beta 42, and found a very similar relationship, such that we can conclude that the increases in A-beta 42 predict changes in cognitive endpoints at least as well as the decreases in amyloid. So, in summary, I hope to have presented here some evidence to conclude that most presenilin-1 mutations reduce gamma secretase activity and lower A-beta-42 production. The lowered gamma secretase activity and A-beta-42 levels in presenilin-1-associated familial Alzheimer’s disease correlate with earlier symptom onset, faster progression, worse cognition, and more atrophy. The gamma secretase inhibitors and modulators lower A-beta-42 and worsen cognitive decline compared to placebo in randomized clinical trials. That increases in A-beta-42 may in fact be the mechanism by which anti-A-beta monoclonal antibodies work. Since reduced gamma secretase activity and A-beta-42 levels are detrimental, then it behooves us to think of strategies that would do the opposite, to increase gamma-secretase activity and A-beta-42 levels. And although having gone through this, there are already therapies out there that are known to increase A-beta-42, and that might be the mechanism by which, at a population level, they are associated with a reduced incidence in the development of Alzheimer’s disease. Thank you.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.