Late last week came this report in Science about doctored images in a series of very influential papers on amyloid and Alzheimer’s disease. That’s attracted a lot of interest, as well it should, and as a longtime observer of the field (and onetime researcher in it), I wanted to offer my own opinions on the controversy.
First off, I’ve noticed a lot of takes along the lines of “OMG, because of this fraud we’ve been wasting our time on Alzheimer’s research since 2006”. That’s not really the case, as I’ll explain. But don’t get your hopes up: from one point of view, the main inaccuracy in that statement is that we’ve been actually been wasting our time in Alzheimer’s research for even longer than that. So that’s not very comforting, either. We’ll start with some background and history, the better to appreciate the current furor in context. Those of you who know the field can skip ahead to later sections as marked, but your admission ticket is valid for the entire length of the ride if you want to get on here.
The Amyloid Hypothesis
An association between Alzheimer’s disease and amyloid protein in the brain has been around since. . .well, ever since the early 1900s, when Alois Alzheimer (and Oskar Fischer, independently) recognized some odd features in the brains of people who had died with memory loss and dementia. There were dark plaques among the neurons, which stained in a way that suggested they were some sort of amyloid protein. The association of these plaques with dying neurons made a compelling case that they were involved in the disease, although it was recognized at the same time that there were “neurofibrillary tangles” that were also present as a sign of pathology. A classic AD plaque has a dense core of precipitated protein, surrounded by a less dense halo of abnormal protein around it, and also surrounded by a lot of rather unhealthy-looking extensions of nearby neurons. Activated astrocytes and microglia are present as well, suggesting that some degenerative process has been taking place for some time and that the usual repair mechanisms have not been up to the job.
In the mid-1980s, the main protein in the plaques was conclusively identified as what became known as beta-amyloid, a fairly short (36 to 42 amino acid) piece that showed a profound tendency to aggregate into insoluble masses. Pure “A-beta” is not a lot of fun to work with or even to synthesize; it really does gum things up alarmingly. A lot of proteins can do that to some degree, with various types of amyloid high on that list. The same protein was found in the (similar) lesions that develop in the brain tissue of Down’s syndrome patients, who often show memory loss and AD-like symptoms even earlier than usual. By the early 1990s, the “amyloid cascade hypothesis” of Alzheimer’s was the hot topic in the field. Beta-amyloid had been found to be a cleavage product from inside the sequence of a much larger species (APP, or amyloid precursor protein), and the cascade hypothesis was that excess or inappropriately processed beta-amyloid was in fact the causative agent of plaque formation, which in turn was the cause of Alzheimer’s, with all the other neuropathology (tangles and so on) downstream of this central event. And for the last thirty years, this has been the reigning idea in the field, although there are others (such as tau protein, which is more involved with the neurofibrillary tangles). It has not been a smooth ride, though.
Before getting to that part, please keep in mind that there’s a lot of support for the amyloid hypothesis itself, and I say that as someone who has been increasingly skeptical of the whole thing. For example, mutations in APP that lead to easier amyloid cleavage also lead to earlier development of Alzheimer’s symptoms, and that’s pretty damn strong evidence. There are human families around the world (in Sweden, Holland, Mexico, Colombia and more) with “hereditary early onset Alzheimer’s” of this sort, and these things almost always map back to mutations in amyloid processing. If you engineer excess beta-amyloid in cells in culture, you see toxic effects, and the same goes for engineering that in living animals. Now, none of those mice or whatever develop syndromes quite like human Alzheimer’s, to be sure – we’re the only animal that does, interestingly, but excess beta-amyloid is always trouble.
But along with these strong signals there have always been plenty of mysteries, too: no one’s quite sure of all the functions of the APP protein, for starters, and that goes double for whatever the normal functions of beta-amyloid might be. Meanwhile, although there does seem to be a correlation between amyloid plaques and dementia, there are people who show significant amyloid pathology on examination after death who did not show real signs of Alzheimer’s. Progress has been slowed by the longstanding problem of only being able to see the plaques post-mortem (brain tissue biopsies are not a popular technique) – there are now imaging agents that give a general picture in a less invasive manner, but they have not helped settle the debates. The results of the clinical trials of the last twenty years or so have only added to these problems.
Putting It to the Test
Ever since the 1990s, researchers and clinicians have been spending uncountable hours (and uncountable dollars) trying to turn the amyloid hypothesis into a treatment for Alzheimer’s. I would not like to count the number of such attempts, nor even to try to list all of the variations. There have been all sorts of treat-the-symptoms approaches, for sure, but also a number of direct shots on goal. The enzymes that cleave beta-amyloid out of the APP protein (beta-secretase and gamma-secretase) have been targeted for inhibition, naturally. Small molecules have been sought that would slow down amyloid aggregation or even to promote its clearance. Most famously, antibodies have been produced against various forms of beta-amyloid itself, in attempts to interrupt their toxicity and cause them to be cleared by the immune system.
Every single one of these interventions has failed in the clinic. Every last damn one. If you look for the best outcome of all, actual reversal of Alzheimer’s symptoms, you never see it. No one has, and given the level of neuronal damage, it’s quite possible that no one ever will, unfortunately. What about just slowing down the inexorable progress that the disease seems to show in so many patients? No luck there, either. Compared to control patients, none of these therapies have shown meaningful effects on the rate of decline. One of the gamma-secretase inhibitor trials actually seemed to speed it up, for reasons yet unknown.
The antibody trials have been the most disconcerting. Some of them have actually shown real reductions in amyloid levels in the brains of the patients, which should be good news, but at the same time these reductions have not led to any real improvements in their cognitive state. Not even a slowdown in the rate of developing Alzheimer’s symptoms. If you had time-traveled back to the mid-1990s and told people that antibody therapies would actually have cleared brain amyloid in Alzheimer’s patients, people would have started celebrating – until you hit them with the rest of the news. Of course, as I’ve often said, if you’d time-traveled back 30 years and told *me* (while I was working in the field) that we’d still be arguing about the amyloid hypothesis itself in 2022, I would have been profoundly displeased, to put it judiciously.
These failures have led to a whole list of explanatory, not to say exculpatory hypotheses: perhaps the damage had already been done by the time people could be enrolled in a clinical trial, and patients needed to be treated earlier (much, much earlier). Or we had somehow picked the wrong kind of Alzheimer’s patients – the disease might well stratify in ways that we couldn’t yet detect, and we needed to wait for better ways to pick those who would benefit. Maybe the different forms of beta-amyloid (different lengths and different aggregation/oligomerization states) were not being targeted correctly: we had raised antibodies to the wrong ones, and when we zeroed in on the right one we would see some real clinical action.
Oligomers As An Explanation, And the *56 Species
That’s where the currently contested work really comes into the picture. There had been a lot of work (and a lot of speculation) about the possibility of there being hard-to-track-down forms of amyloid that were the real causative agent of Alzheimer’s. You can find plenty of papers from the late 1990s and early 2000s on the idea of pathogenic soluble amyloid oligomers, oligimerization state as correlated with disease, all that sort of thing. It was certainly not an unexplored idea.
But this 2006 paper did indeed get a lot of attention, because it took the idea further than many other research groups had. It was from the lab of Karen Ashe at Minnesota, highlighting work from Sylvain Lesné in her lab on a form of amyloid called AB*56. Isolating this species from a transgenic mouse model and injecting it into young rats caused them to start exhibiting memory defects in turn. Now that’s the kind of smoking gun you want to see, especially in a field as murky and tangled as this one. That paper has been cited well over 2000 times since then, and Lesné has since produced a large number of papers following up on this idea and its ramifications. He’s running his own research group now, naturally, and Ashe’s group has also continued to work on amyloid oligomers, as have (by now) many others.
But some of these later Lesné papers had been flagged on the PubPeer site for potential image doctoring, and the Science report linked to in the first paragraph of this blog details the efforts of neuroscientist Matthew Schrag at Vanderbilt to track these things down. He was originally hired by two other neuroscientists who also sell biopharma stocks short – my kind of people, to be honest – to investigate published research related to Cassava Sciences and their drug Simufilam, and that work led him deeper into the amyloid literature.
Now Cassava is a story of their own, and I have frankly been steering clear of it, despite some requests. To me, it’s an excellent example of a biotech stock with a passionate (and often flat-out irrational) fan club. In such cases, if you say bad things about a beloved stock then plenty of helpful strangers will point out that you are an idiot, a shill, an evil agent of the moneyed interests, and much, much more. There’s a detailed sidebar in the Science article on Cassava and on simufilam, which I recommend to anyone who wants to catch up on that aspect. That compound is supposed to restore the function of the protein Filamin A, which is supposed to be beneficial in Alzheimer’s, and my own opinion is that neither the published work on this compound nor the conduct of the company inspires my trust. There’s an ongoing investigation into the work at CUNY, and perhaps I’ll return to the subject once it concludes.
But Schrag’s dive into the Alzheimer’s literature put him onto allegations of image manipulation in the amyloid field as well, and that’s why we find ourselves in the current situation. Schrag (and others on PubPeer) have found what looks like a long trail of image manipulation in Lesné’s papers, particularly the ever-popular duplication of Western blots to produce bands where you need them. The Science article illustrates some of these, and it looks bad: protein bands showing up in different places with exactly the same noise in their borders, apparent copy-and-past border lines, etc. Schrag has filed a long, detailed whistleblower report with the NIH and alerted numerous journal editors, and some of the papers are already being flagged with Expressions of Concern while the matter is being investigated.
Science had Schrag’s findings re-evaluated by several neuroscientists, by Elisabeth Bik, a microbiologist and extremely skilled spotter of image manipulation, and by another well-known image consultant, Jana Christopher. Those last two even found more examples that Schrag himself had missed. Apparently everyone agrees that Lesné’s work is full of trouble. Phrases like “shockingly blatant” and “highly egregious” are quoted, and it looks like key parts of the experimental evidence in these papers is nothing more than cut-and-paste jobs assembled to show the desired result.
What Did the *56 Paper Lead To?
So what’s happened with the other people working on AB*56? As the article notes, well-known Alzheimer’s researcher Dennis Selkoe at first co-authored with Lesné, but later published work where he and his co-workers failed to find the *56 oligomer species in transgenic mouse lines or even human fluids and human tissues. It appears that other research groups had similar problems. But the whole amyloid-oligomer idea (which, as shown above, predated the AB*56 work) has continued to be the focus of a huge amount of research. As detection methods became better and better, it turned out that you could find huge numbers of different sorts of amyloid species in the tissues and fluids of animal models and human samples, especially when you get down to nanomolar levels. There are all sorts of different cleavages leading to different short amyloid-ish proteins, different oligomerization states, and different equilibria between them all, and I think it’s safe to say that no one understands what’s going on with them or just how they relate to Alzheimer’s disease.
I could be wrong about this, but from this vantage point the original Lesné paper and its numerous follow-ups have largely just given people in the field something to point at when asked about the evidence for amyloid oligomers directly affecting memory. I’m not sure how many groups tried to replicate the findings, although (as just mentioned) when people did it looks like they indeed couldn’t find the *56 oligomer. And judging from the number of faked Westerns, that’s probably because it doesn’t exist in the first place. But my impression is that a lot of labs that were interested in the general idea of beta-amyloid oligomers just took the earlier papers as validation for that interest, and kept on doing their own research into the area without really jumping directly onto the *56 story itself. The bewildering nature of the amyloid-oligomer situation in live cells has given everyone plenty of opportunities for that! The expressions in the literature about the failure to find *56 (as in the Selkoe lab’s papers) did not de-validate the general idea for anyone – indeed, Selkoe’s lab has been working on amyloid oligomers the whole time and continues to do so. Just not Lesné’s oligomer.
What About Peer Review, Damn It All?
Yeah, there’s that. The Lesné stuff should have been caught at the publication stage, but you can say that about every faked paper and every jiggered Western blot. When I review a paper, I freely admit that I am generally not thinking “What if all of this is based on lies and fakery?” It’s not the way that we tend to approach scientific manuscripts. Rather, you ask whether the hypothesis is a sound one and if it was tested in a useful way: were the procedures used sufficient to trust the results and were these results good enough to draw conclusions that can in turn be built upon by further research? Are there other experiments that would make things stronger? Other explanations that the authors didn’t consider and should address? Are there any parts where the story doesn’t hang together? If so, how would these best be fixed?
There is a good-faith assumption behind all these questions: you are starting by accepting the results as shown. But if someone comes in with data that have in fact been outright faked, and what’s more, faked in such a way as to answer or forestall just those sorts of reviewing tasks, there’s a good chance that these things will go through, unfortunately. That is changing, slowly, in no small part due to sites like PubPeer and a realization of how many times people are willing to engage in such fakery.
It’s fair to ask “But isn’t science supposed to be self-correcting, as people try to reproduce the results?” That really is the case, but it’s not the case for every single paper and every single result. A lot of work never gets reproduced at all – there is just so much of it, and everyone’s working on their own ideas. But that said, the Lesné situation is a black mark on the whole amyloid research area. People were already excited by the amyloid-oligomer idea (which, as mentioned, is a perfectly good one, or was at first). Probably not too many people directly tried to replicate the AB*56 work and those who tried and couldn’t probably just said “Oh well, that’s neuroscience, there are so many variables involved” and kept on going with their own projects.
But the faked Westerns in this case were already being noticed on PubPeer over the last few years. Prof. Schrag’s deep dive through Lesné’s work could have been done years ago, and journal editors could have responded to the concerns that were already being raised.
Did the *56 Work Lead to Clinical Trials?
That’s a question that many have been asking since this scandal broke a few days ago. And the answer is that no, I have been unable to find a clinical trial that specifically targeted the AB*56 oligomer itself (I’ll be glad to be corrected on this point, though). What there have been are trials that (to a greater or lesser extent) tried to target the whole amyloid-oligomer hypothesis in general, but I have to think that those would have happened anyway. In case you’re wondering, Biogen’s recent aducanumab antibody seems to hit both aggregated amyloid in plaques and some types of oligomers, not that it mattered in the end. The idea of amyloid oligomers as a key driver of AD is not a crazy one in any way, and people were going to put it to the test whether the *56 paper came out or not.
Most of these have been antibodies, as that last link shows. A non-antibody approach is ALZ-801, which is the small molecule homotaurine and is claimed to inhibit amyloid oligomer formation in general. The problem is that this drug has already failed an Alzheimer’s Phase III, and I’m not convinced by the attempts at post-hoc analysis being done to revive its prospects. But antibody or small molecule, though, nothing has worked.
All this said, the excitement over the AB*56 work surely did accelerate things. We probably got more clinical trials, sooner, than we would have otherwise. Perhaps in a way this might have helped to bury the hypothesis even more quickly than otherwise? There’s no way of knowing.
What Now?
Well, as the world well knows, every single Alzheimer’s trial to date has failed. I know, I know, there are all sorts of special pleadings for aducanumab and what have you, if you look at the data sideways with binoculars you can start to begin to see the outlines of the beginnings of efficacy, sure, sure. I’m not having it. Every single disease-modifying trial of Alzheimer’s has failed.
The huge majority of those have addressed the amyloid hypothesis, of course, from all sorts of angles. Even the truest believers are starting to wonder. Dennis Selkoe’s entire career has been devoted to the subject, and he’s quoted in the Science article as saying that if the trials that are already in progress also fail, then “the A-beta hypothesis is very much under duress”. Yep.
The Big Picture
OK, let me try to summarize and give people sometime to skip ahead to. Here’s how I see the situation:
1. beta-Amyloid has been the dominant explanation for Alzheimer’s for decades. There are a lot of good reasons for thinking that, but every attempt to target it and slow the disease has failed in the clinic.
2. These failures, combined with the still-compelling reasons to think that amyloid is indeed a major part of the disease, have led to hypotheses that would square all these conflicting findings: perhaps amyloid really is the cause of Alzheimer’s, but not the form of amyloid we’ve been looking at. The real cause could be well upstream, in small soluble oligomers of the protein that are the earlier bad actors in the disease.
3. That was already a major hypothesis before the Lesné work on AB*56. But that one was reported (in 2006) as just such a soluble oligomer which had direct effects on memory when injected into animal models. It was big news, and the paper has been cited very widely indeed in years since. People already working on oligomers redoubled their efforts, and more researchers joined the field.
4. Lesné’s work now appears suspect across his entire publication record. AB*56 itself does not seem to exist. Other researchers had failed to find it even in the first years after the 2006 publication, but that did not slow the beta-amyloid-oligomer field down at all. It was going to grow anyway, but it’s for sure that the AB*56 stuff turbocharged it, too. Amyloid oligomers are a huge tangled area with all kinds of stuff to work on, and while no one could really prove that any particular oligomeric species was the cause of Alzheimer’s, no one could prove that there wasn’t such a causative agent, either, of course.
5. The failure to notice and act on the faked data in the Lesné papers is still a disgrace, and there’s plenty of blame to go around among other researchers in the field as well as reviewers and journal editorial staffs. I’m not sure how often we have to learn this lesson about dealing with these things more quickly and more seriously. Apparently very often indeed.
6. The AB*56 work did not lead directly to any clinical trials on that amyloid species, and the amyloid oligomer hypothesis was going to lead to such trials anyway at some point. But it certainly did raise the excitement and funding levels in the area and gave people more reason to believe that yes, targeting oligomers could really be the way to go. It’s definitely fair to say that the Lesné work caused these trials to happen more quickly and probably in greater number than they would have otherwise.
7. Those trials have failed. But every single Alzheimer’s trial has failed. I think that any ultimate explanation of Alzheimer’s disease is going to have to include beta-amyloid as a big part of the story – but if attacking the disease from that standpoint is going to lead to viable treatments, we sure as hell haven’t been seeing it. We have to put money and effort down on other hypotheses and stop hammering, hammering, hammering on beta-amyloid so much. It isn’t working.
