This ended up being another one of those books I had to read over an extended period of time due the stress it gave me. This book shares some similarity to John Carreyrou’s Bad Blood in a way, but it covers a field that I am somewhat less familiar with, which also contributed to it becoming a stressful read. In addition, unlike the Theranos story, you also can’t just lay blame here at the feet of a small number of individuals. The issues here are much more complex, and systemic.
So what’s the deal right now with Alzheimers research? To keep it very simple, the focus of the last few decades has been on the amyloid cascade hypothesis, which proposes that the symptoms we associate with the disease are caused by a certain protein—amyloid-beta (Aβ)—misfolding and accumulating in the brain. You can see these amyloid plaques in the brains of patients with neurodegenerative disorders who have died. It seems logical that their presence would have something to do with the disease’s pathology; they make for a very striking biomarker, after all. So it is perhaps unsurprising that most recent treatments for Alzheimers disease have focused almost entirely on amyloid plaques. But what if we’re all been chugging along on the wrong track all this time? And was it research fraud that lead to us hitting the wrong railroad switch?
In Doctored, Pillar doesn’t just focus on one case here, there are multiple cases of questionable activities that have cast a shadow on Alzheimers research. But I’ll start with one of the big ones: There are multiple forms of Aβ, and the variant that has been thought to be the most ‘toxic’ is the perhaps awkwardly named Aβ*56. The effects of Aβ*56 on memory was first characterized in 2006 by Karen Ashe and Sylvain Lesné at University of Minnesota, and their work received an enormous amount of attention.
Of course, the pair went on to publish many follow up papers. But while the original paper has been cited thousands of times, not everyone has been able to even detect this variant protein. Let alone replicate the results. Then, in 2022, attention was drawn to the fact that several of Lesné’s papers had what appeared to be doctored images. Last year, Nature retracted the original paper.
Woof. How much money, do we think, was spent chasing a protein variant that may not exist?
This is by far not the only Alzheimers controversy; the earlier case of simufilam also turned into another egregious example of fraud. Simufilam, backed by the biopharmaceutical company Cassava Sciences, was meant to bind to a deleterious form of filamin A, another protein in the brain. This is still linked to the amyloid hypothesis, by the way; another variant of Aβ is meant to cause this change in filamin A, and Simufilam is meant to reverse this.
But again, the studies here were questionable. There were multiple issues; one is that Cassava’s theory was a little weak and watery. Simufilam was not designed as an Alzheimer treatment—it was a co-opted one. And, just like the example of Aβ*56, there were allegations of fraud against one of the lead researchers behind the drug, Hoau-Yan Wang. Again, these accusations involved doctored images.
But wait, there’s more…
First, more image manipulation? Apparently so. The chapter “More Blots”, in reference to the protein detection technique known as the western blot, is particularly eye opening. Thanks to the work of neurologist Matthew Schrag and a number of other image detectives, even more doctored images were uncovered in a number of high impact papers on the disease, hitting many high profile researchers.
And it’s not just the researchers either; it’s the regulators. There have been three anti-amyloid antibody-based drugs so far cleared (at some point) by the FDA. Aducanumab was the first, and in the eyes of many—including most members of the FDA advisors committee—it shouldn’t have been. Three people resigned in protest. Expensive, potentially risky and with little overall therapeutic benefit, it’s obvious why there were serious reservations about the drug. But it was approved anyhow, first in its class. Two additional drugs, Lecanemab and Donanemab have followed more recently; both still not terribly impressive (statistical significance and clinical significance are not the same thing) and both under a cloud of suspicion.*
So it almost goes without saying that Charles Piller has put in an extraordinary amount of work at exposing this misconduct. And Doctored goes into a considerable degree of detail. There are a lot of researchers covered in this book and it is very information dense. Perhaps too much so. While sometimes interesting, I don’t think I needed as many biographical details on some of the people involved Additionally, with so many moving parts, the book can feel a little disjointed. I myself have not addressed the main events here in the same order they were presented, for example, because I didn’t agree with the ‘narrative’ flow.
I did have a number of other concerns reading the book. Unsurprisingly, with all the research misconduct that has been uncovered, the amyloid cascade hypothesis is looking far more shaky than it was a few years ago. Perhaps misfolding Aβ is just another symptom and not a cause. But there’s just one line of reasoning that I cannot get behind; the idea that the fact that we still don’t have a good drug after all of this time and all this money itself is evidence that the hypothesis must be incorrect. This is something that I see a little too often when it comes to the clinical side of research—the fact people (or maybe its their funders) forget that it is still research, you can still be thrown curveballs, and your outcomes are not proportional to the time and money you pour in. Shit is not linear. You can still be on the right track theoretically and fuck up at the drug design and delivery stage—especially when dealing with the brain. Passing the blood brain barrier and then not upsetting the grey matter within is no simple task.
So, then, where do I fall regarding the amyloid cascade hypothesis? Fair disclaimer, my background is not in the neuro-side of biology. I could be talking out of my arse, after all. But I have worked on dissecting the basis of complex genetic traits, so, with that in mind, my first response would be to go: shit’s complicated. I can’t completely exclude a role for Aβ based off what I know, but I know of a number of other genetic risk factors and suspect there are multiple environmental cues involved. The presentation of the disease is very heterogenous; I would not be surprised if the cause is the same. Maybe these other factors will get a little more attention now.
Another concern I had, which Pillar mentions near the end, is how we deal with scientific misconduct. It’s clear that ‘the establishment’ is not handling these cases well at all—or even identifying them. This is something we have to reckon with. Many of the papers with doctored images were instead uncovered—as detailed in the ‘More Blots’ chapter—by a set of image detectives, many of whom worked online. And sometimes anonymously.
I see two issues here: firstly, this makes it hard to credit the right people when they do put in the effort to catch what ‘the establishment’ has missed. Anonymous whistleblowers can’t always build reputations or careers from their work, and institutions rarely acknowledge these outside contributions, even when they lead to retractions. I don’t know about the print copy of the book, but the audio version I listened to here did not list these contributors names, to give an example. We’re essentially relying on nameless people to perform a function that should be handled by ‘the establishment’.
But secondly, and forgive me for being a bit cynical here, regarding the state of social media and all, we could be having the opposite effect with callouts and accusations that can’t be verified. When anyone can post allegations without accountability, we risk creating an environment where legitimate researchers face damaging accusations based on what might perhaps be simple errors. Not all images initially appear ‘doctored’ actually are, and not all errors in data are necessarily deliberate or malicious. People do make honest mistakes. This—as Pillar points out—is something that might impact early career researchers than those in more established positions. They’d have a harder time defending themselves.
None of this is ideal, to say the least.
Again, reading this book stressed the hell out of me—and made me depressed. I really don’t know how to deal with all this. but least I can take pride in self restraint here; unlike other instances (Bad Blood) I resisted the urge to pull my rant-ie panties on when writing this up.
Maybe because in this instance, the fraud involves a lot of regular seeming people. There isn’t much in the way of cartoonishly weird Elizabeth Holmes-types here.**
For cbr17bingo, this is Rec’d. A coworker recommended this book to me back in July. I suspect I’m going to have to message her with my thoughts about it at some point.