A neuroscience image sleuth has found signs of fabrication in scores of Alzheimer’s articles, threatening a reigning theory of the disease, and uncovering what could be proof that manipulated’ Alzheimer’s data may have misled research for 16 years.
In August 2021, reports Charles Piller in Science, Matthew Schrag, a neuroscientist and physician at Vanderbilt University, was asked by a colleague to contact an attorney investigating an experimental drug for Alzheimer’s disease called Simufilam.
The drug’s developer, Cassava Sciences, claimed it improved cognition, partly by repairing a protein that can block sticky brain deposits of the protein amyloid beta (Aβ), an Alzheimer’s hallmark. The attorney’s clients, two prominent neuroscientists, believed some research related to Simufilam may have been “fraudulent”, according to a petition later filed on their behalf with the US Food and Drug Administration (FDA).
Schrag, a junior professor, had already gained some notoriety by publicly criticising the controversial FDA approval of the anti-Aβ drug Aduhelm. His own research also contradicted some of Cassava’s claims. He feared volunteers in ongoing Simufilam trials faced risks of side effects with no chance of benefit.
So he applied his technical and medical knowledge to interrogate published images about the drug and its underlying science. He identified apparently altered or duplicated images in dozens of journal articles. The attorney reported many of the discoveries in the FDA petition, and Schrag sent all of them to the National Institutes of Health (NIH), which had invested tens of millions of dollars in the work. (Cassava denies any misconduct).
But Schrag’s sleuthing drew him into a different episode of possible misconduct, leading to findings that threaten one of the most cited Alzheimer’s studies of this century and numerous related experiments.
As reported in MedicalBrief last week, the first author of that influential study, published in Nature in 2006, was an ascending neuroscientist: Sylvain Lesné of the University of Minnesota (UMN), Twin Cities. His work underpins a key element of the dominant yet controversial amyloid hypothesis of Alzheimer’s – that Aβ clumps (plaques), in brain tissue, are a primary cause of the illness, which afflicts tens of millions globally. In what looked like a smoking gun for the theory and a lead to possible therapies, Lesné and his colleagues discovered an Aβ subtype and seemed to prove it caused dementia in rats. If Schrag’s doubts are correct, Lesné’s findings were an elaborate mirage.
Schrag, who had not publicly revealed his role as a whistle-blower until this article, avoids the word “fraud” in his critiques of Lesné’s work and the Cassava-related studies and does not claim to have proved misconduct. That would require access to original, complete, unpublished images and in some cases raw numerical data. “I focus on what we can see in the published images, and describe them as red flags, not final conclusions,” he says. “The data should speak for itself.”
A six-month investigation by Science provided strong support for Schrag’s suspicions and raised questions about Lesné’s research. A leading independent image analyst and several top Alzheimer’s researchers, including George Perry of the University of Texas, San Antonio, and John Forsayeth of the University of California, San Francisco (UCSF), reviewed most of Schrag’s findings at Science’s request. They concurred with his overall conclusions, which cast doubt on hundreds of images, including more than 70 in Lesné’s papers. Some look like “shockingly blatant” examples of image tampering, says Donna Wilcock, an Alzheimer’s expert at the University of Kentucky.
“The authors appeared to have composed figures by piecing together parts of photos from different experiments,” says Elisabeth Bik, a molecular biologist and well-known forensic image consultant. “The obtained experimental results might not have been the desired results, and those data might have been changed to … better fit a hypothesis.”
Early this year, Schrag raised his doubts with NIH and journals including Nature; two have published expressions of concern about papers by Lesné. Schrag’s work, done independently of Vanderbilt and its medical centre, implies millions of dollars may have been misspent on the research – and more on related efforts. Some Alzheimer’s experts now suspect Lesné’s studies have misdirected Alzheimer’s research for 16 years.
“The immediate damage is wasted NIH funding and wasted thinking in the field because people use these results as a starting point for their own experiments,” says Stanford University neuroscientist Thomas Südhof, a Nobel laureate and expert on Alzheimer’s and related conditions.
Lesné did not respond to requests for comment. A UMN spokesperson says the university is reviewing complaints about his work.
To Schrag, the two disputed threads of Aβ research raise questions about scientific integrity in the struggle to understand and cure Alzheimer’s. Some adherents of the amyloid hypothesis are too uncritical of work that seems to support it, he says. “Even if misconduct is rare, false ideas inserted into key nodes in our body of scientific knowledge can warp our understanding.”
One of biggest mysteries about Alzheimer's disease is also its most distinctive feature: the plaques and other protein deposits that German pathologist Alois Alzheimer first saw in 1906 in the brain of a deceased dementia patient.
In 1984, Aβ was identified as the main component of the plaques. And in 1991, researchers traced family-linked Alzheimer’s to mutations in the gene for a precursor protein from which amyloid derives. To many scientists, it seemed clear that Aβ buildup sets off a cascade of damage and dysfunction in neurons, causing dementia. Stopping amyloid deposits became the most plausible therapeutic strategy.
Hundreds of clinical trials of amyloid-targeted therapies have yielded few promising glimmers, however; only the underwhelming Aduhelm has gained FDA approval. Yet Aβ still dominates research and drug development. NIH spent about $1.6bn on projects that mention amyloids in this fiscal year, about half of its overall Alzheimer’s funding.
Scientists who advance other potential Alzheimer’s causes, like immune dysfunction or inflammation, complain about being sidelined by the “amyloid mafia”. Forsayeth says the amyloid hypothesis became “the scientific equivalent of the Ptolemaic model of the solar system”, in which the sun and planets rotate around earth.
By 2006, the centenary of Alois Alzheimer’s epic discovery, a growing cadre of sceptics wondered whether the field needed a reset.
Then, a Nature paper entered the breach.
It emerged from the lab of UMN physician and neuroscientist Karen Ashe, who had already made a remarkable series of discoveries. As a medical resident at UCSF, she contributed to Nobel laureate Stanley Prusiner’s pioneering work on prions, infectious proteins that cause rare neurological disorders. In the 1990s, she created a transgenic mouse that churns out human Aβ, forming plaques in the animal’s brain. The mouse also shows dementia-like symptoms. It became a favoured Alzheimer’s model.
By the early 2000s, “toxic oligomers”, subtypes of Aβ that dissolve in some bodily fluids, had gained currency as a chief culprit for Alzheimer’s – potentially more pathogenic than the insoluble plaques. Amyloid oligomers had been linked to impaired communication between neurons in vitro and in animals, and autopsies have shown higher levels of the oligomers in people with Alzheimer’s than in cognitively sound individuals. But no one had proved that any of the many known oligomers directly caused cognitive decline.
In the brains of Ashe’s transgenic mice, the UMN team discovered a previously unknown oligomer species, dubbed Aβ*56 (pronounced “amyloid beta star 56”) after its relatively heavy molecular weight compared with other oligomers. The group isolated Aβ*56 and injected it into young rats. The rats’ capacity to recall simple, previously learned information, like the location of a hidden platform in a maze, plummeted. The 2006 paper’s first author, sometimes credited as the discoverer of Aβ*56, was Lesné, a young scientist Ashe had hired straight from a PhD programme at France’s University of Caen Normandy.
Ashe touted Aβ*56 on her website as “the first substance ever identified in brain tissue in Alzheimer’s research shown to cause memory impairment”. An accompanying editorial in Nature called Aβ*56 “a star suspect” in Alzheimer’s. A fortnight later, Ashe won the Potamkin Prize for neuroscience, partly for work leading to Aβ*56.
The Nature paper has been cited in about 2,300 scholarly articles, more than all but four other Alzheimer’s basic research reports published since 2006. Since then, annual NIH support for studies labelled “amyloid, oligomer, and Alzheimer’s” has risen from near zero to $287m n 2021. Lesné and Ashe helped spark that explosion, experts say.
The paper provided an “important boost” to the amyloid and toxic oligomer hypotheses when they faced rising doubts, Südhof says. “Proponents loved it – it seemed to be an independent validation of what they have been proposing for years.”
As Ashe’s star burned more brightly, Lesné’s rose. He joined UMN with his own NIH-funded lab in 2009. Aβ*56 remained a primary research focus. Megan Larson, who worked as a junior scientist for Lesné and is now a product manager at Bio-Techne, a biosciences supply company, calls him hardworking, and charismatic. She and others in the lab often ran experiments and produced Western blots, Larson says, but in their papers together, Lesné prepared all the images for publication.
He became a leader of UMN’s neuroscience graduate programme in 2020, and in May 2022, four months after Schrag delivered his concerns to NIH, Lesné received a coveted R01 grant from the agency, with up to five years of support. The NIH programme officer for the grant, Austin Yang, a co-author on the 2006 Nature paper, declined to comment.
In December 2021, Schrag visited PubPeer, a website where scientists flag possible errors in published papers. Many of the site’s posts come from technical gumshoes who deconstruct Western blots for telltale marks indicating that bands representing proteins could have been removed or inserted where they don’t belong. Such manipulations can falsely suggest a protein is present, or alter the levels at which a detected protein is apparently found. Schrag, still focused on Cassava-linked scientists, was looking for examples that could refine his own sleuthing.
In a PubPeer search for “Alzheimer’s,” postings about articles in The Journal of Neuroscience caught Schrag’s eye. They questioned the authenticity of blots used to differentiate Aβ and similar proteins in mouse brain tissue. Several bands seemed to be duplicated.
Using software tools, Schrag confirmed the PubPeer comments and found similar problems with other blots in the same articles. He also found some blot backgrounds that had apparently been improperly duplicated.
Three of the papers listed Lesné, whom Schrag had never heard of, as first or senior author. Schrag quickly found that another Lesné paper had also drawn scrutiny on PubPeer, and he broadened his search to Lesné papers that had not been flagged there. The investigation “developed organically”, he says, as other apparent problems emerged.
Eventually Schrag ran across the seminal Nature paper, the basis for many others. It, too, seemed to contain multiple doctored images.
Science asked two independent image analysts, Bik and Jana Christopher, to review Schrag’s findings about that paper and others by Lesné. They say some supposed manipulation might be digital artifacts that can occur inadvertently during image processing, a possibility Schrag concedes. But Bik found his conclusions sound. Christopher concurred about the many duplicated images and some markings suggesting cut-and-pasted Western blots flagged by Schrag. She also identified additional dubious blots and backgrounds he had missed.
In the 16 years after the landmark paper, Lesné and Ashe, separately or jointly, published many articles on their stellar oligomer. Yet only a handful of other groups have reported detecting Aβ*56.
Citing the ongoing UMN review of Lesné’s work, Ashe declined via email to be interviewed or to answer questions from Science, which she called “sobering.” But she wrote: “I still have faith in Aβ*56,” noting her ongoing work studying the structure of Aβ oligomers. “We have promising initial results. I … believe it has the potential to explain why Aβ therapies may yet work despite recent failures targeting amyloid plaques.”
But even before Schrag’s investigation, the spotty evidence that Aβ*56 plays a role in Alzheimer’s was questioned. Wilcock has long doubted studies that claim to use “purified” Aβ*56. Such oligomers are notoriously unstable, converting to other oligomer types spontaneously.
Multiple types can be present in a sample even after purification efforts, making it hard to say any cognitive effects are due to Aβ*56 alone, she notes – assuming it exists. In fact, Wilcock and others say, several labs have tried and failed to find Aβ*56, although few have published those findings. Journals are often uninterested in negative results, and researchers can be reluctant to contradict a famous investigator.
An exception was Harvard University’s Dennis Selkoe, a leading advocate of the amyloid and toxic oligomer hypotheses, who has cited the Nature paper at least 13 times. In two 2008 papers, Selkoe said he could not find Aβ*56 in human fluids or tissues.
Selkoe examined Schrag’s dossier on Lesné’s papers at Science’s request, and says he finds it credible and well supported. He did not see manipulation in every suspect image, but says: “There are at least 12 or 15 images where I would agree that there is no other explanation” than manipulation. One – an image in the Nature paper displaying purified Aβ*56 – shows “very worrisome” signs of tampering, Selkoe says.
The same image reappeared in a different paper, co-authored by Lesné and Ashe, five years later. Many other images in Lesné’s papers might be improper, more than enough to challenge the body of work, Selkoe adds.
A few of Lesné’s questioned papers describe a technique he developed to measure Aβ oligomers separately in brain cells, spaces outside the cells, and cell membranes. Selkoe recalls Ashe talking about her “brilliant postdoctoral fellow” who devised it. He was sceptical of Lesné’s claim that oligomers could be analysed separately inside and outside cells in a mixture of soluble material from frozen or processed brain tissue.
“All of us who heard about that knew in a moment it made no biochemical sense. If it did, we’d all be using a method like that,” Selkoe says. The Nature paper depended on that method.
Selkoe himself co-authored a 2006 paper with Lesné in the Annals of Neurology. They sought to neutralise the effects of toxic oligomers, although not Aβ*56. The paper includes an image that Schrag, Bik and Christopher agree was reprinted as if original in two subsequent Lesné articles. Selkoe calls that “highly egregious”.
Given those findings, the scarcity of independent confirmation of the Aβ*56 claims seems telling, Selkoe says. “In science, once you publish your data, if it’s not readily replicated, then there is real concern it’s not correct or true. There’s little clear-cut evidence that Aβ*56 exists, or if it exists, correlates in a reproducible fashion with features of Alzheimer’s – even in animal models.”
In all, Schrag or Bik identified more than 20 suspect Lesné papers; 10 concerned Aβ*56. Schrag contacted several of the journals early this year, and Lesné and his collaborators recently published two corrections. One for a 2012 paper in The Journal of Neuroscience replaced several images Schrag had flagged as problematic, writing that the earlier versions had been “processed inappropriately”. But Schrag says even the corrected images show numerous signs of improper changes in bands, and in one case, complete replacement of a blot.
A 2013 Brain paper in which Schrag had flagged multiple images was also extensively corrected in May. Lesné and Ashe were the first and senior authors of the study, which showed “negligible” levels of Aβ*56 in children and young adults, more when people reached their 40s, and steadily increasing levels after that. It concluded that Aβ*56 “may play a pathogenic role very early in the pathogenesis of Alzheimer’s disease”. The authors said the correction had no bearing on the study’s findings.
Schrag isn’t convinced. Among other problems, one corrected blot shows multiple bands that appear to have been added or removed artificially, he says.
Selkoe calls the apparently falsified corrections “shocking”, particularly in light of Ashe’s pride in the 2006 Nature paper. “I don’t see how she would not hyper-scrutinise anything that subsequently related to Aβ*56,” he says.
After Science contacted Ashe, she separately posted to PubPeer a defence of some images Schrag had challenged in the Nature paper. She supplied portions of a few original, unpublished versions that do not show the apparent digital cut marks Schrag had detected in the published images. That suggests the markings were harmless digital artifacts. Yet the original images reveal something Schrag and Selkoe find even more incriminating: unequivocal evidence that, despite the lack of obvious cut marks, multiple bands were copied and pasted from adjacent areas.
Schrag could find no explanation for a two-decade litany of oddities. In experiment after experiment using Western blots, microscopy, and other techniques, serious anomalies emerged. But he says he has not examined the original, uncropped, high-resolution images.
Questions about Lesné’s work are not new. Cell biologist Denis Vivien, a senior scientist at Caen, co-authored five Lesné papers flagged by Schrag or Bik. Vivien defends the validity of those articles, but says he had reason to be wary of Lesné.
Toward the end of Lesné’s time in France, Vivien says they worked together on a paper for Nature Neuroscience involving Aβ.
During final revisions, he saw immunostaining images, in which antibodies detect proteins in tissue samples, that Lesné had provided. They looked dubious to Vivien, and he asked other students to replicate the findings. Their efforts failed. Vivien says he confronted Lesné, who denied wrongdoing. Although Vivien lacked “irrefutable proof” of misconduct, he withdrew the paper before publication “to preserve my scientific integrity”, and broke off all contact with Lesné. “We are never safe from a student who would like to deceive us and we must remain vigilant.”
Schrag spot checked papers by Vivien or Ashe without Lesné. He found no anomalies –suggesting Vivien and Ashe were innocent of misconduct.
Yet senior scientists must balance the trust essential to fostering a protégé’s independence with prudent verification, Wilcock says. If you sign off on images time after time, claim credit, speak publicly, and win awards for the work, as Ashe has done, you have to be sure it’s right, she adds.
“Ashe obviously failed in that very serious duty” to ask tough questions and ensure the data’s accuracy, Forsayeth says. “It was a major ethical lapse.”
In his whistle-blower report to NIH about Lesné’s research, Schrag made its scope and stakes clear: “(This) dossier is a fraction of the anomalies easily visible on review of the publicly accessible data,” he wrote. The suspect work “not only represents a substantial investment in (NIH) research support, but has been cited … thousands of times and thus has the potential to mislead an entire field of research”.
The agency’s reply, which Schrag shared with Science, noted that complaints deemed credible and will go to the Department of Health and Human Services Office of Research Integrity (ORI) for review. That agency could then instruct grantee universities to investigate before a final ORI review, a process that can take years and remains confidential absent an official misconduct finding. To Science, NIH said it takes research misconduct seriously, but otherwise declined to comment.
A spokesperson for Nature, which publishes image integrity standards, says the journal takes concerns raised about its papers seriously, but otherwise had no comment. Days after an inquiry from Science, Nature published a note saying it was investigating Lesné’s 2006 paper and advising caution about its results.
The Journal of Neuroscience stands out with five suspect Lesné papers. A journal spokesperson said it follows guidelines from the Committee on Publication Ethics to assess concerns, but otherwise had no comment.
Holden Thorp, editor-in-chief of the Science journals, said the journals have subjected images to increasing scrutiny, adding that “2017 would have been (near) the beginning of when more attention was being paid to this – not just for us, but across scientific publishing”. He cited the Materials Design Analysis Reporting framework developed jointly by several publishers to improve data transparency and weed out image manipulation.
Like other anti-Aβ efforts, toxic oligomer research has spawned no effective therapies. “Many companies have invested millions and millions of dollars, or even billions … to go after soluble Aβ (oligomers). And that hasn’t worked,” says Daniel Alkon, president of the bioscience company Synaptogenix, who once directed neurologic research at NIH.
Schrag says oligomers might still a play role in Alzheimer’s. After the Nature paper, other investigators connected combinations of oligomers to cognitive impairment in animals. “The wider story (of oligomers) potentially survives this one problem,” Schrag says.
Selkoe adds that the broader amyloid hypothesis remains viable. “I hope that people will not become faint-hearted as a result of what really looks like an egregious example of malfeasance that’s squarely in the Aβ oligomer field,” he says. But if current phase 3 clinical trials of three drugs targeting amyloid oligomers all fail, he notes, “the Aβ hypothesis is very much under duress”.
Schrag’s sleuthing implicates work by Cassava Senior Vice President Lindsay Burns, Hoau Yan Wang of the City University of New York (CUNY), and Harvard University neurologist Steven Arnold. Wang and Arnold have advised Cassava, and Wang collaborated with the company for 15 years.
None agreed to answer questions from Science. Cassava CEO Remi Barbier also declined to answer questions or to name the company’s current scientific advisers. He said in an email that Schrag’s dossier is “generally consistent with prior allegations about our science … such allegations are false.” Cassava hired investigators to review its work, provided “nearly 100,000 pages of documents to an alphabet soup of outside investigative agencies”, and asked CUNY to investigate, he added. That effort “has yielded an important finding to date: there is no evidence of research misconduct”. (CUNY says it takes allegations of misconduct seriously, but otherwise declined to comment because of its ongoing investigation.)
Last year, Schrag contacted most of the journals that published questioned papers. Seven were retracted, including five by PLOS ONE in April. Three others received expressions of concern; in each case, the editors said they were awaiting completion of the CUNY investigation. In a few cases, the editors told him, reviews were under way.
Cassava has said editors of two suspect papers dismissed misconduct concerns. Last year, the editors of a 2005 Neuroscience paper co-authored by Wang, Burns, and others found no improper manipulation of Western blots, but said in an editorial note they would review any concerns from an “institutional investigation”, apparently CUNY’s probe. They did not respond to additional findings Schrag raised this year.
Another paper that purportedly validated science behind Simufilam, also by Wang, Burns, and colleagues, appeared in 2012 in The Journal of Neuroscience. In December 2021, the editors corrected one figure. Barbier said in a statement they told him they had found no manipulation. But in January, after Schrag and others raised additional doubts, the editors issued an expression of concern, reserving judgment until CUNY completes its investigation.
The most influential Cassava-related paper appeared in The Journal of Clinical Investigation in 2012. The authors – including Wang; Arnold; David Bennett, who leads a brain-tissue bank at Rush University; and his Rush colleague, neuroscientist Zoe Arvanitakis – linked insulin resistance to Alzheimer’s and the formation of amyloid plaques. Cassava scientists say Simufilam lessens insulin resistance. They relied on a method in which dead brain tissue, frozen for a decade and then partially thawed and chopped, purportedly transmits nerve impulses.
Schrag and others say it contradicts basic neurobiology. Schrag adds that he could find no evidence that other investigators have replicated that result. (None of the authors agreed to be interviewed for this article.)
That paper supported the science behind Simufilam, Schrag says, “and spawned an entire field of research in Alzheimer’s, ‘diabetes of the brain’.” It has been cited more than 1,500 times. Schrag sent the journal’s editor his analysis of more than 15 suspect images.
In an email that Schrag provided to Science, the editor said the journal had reviewed high-resolution versions of the images when they were originally submitted and declined to consider Schrag’s findings.
See more from MedicalBrief archives: