Stanford University president Marc Tessier-Lavigne, a senior author of a study published in the journal Nature in 2008, has been accused by former scientists and colleagues at US biotechnology company Genentech, where he used to be a top executive, of falsifying data in the article.
The Stanford Daily student newspaper hat an internal review knew about the fraud but never exposed it, and neither did Tessier-Lavigne retract it.
At least five other scholarly papers co-authored by Tessier-Lavigne are already under scrutiny, including three published in 2001 and 1999, in which he is a senior author. Based on earlier reporting by the Stanford Daily, which in November identified four of the papers in question, the Stanford trustees hired a special committee to look into the accuracy of the research.
The 2009 paper in question created international interest on publication, claiming to have found the potential cause for brain degeneration in Alzheimer’s patients.
“Because of this research,” read Genentech’s annual letter to shareholders, “we are working to develop both antibodies and small molecules that may attack Alzheimer’s from a novel entry point and help the millions of people suffering from this devastating disease.”
But after several unsuccessful attempts to reproduce the research, the paper became the subject of an internal review by Genentech’s Research Review Committee (RRC), say four high-level Genentech employees at the time; two were senior scientists and two were scientists also serving as executives.
The scientists said the inquiry discovered falsification of data in the research, and that Tessier-Lavigne kept the finding from becoming public.
Tessier-Lavigne denies both allegations. Genentech said in a statement that “as part of our diligence related to these allegations, we reviewed the records from that November 2011 RRC meeting and saw no allegations of fraud or wrongdoing”.
After the review, which began in 2011, Genentech cancelled research based on the paper’s findings.
Till Maurer, a senior scientist at the company from 2009-2018 assigned to develop drugs based on the 2009 paper, told The Daily his superior told him “the project is being cancelled … because they found falsified data”.
Tessier-Lavigne, who became Stanford’s president in 2016, has been under investigation by the Stanford Board of Trustees since late November, after The Daily revealed concerns that several other papers he had co-authored contained altered imagery.
But these latest allegations, about a different paper, are more serious because they involve what was once considered a promising treatment target for Alzheimer’s disease, and because people involved in the review allege that Tessier-Lavigne tried to conceal its findings.
Tessier-Lavigne declined multiple requests for interviews, and Stephen Neal, chairman emeritus of Cooley, a prominent law firm representing the president, responded in writing to questions sent to Tessier-Lavigne.
“Dr Tessier-Lavigne is not aware of any internal investigation of the paper,” wrote Neal. “Given that there was no investigation of the paper, he disputes any allegation … that he ‘covered up’ any findings or was opposed to allowing such (non-existent) findings to become public.”
Genentech, in a written statement to The Daily, confirmed a “routine” internal review took place in 2011, a fact not previously public. However, the scientist whom The Daily confirmed belonged to the research review committee, denied this happened.
“There have not been any formal investigations, allegations, claims or complaints regarding scientific fraud or misrepresentation involving the Nature 2009 paper,” wrote Susan Willson, a Genentech spokesperson.
The research review committee authorised subsequent experiments to further analyse the paper’s conclusions, she wrote, and “based on the results of the genetic experiments… terminated the Genentech research project in 2012”.
Matthew Schrag, an Alzheimer’s expert with no relation to Genentech or Tessier-Lavigne, who reviewed the scientific literature at the request of The Daily, concluded that parts of the paper implicating a specific pathway in Alzheimer’s “were found in later studies to be inaccurate”.
The four senior scientists said the committee found the fundamental science underpinning the 2009 study’s conclusion was fabricated. In the face of this, retraction of the paper would have been the expected outcome, according to Nature’s policies.
The Committee of Publication Ethics, a non-profit that supports worldwide journal editors, recommends retraction in the case of “clear evidence that the findings are unreliable, either as a result of major error (e.g, miscalculation or experimental error), or as a result of fabrication (e.g, of data) or falsification (e.g, image manipulation)”.
But Tessier-Lavigne “was unwilling to clean up the mess,” said the scientist and former Genentech executive who participated in a series of 2011 meetings about the paper.
Neal, Tessier-Lavigne’s lawyer, asserted the conclusions of the 2009 paper that did not hold up. But he said: “No one involved in the experiments … forged gels, falsified assays or fabricated experiments.”
A “miracle result” that didn’t pan out
Two days after the 2009 paper came out, Genentech’s letter to shareholders called it “groundbreaking research about an entirely new way of looking at the cause of Alzheimer’s”.
Paul Greengard, a Nobel Laureate, called it “an exciting paper,” that would have a “major impact on the Alzheimer’s field.”
And Nature published an article titled, Alzheimer’s theory makes a splash.
The study was co-authored with Genentech postdoctoral student Anatoly Nikolaev and two Salk Institute scientists.
After publication, Nikolaev was hired by Genentech as a scientist and Tessier-Lavigne, then executive vice-president of research drug discovery, was promoted to chief scientific officer, overseeing 1 400 scientists.
The paper received 1 245 citations in a field where most scarcely get 10. Soon after it was released, Tessier-Lavigne and Nikolaev submitted a 187 page patent to the World Intellectual Property Organisation with the title Method for inhibiting neurodegeneration.
Simultaneously they filed applications for patents in the US, Brazil, Taiwan, Israel, Canada and Australia.
But the research has not turned into an Alzheimer’s treatment. And several of the patent applications, including the US patent application, were abandoned after the internal review.
By 2011, Alzheimer’s experts had raised suspicions over the paper, both inside the company and outside. (Genentech declined to say whether any concerns had been raised, though the company denied “any formal … complaints regarding fraud or misrepresentation”).
Independent experts who analysed the published works questioned several aspects of the paper. For example, it did not include a robust statistical analysis of its findings, included a low number of samples and, according to Schrag, the Alzheimer’s expert, “the degree of variation (between experiments) is extremely low”.
Genentech’s research review committee, to which Tessier-Lavigne belonged until his departure from the company, conducted an internal inquiry into the underlying experiments of the paper, said the scientist from the review committee.
The review was “a deliberate attempt by some experienced biologists to reproduce the key findings in the paper” and “a re-examination of the paper’s key findings”.
Genentech, in its written statement, called the inquiry a “normal review process.”
Till Maurer, the senior scientist tasked with drug discovery based on the research, said that after the paper’s publication, he was “invited to a meeting with the new star scientist at Genentech – Anatoly”.
Maurer’s supervisor told him Nikolaev had found “the involvement of a specific protein in Alzheimer’s disease”, and “the company immediately initiated a drug discovery programme around that protein.”
But the team assembled to target the protein was later abruptly re-assigned, said Maurer.
The review committee found that “lab data had been falsified,” Maurer was told.
The paper’s specific issues related to its central conclusions: it included several experiments claiming to show the amino-terminal fragment of the amyloid precursor protein (N-APP) binds to death receptor six (DR6), this bind causing neurodegeneration.
But later-published studies by Tessier-Lavigne’s lab and other research groups showed this was inaccurate. A crystal structure – a type of labour-intensive experiment determining the exact arrangement of atoms – published in 2015 proved that APP, the complete protein, binds to DR6 at the E2 site.
The N-APP fragment used in the 2009 research did not include the E2 site, meaning DR6 could not have bound to the fragment and caused the results described in the paper.
“It was a shock to every scientist,” said Maurer of the alleged findings.
Genentech, in its written statement to The Daily, said “no claims or complaints of scientific fraud or misrepresentation were ever made by or to the RRC”.
Nikolaev, responding to questions by email, first said “nothing in our paper is ‘false’ or ‘falsified’,” and subsequently, “I can only speak for myself… I did not do anything wrong w at Genentech.”
Neal, Tessier-Lavigne’s lawyer, wrote that “the paper’s original results were accurately reported.” On whether, even aside from scientists’ concerns that the results of experiments had been falsified, four panels of the paper appeared to contain visible duplications and whether this constituted inaccurate reporting,
Neal wrote: “Dr Tessier-Lavigne first saw reference to those allegations when he saw the new post in PubPeer,” an online site where scientists identify issues in published papers.
“He is looking into those issues.”
Schrag, in reviewing the 2009 paper, noted several panels that were supposed to represent different experiments but appeared to contain the same result. He posted his concerns on PubPeer.
Elisabeth Bik, a prominent research misconduct investigator, independently identified the same alleged duplications when she reviewed the paper.
The executive who was part of the committee reviewing the paper, said the thorough inquiry left little room for doubt.
Lab technicians and assistants were interviewed while scientists independent of the lab attempted to verify the study’s findings.
“None of (the research review committee members) believed these data were true by the time people had attempted to reproduce it,” the executive said.
He said the understanding was that the paper’s supposed finding of N-APP’s role in Alzheimer’s had been “faked”.
Tessier-Lavigne: “I understand Genentech has also communicated to you that (the review) did not raise issues.”
Genentech did not say that the review raised no issues and declined to answer questions about whether the review had uncovered any issues.
The internal review’s aftermath
Fabrication constitutes the most troubling form of manipulation in the scientific world.
Tessier-Lavigne, who had just let Genentech to become president of Rockefeller University in New York when the review was instigated in 2011, flew back to California to “clean it up,” said one of the senior scientists.
Faced with such a finding, Genentech leadership urged Tessier-Lavigne to retract the paper and its conclusions. But neither a correction nor a retraction was issued, and the paper stands to this day.
The scientists said the journal had not been informed of any potential issues.
Nature’s editor in chief, Magdalena Skipper, said that “while we do not usually comment on details of individual cases for confidentiality reasons… we have not received a report relating to an internal investigation of this paper”.
Nikolaev left Genentech in 2011 and now works in Florida as a radiation oncologist. When contacted, he denied a review had occurred and also said “the (2009) paper doesn’t talk about Alzheimer’s disease” and that Alzheimer’s was not “even mentioned”.
But the word “Alzheimer’s” appears 31 times, including in the abstract.
Alzheimer’s was also routinely referenced in comments by Tessier-Lavigne and Genentech upon release of the research, and in their responses to this article.
Those with knowledge of the review were sworn to secrecy, according to the scientists, several of whom cited confidentiality agreements.
“I don’t think anyone at Genentech would be interested in this becoming public,” said Maurer.
No retraction was issued, but Tessier-Lavigne walked back the research in several subsequent publications, said the scientists.
His lawyer wrote that the “revisions to the conclusions” were not “nefarious”, “this is how science works.”
Specific conclusions in the original paper were first revised by a 2012 Journal of Neuroscience paper that included all co-authors of the original 2009 study except Nikolaev (and for which Tessier-Lavigne served as senior author).
In a seemingly direct rebuke to the Alzheimer’s hypothesis in the 2009 paper, the title of a 2014 paper also published in the Journal of Neuroscience by all of the original authors except Nikolaev says the molecule DR6 identified in the 2009 paper “does not contribute to Alzheimer’s disease-related pathophysiology in murine models”.
Schrag, the Alzheimer’s researcher said that “to his credit, Tessier-Lavigne authored several of the later studies, revising the findings of his 2009 paper”.
The Genentech scientists interpreted the subsequent papers differently, one saying the original “absolutely should’ve been retracted” based on what had been proven inaccurate.
The top Genentech executive who sat on the review committee observed: “If everything were true in the original paper, (Tessier-Lavigne) would have won the Nobel.”
Aidan Ryan, a spokesperson for the special committee of the Stanford Board of Trustees charged with investigating Tessier-Lavigne’s research, would not respond on whether the committee was aware of the allegations and whether the board had been informed of any such concerns when originally considering Tessier-Lavigne for the presidency.
Study details
APP binds DR6 to trigger axon pruning and neuron death via distinct caspases
Anatoly Nikolaev, Todd McLaughlin, Dennis O’Leary, Marc Tessier-Lavigne.
Published in Nature on 19 February 2009
Abstract
Naturally occurring axonal pruning and neuronal cell death help to sculpt neuronal connections during development, but their mechanistic basis remains poorly understood. Here we report that beta-amyloid precursor protein (APP) and death receptor 6 (DR6, also known as TNFRSF21) activate a widespread caspase-dependent self-destruction program. DR6 is broadly expressed by developing neurons, and is required for normal cell body death and axonal pruning both in vivo and after trophic-factor deprivation in vitro. Unlike neuronal cell body apoptosis, which requires caspase 3, we show that axonal degeneration requires caspase 6, which is activated in a punctate pattern that parallels the pattern of axonal fragmentation. DR6 is activated locally by an inactive surface ligand(s) that is released in an active form after trophic-factor deprivation, and we identify APP as a DR6 ligand. Trophic-factor deprivation triggers the shedding of surface APP in a beta-secretase (BACE)-dependent manner. Loss- and gain-of-function studies support a model in which a cleaved amino-terminal fragment of APP (N-APP) binds DR6 and triggers degeneration. Genetic support is provided by a common neuromuscular junction phenotype in mutant mice. Our results indicate that APP and DR6 are components of a neuronal self-destruction pathway, and suggest that an extracellular fragment of APP, acting via DR6 and caspase 6, contributes to Alzheimer's disease.
Nature article – Alzheimer’s theory makes a splash (Open access)
See more from MedicalBrief archives:
Probe into possible data manipulation in Stanford president’s research
Leading Alzheimer’s study under investigation over possible manipulation
The red flags that led to a massive Alzheimer’s research scandal