Review questions key work by University of Minnesota on Alzheimer's

Allegations of duplicated or manipulated images have cast suspicion on University of Minnesota discoveries that have been fundamental to Alzheimer's disease research.

An investigation published on Thursday by Science questioned the U discoveries, primarily by researcher Sylvain Lesné, about the role of a protein in inhibiting memory and contributing to Alzheimer's-related dementia.

Images in a key study showed growth in a protein, known as Aβ*56 (or alpha-beta star 56), at the same time as mice aged and experienced dementia symptoms. The investigation called the images "apparently falsified," raising questions about the link between the protein and symptoms.

The report was based on concerns by Dr. Matthew Schrag, who reviewed the images outside his role at the Vanderbilt Memory and Alzheimer's Center in Tennessee. He reported the concerns to the National Institutes of Health, which funded much of the U research.

"What I saw was a pervasive pattern through quite a lot of articles," Schrag said Thursday. "It turned out a lot of those articles were very high-profile, high-impact work that had a lot to do with how many people formulated the problem of Alzheimer's disease."

The scrutiny prompted the scientific journal Nature last week to investigate a key 2006 paper by Lesné and U colleagues, and to encourage readers "to use caution when using results" in the meantime.

A spokesperson for the U Medical School declined comment Thursday other than to say that the institution was aware of the claims and was following its standard process for reviewing them. An e-mail to Lesné wasn't immediately returned Thursday.

The allegations strike at one of the medical school's signature research achievements of the past quarter-century — defining the role of amyloid-beta proteins in the process of dementia and giving scientists a much-needed target for treatment of Alzheimer's disease.

Much of the work has centered on mouse-maze research by Karen Ashe, a distinguished professor considered by many to be on the short list for a Nobel Prize for her work. She was a co-author of many of the disputed papers.

Schrag said he went public with his concerns to hasten the review and sort out what is accurate so patients can confidently enter future clinical trials of Alzheimer's treatments. His concerns emerged late last year on PubPeer, a website in which scientists raise concerns about colleagues' work — including images of Western blot laboratory tests that identify protein levels in blood or tissue.

Schrag magnified, colorized and inverted those images from Lesné's studies in ways that revealed concerns about their authenticity.

Ashe responded to some of the concerns on PubPeer. In one instance, Schrag identified splice marks indicating that blot results were cut and pasted onto an image. Ashe said the marks didn't exist in the manuscript submitted to Nature.

"I conclude that the linear splice marks were introduced during the processing of the manuscript," she replied. "Importantly, I hope you will agree that my analysis indicates that the 'splice marks' are artifacts and that published data are valid."

The Science article built on Schrag's concerns by reviewing the images with other leading dementia researchers. The experts included Harvard University's Dr. Dennis Selkoe, who shares many of the U researchers' hypotheses about the origins of Alzheimer's and has cited the 2006 Nature paper in his work.

Selkoe told Science that he disagreed with allegations of manipulation for some of the images, but there "are certainly at least 12 or 15 images where I would agree that there is no other explanation."

The concerns wouldn't scuttle the entire theory of amyloid proteins having neurotoxic effects that lead to Alzheimer's-related dementia, but Selkoe told Science that they undercut the existence of the Aβ*56 protein that is central to Lesné's research.

Regarding Ashe, he told Science, "I don't see how she would not hyperscrutinize anything that subsequently related to Aβ*56."