Just days after Science fully retracted the controversial 2009 paper suggesting that a virus called XMRV plays a role in chronic fatigue syndrome (CFS), the only other paper supporting a link between a mouse-related virus and the mysterious condition has been officially stricken from the scientific record. Yesterday, the Proceedings of the National Academy of Sciences (PNAS) published a retraction notice, signed by all seven authors, for the 2010 study, which, like the Science paper, had come under fire from virologists.
"In my mind, they should have done this months ago," says Jonathan Stoye of the Medical Research Council National Institute for Medical Research in London, who co-authored a study in PLoS ONE in May that questioned the results of the PNAS paper.
The PNAS study, whose lead researcher was Shyh-Ching Lo of the Food and Drug Administration (FDA), played a peculiar role in the 2-year-old saga that erupted over XMRV's purported link to CFS. News about the study first appeared on the Web after one of the authors, Harvey Alter of the Clinical Center of the National Institutes of Health (NIH), discussed the work at a meeting in Croatia in May 2010. At the time, the Science paper on XMRV, authored by Judy Mikovits of the Whittemore Peterson Institute in Reno, Nevada, and colleagues, was under siege because a raft of other studies had failed to replicate the data. Alter, a highly respected virologist and winner of an Albert Lasker award, said that Mikovits was right after all, creating elation among patients eager to find the cause of their elusive disease.
But when PNAS finally published the paper in August 2010, some retrovirologists disputed the authors' claim that their study supported Mikovits's work. The viral sequences detected in CFS patients by Lo, Alter, and their colleagues did not belong to XMRV but to another broad group of viruses, the murine leukemia virus (MLV)-related viruses.
Still, the work did seem to hint at a viral link to the debilitating condition. Using PCR, the team found MLV-related DNA sequences in 32 of 37 CFS patients, and in only three of 44 healthy controls. The blood samples from the CFS patients dated back to the 1990s, but the team was able to draw fresh blood from eight of the patients and found DNA evidence of MLVs in seven of those. Over the 15 years, the virus appeared to have evolved, the researchers wrote, which is what would be expected during a long-lasting retroviral infection—but not if the findings were the cause of contamination.
But the latter finding actually proved to be an Achilles' heel. In the PloS ONE paper, Stoye and others argued that the viral DNA sequences identified in the fresh samples were very unlikely to have evolved from the viruses found 15 years before. A more elaborate phylogenetic analysis from a team led by Greg Towers of University College London, published in the Journal of Virology last October, argued the same thing. "The only realistic explanation," that paper concluded, was that the patient samples or PCR reagents "were contaminated with mouse DNA."
Towers says that Lo, Alter, and their co-authors never publicly responded to his team's paper. But they appear to have accepted its conclusions, citing Towers's phylogeny as one of the reasons for the retraction.
The retraction notice cites other reasons as well. There wasn't enough left of the original patient samples to have them tested by independent researchers, the team writes, and additional work to find antiviral antibodies in patients and to isolate the actual virus failed. Moreover, the researchers themselves weren't able to find MLVs in a blinded, multilab study of the so-called Blood XMRV Scientific Research Working Group, in which they participated, and which included samples from five of their original CFS patients. The group published its results in Science in September.
Lo and Alter did not respond to interview requests. Press officers at NIH said Alter was on vacation and sent ScienceInsider a statement "ïn lieu of interviews," which closely resembled the text of the retraction and contained no additional details.
The retraction takes away the only paper still left suggesting a role for murine viruses in CFS. Support for Lo and Alter's study has also come from Cornell University's Maureen Hanson, who at meetings has also reported finding MLV-like sequences in CFS patients. In an e-mail to ScienceInsider, Hanson writes that she has not submitted those results for publication, "because we cannot determine whether or not these findings were due to contamination." Hanson believes a second large multilab study, led by Ian Lipkin of Columbia University, will provide the final answer. "I am reserving judgment until it is completed," she says.