From ‘Retraction Watch', 22 September 2011
If past experience is any indication, billions of pixels will be spilled in the coming days as scientists and advocates debate the latest twist in the story of XMRV, or xenotropic murine leukemia-related virus, and chronic fatigue syndrome (CFS). Today’s news is that Science is partially retracting a 2009 paper by Judy Mikovits and colleagues, including Vincent Lombardi, purporting to show a link between the virus and the syndrome — a paper about which they issued an Expression of Concern in May. The retraction is of a table and a figure — more on that in a bit.
In an excellent blow-by-blow account in Science of the nearly 20-year-long saga, also out today, Jon Cohen and Martin Enserink review the unusual circumstances of that Expression of Concern.
Science editor-in-chief Bruce Alberts and Science Executive Editor Monica Bradford had first suggested that Mikovits and her co-authors retract the paper voluntarily. “Science feels it would be in the best interest of the scientific community,” they wrote in a 26 May letter. Mikovits was livid and questioned Alberts’s motives. “Who wrote that letter? I don’t think it was Science,” she says. The co-authors thought the retraction request was premature, too. “What if we walk away from this based on contamination and it’s not contamination?” Lombardi asked. “You’ve got to give us time to figure this out.”
Alberts stresses that they floated the retraction idea because Science already planned to publish the Expression of Concern. “It wasn’t a public call for retraction,” he notes, emphasizing that the recipients shared it with the media. He also does not think it would have been premature, although he says it’s often a tough call whether to retract a paper. “Ultimately, it requires expert judgment and a lot of sensitivity to the issues,” he says. “We had lost confidence in the results.”
As Science noted in May, two studies accompanied the May Expression of Concern
that strongly support the growing view that the association between XMRV and CFS described by Lombardi et al. likely reflects contamination of laboratories and research reagents with the virus.
Today’s partial retraction is also accompanied by a paper casting doubt on the 2009 study, this one by Michael Busch and colleagues — including Mikovits and Lombardi — from the Blood Working Group. The authors took blood samples from 15 people who had been found to be XMRV-positive, 14 of which had CFS, and 15 donors negative for the virus, and had 9 labs test them blindly. As their abstract concludes:
Only two laboratories reported evidence of XMRV/MLVs; however, replicate sample results showed disagreement and reactivity was similar among CFS subjects and negative controls. These results indicate that current assays do not reproducibly detect XMRV/MLV in blood samples and that blood donor screening is not warranted.
In today’s Science feature, Busch is quoted praising Mikovits and her one of her co-authors on the 2009 Science paper, Francis Ruscetti of the National Cancer Institute, as well as the National Institute’s of Health’s Harvey Alter and the FDA’s Shyh-Ching Lo — authors of a 2010 paper in the Proceedings of the National Academy of Sciences linking XMRV and CFS:
Blood Working Group member Michael Busch, head of the Blood Systems Research Institute in San Francisco, California, says Ruscetti, Mikovits, Lo, and Alter deserve kudos for participating fully in the study. “I commend them on their scientific integrity and commitment to the scientific process,” Busch says. “This has been a difficult and disappointing process for them and for CFS patients, but hopefully we have all learned lessons that will guide future research and lead to discovery of the cause and cure of this disease.”
But it turns out that the reason for the partial retraction — an unusual move we’ll have more on in a second — is that not all of the authors of the 2009 Science paper were as forthcoming when that work was being checked. Bradford, Science‘s executive editor, tells Retraction Watch that two of those authors, Robert Silverman and Jaydip Das Gupta of the Cleveland Clinic:
determined that the experimental results contributed by them to the Lombardi, et al. paper were the result of contamination, and they requested that we retract those parts of the paper. While we were aware that other co-authors had tested samples and claimed to not find evidence of plasmid contamination, those co-authors were unwilling to provide their data for examination so we were unable to comment on the validity of the other experiments. Since all of the other co-authors agreed with Silverman’s conclusion that his samples were contaminated, we felt it was in the best interest of our readers and the CFS community to publish this information and to label the notice to make it clear that part of the original paper was not valid.
Here’s the text of today’s partial retraction:
In our 23 October 2009 Report, “Detection of an infectious retrovirus, XMRV, in blood cells of patients with chronic fatigue syndrome” (1), two of the coauthors, Silverman and Das Gupta, analyzed DNA samples from chronic fatigue syndrome (CFS) patients and healthy controls. A reexamination by Silverman and Das Gupta of the samples they used shows that some of the CFS peripheral blood mononuclear cell (PBMC) DNA preparations are contaminated with XMRV plasmid DNA (2). The following figures and table were based on the contaminated data: Figure 1, single-round PCR detection of XMRV sequences in CFS PBMC DNA samples; table S1, XMRV sequences previously attributed to CFS patients; and figure S2, the phylogenetic analysis of those sequences. Therefore, we are retracting those figures and table.
We’ve asked Mikovits to explain the circumstances surrounding the samples, and will update with anything we hear back.
Partial retractions are rare. We haven’t seen any since starting Retraction Watch in August 2010, although comments by a Duke official suggest a number are on the way in the Anil Potti case. The closest we’ve come is a Nature Medicine paper that might have been partially retracted, based on the notice, except that the author told us that particular journal doesn’t do partial retractions.
Partial retractions are rare at Science, too. Bradford tells Retraction Watch this is only the third since she joined the journal’s staff in 1989.
The Committee on Publication Ethics recommends against partial retractions:
Partial retractions are not helpful because they make it difficult for readers to determine the status of the article and which parts may be relied upon.
Still, we can see why Science handled this one the way they did. And as Cohen and Enserink’s feature makes clear, this is unlikely to be the end of this saga.
A little run down on the BWG stage 3 result:
In this study the WPI found 2 plasma gag samples positive. Both were from the negative controls.
These 2 were sequenced and are 1 to 3 nucleotides different to VP62, which makes them the same type Lo et al found.
“However, two plasma clinical aliquots were reported as positive in the WPI nested RT-PCR gag assay. These samples were from two different negative controls, and only one out of the three replicates was positive in each case. Sequencing of the excised bands revealed 1-3 base changes compared to XMRV derived from 22Rv1 (supporting online text).”
Ruscetti didn’t take part in this section of the study and Lo change assay, to one using VP62 primers.
If others used VP62 to validate plasmas as negative, they would not have picked up anything because plasma RNA has sequence variability. Which explains why Lo cannot now detect the type of virus found in Lo et al. So the VP62 assays could not find integrated MLVs. (not the mouse sort) What they can find is free floating virions in plasma.
As the WPI use lower annealing temperatures their primers would pick up both.
So I would ask, were the plasmas validated negative using VP62 as positive controls to optimise the PCR and did those assays use high annealing temperatures? because using a free floating clone, which does not exist in nature, will never give a PCR the clinical sensitivity to pick up a methylated provirus.
High stringency primers will also not detect virion RNA in plasma because, as shown by the macaque studies, the RNA has sequence variability.
The virus that was discovered by the WPI and NCI/Ruscetti was given the wrong name because Silverman incorrectly sequenced 3 of the 68 positives incorrectly. He has the VP62 clone in his lab and it was this he picked up.
The virus found by those labs in Lombardi et al. is the same as found by Lo and Alter in Lo et al. and Hanson in her study. All 3 studies have found human gammaretroviruses (HGRVs) and infecting people with ME/cfs.
This is great news.
All the negative studies were looking for the wrong virus, and using the VP62 clone, which doesn’t actually exist, as this clone was put together from 3 separate sources. The Blood working group did the same, including the Lo lab. The Lo lab didn’t use the assay that had worked in Lo et al. he used the one that detected nothing and was calibrated to VP62. I’m thinking he gave bad instructions to his team.
So the only labs with good assays were WPI and NCI/Ruscetti.
But NCI/Ruscetti didn’t do any PCR, and the WPI did not do culture.
Not all the “negative” controls in the study were screened by all the labs and the positive patients were all on medicine’s that would give false negative results.
Busch should have noticed these issues. I’m sure he and the other are more than upset by this. False negatives and the mistake of not getting all labs to screen all controls.