Followup: More Evidence against the XMRV Virus as a Cause of Chronic Fatigue Syndrome

A mouse leukemia retrovirus, xenotropic murine leukemia virus-related virus (XMRV retrovirus), has been under consideration as a possible cause of chronic fatigue syndrome (CFS, and also prostate cancer). In a study published in Science in October 2009, Lombardi et al. found XMRV in 67% of CFS patients and 3.7% of controls. Several subsequent studies in the UK, the Netherlands, and the US — by lead authors Erlwein, van Kuppleveld, Groom , Switzer and Henrich — failed to find XMRV at all.

Now a new study published in Retrovirology by Hue et al. shows that the original positive findings were likely erroneous and due to contamination in the lab. The complete article is available online.

We provide several independent lines of evidence that XMRV detected by sensitive PCR methods in patient samples is the likely result of PCR contamination with mouse DNA and that the described clones of XMRV arose from the tumour cell line 22Rv1, which was probably infected with XMRV during xenografting in mice. We propose that XMRV might not be a genuine human pathogen.

Similarity of the XMRV genome in patients from widely separated geographic areas is inconsistent with the way a retrovirus typically changes as an infection spreads.  Human cell lines are commonly contaminated with mouse viruses, and polymerase chain reaction (PCR) contamination has previously been found to underlie erroneous association between retroviruses and human disease. The genetic similarity of the XMRVs detected in patients is more consistent with contamination than with human infection. Primers previously used to detect XMRV were not specific to that virus, and previous attempts to assess contamination by detection of murine DNA alone were inadequate.

Contamination may have caused spurious results in studies of prostate cancer as well. Bryan et al. found XMRV in 22% of patients with prostate cancer. XMRV has been found in prostate cancer and in 1–6% of controls in other studies. Examination of infected prostate tumors showed that not all the tumor cells were infected with XMRV, suggesting that it did not cause those cancers.

The authors of the new study provide suggestions to prevent contamination or identification errors in future studies:

We conclude that future screens for MLV-related sequences use more rigorous PCR containment procedures, such as those used to reliably recover ancient DNA, or manage contamination by controlling for its inevitable frequency, for example by screening equal numbers of controls prepared and stored identically, together with test samples. Positive samples must be sequenced and those that are identical to known endogenous murine sequences, or plasmids present in the host laboratory, should be treated with caution.

When the first press release came out about the original Lombardi study, even before the article itself was available, our own Dr. Wallace Sampson commented on it; he was skeptical even about the basic idea that a virus could cause CFS and prognosticated:

Funny things happen. I’m betting that either the work will not be replicated, or that some other innocent explanation such as an artifact of lab technique accounting for viral presence will be found. I estimate the odds at 80:20 (4 to 1).

He won that bet, on both counts.

Last February I wrote  about it after the first two subsequent studies had failed to confirm Lombardi’s results.  I cautioned against jumping the gun to offer testing and treatment based on a reported but unconfirmed and questionable correlation that in no way proved causation.

In Sept 2010, Dr. Mark Crislip chimed in with his own speculations about possible viral etiologies for CFS. He said:

The nice thing is that, while it is fun to speculate, eventually the science will sort it out.

That sorting has now advanced another step.

This new study will not be the end of the debate and probably won’t even dissuade desperate CFS patients from prematurely requesting viral testing and anti-viral treatments. But the scientific process is working as it should. While a viral cause of CFS has not been conclusively ruled out, the evidence against XMRV far outweighs the evidence for it.

Posted in: Basic Science

Leave a Comment (33) ↓

33 thoughts on “Followup: More Evidence against the XMRV Virus as a Cause of Chronic Fatigue Syndrome

  1. superdave says:

    It’s disappointing that CFS patients lose out on a possible avenue of treatment but as we are all aware, it’s better than gaining a dubious one.

  2. Recovering Cam User says:

    While I agree there is still a long way to go to sort out what is and isn’t happening with XMRV, and testing and treatment is still premature, it’s incorrect to assume that these papers prove the original results were due to contamination.

    Noted Virologist Dr. Rancaniello recently stated in his blog “XMRV and CFS- It’s Not the End” :

    “Upon re-reading three of the four Retrovirology papers it became clear to me that they show that identification of XMRV can be fraught with contamination problems, but they do not imply that previously published studies are compromised by these findings. Clearly any new studies done on XMRV should keep in mind the potential for contamination from PCR kits and murine nucleic acids.”


    “I was initially more troubled by the fourth paper by Hue and colleagues. There are four major findings in this paper (gag PCR primers are not specific for XMRV; the virus is present in 5 human tumor cell lines; two XMRV isolates are nearly identical to a virus from the human prostate cell line and also contain an insertion from the murine retrovirus MoMLV; and there is more nucleotide diversity in viral sequences from 22Rv1 cells than in all the patient XMRV sequences). The fact that two XMRV isolates seem to be laboratory contaminants – judged by the presence of MoMLV sequences – was initially unsettling until it became clear that other XMRV isolates do not have this insertion. That leaves the fourth finding – that XMRV from 22Rv1 cells appears ancestral to, and more diverse than, all the human XMRV sequences. I decided that this result was less troublesome than I had originally believed, in part because it is not clear that the differences among the 22Rv1 viruses did not arise during PCR amplification.”

    As the WPI has repeatedly pointed out, the fact that they were able to demonstrate antibody response against XMRV argues heavily against contamination.

    In addition, previously published negative studies have failed to duplicate the WPI’s methods exactly, and some have demonstrated an inability to find XMRV at all.

    Until the original Science study is exactly reproduced, it is far too premature to claim Dr. Sampson “won that bet.”

    The stakes for everyone are far too high.

  3. Recovering Cam User says:


    “A strong argument that the novel human retrovirus XMRV is not a laboratory contaminant is the the finding that viral DNA is integrated in chromosomal DNA of prostate tumors. Why does this result constitute such strong proof of viral infection?”

  4. daedalus2u says:

    RCU, I don’t think that is strong evidence at all.

    I think the strongest evidence against XMRV being from an actual infection is how virtually identical the sequences are from very diverse locations. Retroviruses don’t have that good a fidelity of DNA replication. Their fidelity of replication is nothing like that of their mammalian hosts.

    Virtually all mice have essentially the same ERV sequences because replication of chromosomal DNA by mice is extremely high fidelity, even when that DNA is ERVs. Viruses don’t replicate that precisely.

    Antibody reactions are not specific at all. PCR is many orders of magnitude more sensitive and more specific (both). A positive antibody reaction and a negative PCR means the antibody reaction is an artifact 99.999%+ percent of the time.

  5. Angora Rabbit says:

    David Tuller of the NY Times has an interesting article about the studies here:

    My research lab does a lot of PCR (both flavors) and genomic analysis so we are familiar with the technique and its pitfalls. For conclusions of this clinical magnitude, the authors are right in calling for alternate approaches to verify the PCR data, such as sequencing the amplified fragment and confirming it didn’t originate from plasmid contamination.

    But we all know that “could happen” does not equate with “it did happen.” As RCU points out above, the original paper had multiple lines of evidence. Quoting from the NYT article: “Using other technologies, however, they also documented an antibody response in some chronic fatigue syndrome patients, and reported that XMRV in human blood could infect other human cell lines. ”

    In the interest of full disclosure, my step-mother suffers from this, and to be frank I don’t know that I believe her or that CFS exists. Nonetheless, I’m unconvinced that these most recent papers close the case. It will be interesting to see how this plays out. As one possible outcome, if the association proves to be real, perhaps the XMRV is not causative and instead takes advantage of altered immunocompetence of the human host.

  6. Angora Rabbit says:

    “Antibody reactions are not specific at all. PCR is many orders of magnitude more sensitive and more specific (both). A positive antibody reaction and a negative PCR means the antibody reaction is an artifact 99.999%+ percent of the time.”

    As someone who did an entire Ph.D. heavily based on antibody reactions and development of our own antibody reagents, I don’t understand this statement. Antibodies can be messy or they can be highly selective. As we used to say in lab, “you get what you select for.”

    Conversely I can think of several instances where PCR failed to detect and it turned out to be operator error. :)

  7. Harriet Hall says:

    This reminds me of what I recently read in The Great Influenza about the early researchers trying to find the bacterium that caused influenza. Once B. influenzae had been identified, when multiple other labs failed to find it they were accused of not being competent at growing and identifying the finicky bacteria. Or the excuse was given that the patients they studied must not have really had influenza. The true explanation was that influenza is not caused by bacteria, but by a virus. The B. influenzae was present in some flu victims, but it was an innocent bystander or the cause of a secondary infection.

    One study found XMRV in CFS patients; five studies did not. Either the first study was wrong because of contaminants or some other factor, or all the other five studies failed to find the virus because their lab techniques were incompetent. Which is more likely?

  8. geo says:

    While it would be surprising if a single virus was the cause of a dustbin diagnosis like CFS, it’s not surprising that once again people feel willing to go beyond the evidence and make exaggerated claims about this condition.

    The nature of their research does not allow for statements such as these: “Our conclusion is quite simple: XMRV is not the cause of chronic fatigue syndrome,” said Professor Greg Towers, a Wellcome Trust senior research fellow at University College, London, who led the research.

    Wallace Sampson has made comically unreasonable claims about CFS patients in the past too… unless his precognitive skills are rather more reliable than I feel able to assume.

    Here’s a new Virology post on (old) evidence of XMRV integration :

    It’s entirely possible that Towers is right, and that the sequences identified as XMRV are actually the result of contamination rather than a human pathogen, but the evidence seems rather less clear-cut than he believes.

  9. passionlessDrone says:

    Hello friends –

    When the original two dueling studies were released, there seemed to be some big differences in the immunological profiles of the participants; i.e., the WPI positives had skewed cytokine profiles, while the other study specifically excluded individuals who showed immune disturbances. It shouldn’t have been much of a surprise that their other findings were different.

    I have no idea what pattern, if any, the studies here contained.

    This doesn’t rule out contaimination, of course, but I think that proper participant matching is a proper step in sorting this mess out.

    - pD

  10. Robin says:

    One study found XMRV in CFS patients; five studies did not. Either the first study was wrong because of contaminants or some other factor, or all the other five studies failed to find the virus because their lab techniques were incompetent. Which is more likely?

    That would be a compelling argument if all of the studies were done on the same patients, using the same methods and materials.

    However, that’s not the case. Some used the similar primers and reagents, others didn’t. Some looked in different types of blood product. Different types of tests were used. Some used loosely defined “fatigued” patients, others used patients with a more established pattern of symptoms including sudden onset disability, tender lymphadenopathy, exercise intolerance, cognitive problems, etc.

    Further, because this is an emerging research issue, you should probably take into account the broader picture. Pilot studies by other researchers reported at conferences and the Blood Working Group on XMRV. Some labs have found it, others have not — while the contamination issue is a strong possibility it is obviously premature to conclude that there is no virus in CFS patients.

  11. daedalus2u says:

    Angora, antibodies can’t be as specific as DNA. If you do PCR and recover something, you really need to sequence it to see if it actually is what you think it is.

    Antibodies bind through a protein-protein interaction. The contacting surfaces are relatively small, they do not extend over hundreds of bases the way that DNA sequences do. Antibodies look at a protein’s shape. Multiple substitutions can be made that still leave the folded shape the same.

    You can find one sequence difference in hundreds of bases if you sequence it. I don’t think that antibodies have the differential sensitivity to be able to differentiate between two proteins different by only one amino acid.

  12. Harriet Hall says:

    Robin said “Different types of tests were used. Some used loosely defined “fatigued” patients, others used patients with a more established pattern of symptoms including sudden onset disability, tender lymphadenopathy, exercise intolerance, cognitive problems, etc.”

    If XMRV were the cause of CFS, one would expect to find it consistently and to have it show up on different kinds of tests. Even if some subjects did not meet the most rigorous criteria for CFS, if ANY of them did have CFS you would not expect the study to fail to find XMRV in anyone.

    While there may still be room for doubt, the probability of XMRV being the cause of CFS is looking smaller and smaller.

  13. Robin says:

    If XMRV were the cause of CFS, one would expect to find it consistently and to have it show up on different kinds of tests.

    Not necessarily. Animal studies suggest active infection occurs in tissue and that viral load is very dilute in blood. Researchers are working at limits of detection and have reported discordant results from the same patients on different days. (Early HIV tests were not consistent either.)

    Remember that there is no standardized assay, whether through PCR, antibody, or culture for XMRV or related MLVs. Until researchers are satisfied that they have a sensitive enough assay and that contamination is ruled out, any work testing for XMRV/MLVs in CFS should be taken with a grain of salt.

    …the probability of XMRV being the cause of CFS is looking smaller and smaller.

    Agreed, but I think your reasoning oversimplified.

  14. cinderkeys says:

    …the probability of XMRV being the cause of CFS is looking smaller and smaller.

    None of the authors of the positive studies have claimed that XMRV causes ME. The studies suggest a link. Other commenters have already explained why the negative studies haven’t disproven the link.

    Can laboratory contamination explain why XMRV-positive subjects produced human antibodies to gamma retroviruses?

  15. daedalus2u says:

    cinderkeys, yes, very often antibodies have cross-reactivity to other antigens. That is why vaccination with a virus that produces little or no disease can induce antibodies that are protective against a similar virus that causes a disease.

    Measles, mumps, rubella, chickenpox, smallpox are examples of diseases that are controlled because innoculation with a virus that is similar enough to produce antibodies that are protective doesn’t cause disease.

  16. Harriet Hall says:


    1. “None of the authors of the positive studies have claimed that XMRV causes ME.” But some patients are convinced that it does and are requesting anti-viral treatment.

    2. “The studies suggest a link.” Only one study (the first study) did.

    3. “negative studies haven’t disproven the link” No, but they have certainly shed doubt on it.

    4. “Can laboratory contamination explain why XMRV-positive subjects produced human antibodies to gamma retroviruses?” They found “antibodies to XMRV related proteins” and we’re not sure what that means. It doesn’t necessarily mean actual infection by XMRV virus. Could be cross-reactivity with something else.

  17. Recovering Cam User says:

    There is more than one positive study. While it did not detect XMRV specifically, the Lo/Alter study done by the NIH/FDA also supports the idea of an infectious mouse virus in CFS patients:

    “We examined 41 PBMC-derived DNA samples from 37 patients meeting accepted diagnostic criteria for CFS and found MLV-like virus gag gene sequences in 32 of 37 (86.5%) compared with only 3 of 44 (6.8%) healthy volunteer blood donors. No evidence of mouse DNA contamination was detected in the PCR assay system or the clinical samples. Seven of 8 gag-positive patients tested again positive in a sample obtained nearly 15 y later….Further studies are needed to determine whether the same strong association with MLV-related viruses is found in other groups of patients with CFS, whether these viruses play a causative role in the development of CFS, and whether they represent a threat to the blood supply.”

  18. daedalus2u says:

    RCU, finding increased levels of bacteria and viruses in the blood of people with CFS is not new. There is quite good data that people with CFS have higher levels of various viruses and bacteria in their blood, including mycoplasms. These are not “infections”, they are elevated levels. Usually blood is completely sterile. Bacteria and viruses are cleared very rapidly. For some reason in CFS the blood is cleared less rapidly. Because there is a large variety of bacteria found, it is thought that the presence of the bacteria is a sign of slower clearing, rather than an infection.

    Detecting “the same” virus in the same patient 15 years later is extremely suspicious. ERVs don’t replicate with very high fidelity except while they are in genomic DNA. If XMRV is actively infecting blood cells, it is very likely that there are none of the same blood cells left after 15 years, they have likely all been replaced. Are those 15 year later tests on cells that have been infected, produced virus, infected new cells, produced virus and so on? How many cycles of virus replication have there been in 15 years? If there are many, then the virus should show sequence changes because viruses don’t have good DNA replication fidelity. If there is no sequence change, then how did the virus hang around for 15 years without changing? No changes implies that it was not replicating. Was it not replicating as virus in the patient? Not replicating as an ERV in genomic DNA? Not replicating as a contaminant in the reagents?

  19. asleep1 says:


    I believe Harvey Alter himself stated that the virus isolated 15 years later showed the genetic mutation that one would expect in a real retroviral infection (can’t find his quote at the moment). WSJ mentions exactly this: “Seven of the eight remained positive, and the virus had changed slightly over time, a characteristic of a retrovirus.”

  20. asleep1 says:

    @Harriet Hall

    > “2. “The studies suggest a link.” Only one study (the first study) did.”

    Harvey Alter says he believes his study to be confirmatory of Ruscetti et al, noting that his study’s finding of greater genetic diversity is further evidence of an infectious retrovirus (he notes WPI reported more diversity after their Science publication). Discussion found here:

    To say that this paper doesn’t show a link is brutally dishonest.

    > “3. “negative studies haven’t disproven the link” No, but they have certainly shed doubt on it.”

    Actually, no. Again, wildly dishonest. Most of these studies have very little to say beyond the inability of their authors to find XMRV using their particular assays under their particular testing conditions.

    This is not very damning evidence, especially since the authors of the Science paper made it clear very early on that XMRV is difficult to find in the blood using only PCR. As for why, many hypotheses have been advanced (collection/storage issues, low viral presence in blood, assays not capable of detecting wild-type strains), but currently the published literature does little to answer these questions [one might say: largely thanks to the apparent zeal of numerous researchers to rush out yet another "me too" study showing their personal inability to find XMRV rather than attempt to replicate the WPI's methods].

    The only potentially damning *hypothesis* in play is contamination. And as Dr. Racaniello and others have realized, the recent studies were actually much weaker than they appeared at first glance. They actually don’t do much at all to cast doubt on either Ruscetti et al or Lo et al. I would be more than happy to walk you through the reasons why, but for now I’ll assume (likely incorrectly) that you don’t need my hand-holding since this “science based” blog is clearly a self-purported bastion of objective reasoning.

    Lastly, arriving at scientific conclusions or probabilities by counting positive vs negative studies (w/ no regard to their details) is to real science what painting by numbers is to real painting.

  21. Recovering Cam User says:

    From Annette Whitmore of the WPI:

    “XMRV: A Human Retrovirus with Unknown Pathogenic Potential, Not a Lab Contaminant.

    The recent proclamation that “XMRV is not the cause of CFS,” came from an individual who did laboratory experiments to show how PCR experiments can become contaminated. These results have nothing to do with the reality of a disease or the methods used by those who have detected XMRV in the blood and tissue of patients found to be infected. The positive studies, which cannot be explained away by PCR experiments, are those which have used multiple methods to show that XMRV is a live replicating gamma retrovirus in human blood and tissue samples using the gold standard methods of viral isolation and antibody testing, in addition to PCR. Unsupported conclusions, such as the one offered by the Wellcome Trust spokesman, often create sensational headlines but do little to move science forward. Authors of the positive XMRV studies have been extremely careful not to claim causality, realizing that more scientific research is required to make such a statement. However, one fact still remains clear. Not one of the negative studies changes the results of the scientific research done by Lombardi et al., Lo et al., Urisman et al., and Schlaberg et al.

    The WPI-led scientific study, which rigorously ruled out contamination, revealed high associations of gamma retroviruses with physician-diagnosed CFS patients, using four different methods of detection. Recent commentary associated with the negative research papers on XMRV, which used only one testing method, claimed that these studies proved that XMRV was not the cause of human disease. On the contrary, what the authors of the “contamination studies” confirmed is something that most experienced scientists already know; there are risks associated with using PCR if one does not properly control for contamination. They cannot conclude that other research groups had the same problems or that “XMRV is not the cause of CFS”.

    Most significantly, the recent Retrovirology publications failed to address the most important pieces of scientific evidence of human infection in the previous XMRV studies, including the fact that XMRV positive patients produce human antibodies to gamma retroviruses, XMRV integrates into human tissues, and infectious virus has been cultured from the blood of hundreds of patients with a diagnosis of Chronic Fatigue Syndrome and M.E. Humans do not make antibody responses to mouse DNA sequences from contaminated lab experiments. The Retrovirology studies only point out that XMRV research cannot be done in a mouse laboratory without extreme caution and should not rely solely on PCR methods.”

  22. bryterlayter says:


    Lo and Alter found gag sequences from MLV-related viruses in the blood of people with CFS and detected variations in those sequences over time which they believe to be consistent with retroviral infection.

    Do their findings support the idea that mousey-type retroviruses might be infecting people? I dunno. What do you think?

    …..a new set of blood samples was recently obtained from
    8 of the original 25 patients followed in an academic medical
    center. Testing of the repeat blood samples, ~15 y after the index
    sample, showed that seven of eight patients examined had detectable
    MLV-gag gene sequences. Significant variations of
    MLV-like virus gag gene sequences amplified from the freshly
    obtained blood samples were identified as would be expected in
    retroviral infections, but not from contamination.

    Detection of MLV-related virus gene sequences in blood of patients with chronic fatigue syndrome
    and healthy blood donors

    Shyh-Ching Lo et al

    PNAS 2010 Sep 7;107(36):15874-9

  23. cinderkeys says:

    Others have stated most of the counterarguments better than I could, so I’ll just address #1:

    1. “None of the authors of the positive studies have claimed that XMRV causes ME.” But some patients are convinced that it does and are requesting anti-viral treatment.

    Convinced? Not necessarily. Willing to give it a try just in case it helps? Yes. And I don’t blame them a bit. If I were that sick, and there were no approved treatments, I’d do the same thing.

    Anecdotally, results have been mixed. We need real clinical trials.

  24. asleep1 says:

    Yet more evidence that XMRV is a true infectious retrovirus:

    It should be painfully obvious to anyone still paying attention here that this article’s gleeful disdain is the result of either the author’s spectacularly selective or misinformed understanding or the author’s blatant and palpable bias.

    This kind of so-called “skepticism” is just shameful.

  25. geo says:

    I don’t think we can be so confident either way. There are a number of things which make it seem unlikely that XMRV is just a contaminant rather than a real human pathogen, but it’s not been consistently detectable and these new papers make clear just how easy it is for contamination to cause problems. Then there’s the additional uncertainty that surrounds its potential role in CFS.

    Best to sit back and wait for more data imo.

  26. Harriet Hall says:

    I presented the facts: several studies have failed to find XMRV in CFS patients and the possibility of contamination in the original study has been raised and has not been ruled out. Commenters have responded by citing studies about prostate cancer, studies that are not specific to XMRV, and blog opinions by people who are biased in favor of a viral etiology. Several commenters have insulted me personally and I think quite unjustifiably. None of that elucidates the possibility of a correlation of XMRV infection with CFS, much less a causal link. I said that a viral cause has not been ruled out. Science is working as it should. We can argue all we want, but the data are not sufficient at this time to settle the argument.

    Cinderkeys thinks it’s reasonable to try antiretroviral treatment just in case; I don’t, because of the significant possibility of adverse effects.

  27. geo says:

    Virology is not a blog written by someone biased in favour of a viral etiology, and the only other link to a blog was one which presented a letter from the WPI without comment, biased or otherwise.

    The prostate cancer studies are relevant because the Towers paper would, if true, undermine the role of XMRV in any human disease, and it was these authors who were most confident in dismissing the link with CFS. Also, if the prostate cancer studies finding XMRV in 1-4% of healthy controls are accurate then that would lend support to the CFS studies finding likewise, and challenge those studies which do not find it at all.

  28. Harriet Hall says:

    “link to a blog was one which presented a letter from the WPI without comment, biased or otherwise”

    That blog itself is self-admittedly biased: it was created by a CFS patient for the sole purpose of discussing XMRV. And the statement it copies from the WPI (Whittemore Peterson Institute) is open to question because the WPI institute was founded by a couple (the Whittemores) whose daughter had CFS and who was treated by Dr. Peterson with an experimental antiviral drug. They are clearly biased towards finding a viral etiology. Not only that, but they own a company that sells an unvalidated $450 test for XMRV and plows its profits back into the WPI.

    The prostate studies may be marginally relevant, but my point was that they do not directly support the claim of a link between XMRV and CFS.

  29. geo says:

    I don’t think anyone was claiming that studies examining a possible link between XMRV and prostate cancer directly supported any link between CFS and XMRV. Without some additional work on CFS patients it would be difficult to imagine that any prostate cancer study, no matter how thorough, could show that XMRV caused CFS.

    The WPI’s letter can be safely assumed to be biased in favour of the WPI, but that does not mean that their response will not be of interest to readers here. There’s rather less said, but here’s the Wellcome Trust’s view of things too:

    Whether the WPI’s letter was sourced from a blog created by a CFS patient in order to discuss XMRV, or a blog set up to discredit any possible viral cause for CFS, makes no difference so long as the letter itself has not been altered. It was not a link to blog opinions of either sort, but to a primary document which was presented without comment, so complaints of bias would seem to miss their mark.

  30. urbantravels says:

    “The prostate studies may be marginally relevant, but my point was that they do not directly support the claim of a link between XMRV and CFS.”

    The prostate studies, specifically the Kim et. al paper showing the integration site in prostate cancer:

    are directly relevant to the hypothesis proposed in Hue et. al.: “We propose that XMRV might not be a genuine human pathogen.” That radical and unproven hypothesis, if true, would serve to discredit *all* the previous work on XMRV in humans, regardless of the condition being studied.

    It is interesting, from a political viewpoint, to note how the above-quoted hypothesis put forth in the actual Hue paper became, in the Wellcome Trust’s press release, (1) confidently presented as a proven fact rather than a tentative possibility and (2) ENTIRELY about discrediting proposed links between chronic fatigue syndrome and XMRV. (See if you can find the one reference to prostate cancer in this press release, compared to the way chronic fatigue syndrome is discussed.)

    Regardless of how the science proves out in the end, this kind of misrepresentation of their own work by researchers strikes me as highly unethical and irresponsible. It’s unfortunate that claims made in press releases are not subject to peer review – or, all too often, to critical scrutiny by knowledgeable journalists.

    As for whether patients should be taking anti-retrovirals at this point, I’m not in favor of it myself: but only because properly controlled clinical studies are not yet available. The notion that such trials should be started sooner rather than later is not some fringey crazy-patient notion that needs all our quack-busting energy directed at it: it was proposed in the commentary by Valerie Courgnauda, Jean-Luc Battinia, Marc Sitbona, and Andrew L. Mason that accompanied the Lo et. al. paper in PNAS:

    “As we currently lack postulates to prove a causal association with a prevalent agent and a chronic disease with genetic predisposition, it would also be appropriate to conduct interventional studies. Indeed, the Helicobacter pylori hypothesis of peptic ulcer disease was only accepted after Barry Marshall showed that bacterial eradication with antibiotics cured peptic ulcer disease (21). Studies to gain proof of principle have been performed with antivirals in other chronic, idiopathic diseases linked to retroviral infection, such as primary biliary cirrhosis associated with mouse mammary tumor virus, another possible murine zoonosis (22). Trials using a combination of reverse transcriptase inhibitors led to significant improvements in clinical, histological, and biochemical outcomes in these patients, albeit with some evidence of viral resistance to therapy (23). Such studies are now feasible for CFS, because reverse-transcriptase inhibitors, such as tenofovir and emtracitabine, and the integrase inhibitor raltegravir can inhibit XMRV (24).

    “The caveats for conducting clinical trials in patients with CFS and MLV infection are that the potential benefits of treatment should outweigh the risks; also, studies should be conducted as randomized controlled trials with meaningful and feasible endpoints using robust therapies. At this juncture, studies to establish proof of principle are justified to determine whether safe antiviral regimens can impact on CFS and to determine whether xenotropic or polytropic MLV is causally associated with this debilitating disease.”

    I personally don’t feel that patients doing their own one-off antiretroviral treatments is such a hot idea, but mainly because there’s no useful knowledge to be gained thereby as it would be happening outside the context of properly controlled trials. There is much pearl-clutching concern being expressed in various places about the horrors of antiretroviral drugs and how CFS patients don’t have a full understanding of how serious these drugs are. I am not terribly moved by the overstatement of these concerns, and not fully convinced they are motivated only by a concern for patients’ well-being.

    In the first place, antiretroviral regimes for HIV have been developed to a point of being fairly well-tolerated, thanks to the many years of diligent research and funding that have gone into them: a level of research and funding of which CFS treatments have never seen the barest fraction. In the second place, I do believe there is a widespread misunderstanding of the severity and disease burden of chronic fatigue syndrome. It might be unthinkable to take serious anti-retroviral drugs for a condition that only consisted of some “fatigue,” but many investigators have compared the level of illness and debility suffered by CFS patients to that suffered by end-stage AIDS patients in the last two months of life, heart failure patients, and those currently undergoing chemotherapy. In the US., the disease truly has a misleading and belittling name, that only hints at the larger misconceptions about its seriousness.

    As a long-time reader of this blog, I am deeply disappointed at the low caliber of the discussion and analysis being provided by Harriet Hall. Providing retorts along the lines of “Well there are more studies showing X than there are showing Y, so X is more likely,” without any seeming insight into the methodological problems being encountered at the limits of detection in the current science, is somewhat lower than the standard I expect from this blog. Readers who really want an informed opinion about what issues the new “contamination” studies may or may not raise about past XMRV work would be well-advised to consult Vincent Racaniello’s blog instead. His reputation as a retrovirologist is impeccable, and he combines a real ability to evaluate the science with what seems to be genuine compassion for the pllght of patients. He demonstrates that such compassion is not incompatible with evaluating the science knowledgeably, fairly and objectively.

  31. Recovering Cam User says:

    Dr. Hall, you say that you have simply stated the facts. Yet in your post, you say that these new studies “show that the original positive findings were likely erroneous and due to contamination in the lab.”

    If there is one thing I have learned from reading SBM, it’s to be extremely wary of statements that cannot be supported by the evidence at hand. As other commentors have already pointed out, the fact that contamination is possible doesn’t mean it occurred in the original study. Professor Tower’s conclusion that “XMRV is not the cause of chronic fatigue syndrome” is far overstepping what these results actually show and is designed for the press, not the facts.

    In addition, I have learned from SBM to look closely at the quality of disparate studies. The original Science paper was incredibly strong, and experts in the field consider the Lo/Alter paper to support that finding. In contrast, the negative studies were nowhere near as comprehensive, using one technique instead of four, and different methods that may very well have impacted their negative results. In addition, at least two of the negative studies failed to disclose that they were unable to find XMRV in a confirmed positive sample, throwing their fully negative results even further into doubt.

    You imply that the WPI’s results are tainted by the fact of their belief in a viral cause, yet you fail to point out that two of the negative studies were done by parties who have staked their entire careers on the theory that CFS is a psychological disorder. It is unrealistic to expect any scientist to be completely bias-free, but to mention one side without the other seems to this lay observer to reflect bias on your own part.

    I agree with you that the evidence regarding XMRV and CFS is inconclusive and does not warrant experimental use of antiretroviral drugs by individual patients. But I hope you can understand that we CFS patients have good reason to be touchy – we have been marginalized for decades, our disease languishing third from the bottom in government research funding, with the vast majority of even that minimal amount having been wasted on useless research into psychological issues that ignores the well-documented biological dysfunction we share.

    It is therefore crucial to us that XMRV be given the full examination it deserves. Until someone exactly duplicates the WPI’s research, it is unfair to both patients and researchers to imply that this game is almost over. You are of course entitled to your opinion, but what I and others who disagree with you here are pointing out is that your opinion seems to have been formed without access to all the relevant information, as well as impacted by the kind of cognitive errors this blog is designed to teach people not to make.

  32. jace says:

    A consideration of Hue et al, 2010, posted with permission from GM:

    The paper states “To consider the provenance of XMRV we sequenced XMRV from the cell line 22Rv1, which is infected with an MLV-X that is indistinguishable from patient derived XMRV.”

    The virus expressed by the 22Rv1 cell line  has been sequenced and has the unique nucleotide sequences in the env and gag regions which identify it as XMRV 123, therefore it is the same virus. The cell line originates from a patient suffering from prostate cancer in 1993. Taking the evidence as a whole the explanation that the origin of  XMRV expressed by these cells is that patient, in which case the difference in sequence was not derived from replication in an immortalised cell line but was representative of wild type XMRV in 1993 and yet Towers et al seem determined to adhere to their speculative viewpoint that this is a virus acquired by these cells, via passage through mouse cell lines. This seems irrational bearing in mind that there is no known MuLV virus which has the characteristic 24 nt deletion in the gag leader or the same sequences in the env su region(1)

    It continues:
    “Bayesian phylogenies clearly show that XMRV sequences reportedly derived from unlinked patients form a monophyletic clade with interspersed 22Rv1 clones (posterior probability >0.99). The cell line-derived sequences are ancestral to the patient-derived sequences (posterior probability >0.99).”

    This is clearly been represented as objective scientific evidence when in fact it is neither and directly contradicts experimental scientific evidence. This is an example of Baysian inference based on a computer model and should not be portrayed as objective or scientific (5, 6). Baysian mathematics is a probability estimate of a hypothesis being true. The Baysian methodology rests on the idealised assumption  that human beings make decisions on a purely rational basis and that knowledge can be gained by using reason and inductive logic alone. In a Baysian analysis an ideally rational (free of political or emotional biases) individual assigns a probability of truth to the hypothesis in question at the beginning of the analysis which is called a prior position. There follows an analysis via the input of new information and the posterior probability of the hypothesis being true (from the perspective of the person(s) conducting the analysis) is inferred. In other words the prior view can heavily influence if not determine the posterior inferential probability of truth.

    The model does not work at all if the person constructing the prior has an intransigent view regarding the hypothesis in question. A rational person (for example) would not simply add information into the analysis which favours his/her initial viewpoint and exclude that which does not. If this does happen then the outputs, as well as neither objective or scientific, are not even correct in Baysian reasoning. The following summary re Baysian methodology is provided for the reader (5).

    *Prior information may not be accurate – generating misleading conclusions.
    *The way of inputting prior information (choice of prior) may not be correct, and highly subjective
    *Prior data is a construct made by judgements regarding relevance.
    *There is no one “correct way” of inputting prior information and different approaches can give different results.
    *Results aren’t objective and cannot stand by themselves. We will turn to the information omitted from this analysis later.

    “Furthermore, pol sequences apparently amplified from PC patient material (VP29 and VP184) are recombinants of XMRV and Moloney MLV (MoMLV) a virus with an envelope that lacks tropism for human cells.”

    This is clearly presented as evidence that XMRV is a contaminant. It is known that the Pol gene of XMRV consists of polytropic and xenotropic sequences (1, 17). It is also known that the human genome plays host to Moloney Murine leukaemia Pol sequences (7) and that acquisition of such sequences would actually be evidence of a recombinant which would be expected if XMRV was a human replicating MuLV class virus (8). XMRV has been compared to MoMulv by different researchers and found to be very similar, but an entirely different virus (1,9). In fact the env proteins of XMRV and MoMulv only have a homology of 54% (9). The reverse transcriptase of XMRV has different biochemical activity and three dimensional structure when compared to MoMulv (10).

    “Considering the diversity of XMRV we show that the mean pairwise genetic distance among env and pol 22Rv1-derived sequences exceeds that of patient-associated sequences (Wilcoxon rank sum test: p=0.005 and p<0.001 for pol and env, respectively). Thus XMRV sequences acquire diversity in a cell line but not in patient samples."

    The hypothesis here seems to be that somehow a sequence acquires sequence diversity in cell lines while not doing so in the same cells in a living human being. No mechanism is proposed. The authors who first isolated and sequenced XMRV concluded that the sequence variability in the POL genes between the three sequenced clones was 5 orders of magnitude higher than could be accounted for by PCR assembly and could only be explained by the viruses being independently acquired by the patients studied (1).

    "These observations are difficult to reconcile with the hypothesis that published XMRV sequences are related by a process of infectious transmission."

    This is a particularly misleading statement. These observations can easily support the other scientific evidence (omitted by this computer simulation). As already mentioned if the XMRV in the cell line pre-existed in the patient with prostate cancer then all the observations support that XMRV is an active oncovirus. Astonishingly Towers and others are clearly assuming that such a virus can only be transmitted horizontally despite a wealth of scientific evidence to the contrary. Gammaretroviruses are transmitted from one generation to the other (11,12) and  cause cancer by inserting into the gpg regions of tumour suppressing genes leading to their silencing (13,14).

    XMRV was originally discovered integrated into Human DNA (1) and has shown to be integrated into the tumour suppressing gene (NFATc3)(2) and has demonstrated a distinct preference for integrating into genes associated with cancer (15).

    Finally we look at the rationality of the analysts themselves both in terms of the evidence they omitted from the analysis and how they came to formulate the hypothesis in the first place.

    XMRV was first isolated by Urisman and others in 2006 (1), when it was isolated from polyadenylated RNA and also detected in samples taken directly from embedded tissue samples. Now polyadenylated RNA is extremely pure so would not have contained any viral DNA. They amplfied either viral mRNA and or genomic RNA. either way a replicating virus must be involved as naked genomic RNA is rapidly degraded. The authors observed viral nucleic acid in the nucleus and saw it integrated into dna and saw nucleic acid clustered just outside the nucleus .This is either cDNA or MRNA. They also detected viral proteins.

    The genetic makeup of the person was a variable linking the presence or absence of XMRV which was only found in stromal cells and not in any epithelial cells thus displayed cellular tropism. The sequence variability in pol was greater than that for gag and 5 orders of magnitude higher than could be accounted for by that created by PCR. How does the theory that xmrv is a contaminant account for these observations?

    XMRV has been found integrated into genes associated into cancer causing genes (2) and indeed shows a distinct preference for such sites (15). Antibodies to XMRV isolated from patients with PCR have been found using serological methods specific to that isolated virus (9) (16)

    PCR is the least sensitive method for isolating XMRV (or polytropic variants) (16) (9) (3). XMRV has been detected with multiple assay methods many used directly on patient samples.

    LnCAP (the cells used in the Lombardi et al 2009 study) do not express XMRV cells naturally(       )

    XMRV was not isolated by PCR but by DNA/RNA hybridisation microarrays (virochip). Its sequence is unique compared to other MuLV viruses be they endogenous or exogenous which have their sequences recorded in repositries.

    Recovery of wild type XMRV by PCR using primers constructed to detect the gag gene sequences of the VP62 clone is impossible even in people known to be positive using other PCR primers (19).

    The studies which isolated XMRV did not use mouse cell lines unlike the studies which examined the issue of contamination.

    The XMRV from 22RV1 cells cannot be the source of wild type XMRV. The 22RV1 cell line originated from a prostate cancer patient in 1993 (4). XMRV was isolated from a patient whose blood sample was taken in 1984! (3, 17)

    Thus was the hypothesis formed by this group a product of scientific observation?Clearly not. It was a belief, and cannot be the product of rational decision making. Moreover the bulk of evidence relating to the hypothesis of XMRV being a contaminant was omitted from the analysis. All this evidence supports the view that XMRV is an infectious retrovirus. Is this omission consistent with rational analysis? I would submit that it is not. Therefore this analysis at its best is neither objective or scientific does not even adhere to the principles of conducting a Baysian computer analysis.

    it would be a sad day indeed if computer simulations were given preference over experimental scientific evidence. It is also sad that a peer reviewer does not seem to understand the difference. If a paper of this quality is used as ammunition by those who wish to have research into the role of XMRV in breast cancer, prostate cancer and neuroendocrine immune diseases halted the millions of people may die or suffer horrendous levels of disability quite unnecessarily.

    1. Urisman A, Molinaro RJ, Fischer N, Plummer SJ, Casey G, Klein EA, Malathi K, Magi-Galluzzi C, Tubbs RR, Ganem D, et al: Identification of a novel Gammaretrovirus in prostate tumors of patients homozygous for R462Q RNASEL variant. PLoS Pathog 2006, 2:e25.

    2. Dong B, Kim S, Hong S, Das Gupta J, Malathi K, Klein EA, Ganem D, Derisi JL, Chow SA, Silverman RH: An infectious retrovirus susceptible to an IFN antiviral pathway from human prostate tumors. Proc Natl Acad Sci U S A 2007, 104:1655-1660.

    3. Lombardi, V.C.; Ruscetti, F.W.; Das Gupta, J.; Pfost, M.A.; Hagen, K.S.; Peterson, D.L.; Ruscetti, S.K.; Bagni, R.K.; Petrow-Sadowski, C.; Gold, B.; Dean, M.; Silverman, R.H.; Mikovits, J.A. Detection of an infectious retrovirus, XMRV, in blood cells of patients with chronic fatigue syndrome. Science 2009, 326, 585-589.

    4. Pretlow, T. G., S. R. Wolman, M. A. Micale, R. J. Pelley, E. D. Kursh, M. I. Resnick, D. R. Bodner, J. W. Jacobberger, C. M. Delmoro, J. M. Giaconia, and T. P. Pretlow. 1993. Xenografts of primary human prostatic carcinoma. J. Natl. Cancer Inst. 85:394-398.

    5. Brian Dennis; Statistics and the Scientific Method in Ecology; Nat. of Scientific Evidence: Statistical, Philosophical, and Empirical Considerations: pp 327-378

    6. KT Kelly, C Glymour; Why Bayesian Confirmation Does Not Capture the Logic of Scientific Justification; Tech Rep CMU-PHIL-138; Carnegie Mellon.

    7. D L Mager and J D Freeman; Human endogenous retroviruslike genome with type C pol sequences and gag sequences related to human T-cell lymphotropic viruses; J Virol. 1987 December; 61(12): 4060–4066.

    8. H van der Putten, W Quint, J van Raaij, ER Maandag, IM Verma, A Berns; M-MuLV-induced leukemogenesis: integration and structure of recombinant proviruses in tumors. Cell (1981) 24: 729-39.

    9. Schlaberg R, Choe DJ, Brown KR, Thaker HM, Singh IR: XMRV is present in malignant prostatic epithelium and is associated with prostate cancer, especially high-grade tumors. Proc Natl Acad Sci U S A 2009, 106:16351-16356.

    10. T. Ndongwe, SG Saralianos et al; Biochemical, Structural, and Inhibition Studies of XMRV Reverse Transcriptase; 11th Annual Symposium on Anitviral Drug Resistance, poster 5   10

    11. Panthier JJ, Gounon P, Condamine H, Jacob F.; Pattern of expression of ecotropic murine leukemia virus in gonads of inoculated SWR/J mice; J Virol. 1989 May;63(5):2134-42.

    12. A H Sharpe, J J Hunter, R M Ruprecht, and R Jaenisch; Maternal transmission of retroviral disease and strategies for preventing infection of the neonate; J Virol. 1989 March; 63(3): 1049–1053.

    13. F Martin, S Ladoire, G Mignot, L Apetoh and F Ghiringhelli; Human FOXP3 and cancer;  Oncogene, (24 May 2010) | doi:10.1038/onc.2010.174

    14. Emily C. Knouf, Michael J. Metzger, Patrick S. Mitchell, Jason D. Arroyo, John R. Chevillet, Muneesh Tewari, and A. Dusty Miller; Multiple Integrated Copies and High-Level Production of the Human Retrovirus XMRV (Xenotropic Murine Leukemia Virus-Related Virus) from 22Rv1 Prostate Carcinoma Cells; Journal of Virology, July 2009, p. 7353-7356, Vol. 83, No. 14

    15. Kim, S., N. Kim, B. Dong, D. Boren, S. A. Lee, J. Das Gupta, C. Gaughan, E. A. Klein, C. Lee, R. H. Silverman, and S. A. Chow. 2008. Integration site preference of xenotropic murine leukemia virus-related virus, a new human retrovirus associated with prostate cancer. J. Virol. 82:9964-9977.

    16. Arnold RS, Makarova NV, Osunkoya AO, Suppiah S, Scott TA, Johnson NA, Bhosle SM, Liotta D, Hunter E, Marshall FF, et al: XMRV infection in patients with prostate cancer: novel serologic assay and correlation with PCR and FISH. Urology 2010, 75:755-761

    17. Lo, S.C.; Pripuzova, N.; Li, B.; Komaroff, A.L.; Hung, G.C.; Wang, R.; Alter, H.J. Detection of MLV-related virus gene sequences in blood of patients with chronic fatigue syndrome and healthy blood donors. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 15874-15879.

    18. Wong KS, Li YJ, Howard J, Ben-David Y.  Loss of p53 in F-MuLV 14induced-erythroleukemias accelerates the acquisition of mutational events that confers immortality and growth factor independence.  Oncogene. 1999 Sep 30;18(40):5525-34.

    19. BP Danielson, GE Ayala and JT Kimata; Detection of Xenotropic Murine Leukemia Virus-Related Virus in Normal and Tumor Tissue of Patients from the Southern United States with Prostate Cancer Is Dependent on Specific Polymerase Chain Reaction Conditions; J Infect Dis. (2010) 202 (10): 1470-1477. doi: 10.1086/656146

  33. jace says:

    For anyone wishing to get involved in this area, the following extract from Danielson et al, 2010 is required reading

    “ We have found that detection of XMRV required rather specific conditions. For instance, at least 600 ng of prostate tissue DNA was necessary for reliable detection with our PCR assay.

    XMRV was detected in 3.2% of the patients when we nitially used 100–140 ng of prostate tissue DNA, compared with 22.2% of the patients when we used 650 ng.

    Additionally, we found that detection of XMRV from patient specimens, but not
    from LNCaP cells infected in vitro, depended on the gene targeted in the PCR assay.

    We were unable to detect XMRV in the patienttissue samples by nested PCR with primers specific for the gagand pol genes, regardless of whether 100 or 650 ng of DNA wasused as template. We found the gag primers to be at least 10-fold less sensitive than the env primers, and the pol primers tended to amplify a competing region from the human genome (data not shown).

     It is unclear whether these deficiencies account for the inability to detect XMRV in patient samples or whether XMRV is mainly present as an incomplete provirus in the cells of these patients.

     Nonetheless, the difficulty associated with detecting XMRV in patient samples may perhaps explain studies that do not detect the virus among large cohorts.

    this is perhaps even more damning

    To test for VP62 plasmid contamination in prostate specimens,a set of 4 primers were designed for nested PCR. The 2 forward primers (5TCTGGCTAACTAGAGAACCCACTG-3and 5AATACGACTCACTATAGGGAGACC-3) were specific to the multiple cloning site of pCDA3.1(-) (Invitrogen). The 2 reverse primers
    5-GTTACGGTCTGTCCCATGATCTC-3) were specific to the 5 terminus of VP62 gag. The VP62 nested PCR assay was foundto be capable of detecting 10 plasmids diluted in 600 ng of uninfected LNCaP DNA in 3 of 3 samples and 1 plasmid in 600 ng of uninfected LNCaP DNA in 1 of 3 samples (data not shown).

    In other words a PCR assay capable of detecting very low copy number of VP62 gag cannot detect XMRV in a known clinically positive sample.”

    BP Danielson, GE Ayala and JT Kimata; Detection of Xenotropic Murine Leukemia Virus-Related Virus in Normal and Tumor Tissue of Patients from the Southern United States with Prostate Cancer Is Dependent on Specific Polymerase Chain Reaction Conditions; J Infect Dis. (2010) 202 (10): 1470-1477. doi: 10.1086/656146

Comments are closed.