From Vincent Racaniello's ‘Virology Blog', 27 September 2011
Whether or not the retrovirus XMRV is a human pathogen has been debated since the virus was first described in 2006. The answer is now clear: the results of Blood XMRV Scientific Research Group, along with a partial retraction of the 2009 Science paper describing identification of the retrovirus in patients with chronic fatigue syndrome (CFS) show that detection of XMRV in patient samples is a result of contamination.
The Blood XMRV group obtained new blood samples from 15 individuals previously shown to be positive for XMRV (Lombardi et al., 2009) or MLV (Lo et al., 2010) ; 14 of these were from CFS patients. Fifteen blood samples were also obtained from healthy donors. The samples were coded and sent to 9 laboratories for analysis. These laboratories (Abbott Molecular, Abbott Diagnostics, CDC, FDA/Lo, FDA/Hewlett, Gen-Probe, NCI/DRP, and WPI) conducted validated assays for viral nucleic acid, viral replication, or viral antibodies. Positive contra samples were also included that were ‘spiked’ with XMRV, in the form of cell culture fluids from the cell line 22Rv1. Each laboratory was at liberty to choose which assays to carry out.
Two laboratories reported evidence of XMRV in the coded samples. Only WPI identified positive specimens by PCR: two from negative controls, and one from a CFS patient. The FDA/Lo laboratory did not detect any positives by PCR, using the same nested assay that they had previously reported in their published study. The samples tested included 5 specimens that were positive in the Lo et al. study.
Lombardi and colleagues have previously concluded that viral culture is the most sensitive method for detecting XMRV; however the FDA/Hewlett laboratory failed to culture virus from CFS samples. This laboratory did culture virus from positive control specimens, demonstrating the sensitivity of their methods. The FDA/Ruscetti laboratory recovered virus from 3/15 CFS samples but also from 6/15 negative control specimens. WPI did not carry out viral culture assays due to contamination of their cell lines with mycoplasma.
Four laboratories tested the samples for the presence of antibodies that react with XMRV proteins. Only WPI and NCI/Ruscetti detected reactive antibodies, both in CFS specimens and negative controls. There was no statistically significant difference in the rates of positivity between the positive and negative controls, nor in the identity of the positive samples between the two laboratories.
These results demonstrate that XMRV or antibodies to the virus are not present in clinical specimens. Detection of XMRV nucleic acid by WPI is likely a consequence of contamination. The positive serology reported by WPI and NCI/Ruscetti laboratories remained unexplained, but are most likely the result of the presence of cross-reactive epitopes. The authors of the study conclude that ‘routine blood screening for XMRV/P-MLV is not warranted at this time’.
One of the authors on Lombardi et al., Robert Silverman, decided to reexamine some of the DNA preparations from CFS patients that were originally used to detect XMRV DNA by PCR. He found that all the positive specimens from CFS patients were contaminated with XMRV plasmid DNA. Therefore the authors of the original study have retracted Figure 1 (single-round PCR detection of XMRV in CFS PBMC DNA); table S1, XMRV sequences, and figure S2, phylogenetic analysis of XMRV sequences.
A puzzling aspect of Silverman’s results is that XMRV plasmid DNA was detected only in samples from CFS patients, not healthy controls. This pattern would not be expected if the specimens were properly blinded, that is, coded so that the investigators did not know which were controls and which were from CFS patients. The authors offer no explanation of these findings.
The paper reporting contamination of samples with XMRV is entitled ‘Partial Retraction‘. It’s not clear to me why the entire paper has not been retracted. After removing the PCR and nucleic acid sequencing results, there is no evidence indicating the presence of XMRV in the patient samples. The remaining experiments show detection of a retrovirus by cell culture experiments, and the presence of viral proteins or antibodies to the virus in clinical specimens. None of these findings prove that what is being studied is XMRV. The title of the original paper ‘Detection of an infectious retrovirus’, XMRV, in blood cells of patients with chronic fatigue syndrome‘, is unsupported.
In an accompanying article on the XMRV story entitled ‘False Positive‘, Judy Mikovits of WPI notes that “Anyone who says this is a lab contaminant has drawn the wrong conclusion and has done a disservice to the public”. She goes on to imply that a gammaretrovirus is likely involved in CFS. On the contrary, pursuing the CFS-gammaretrovirus hypothesis is a disservice to those with CFS, and detracts from efforts to solve the disease. There are no data to support such an association, and to suggest that a lab contaminant, XMRV, has pointed the way to a bona fide etiologic agent seems
XMRV does not cause CFS. The virus arose in mice between 1993-96, and its detection in patient samples is clearly a result of contamination. Reaching these conclusions has required a long and often contentious journey that has highlighted the best and worst aspects of scientific research. There are many lessons to be learned from XMRV, but an important one is that science progresses not from the work of a single investigator, but from the collective efforts of many laboratories.
XMRV reminds us to trust science, not scientists.