By Suzanne Vernon, PhD, Scientific Director
The 2009 Science paper by Lombardi et al. was a seminal investigation that mobilized many in virology, blood safety and the media to understand XMRV and determine its association with CFS. In today’s issue of Science, a partial retraction of the 2009 Lombardi et al. paper is published with Robert H. Silverman of the Cleveland Clinic as the lead author. This issue of Science also features the long-awaited results of the Blood XMRV Scientific Research Working Group (SRWG) in a paper titled, “Failure to confirm XMRV/MLVs in the blood of patients with CFS: a multi-laboratory study” with Graham Simmons of Blood Systems Research Institute and University of California, San Francisco as the lead author. I’ll break down these papers in two parts.
Part 1: The Partial Retraction
An eight-page feature story titled “False Positive” accompanies these papers and chronicles the events that led to the publication of the 2009 Science paper and backgrounds of many of the key players. Silverman, who discovered XMRV, collaborated with the Whittemore Peterson Institute (WPI) team and provided them with XMRV reagents so they could pursue what seemed like a promising avenue of CFS research. With repeated failures to replicate the initial CFS/XMRV report, the incongruent results reported from Phase II of the SRWG and the possibility that XMRV originated from passage through laboratory mice, Silverman and his colleague Jaydip Das Gupta set out to reexamine CFS and healthy control DNA samples originally received from the WPI.
Structure of XMRV, courtesy of Nature Reviews Urology
Using immaculate laboratory and PCR procedures, Silverman and Das Gupta detected the XMRV VP62 plasmid in 7 of the 15 CFS patient DNA samples and 0 of 17 healthy control DNA samples. This XMRV VP62 plasmid is a full-length XMRV molecular viral clone that can produce infectious XMRV virus when transfected into LNCaP cells. Plasmids can be thought of as DNA transfer vehicles, containing a few other genes that are not XMRV-specific but help the plasmid in producing high levels of infectious virus. Silverman and Das Gupta first used single round PCR for the XMRV env gene and detected this XMRV gene in 6 of the 15 CFS patient DNA samples and 0 of 17 healthy controls. The six CFS XMRV env positive samples were also positive for a piece of the VP62 plasmid DNA specific to the plasmid and not XMRV, called neo. None of the healthy control DNA samples had neo plasmid DNA amplified. Further confirmation of plasmid contamination was demonstrated by single round PCR of the CMV promoter region of the plasmid in the same 6 CFS DNA samples and none of the controls. They amplified and sequenced DNA from another CFS sample identified as WPI-1124. This step verified that seven of the CFS patient samples received from the WPI and shown in Figure 1 of the 2009 Science paper were positive for the XMRV VP62 plasmid. DNA sequence and analysis from the positive CFS patient samples also indicated that data in Table S1 and Figure S2 of the Science 2009 paper was likely “spurious and due to contamination with the XMRV VP62 plasmid DNA,” the retraction states.
This work by Silverman and Das Gupta indicates the source of XMRV in the CFS patient samples as the original full-length XMRV plasmid rather than naturally occurring XMRV. They reached this conclusion based on the fact that they easily amplified portions of the plasmid DNA that is not naturally occurring DNA. This partial retraction of data from the Science 2009 paper, signed by all of the authors on the original paper, casts further doubt on the validity of the original report and the likelihood that naturally occurring XMRV is associated with CFS. The editors of Science have indicated that they are discussing “next steps” with the authors and that they stand by the Expression of Editorial Concern published in May.
Part 2: The Blood Safety Study
[Note: Study authors provided a webinar presentation about this study for the Association. Recording, slides and more info: http://www.research1st.com/2011/10/14/xmrv-updates/]
In the Science 2009 paper, Lombardi et al. stated that XMRV could be detected in the blood of 8 of 218 healthy controls. This raised the possibility that XMRV could be a threat to the blood supply. This type of threat is not new to the community of scientists who deal daily with maintaining a safe blood supply and within a month, the Blood XMRV Scientific Research Working Group (SRWG) was formed. Its primary purpose was to design and carry out a four-phase study to determine whether XMRV posed a threat to blood safety. An overview of this four-phase study was published in March 2011 in Transfusion. Phase I and II focused on assay development, optimization and standardization. The aim of Phase III was to develop coded panels of clinical samples from CFS patients who had tested positive for XMRV in the 2009 Science paper and P-MLVs as described in the 2010 paper by Lo et al., pedigreed negative controls (samples tested repeatedly and consistently shown by all participating laboratories to be negative for XMRV and P-MLVs) and experimentally prepared positive controls also known as “spiked” controls. Phase IV would use optimized assays and conditions to test for XMRV and P-MLVs in a large number of samples collected from healthy blood donors.
The paper published today in Science titled, “Failure to confirm XMRV/MLVs in the blood of patients with chronic fatigue syndrome: A multi-laboratory study,” by Simmons et al., finishes the work of the SRWG. Nine independent laboratories (see figure at right) using 19 optimized and highly sensitive assays were unable to reproducibly detect XMRV or MLVs in the coded panel of samples. The coded panel consisted of blood samples from five CFS patients who were positive for P-MLVs as described in the Lo et al. paper and 10 CFS patients recruited by the WPI who had previously tested positive for XMRV by at least one of a variety of test methods. Samples from six of these CFS patients were also used in the original study published in Science in 2009. The study design as described in Transfusion called for a larger panel of 30 previously-positive CFS subjects; however, Simmons et al., note that this smaller group was, “the maximum number of subjects who could be recruited by the cohort investigators.”
As noted above, the coded panel included pedigreed negative controls collected from three laboratory personnel and 12 local blood donors from Blood Centers of the Pacific, San Francisco. These negative control blood samples were processed into various blood components needed for the assays and distributed to participating laboratories for XMRV/MLV. All participating laboratories were in unanimous agreement that these controls were pedigreed negative. Finally, three experimental positive controls were prepared by spiking samples with cells or viral RNA from 22Rv1 cells to serve as an XMRV positive controls. All samples were coded and included as duplicates or triplicates in the panel.
Samples were collected, processed and prepared under stringently controlled, identical procedures. Sources of contamination of supplies and materials were carefully eliminated. All of the samples were coded and tested under blinded conditions, meaning that the laboratory personal doing the tests did not know whether they were testing samples from CFS patients, blood donors, pedigreed negatives or spiked positive controls. Each of the laboratories was able to define the assays it used to test samples. In total, there were 19 different assays used that fell into three categories: polymerase chain reaction (PCR), antibody reactivity and culture. The WPI has stated that culture is its most sensitive means of detecting XMRV and other gammaretroviruses; however, a laboratory problem with mycoplasma contamination prevented it from testing samples using this method. Dr. Francis Ruscetti’s lab at the National Cancer Institute (NCI) and Dr. Indira Hewlett’s lab at the Food and Drug Administration (FDA) used methods identical to WPI, as described in the supporting materials for the study.
Phase III results: red indicates positive results; orange reflects indeterminate results
The labs completed their work, submitted results and the code was broken by staff at the Blood Systems Research Institute in August. Only two of the nine laboratories, the WPI and Dr. Francis W. Ruscetti’s laboratory at the NCI, detected XMRV/MLV in any of the clinical samples contained in the coded panel. The remaining seven laboratories, including Dr. Lo’s FDA lab, only detected XMRV/MLV in the spiked positive controls, each with high accuracy. Only the WPI detected XMRV/MLVs in the CFS samples using PCR assays. The WPI assays were the least sensitive for detecting viral RNA as illustrated by the inability to detect RNA in 2 of 5 spiked plasma samples and 1 of 5 PBMC-spiked samples. The WPI and NCI/Ruscetti laboratories detected XMRV/MLVs at the same rate in CFS patient samples as in healthy blood donor samples. Further, even though each sample was represented in duplicate or triplicate, WPI and NCI/Ruscetti frequently only detected XMRV/MLVs in one of the duplicate or triplicate samples from the same subject. Sequence analysis of DNA amplified from the positive assays showed a high degree of similarity to XMRV derived from 22Rv1. Mouse DNA contamination was not found in any of the positive samples. Using the kappa coefficient as a statistical measure of agreement, it was found that even though both WPI and NCI/Ruscetti detected XMRV/MLV in some of the samples, the samples that were positive were often different samples, indicating there was no agreement on what could be called an XMRV/MLV positive sample.
The conclusion of this nearly two-year study is that XMRV/P-MLVs are not reproducibly detected in clinical samples (15 from individuals who had previously tested positive) using current, state-of-the art assays. The authors state, “our findings are reassuring with respect to blood safety and indicate that routine blood donor screening for XMRV/P-MLV is not warranted at this time.” Phase IV will not proceed, although there are results from studies of large numbers of blood donor samples pending publication.
With this closing statement, it appears that the blood safety community is satisfied with the scientific rigor of the SRWG study and that XMRV/P-MLVs are not a threat to the blood supply. It is remarkable to see this type of study done with government, commercial and private laboratories working in parallel to address a potentially serious public health issue. Even though the SRWG studied was designed as a blood safety study, it helped bring CFS into the light. Many of investigators on the SRWG previously knew little about CFS. Now they are familiar with the enormous suffering that CFS inflicts, the problems with the CFS case definition for research and the anger and passion that comes with an illness that has so long been in the dark shadows, unsolved.
Suzanne Vernon, PhD, is the Association’s scientific director. She has nearly two decades of experience as a microbiologistSeptember 22, 2011