Monday 7 November 2011

No evidence that a strain of XMRV was created in a cell line in the 1990s

Update on the paper Paprotka et al.

The argument in Paprotka et al. (a study led by Pathak and Coffin) is that the sequence isolated by Silverman, which he claimed came from patient VP-62, is a result of a unique recombination event. This unique recombination event occurred between a virus, found in many species of mice, and a hitherto undiscovered replication defective erv, when the xenograft which was involved in the creation of the 22Rv1 line was passed through mice.

They claim, but cannot prove that these mice were NU/NU and or Hsd mice, as they contain the viruses or at least the virus fragments they deem to be necessary to further their hypothesis.

They argue that their assays did not amplify XMRV in the early xenografts, but after several passages of the cells through the proposed mice their assays were able to amplify XMRV. Thus XMRV must have been absent from the early xenografts and created by this unique recombination event.

So we have to judge between two explanations. One is that the VP-62 sequence was generated by a unique recombination event, or that the PCR assays used lacked the clinical sensitivity needed to detect low copy VP-62 sequences in the early xenografts.

Given that a unique event is so very unlikely they must provide very strong evidence that VP-62 equences were indeed absent from the early xenografts and that their apparent absence was not caused by the failure of their PCR assays. Absence of evidence is not evidence of absence.

If one focuses on the initial quantitative PCR assay using one envelope primer it becomes clear that there is no evidence provided regarding the copy number of VP-62 sequences per 100 cells in the NU/NU or Hsd mice, later xenografts or early xenografts. This is because the PCR amplified what were known to be VP-62 sequences from the DNA taken from 22Rv1 cell line and the CWR-R1 cell line. The problem here is that the copy number of VP-62 in these cells is huge, being 2000 copies per 100 cells and 3000 copies per 100 cells respectively. We do know that this assay could detect ERV sequences in the mice and the early xenografts, but we don’t know what the copy number of ERV sequences is compared to the copy numbers of VP-62. Thus it is quite possible for VP-62 sequences to be present but undetectable, due to a lower copy number for example. There are other more complex reasons. In any event absence of evidence is not evidence of absence. We also have the problem that this assay with the env primer was not used to screen other species of lab mice or wild derived mice. Had this assay being able to detect either XMRV or ERV sequences in these mice the paper would have collapsed. Certain vested interests have argued that it would be silly to use this assay with this primer to screen the other mice because the primer could amplify both VP-62 and ERV sequences. In that case why use that primer to screen NU/NU and Hsd mice, and indeed why use that primer at all. I wonder what the defense is for not using this primer and assay to any screen wild mice. Wild derived mice are not wild mice.

There is no evidence that this assay can amplify XMRV specific env sequences other than in the DNA taken from 22Rv1 and CWR-R1 cells. There is no evidence that this assay with this primer could detect VP-62 sequences in the DNA extracted from lab mice, the early xenografts or the later xenografts. The later xenografts were not screened using this assay. In fact figure S3 in the SOM (Supporting Online Material) shows that provided this primer was used at all to screen the later xenografts using the second single round PCR, then that PCR assay could not amplify any product. As this primer was used to screen the early xenografts with the second PCR assay (Figure S3), not to use it to screen the later xenografts would be a serious departure from the scientific method.

Next we examine the second PCR single round assay with a gag primer, which is only capable of detecting VP-62 gag sequences. Figure 2D and 2E suggest that this assay, with this primer, was used to screen the early xenografts. This is however in direct conflict with the text, which states that all the primers used with the single round assay were capable of amplifying both VP-62 and ERV sequences. This is prima facie evidence of misfeasance and at the very least should be investigated. In any event, it seriously weakens if not invalidated their claim that gag sequences were not present in the early xenografts. Indeed we once again have the scenario that there is no evidence that this single round assay can detect VP-62 gag sequences in anything other than the DNA taken from the 22Rv1 or CWR-R1 cell line either. So they cannot provide evidence that their assay could detect VP-62 GAG sequences in mice or later xenografts let alone the early xenografts.

Thus I will leave it to those without vested interests to consider whether these authors have provided convincing evidence that VP-62 sequences are absent from the early xenografts, bearing in mind that absence of evidence is not evidence of absence and that the alternative explanation is a unique recombination event, which cannot have happened in the past and its future occurrence is a virtual impossibility.

Finally the paper does not fulfill the minimum requirements of surviving a peer review process. There is not enough evidence to enable the scientific merit of the study to be judged or enough data in the methodology section to enable replication. The claim that it is custom and practice among a certain group of retrovirologists not to include such details unless they deem it necessary merely means that none of their studies should have passed the Science peer review process or indeed any rigorous peer review process. Contrary to statements made by certain vested interests the Science magazine is not published merely for the perusal of certain retrovirologists but also other scientists from different disciplines. I would remind everyone what the requirements of science are:

"Data and materials availability
All data necessary to understand, assess, and extend the conclusions of the manuscript must be available to any reader of Science."

Paprotka et al., ‘Recombinant Origin of the Retrovirus XMRV’ does not meet the requirements as specified by Science. 


Permission to repost 


Thank you to Pumpkin for the post.

No comments:

Post a Comment