Conclusions not supported by the data

A Dusty Miller   (2012-01-13 10:25)  Fred Hutchinson Cancer Research Center

This report claims two major conclusions (last paragraph of Abstract):

1. XMRV replicates efficiently in prostate epithelial cells by downregulating A3G expression.

2. Our data suggest a novel mechanism by which retroviruses can counteract the antiviral effects of A3G proteins.

Neither conclusion is supported by the data.

First, some background. Many publications to date have shown that XMRV can be mutated by APOBEC3 proteins (A3A - A3G) present at various levels in human cells. Sequencing of the RNA genomes of XMRV viruses produced from particular human cells shows that some genomes are intensely mutated (hypermutated) while others show only background mutation rates likely due to errors in reverse transcription. Rates of hypermutation vary for viruses produced from different human cells, from almost none for XMRV produced by 22Rv1 prostate cancer cells (Paprotka et al., 2010), to ~25% for DU145 prostate cancer cells (Paprotka et al., 2010), to almost 100% for virus produced by human peripheral blood mononuclear cells (PBMC) (Chaipan et al., 2011). The results obtained for human PBMC clearly show that XMRV is not able to circumvent the effects of A3 restriction in human cells.

So, what about the current authors' conclusion that XMRV replicates efficiently in prostate epithelial cells (LNCaP and DU145 cells) by downregulating A3G? The first problem with this claim is the lack of a definition or assay for 'efficient replication'. No measurements of XMRV replication (or hypermutation) were performed by the authors. Previous reports cited by the authors in support of 'efficient' replication (manuscript refs. 3 and 18) documented XMRV replication by showing that XMRV proteins and reverse transcriptase increased with time after exposure of prostate cancer cell lines to XMRV, but provided no evidence that XMRV replication was more or less efficient than that of any other viruses. Next, we already know that XMRV virus produced by DU145 cells is ~25% hypermutated (Paprotka et al., 2010), so XMRV is not completely resistant to at least one of the A3 proteins present in DU145 cells. Even so, XMRV has been shown to 'efficiently' spread in DU145 cells (ref. 3). This is not unexpected, because ~75% of the virus produced from DU145 cells is infectious. Thus, in cells like DU145 that make low levels of A3 proteins, it is unnecessary for XMRV to completely downregulate A3 proteins to replicate 'efficiently'.

The authors do provide some evidence that A3G protein levels are reduced by ~50% in XMRV-infected LNCaP and DU145 cells compared to the parental cell lines. But, could a two-fold reduction in A3G levels mediated by XMRV be responsible for 'efficient' XMRV replication in these cells? In the case of XMRV-infected DU145 cells, where we know that ~25% of the virus produced is hypermutated and ~75% is active, would it matter if there was 2-fold more A3G, resulting in ~50% hypermutation and ~50% active virus production? We don't know because the authors have not measured XMRV replication rates in cells with different A3G levels, but one would expect the virus to replicate well under either condition.

Lastly, there is a simple explanation for the decrease in A3G levels the authors report; that A3G packaging into virions, which is required for virus hypermutation, is responsible for the decrease in cellular levels of A3G. This possibility would be easy to address by assay for A3G associated with XMRV virions in the cell culture medium. If so, there is no reason to propose an A3G regulatory mechanism involving XMRV.

What about the authors' second major conclusion, that their findings suggest a novel mechanism by which retroviruses can counteract the antiviral effects of A3G proteins? The mechanism, if any, is certainly not robust, and its existence is not supported by the data provided. At the very least, the authors should measure A3G mRNA levels in infected and uninfected cells to see if XMRV might be regulating A3G transcription, which would suggest the production of some transcription factor by XMRV, and might provide some support for the suggested regulatory mechanism. This experiment is critical for the authors' claim that XMRV downregulates A3G expression.

On a final note, the representative A3G protein data presented for LNCaP cells +/- XMRV in Fig. 3B don't support the average results from three experiments shown in Fig. 3C. In Fig. 3B, the ratio of A3G in LNCaP+XMRV cells to that in LNCaP cells appears to be about 1:5, the ratio of b-actin about 2:1, so the overall ratio of A3G, normalized to b-actin, is about 1:10 or 10%. The average ratio shown in Fig. 3C is 40% with a very tight error bar, inconsistent with the results shown in Fig. 3B being included in this average. In contrast, the results in Fig. 3B and 3D for the DU145 cells appear consistent.

Publications cited:

Chaipan, C., Dilley, K.A., Paprotka, T., Delviks-Frankenberry, K.A., Venkatachari, N.J., Hu, W.-S., and Pathak, V.K. (2011). Severe restriction of xenotropic murine leukemia virus-related virus replication and spread in cultured human peripheral blood mononuclear cells. J Virol 85, 4888-4897.

Paprotka, T., Venkatachari, N.J., Chaipan, C., Burdick, R., Delviks-Frankenberry, K.A., Hu, W.S., and Pathak, V.K. (2010). Inhibition of xenotropic murine leukemia virus-related virus by APOBEC3 proteins and antiviral drugs. J Virol 84, 5719-5729.

Competing interests

none

top