Determining whether serological evidence of immune responses to gammaretroviruses in humans8,21
. is an indication of authentic infection or just non-specific cross-reactivity is an important final step in the XMRV saga. In this study, we generated a robust, high-throughput micro-neutralization assay for the screening of large numbers of subjects for serological evidence of XMRV and MLV infection based on the DEP assay system we recently developed22
. This assay includes an internal control pseudovirus which is very useful for avoiding nonspecific inhibition and also controls for cytotoxicity. This method provides a reproducible high-throughput micro-neutralization research assay for large scale testing for evidence of XMRV and MLV infection.
Currently, enzyme immuno-assays (EIAs) and Western blot (WB) are the two most common serological methods utilized for viral diagnosis47,48
. WB is limited to the recognition of linear epitopes and is prone to high-background rates, while EIA can be restricted by the quality of the antigens, antibodies and detection methods. Instead of directly detecting the existence of antiviral antibodies in the sera, the DEP-based micro-neutralization assay is based on the ability of a serum to neutralize pseudovirus infection. Compared with standard assays such as EIAs, the micro-neutralization assay has fewer steps and can be performed by automated liquid handling equipment, which may generate less standard deviation. The disadvantage is a two-day incubation period during the assay, which impacts the clinical usefulness of the assay.
A recent study identified neutralizing activity against XMRV in about 14% of blood donor samples10
. although in this instance many of these sera neutralized control viruses in addition to XMRV. In contrast, while we identified 23/354 blood donors (6.5%) able to moderately neutralize XMRV Env-mediated infection, control and other MLV envelopes were poorly or not at all neutralized. None of the samples tested showed any evidence of a serologic response to XMRV by WB testing. Furthermore, all 23 seroreactive samples were negative for XMRV and MLV sequences using PCR or virus culture. These PCR and culture assays were designed to detect a broad range of gammaretroviruses, as well as XMRV specifically, thus, excluding XMRV/MLV and other gammaretroviruses as a source of the non-specific reactivity. The finding that neutralization by the 23 blood donors was specific to XMRV envelopes, but not other MLV envelopes was surprising. Pairwise comparison of the amino acid (aa) sequence of the envelope region between XMRV, and MLV-P or MLV-E shows the aa similarity is about 89% and 68% respectively.
Given that the true XMRV neutralizing responses raised in animals were more broadly neutralizing (), this result strongly argues against specific neutralization, but rather suggests the moderate neutralization observed was mediated by other non-specific means. This could be cross-reactive antibodies raised against endogenous retroviral elements, completely unrelated proteins, or other non-antibody serum factors. Human serum potently inhibits XMRV14
, however, this is largely complement-driven, and in our assay serum complement was inactivated by heating and did not influence our test results. The relatively high level of non-specificity is greater than that seen with other microneutralization assays 20,49
, and is partly due to the lack known of human positive cases that can be used in order to accurately set cut-offs for defining specific neutralization. Our results likely also explain other reported XMRV neutralization results in human samples 21
In addition to the initial association of XMRV to CFS made by Lombardi et al.8
, a second publication by Lo and colleagues40
, based only on PCR analysis, also yielded a strong association between CFS and MLV-like viruses40
. These subsequent viruses demonstrated a far greater degree of sequence variation than XMRV, with the majority of sequences resembling polytropic MLV (P-MLV). Although Lo et al. reported very stringent measures to minimize contamination40
, the most parsimonious explanation, given the extent of reported contamination of laboratory reagents, is that their PCR results are false positives resulting from reagent contamination. Indeed, Lo et al. used Platinum Taq (Invitrogen) for PCR amplification, which several groups have convincingly demonstrated is contaminated with mouse DNA14,15,50
due to the use of a mouse monoclonal antibody in the enzyme mix. Furthermore, recent detailed phylogenetic analysis of the longitudinal polytropic MLV sequences reported by Lo et al. showed that these sequences are inconsistent with retroviral evolution51
. Nonetheless, the findings of Lo et al. raised the hypothesis that while XMRV itself is clearly a laboratory contaminant, the serological responses detected in Lombardi et al. may be due to infection by other MLVs or gammaretroviruses. The serological assay used by Lombardi et al. relies on antibody binding to the MLV spleen focus-forming virus (SFFV) Env expressed on the surface of cells. The logic of this assay is that conformationally-dependent cross-reactive epitopes shared between this mouse gammaretrovirus and XMRV would bind XMRV antibodies which would then be detected in a flow cytometry-based assay. However, it is likely that, as with our micro-neutralization assay, mammalian cell culture-based expression of an unrelated retrovirus Env would be highly prone to non-specific cross-reactivity that can confound the testing and which requires clarification by WB analysis using purified antigen. Indeed, when the Lombardi et al. flow-based assay was used by two laboratories on plasma specimens in a blinded study, high levels of non-specific reactivity was observed15
In conclusion, we developed a robust, high-throughput micro-neutralization assay in order to conduct studies looking for evidence of infection with XMRV and MLV. Although a small proportion of blood donors demonstrated the ability to block XMRV-mediated infection, we found no evidence that this inhibition was mediated by specific antibodies elicited by exposure to XMRV or related MLVs. It is likely that this moderate neutralization is mediated through another, non-specific mechanism. Our findings also explain further the highly non-reproducible and non-specific serological responses detected with other assays8,17
. In addition, this micro-neutralization assay system can be easily adapted to screen donor samples against other viruses with careful selection of matching partner virus envelopes, which will provide important information for neutralizing antibody responses and infectious disease profiles.