Research Article: Endogenous Murine Leukemia Viruses: Relationship to XMRV and Related Sequences Detected in Human DNA Samples

Date Published: October 24, 2011

Publisher: Hindawi Publishing Corporation

Author(s): Oya Cingöz, John M. Coffin.


Xenotropic-murine-leukemia-virus-related virus (XMRV) was the first gammaretrovirus to be reported in humans. The sequence similarity between XMRV and murine leukemia viruses (MLVs) was consistent with an origin of XMRV from one or more MLVs present as endogenous proviruses in mouse genomes. Here, we review the relationship of the human and mouse virus isolates and discuss the potential complications associated with the detection of MLV-like sequences from clinical samples.

Partial Text

Retroviruses are unique in their requirement, as a natural step in their replication cycle, for integration into the genomes of the host cells they infect [1]. Through infection of the germ line, exogenous retroviruses can become a part of the host genome, leading to the generation of endogenous proviruses. All vertebrate species examined carry remnants of such prior retroviral infections in their genomes [1, 2]. Humans, for example, carry some 80,000 sequences or about 8% of our total genome, derived from retrovirus infections dating from some 40 million to a hundred thousand or so years ago. Mouse genomes contain a large number of more recently integrated endogenous proviruses; one of the best-studied groups comprises the murine leukemia viruses (MLVs). MLVs are thought to have entered the Mus germ line less than 1.5 million years ago [3, 4], and generation of new endogenous proviruses continues to this day. They can be classified based on their host range and sequence into ecotropic, xenotropic, polytropic and modified polytropic, MLVs [1, 5]. The host range of MLVs is determined by the species distribution of the receptors they use for entry. Ecotropic MLVs (of which there are only a few) use mCAT1, an allele of the cationic amino acid transporter found only in mice and a few other rodents [6]. The more common nonecotropic MLVs, including xenotropic (X-MLV), polytropic (P-MLV), and modified polytropic (MP-MLV) viruses use distinct alleles of Xpr1, a cell surface protein of unknown function, with X-MLVs unable to recognize the allele found in most inbred and a few wild mice, hence the name “xenotropic” (reviewed in [7]). The proviruses that correspond to these different viruses are referred to as Xmv, Pmv, and Mpmv, respectively. Experimentally, they are distinguished by hybridization to oligonucleotide probes specific for a sequence in the SU region of env. P-MLVs can be distinguished from MP-MLVs by a 27 bp insertion in their env genes [1, 5]. Endogenous nonecotropic proviruses are highly polymorphic in their genome location: inbred mice contain about 20 of each type and, on average, share about 30 with any other inbred strain. Of the 150 or so identified proviruses, only a few Xmvs, including Bxv1 (Xmv43), are known to encode infectious virus [8]. No complete, replicating P-MLV or MP-MLV has ever been identified, although their env genes are often found as recombinants [1, 9].

Evidence for the presence of XMRV and other MLV-related viruses in human tissues have so far relied on various methods including fluorescence in situ hybridization (FISH), immunological detection of viral antigens and antibodies against them, isolation of infectious virus from human samples, and, finally, PCR amplification of MLV sequences [12–14]. We will discuss here only the last two issues, focusing on the relationship of XMRV and related sequences to endogenous MLVs and the likely events that gave rise to them. By the term “XMRV,” we refer only to the infectious viruses reported in the original papers [12, 13, 15]; related sequences detected by PCR amplification [14] will be referred to as “MLV-like” (Figure 1). We first focus on XMRV.

In an attempt to replicate the findings of XMRV in CFS patients, Lo et al. analyzed plasma and PBMCs from another cohort of CFS patients as compared to samples obtained at a different time and place from normal blood donors [14]. They reported that they were unable to detect XMRV using a specific PCR assay, but that, with somewhat less selective primers, they could detect fragments of sequence identical or closely related to some endogenous P-MLVs and MP-MLVs. Again, the sequences were detected much more frequently in samples from CFS patients than from the poorly matched controls. The close match of these sequences to endogenous MLVs present in over 100 copies per cell in laboratory and wild mice immediately raised the possibility of contamination of the samples with traces of mouse DNA. To counter this concern, Lo et al. also tested the samples for mouse mitochondrial DNA and, finding none, concluded that the results reflected infection of the patients, but not the controls, with polytropic-like MLV. In further support of this conclusion, they reported, that later samples from some of the same patients also contained MLV-like sequences. These sequences showed genetic differences from the earlier ones, leading them to conclude that there was ongoing virus replication and evolution.

The initial reports of association between endogenous MLV-related viruses and human disease were attractive because of their biological plausibility: MLV-related viruses cause a similar variety of diseases in mouse models [1]; close contact between mice and humans can result in zoonotic infection; the virus isolated is highly infectious for at least some human tumor cell lines [17, 38]. As the studies presented here unfolded, however, a number of experimental issues regarding the possibility of detection of endogenous proviral sequences and their confusion with replicating viruses infecting human patients came to light. There are a number of lessons that should be heeded in the development of future studies.