Background
Fibromyalgia and Chronic fatigue Syndrome (FM/CFS) are characterized by long-lasting debilitating fatigue, often accompanied by widespread muscle pain. Its incidence is rising and ranges between 0.5-6% with high prevalence in females [
1]. The high socioeconomic expense that this disease supposes is worrisome.
Currently the FM/CFS diagnosis is made solely on clinically grounds, as no biological markers associated with the disease have been found. Alteration in cytokine profiling, decreased function of natural killer (NK) cells, presence of autoantibodies and reduced response of T cells to mitogens and other specific antigens have been reported in these patients [
2‐
4]. The observed high level of pro-inflammatory cytokines may explain some of the manifestations such as fatigue and flu-like symptoms and modify NK cells activity. Existence of RNaseL isoforms reflecting an aberrant immune system function is also well documented [
5,
6]. Identification of markers consistently associated with this pathology will enable clinicians to effectively diagnose FM/CFS, follow the progress of the disease, monitor the effects of therapeutic approaches and probably develop preventive programs.
Since many of the symptoms characterizing FM/CFS resemble those of infectious diseases other authors have investigated a possible viral etiology of the disease. In October 2009, Lombardi
et al. reported the finding of a novel gammaretrovirus named xenotropic murine leukemia virus-related virus (XMRV) in about 67% (68/101) of CFS patient DNA samples prepared from peripheral blood mononuclear cell (PBMC) compared to only 3.6% (5/218) of the samples prepared from healthy controls [
7]. This finding, followed by the report on the presence of MLV-related virus sequences sharing 96% of identity with the XMRV in blood of patients with CFS published in 2010 by Lo
et al.[
8] prompted us to examine whether our local population of fibromyalgia patients affected by chronic fatigue also presented evidence of murine-leukemia virus-related infection. Our initial positive amplifications of XMRV sequences coincided with numerous reports attributing XMRV amplification to murine DNA contamination of reagents [
9‐
14], making us investigate the possibility of spurious amplifications.
A second retrovirus, namely the human T-lymphotropic virus type 2 (HTLV-2), had also been linked to chronic fatigue by a single study [
15]. Contrary to XMRV screening studies, evaluation of HTLV-2 in different cohorts of FM/CFS patients has been scarce [
16‐
18]. This latter retrovirus, closely related to the HTLV-1 virus causing adult T-cell leukemia and lymphoma and the demyelinating disease called myelopathy/Tropical spastic paraparesis (HAM/TSP) has not yet been clearly associated with any disease. Since DeFreitas
et al. reported that 83% (10/12) of chronic fatigue patients had a positive amplification of
gag HTLV-2 sequence compared to 0% (0/20) in matched healthy subjects we considered that a screening of HTLV-2 sequences in our patient cohort was pertinent.
In addition, antiretroviral therapy had been shown to be of limited value in controlling HTLV-2 virus expression in patients co-infected with a second retrovirus, at least for co-infections with the HIV virus [
19,
20]. Thus, finding out whether any of the participating patients were co-infected with a second retrovirus could be of relevance to patient care.
In this study we evaluated the presence of both retroviruses: XMRV and HTLV-2 formerly associated to chronic fatigue, and their close relatives, in a cohort of 75 Spanish fibromyalgia patients affected of chronic fatigue and 79 population-matched healthy subjects. This is, to our knowledge, the largest cohort of patients suffering from chronic fatigue evaluated for infection with the HTLV-2 virus.
Discussion
This is the first report to our knowledge that evaluates the presence of the both families of retrovirus formerly associated with chronic fatigue: the gamma retroviruses MLV-related and the delta retrovirus HTLV-2 in a cohort of patients suffering from fibromyalgia and chronic fatigue.
At present enough evidence exists to support the inexistence of a natural form of the XMRV virus in humans [
14,
27,
28]. However, since the XMRV is competent to infect and replicate within human cells [
29] and some of its relatives: the xenotropic-MLV (X-MLV) viruses can infect non-human primates and other species causing disease [
22,
30], the possibility of a human xeno-infection by an MLV-related virus remains.
Opposite to the actively studied relationship of the XMRV to chronic fatigue, the association or non-association of the orphan HTLV-2 retrovirus to the disease has not been clarified. The initial study which used PCR amplification of HTLV-2 proviral sequences followed by probe hybridization assays to find an association of HTLV-2 infection to CFS was followed by only a few negative studies [
16‐
18]. Even though this technique is highly specific and sensitive, amplification by nested PCR may be superior when detection of low levels of infection and/or virus variants is pursued. In addition none of the three negative studies that followed the initial one included more than 30 patient blood samples [
16‐
18], and therefore further investigation including larger cohorts of patients was at need.
This, to our knowledge, constitutes the study with the largest number of patients (n = 75) of fibromyalgia suffering from chronic fatigue examined for infection by HTLV-2. None of the samples assayed showed specific amplification of HTLV-2 or related sequences at any time making us conclude a non-association between HTLV-2 retroviral infection and the chronic fatigue condition of our patients. As we already proposed [
31] HTLV-2 association with CFS reported by DeFreitas [
15] may also be caused by undetected technical problems like it occurred with XMRV and CFS [
7].
It is also interesting to note that the molecular weight of one of the two bands obtained in our first PCR amplification coincides with the endogenous unspecific 440 bp
gag band obtained by Gow
et al. [
17] in the first round PCR step. However, different to them, we could only observe it in 2/154 samples assayed (Figure
3, panel A) and therefore we cannot coin it as endogenous.
Contrary to the negative results obtained in the HTLV-2 screen, our initial screening analysis of XMRV and MLV-related sequences yielded a moderate percentage of positives 20% (30/154) with a slight predominance on the patient subgroup (24% positive patients versus 15% positive healthy controls).
The lack of sequence diversity of our PCR amplified products with respect to the american VP62 XMRV isolate we were using as the positive control made us suspect our samples could be contaminated with a specific template. Even though our lab does not harbor mice work and nor mouse-derived or XMRV-infected cell lines are maintained we proceeded to perform a sensitive IAP test to evaluate whether the env positive amplifications could come from mouse DNA contamination in any of the components used. Different to other researchers we could not find any presence of murine DNA.
If the obtained amplified sequences were really derived from our participant blood samples we should be able not only to reproduce the obtained results from an independently prepared sample from the same patient but also to detect other parts of the viral genome. The fact that the initial positive amplification of the XMRV env sequence could not be reproduced in an independent extraction of gDNA prepared from frozen aliquots of the same exact samples, added to the complete lack of amplification of other regions of the viral genome, specifically the gag sequence, made us conclude that the initial amplification should have come from fortuitous air driven spreading of env specific amplicons, most likely product of the first round of PCR of our VP62 XMRV positive control. This conclusion points out the extreme importance of independently confirming positive results when using highly sensitive assays.
In addition to the two initial reports by Lombardi
et al. and Lo
et al. which showed a striking association between the presence of XMRV or MLV-related sequences and a chronic fatigue condition, both of which are currently retracted, some other groups of researchers have also reported positive amplification of murine gamma retrovirus sequences in human samples. Most amplifications have been attributed to sample contamination with mouse DNA [
9‐
14], other, confirming lack of contaminant mouse DNA, have not been able to find an alternative explanation, but as in our study, their positive results could never be confirmed for a second viral gene in the same sample [
32‐
34]. It is therefore possible that their inconsistent amplifications may come from spurious amplicon contamination similar to the ones we detected. Alternatively, inconsistent PCR amplifications could come from extremely scarce target sequence presence. However, being that most of the studies finding inconsistent viral sequence amplifications, including ours; used 0.5-1 μg of gDNA template in their assays [
32,
35] it is very unlikely this could be the case. Spiking amplification assays determined 1–10 copies to be the lower threshold limit for successful nested PCR amplification in the context of 1 μg of human gDNA, coinciding with previous reports [
35], and since each somatic cell has 6.16 pg of DNA, less than 1 in 3 × 10
5 cells would need to be infected for such a scenario to take place.
Although, great efforts had been directed to look for MLV related infection in CFS patients only a few [
33,
36‐
38] included fibromyalgia patients in the studied groups, and only one of them included a larger cohort of patients that our study [
33].
To ensure that our testing would not miss genetically diverse XMRV, MLV or HTLV-2 strains we decreased the annealing temperature to allow missprimed amplification. Even though some bands were amplified under these non-stringent conditions none of them contained viral sequences, most probably indicating absence of related infections in our patient and control participants.
Because negative PCR results would be obtained from defective samples, we considered necessary to rigorously perform controls to guarantee that lack of amplification was not due to degraded or impure DNA preps. Even though amplification of house-keeping genes which are present at a frequency of 2 copies per cell does not guarantee amplification of less abundant templates, at least rules out the presence of potent inhibitors of PCR amplification.
It was noticed that the GAPDH primer set used in Figure
1 of the Lombardi
et al. study [
7] amplifies a human non-related genomic sequence in chromosome X of a similar size to expected GAPDH RT-PCR product (228 vs 227 bp) and therefore careful interpretation of the results obtained with it should be taken. In our study it served the purpose of detecting amplification of a region of the genome (Additional file
1: Figure S1, panel B) to evaluate the quality of the gDNA preps screened.
Competing interests
The authors declare they have no competing interests.
Authors’ contributions
EO, AMD, VM and GCO conceived and designed the study. VM, AMD, MGE and GCO confirmed patient diagnostic and collected data on participating patients. EO, MGE and AMD performed specimen testing and data analysis. EO and AMD wrote the manuscript. All authors read and approved the final manuscript.