|Home | About | Journals | Submit | Contact Us | Français|
A recent report suggested an association between xenotropic murine leukemia virus-related virus (XMRV) and chronic fatigue syndrome (CFS). If confirmed, this would suggest that antiretroviral therapy might benefit patients suffering from CFS. We validated a set of assays for XMRV, and evaluated the prevalence of XMRV in a cohort of monozygotic twins discordant for CFS. Stored PBMC were tested with 3 separate PCR assays (one of which was nested) for XMRV DNA, and serum/plasma was tested for XMRV RNA by reverse transcription (RT)-PCR. None of the PBMC samples from the twins with CFS or their unaffected co-twins were positive for XMRV, by any of the assays. One plasma sample, from an unaffected co-twin, was reproducibly positive by RT-PCR. However, serum from the same day was negative, as was a followup plasma sample obtained 2 days after the positive specimen. These data do not support an association of XMRV with CFS.
Chronic fatigue syndrome (CFS) is a disabling condition characterized by debilitating physical and mental fatigue lasting at least 6 months, along with associated rheumatologic, infectious, and neuropsychiatric symptoms (Afari & Buchwald, 2003). Estimates for the prevalence of CFS have varied, ranging up to 3% in primary care settings (Bates, et al., 1993; Buchwald, et al., 1995; Reyes, et al., 1997; Wessely, Chalder, Hirsch, Wallace, & Wright, 1997). Despite much research effort, the causes of CFS remain unclear, and the available therapies for affected patients are limited. Previous reports implicating various viral etiologies have not withstood the test of time and reproducibility.
An ongoing methodologic concern in CFS studies is the choice of appropriate control subjects. Various groups have used healthy controls, patients with chronic medical or psychiatric disorders, and sedentary individuals. To address this, one of us (DB) established a registry of twins affected by CFS (Buchwald, et al., 1999). The registry included a number of monozygotic twin pairs discordant for CFS, allowing better control of genetic and familial effects. The Chronic Fatigue Twin Registry has provided critical insights into the biological underpinnings of CFS (Aaron, et al., 2001; Buchwald, et al., 2001; K. H. Claypoole, et al., 2007; Herrell, et al., 2002; Lewis, et al., 2001). While a precise etiology remains elusive, the finding that concordance for idiopathic chronic fatigue is higher in monozygotic compared to dizygotic twin pairs in the registry (55% vs 19%) suggests a major genetic component to CFS (Buchwald, et al., 1999; Buchwald, et al., 2001).
With this background, a recent report by Lombardi et al. has generated considerable interest (Lombardi, et al., 2009). The authors reported detecting xenotropic murine leukemia virus-related virus (XMRV) in 67% of patients with CFS, compared to only 4% of healthy controls. The virus was reported to be transmissible to uninfected cells, and half of patients reported to have XMRV also showed serologic evidence of infection. In addition to identifying a causative agent for a large proportion of CFS cases, these results also logically implied that antiretroviral therapy might be of clinical benefit for CFS patients. Indeed, at least some antiretrovirals show anti-XMRV activity (Paprotka, et al., 2010; Sakuma, Sakuma, Ohmine, Silverman, & Ikeda, 2010; Singh, Gorzynski, Drobysheva, Bassit, & Schinazi, 2010). However, the association of XMRV with CFS has become controversial, with several studies reporting an inability to detect the virus in samples from CFS patients (Erlwein, et al., 2010; Groom, et al., 2010; Henrich, et al., 2010; Hong, Li, & Li, 2010; Switzer, et al., 2010; van Kuppeveld, et al., 2010), while another study suggested that CFS might be associated with a genetically diverse set of murine leukemia virus-related viruses, rather than XMRV alone (Lo, et al., 2010).
Based on the compelling nature of the original report (Lombardi, et al., 2009), we sought to develop a set of independent testing protocols for XMRV. Such testing would allow us to determine the prevalence of XMRV in our own populations, and, if appropriate, to offer testing and potential therapeutic counseling to patients. After validating the performance of our assays, we used them to investigate the prevalence of XMRV in our well-characterized cohort of monozygotic twins discordant for CFS.
A full description of the Chronic Fatigue Twin Registry has been previously published (Buchwald, et al., 1999). Briefly, twin pairs were recruited through advertisements to twin organizations, patient support groups, and general population twin registries, and by letters to clinicians caring for CFS patients and to investigators studying CFS. Twins completed an extensive questionnaire regarding demographics, family history, CFS and other health symptoms, social environment, and neuropsychological measures. Zygosity was assigned via an extensive set of questions shown to provide an accuracy of 95–98% (Buchwald, et al., 1999; Torgersen, 1979). Subjects were classified based on the 1994 Centers for Disease Control and Prevention (CDC) CFS case definition (Fukuda, et al., 1994). This cohort has been extensively evaluated in a number of studies regarding the biology of CFS (Aaron, et al., 2002; Aaron, et al., 2001; Armitage, et al., 2009; Ball, et al., 2004; Buchwald, et al., 2001; K. Claypoole, et al., 2001; K. H. Claypoole, et al., 2007; Herrell, et al., 2002; Lewis, et al., 2001; Mahurin, et al., 2004; Poole, Herrell, Ashton, Goldberg, & Buchwald, 2000; Roy-Byrne, et al., 2002; Watson, Jacobsen, Goldberg, Kapur, & Buchwald, 2004; Watson, et al., 2003), and our group has previously reported on immune parameters associated with CFS (Koelle, et al., 2002; Sabath, et al., 2002). From the original samples obtained in 2001 as described in (Koelle, et al., 2002; Sabath, et al., 2002), material was available for this study from 21 same-sex twin pairs (19 female, 2 male). Matched samples from each of the CFS-affected twins and their infected co-twin were obtained at the same time. After collection, PBMC and plasma samples were stored at −80 C, and serum samples were stored at −20 C. A total of 85 samples were available from the twin pairs, including 67 plasma, 13 PBMC and 5 serum samples.
All laboratory testing was performed by personnel blinded to the CFS status of the specimens. For PBMC samples, DNA was extracted using a Qiagen Blood Kit in a biosafety level 3 laboratory (BL3) along with a control PBMC from an HIV-negative donor and empty blank extraction tubes. Plasma/serum was extracted using a manual method that included a PAW extraction control (Applied Biosystems). Ten microliters of the extracted RNA was run in the X2 RNA assay, while 5 microliters was run in a parallel PAW assay to confirm extraction. Positive controls consisted of RNA or DNA from VP62, a molecular clone of XMRV (Smith, Gottlieb, & Miller, 2010), and was obtained by subjecting VP62-containing cells or 1 ml of VP62 culture supernatant to the same DNA or RNA extraction procedures.
All realtime assays were performed on an ABI 7700. The primer sets used are shown in the Table. The XMRV-F2 and R3 primers targeting the integrase gene were described in a previous study (van Kuppeveld, et al., 2010), while the other primers/probes were designed for this study. The DNA realtime assay was performed using AmpliTaq Gold DNA polymerase (ABI) according to the manufacturer's specifications. Twenty pM of each primer and probe were used per reaction, except for XMRV-2, for which each of the 2 probes were used at 10 pM per reaction. Following an initial 10 min hold at 95°C, the PCR profile was 42 cycles of 95°C for 15 s and 60°C for 30 s.
BIOLASE polymerase (Bioline) was used for the nested PCR reactions according to the manufacturer's specifications. Twenty pM of each primer was used. Ten microliters of sample DNA was used in the first PCR and one microliter of this reaction went into the nested PCR. The profile for each run was a two minute hold at 95°C followed by 35 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 30 s followed by a 7 minute hold at 72°C.
The RNA realtime assay (using the X2 primers/probes in the Table) for plasma/serum specimens was performed using the rTth enzyme (ABI) according to the manufacturers specifications. Probe and primer concentrations were the same as in the DNA assays. Following a 15 minute hold at 50°C and a 2 minute hold at 95°C, the profile was 42 cycles of 95°C for 15 s and 60°C for 30 s. The integrase assay shown in the table used agarose gel/SYBR green detection of product rather than real-time analysis.
The amplicons for the 3 XMRV PCR assays were synthesized and used as templates in PCR reactions, and the resulting material cloned into the TOPO TA vector (Invitrogen). These plasmids were quantitated on a nanodrop spectrophotometer and used as standards in the DNA realtime assays. For RNA standards, the plasmids were linearized with HindIII (New England Biolabs) and RNA synthesized from them using the Ambion MEGAshortscript T7 kit. The quality of the RNA was confirmed by running the sample on an Agilent Bioanalyzer and the quantity was determined with the nanodrop spectrophotometer. The XMRV-2 RNA was then aliquoted and frozen at −80 for use as the standard for the RT-PCR assay.
The lower limit of detection (LLD) was determined for the DNA realtime assay by spiking 50,000 PBMC with plasmid DNA for final concentrations of 2, 10, 20, 30, and 50 copies/reaction. The lowest dilution in which the PCR assay was able to detect DNA in duplicate samples was 20 DNA copies/reaction.
To determine the LLD for the nested PCR assay, the XMRV-2 amplimer was gel-purified, quantitated with a nanodrop spectrophotometer, and serially diluted in a background of 50,000 negative PBMC DNA. The nested assay was able to detect a single copy per reaction.
The LLD for the RNA assay was determined by serially diluting the RNA standards similarly to DNA LLD procedures, and was found to be 20 copies per ml.
PBMC samples were available from 13 twins, including 6 twins affected with CFS and their unaffected co-twins, and one additional unmatched and unaffected co-twin only. The maximum possible volume of DNA per reaction was run in singlet, giving a range of 65–399 ng total input. All samples were subjected to 4 realtime PCR reactions - 3 for XMRV (X1, X2, and X-CSF), and a beta globin assay to confirm the DNA was amplifiable and to confirm the amount of input DNA as estimated by the spectrophotometer values. All samples were negative by all of the real-time XMRV assays.
To improve the assay limit of detection we designed an outside primer set for the X2 PCR assay, allowing us to perform nested PCR. Again, all 13 samples were run with nested PCR and were negative. Negative controls included the HIV-negative donor DNA and no template reactions, which were both negative. The positive control consisted of DNA extracted from cells transfected with the VP62 molecular clone.
Sixty-seven plasma and 5 serum specimens were available from 21 twin pairs (i.e. 21 CFS-affected subjects and their 21 unaffected twins). We did not detect XMRV in any of the samples from CFS-affected subjects using the X2 primer/probe sets. For three samples (one from a CFS-affected subject and 2 from unaffected co-twins), the internal extraction control did not amplify (none of the 3 showed amplifcation of XMRV), and these were therefore not categorized with regard to their XMRV status. Only a single plasma sample, from one unaffected co-twin, had detectable XMRV, at 5000 copies/mL using the X2 RNA assay. A second 10 microliter RNA aliquot of this sample was run and was again positive. However, a serum sample obtained on the same date was negative, and another plasma sample obtained from the same person 2 days later was also negative. Primers for the integrase region confirmed the positive sample (negative controls included with this run were negative). Efforts to sequence the amplicons were not successful, and thus we can not confirm or rule out the possibility that the single positive sample was contaminated with control or other exogenous material.
The identification and confirmation of a viral cause for CFS would have major implications for those afflicted by this condition, which affects approximately 0.2% of the adult population (Jason, et al., 1995). If a viral cause were identified, especially if treatable with current antivirals (as a retrovirus such XMRV presumably might be), this would provide new treatment options directed at the root cause of this condition, potentially benefiting millions of affected individuals.
Given the critical clinical need and the compelling nature of the original report, multiple laboratories, including our own, have sought to confirm the association of XMRV with CFS. To the surprise and disappointment of many, the majority of such studies reported to date have failed to confirm an association of XMRV with CFS. Most of these studies have relied heavily on PCR for detection of virus (Erlwein, et al., 2010; Groom, et al., 2010; Henrich, et al., 2010; Hong, et al., 2010; Switzer, et al., 2010; van Kuppeveld, et al., 2010). One study also used a neutralization assay to evaluate serologic status (Groom, et al., 2010), and also failed to confirm an association of XMRV with CFS. A particularly thorough study was performed by investigators at the US Centers for Disease Control and Prevention (Switzer, et al., 2010), who performed Western blot assays on sera from 121 US blood donors and 26 retrovirus-infected individuals (HTLV 1/2, HIV-1, or dual HIV-1/HIV-2 infection), and on plasma from 51 patients with CFS and 53 controls, all of which were found to be negative for XMRV. Similarly, PCR testing of PMBC from 50 CFS patients, 56 controls, and 41 US blood donors was also negative. The negative status of the PBMC from these CFS patients and controls was also confirmed by an independent laboratory.
In contrast to the preponderance of negative studies, a recent paper from a group at the US National Institutes of Health found evidence, in a cohort of patients diagnosed with CFS, for another genetically diverse set of murine viruses, collectively referred to as murine leukemia virus (MLV)-related viruses (Lo, et al., 2010). MLV-related viral sequences were found in 32 of 37 patients with CFS, compared to only 3 of 44 healthy volunteer blood donors. The sequences of these viruses were more closely related to polytropic mouse endogenous retroviruses than to XMRV. Thus, although not directly confirming the initial report (Lombardi, et al., 2009), these findings suggest the possibility of a retroviral association with XMRV.
The reasons for the disparities between the published studies are unclear, but several possibilities present themselves. First, these studies have been performed on a wide variety of populations, and genetic or geographical variations could explain the differences observed. This is especially true if the findings of Lo, et al. (Lo, et al., 2010) prove correct, and CFS is associated with a genetically diverse set of murine-related viruses. In this scenario, XMRV might represent only one of many potential causative viruses, and it might have been overrepresented in the cohort of the original study (Lombardi, et al., 2009). Thus, PCR assays designed specifically to amplify XMRV might not detect other potentially causative viruses.
Second, the published reports have used a variety of molecular and serologic assays in their searches for XMRV. It is possible that methodological differences in primers, probes, or assay conditions could explain the disparate PCR results. Arguing against this explanation, the published papers come from groups experienced in the detection of human viruses, including retroviruses, and one would be hard-pressed to provide an example of another virus for which a wide variety of sensitive and specific assays can not be designed. A recent study carefully attempted to replicate the methodology of the original report by Lombardi et al (Lombardi, et al., 2009). Using this methodology, Shin et al were unable to detect XMRV in a cohort of 100 CFS-affected patients, or in 14 patients who were evaluated in the original Lombardi study (Shin, et al., 2011). Although some authors have called for the standardization of assays, we would argue that this is premature because no gold standards yet exist; the pressing need at this time is for open sharing of specimens and testing protocols between laboratories to determine the basis for lack of agreement. A final and unfortunate possibility is that the exquisite sensitivity of PCR may be allowing the unintended amplification of murine sequences contaminating the human samples (Hue, et al., 2010; Oakes, et al., 2010; Robinson, et al., 2010; Sato, Furuta, & Miyazawa, 2010; Shin, et al., 2011; Smith, 2010). Because mouse tissues are common in scientific laboratories, this possibility cannot be ignored, and rigorous phylogenetic sequence analysis is imperative.
The main strength of our study was the use of monozygotic twins discordant for CFS status, which we reasoned would allow us to evaluate an association between XMRV and CFS, while minimizing confounding from host genetic factors. In addition, our cohort was recruited from throughout the United States, which could minimize the effect of any hypothesized geographic differences in XMRV prevalence. However, our study was limited by a small sample size (6 CFS-affected twins), and we did not detect XMRV in any samples from the CFS-affected twins. While it is possible that investigation of a larger cohort might have identified XMRV in some CFS-affected twins, our results do not appear to be consistent with the original report of XMRV being present in 67% of patients with CFS (Lombardi, et al., 2009).
Clearly, the question of a role for XMRV in CFS needs to be resolved quickly and definitively. The clinical need for answers for CFS has motivated many laboratories to investigate this issue. Substantial resources have already been expended in this effort, as witnessed by the large number of publications on this topic. It is critical that all interested and involved groups share their results in the scientific literature, which will speed the search for answers. Persons suffering from CFS will benefit when the role of XMRV is resolved, either through the identification and treatment of a causative virus, or though a renewed focus of the scientific community on the search for the underlying cause of this devastating condition.
Supported in part by grants from (R01 AR051524 (N. Afari, Principal Investigator) and U01DK082325 (D. Buchwald, Principal Investigator)
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Potential conflicts of interest: none