Chronic Reactivation of Persistent Human Herpesviruses EBV, HHV-6 and VZV and Heightened Anti-dUTPase IgG Antibodies Are a Recurrent Hallmark in Post-Infectious ME/CFS and is Associated With Fatigue

Abstract

Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a debilitating disease with unknown etiology and heterogeneous symptomology for which there are no validated tests for definitive diagnosis. We examined 873 longitudinal serum samples from ME/CFS patients (n = 40) and 378 from healthy control individuals (n = 16) for differences in human herpesvirus and endogenous retrovirus-K (HERV-K) dUTPase IgG antibodies by ELISA. The results of this study demonstrate a significant increase in dUTPase IgG antibodies to the herpesviruses Epstein-Barr virus (EBV), human herpesvirus-6 (HHV-6) and varicella zoster virus (VZV) in ME/CFS compared to healthy-controls (p < 0.001). Notably, 72.5% (n = 29) of ME/CFS patients simultaneously co-expressed antibodies to multiple herpesvirus and HERV-K dUTPases compared to 31% (n = 5) of the healthy controls. Chi-square test analysis showed strong associations for EBV, HHV-6 and VZV dUTPase antibodies seropositivity (p < 0.001) and Spearman correlation analysis revealed significant positive associations of EBV and HHV-6 dUTPase IgG antibodies with fatigue. Further examination of the distribution of dUTPase antibodies across fatigue severity groups show that heightened dUTPase IgG levels cluster with ME/CFS patients exhibiting moderate and severe fatigue. These findings highlight the importance of examining herpesvirus dUTPase IgG across severity groups in aiding with current challenges for stratifying ME/CFS patients due to the heterogeneity in symptomology.

1 |. Introduction

Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a chronic multi-system illness whose symptoms include post-exertional malaise, long-term fatigue, sleep disturbances or non-restorative sleep, as well as immune and cognitive dysfunction [1]. While numerous hypotheses have been proposed to explain the etiological causes of ME/CFS, they remain unknown. At the present time the general consensus is that there might be multiple etiological agents that trigger disturbances in various physiological processes culminating in the development of symptoms associated with ME/CFS.

While various criteria have been established for diagnosing ME/CFS including the Fukuda, Canadian, International and IOM [1–4], in most cases clinical diagnosis is based on exclusion of other diseases and syndromes. Many ME/CFS studies have reported using classifications based on symptomology such as chronic fatigue or post infectious fatigue, but a closer examination of these studies revealed that most subjects did not meet any established diagnosis criteria for ME/CFS. Diagnosis is further hampered by the lack of available validated biomarkers. It has been estimated that the prevalence of ME/CFS in the US adult population is 1.3% with an economic burden of $18–51 billion annually [1, 5]. Thus, identifying triggers and drivers of ME/CFS is imperative for understanding the molecular mechanisms by which they contribute to the development and/or progression of this illness so that alternative research-based diagnostic tools and more effective therapeutics can be developed.

In over half of ME/CFS cases, disease onset is associated with acute “flu-like” symptoms [6].

There are multiple reports in the literature suggesting a role for viruses as potential triggers of ME/CFS, and particularly the herpesviruses, Parvovirus B19 and enteroviruses [7–9]. Recent reports promoting persistent enterovirus infections as a potential trigger/driver of ME/CFS [10, 11], have not been validated. A causal relationship between a virus and ME/CFS has not been conclusively demonstrated due in part to the heterogeneity of the patient population and to the possibility of multiple etiologies. Furthermore, elucidation of how a virus may contribute to the pathophysiological changes observed in subjects diagnosed with ME/CFS has been hampered by studies primarily focused on detecting an increase in viral load in patients rather than looking for macromolecules produced by these viruses as orchestrators of the pathophysiological disturbances associated with ME/CFS. Recently it was suggested that future studies on ME/CFS should “focus on adaptive immune responses rather than surveillance for viral gene products” [12]. We believe this conclusion is based on the incorrect premise that an increase in viral load is required to indicate virus involvement in disease pathogenesis and ignores modern, more advanced technologies, including single cell analyses, demonstrating that most dynamic viral infections are abortive-lytic in nature and thus, express many virus proteins that could contribute to disease pathogenesis in the absence of an increased in viral load.

Of the viruses reported to be associated with ME/CFS there is accumulating serological and molecular evidence that the human herpesviruses Epstein-Barr virus (EBV) and human herpesvirus-6 (HHV-6) may be involved in the pathogenesis of a subset of ME/CFS patients either as triggers or drivers of the illness. Herpesviruses are very successful pathogens due in part to their ability to establish persistent/latent infections. Primary infection usually occurs within the first 2 years with HHV-6 while 50% of children are infected with EBV by age 5 [13]. In the adult population greater than 95% are persistently infected with these viruses [13]. Varicella-Zoster virus (VZV), another herpesvirus which is the causative agent of chickenpox in adolescents and shingles/zoster in adults is also ubiquitous in nature with greater than 90% of adolescents being infected [13]. Most infections with these herpesviruses (EBV, HHV-6, and VZV) are asymptomatic and routine screening of patients in a clinical setting is not usually done [13].

Multiple factors such as infection, suppression or impairment of the immune system and physiological changes associated with nerve trauma can trigger reactivation of the herpesviruses resulting in the production of early viral proteins, including the deoxyuridine triphosphate nucleotidohydrolase (dUTPase). Reactivation of herpesviruses has been associated with the development and/or pathophysiology of multiple human diseases, including ME/CFS, multiple sclerosis, rheumatoid arthritis, systemic lupus and Alzheimer’s disease [14–19]. Interestingly, the detection of proteins encoded by EBV; (BLLF3, dUTPase) and HHV-6B (U94; OHV3; gB; p41) in several regions of the brain (hippocampus, choroid plexus, mid-brain and amygdala) and nerve root tissue (cervical and lumbar/sacral) in individuals with ME/CFS supports the concept that multiple herpesviruses are capable of simultaneously reactivating in both the peripheral and central nervous systems [20]. Furthermore, a study by Halpin et al [21] found significantly higher serum levels of EBV-, HHV-6-, and VZVdUTPase IgG anti-dUTPase antibodies in ME/CFS cases. Herpesvirus and HERV-K dUTPases have been shown to have immunomodulatory properties, acting as a pathogen-associated molecular pattern proteins for toll-like receptor 2 (TLR2) [22, 23] and TLR4 [24, 25], respectively, and to be potent inducers of pro-inflammatory cytokines via NF-κB activation. They have also been shown to modulate naïve T cell differentiation via activin A and to induce neuroinflammatory mediators as well as the expression of genes involved with innate and adaptive immunity in vitro and in vivo [26–28]. Altogether, the data suggest a potential role for herpesvirus reactivation and virus dUTPase secretion in the pathophysiology of chronic multi-system illnesses such as ME/CFS.

Approximately 8% of the human genome is comprised of human endogenous retroviruses (HERV), which are remnants of ancient retroviral integrations in the germ line. The HERV-K (HML- 2) group, the most recent retrovirus to colonize the human germ line, is closely related to the mouse mammary tumor virus (MMTV), an exogenous beta retrovirus [29]. Although all HML-2 proviruses are defective in at least one gene, many of them possess complete open reading frames (ORFs) with coding capability. HML-2 transcripts and proteins have been detected in healthy tissues including the brain [29–32]. HERV-K, like their exogenous members encode for a dUTPase that is expressed as a trans-frame polypeptide through ribosomal frameshifting [30]. We have previously shown the potential role of HERV-K dUTPase in psoriasis susceptibility and pathology [33, 34], pulmonary arterial hypertension PAH [24, 25, 35]; and most recently HERVs have been proposed to possibly contribute to the pathology of ME/CFS [36–40].

The purpose of this study was to expand on and validate our previous findings showing heightened levels of IgG antibodies against the dUTPase protein of multiple herpesviruses [21] using a different ME/CFS cohort of longitudinal sera samples collected over a 3-month timeline compared to age and gender-matched healthy controls as well as to determine potential associations between serum levels of dUTPase IgG and self-reported fatigue and/or pain severity scores within the study cohorts. Notably, we found significant positive associations between serum levels of EBV and HHV-6 dUTPase IgG antibodies with fatigue severity which has never been described before in ME/CFS patients. This study also confirmed the presence of heightened levels of IgG antibodies to multiple herpesvirus and HERV-K dUTPases in the absence of parallel increases in IFN-γ, suggesting a deficient IFN-γ response in this cohort of ME/CFS patients.

2 |. Methods

2.1 |. ME/CFS and Control Study Cohorts

Serum samples and demographics information were obtained from Dr. Jarred Younger, University of Alabama at Birmingham (UAB). All procedures were approved by the University of Alabama at Birmingham (UAB) institutional review board (IRB-150303004), and all participants gave written informed consent. The study subjects were female individuals between ages 19–65 in the College of Community Health (CCHA) patient database who were pre-screened for ME/CFS using the Fukuda case definition without Reeves modification [2]. Participants were excluded if presenting any of the following: positive rheumatoid factor or anti-nuclear antibody; C-reactive protein ≥ 10 mg/L; ESR > 60; pregnant or plans to become pregnant; autoimmune disorder; major psychological/psychiatric disorder; acute or active chronic infection; blood clotting disorders; use of blood thinning medications; rheumatologic disorder, current participation in a clinical trial, individuals undergoing daily opioid therapy or taking anti-viral or anti-bacterial medications (e.g., valganciclovir, valacyclovir, acyclovir, famciclovir, cidofovir, ganciclovir, and doxycycline). Participants had to wait at least 6 months after experimental ME/CFS therapeutics before participating in the study. All other medications must have been held consistently for at least 2 months prior to study participation, with no plans to change therapies over the course of the study. The mean age of the ME/CFS participants was 42.49 ± 10.13 years (range 20–62 years). Self-reported fatigue and pain mean severity scores were 7.052631579 ± 1.541276558 (range: 4–9) and 5.086956522 ± 2.32618 (range: 0–8), respectively.

2.1.1 |. Healthy Controls

Healthy controls were recruited from an existing database in the Neuroinflammation, Pain & Fatigue lab in the Department of Psychology at University of Alabama, Birmingham (UAB). Healthy controls were selected by Dr. Jarred Younger to match participants with ME/CFS by age. The mean age of the control group was 41.53 ± 4.81 years; (range 24–62 years). Fatigue and pain severity were assessed as described [41]. Self-reported fatigue and pain mean severity scores were 0.5 ± 1.032795559 (range: 0–4) and 0.5 ± 0.631365261, respectively.

2.1.2 |. Serum

Blood was collected and processed from the subjects as described previously [33]. All blood specimens used in this study were collected prior to the start of the SARS-CoV-2 pandemic (2019). Serum was stored at −80°C until use.

2.1.3 |. Purification of Recombinant dUTPase Proteins

Purification of the recombinant dUTPase proteins encoded by EBV (BLLF3), HHV-6A (U45), VZV (ORF8) and were performed as described previously [23, 33, 42, 43]. Purity of recombinant dUTPases was determined by SDS-PAGE and further tested for the presence of contaminants (LPS, peptidoglycan, DNA or RNA) as previously described [23, 24, 33, 42, 43]. The purified recombinant herpesviruses-dUTPase proteins used in these studies were stored at −80°C. Protein solutions were thawed a maximum of 2 times before use. Protein concentration was determined in a Qubit fluorimeter (Invitrogen Carlsbad, CA). The U45 gene from HHV- 6 A strain GS, was the source of the dUTPase used in this study. The recombinant protein exhibits greater than 94% identity with the U45 protein encoded by HHV-6B strain Z29 U45 gene and therefore human antibodies to these proteins are likely cross-reactive. Thus, to avoid confusion we used the terminology HHV-6 dUTPase throughout.

2.1.4 |. Anti-Virus dUTPase Antibody Determination

ELISAs were performed as described [21, 34, 44]. Briefly, 96-microtiter well plates (Nunc-Immuno Plate MaxiSorp Surface) were coated overnight at 4°C with recombinant dUTPase protein at 2.5 μg/mL in phosphate buffered saline (PBS). Plates were washed in a Biotek ELx50 plate washer 3x with PBS/0.05% Tween 20 and blocked with blocking buffer (PBS/2.5% BSA) at room temperature (RT). All serum samples were used at a 1:800 dilution in blocking buffer and incubated for 2 h at RT. Plates were washed 3x with PBS/0.05% Tween-20 followed by incubation with anti-human-IgG horseradish peroxidase (HRP)-conjugated secondary antibody (Sigma Chemical Co., St. Louis, MO) at 1:1000 dilution in blocking buffer at RT for 1 h. Plates were washed 6x with PBS/0.05% Tween 20 and incubated for 15 min with 100 μL of tetramethyl-benzidine (Invitrogen). The reaction was stopped with H₂SO₄ (2 M) and plates were read at 490 and 690 nm for background on a Lab Systems Multiskan MCC/340 plate reader using the Genesis v3.05 Life Sciences Ltd software. The background from the 490 nm uncoated wells and PBS-BSA (negative controls) were subtracted from the mean absorbance of the coated wells. Coefficient of variation (CV) was less than 15%. A positive reaction was defined as a serum sample that led to a signal three times over the background OD of the control serum. If 25% of the participants’ serum samples were positive for antibodies for a particular virus dUTPase, that individual was reported as positive.

2.1.5 |. Cytokine ELISA

Serum concentrations of selected analytes were measured by chemiluminescence using Meso-Scale Discovery (MSD; Rockville, MD) multiplex U-PLEX Custom Immune assay (Human) kits, in accordance with the manufacturer’s instructions. A twofold dilution of serum samples was performed for MSD plates measuring CXCL13, IFN-γ, and IL-21. Analysis of serum levels for these analytes were conducted using a Sector Imager 2400 instrument and the MSD Data Analysis Toolbox software. The Lower limit of detection (LLOD) for the MSD analytes measured in pg/mL are: 1.00 (CXCL13) and 1.7 (IFN-γ). Activin A concentration was determined by ELISA as described by the manufacturer (EHACTIVINA; Invitrogen). LLOD for activin A was 15 pg/mL. The concentration range of these cytokines/chemokines in healthy individuals based on the literature are as follows: activin A: 1270 pg/mL (225.9–1525.6); CXCL13: 100 pg/mL (0–126); IFN-γ: 8.60 pg/mL (2.67–14.46).

2.1.6 |. Statistics

GraphPad Prism 10 software (GraphPad Software, La Jolla, CA) and Pithon 3.8 were used for all statistical analyses. Differences in Abs between two groups were assessed by two-tailed Mann- Whitney U test while differences/relationships between categorical variables were assessed by Kruskal Wallis. Values of p < 0.05 were considered statistically significant.

3 |. Results

3.1 |. Demographics

The ME/CFS cohort consisted of 40 female cases with an average age of 42.49 ± 10.13 years ranging from 20 to 62 years, while the healthy cohort consisted of 16 female participants average age 41.53 ± 11.07 years and a range of 24–62 years.

3.2 |. Heightened dUTPase IgG Antibodies to Multiple Herpesviruses is a Recurrent Hallmark of Infectious ME/CFS

We have previously shown that 55% of ME/CFS patients were positive for dUTPase antibodies (Abs) to one or multiple human herpesviruses [21]. Expanding on these observations we conducted a follow-up dUTPase serology validation study aimed at elucidating for the first time relationships between serum levels of EBV, HHV-6, VZV dUTPase IgG Abs and self-reported pain or fatigue scores, as well as the possible involvement of human endogenous retrovirus K (HERV-K) dUTPase in the pathophysiology of ME/CFS using a large collection of 873 longitudinal serum samples from ME/CFS patients (n = 40) and 378 from healthy controls (n = 16) obtained daily or biweekly over a period of approximately 30-90 days. As shown in Table 1, a striking 72.5% (n = 29) of ME/CFS patients were positive for Abs to two or more herpesvirus dUTPases compared to 31% (n = 5) in the healthy control cohort. Furthermore, 30% (n = 12) of the ME/CFS serum samples examined were positive for Abs to all EBV, HHV-6, VZV and HERV-K dUTPases whereas in the control group only 6% (n = 1) co-expressed Abs to all virus dUTPases. Of the 40 ME/CFS patients examined, 15% (n = 6) were negative for antibodies to all virus dUTPases compared to 50% (n = 8) in the healthy control group. An interesting and unexpected finding was the high prevalence of Abs to HERV-K dUTPase only in the healthy controls (50%, n = 8) compared to just 5% (n = 2) in the ME/CFS cohort (Table 1). The expression patterns of dUTPase IgG Abs to EBV, HHV-6, VZV and HERV-K for individual ME/CFS patients and healthy controls over the study period are presented in Figures 1–4. Overall, individuals seropositive or seronegative for any of the virus dUTPases remained so for the duration of the study.

TABLE 1 |.

Prevalence of anti-EBV, HHV-6, VZV and HERV-K dUTPase antibodies in ME/CFS and Control cohorts.

Virus dUTPase Ab subgroup	Health controls N (% of Total)	ME/CFS N (% of Total)
Negative for all	8 (50%)	7 (17.5%)
Positive for all	1 (6.25%)	12 (35%)
EBV positive only	0 (0.0%)	0 (0.0%)
HHV-6 positive only	0 (0.0%)	1 (2.5%)
VZV positive only	0 (0.0%)	1 (2.5%)
HERV-K positive only	3 (18.75%)	2 (5%)
EBV & HHV-6 positive	0 (0.0%)	2 (5%)
EBV & VZV positive	0 (0.0%)	0 (0.0%)
EBV & HERV-K positive	1 (6.25%)	0 (0.0%)
HHV-6 & VZV positive	0 (0.0%)	1 (2.5%)
HHV-6 & HERV-K positive	0 (0.0%)	3 (7.5%)
VZV & HERV-K positive	1 (6.25%)	0 (0.0%)
EBV, HHV-6 & HERV-K positive	0 (0.0%)	3 (7.5%)
EBV, VZV & HERV-K positive	2 (12.5%)	2 (5%)
HHV-6, VZV & HERV-K positive	0 (0.0%)	1 (2.5%)
EBV, HHV-6 & VZV positive	0 (0.0%)	5 (12.5%)
Total	16 (100%)	40 (100%)

Note: Longitudinal serum/plasma samples from ME/CFS (11–25 serum samples/individual; N = 40 participants, a total of 873 serum samples) or a healthy control cohort (11–25 serum samples/individual; N = 16, a total of 377 serum samples) were examined for anti-HHV-6 A, EBV, VZV and HERV-K dUTPase antibodies by ELISA as described in Methods. Cohort columns indicate the number of subjects (N) and prevalence (% of total) of individuals who are positive for dUTPase Abs against a single or multiple herpesviruses or negative. A positive reaction was defined as a serum sample that led to a signal three times over the background OD of the control serum (99% Confidence interval). An individual was considered positive for antibodies to a specific dUTPase if a positive reaction occurred in 25% of the serum/plasma samples.

FIGURE 1 |.

Herpesvirus dUTPase IgG antibody expression patterns in ME/CFS patients. Longitudinal sera (11–25 samples/participant) collected over a 3-month period from 10 of 40 ME/CFS patients (PIDs: 181, 114, 142, 115, 118, 172, 125, 176, 187, 116) and 4 of 16 healthy controls (CTLs: 9,14, 20, 57) were examined for antibodies to the dUTPase protein of EBV, HHV-6, VZV and HERV-K by ELISA, as described in Methods. (A-C) Graphs of virus anti-dUTPase IgG Abs changes over time in individual ME/CFS patients (A-B) and healthy controls (C). Data represent the mean ± SEM of n ≥ 3 measurements/sample for each recombinant dUTPase protein at each time point. A positive reaction was defined as a serum sample that led to a signal three times over the background OD of the control serum (EBV = 0.392, green dotted line; HHV-6 = 0.327, blue dotted line; VZV = 0.290, red dotted line; and HERV-K = 0.292, pink dotted line; 95% Confidence interval).

FIGURE 4 |.

Herpesvirus dUTPase IgG antibody expression patterns in ME/CFS patients. Longitudinal sera (11–25 samples/participant) collected over a 3-month period from 10 of 40 ME/CFS patients (PIDs: 128, 184, 165, 163, 157, 173, 138, 177, 189, 186) and 4 of 16 healthy controls (CTLs: 35, 39, 51, 15) were examined for antibodies to the dUTPase protein of EBV, HHV-6, VZV and HERV-K by ELISA, as described in Methods. (A-C) Graphs of virus anti-dUTPase IgG Abs changes over time in individual ME/CFS patients (A-B) and healthy controls (C). Data represent the mean ± SEM of n ≥ 3 measurements/sample for each recombinant dUTPase protein at each time point. A positive reaction was defined as a serum sample that led to a signal three times over the background OD of the control serum (EBV = 0.392, green dotted line; HHV-6 = 0.327, blue dotted line; VZV = 0.290, red dotted line; and HERV-K = 0.292, pink dotted line; 95% Confidence interval).

Two-tailed Mann-Whitney U test analysis showed a significant increase in dUTPase IgG antibodies to the herpesviruses EBV, HHV-6 and VZV in ME/CFS compared to healthy-controls (p < 0.001), but no significant difference in HERV-K (p = 0.29). Further Chi square test analysis showed a significant difference in the distribution of herpesvirus dUTPase IgG antibody status (Negative, positive for Abs to single or multiple dUTPases) in ME/CFS patients compared to healthy controls, χ²(2) = 9.8, p = 0.0074.

3.3 |. Heightened Serum Levels of dUTPase IgG Correlate With Fatigue and Pain Severity in ME/CFS Patients

We next sought to determine if there was a relationship between serum levels of EBV, HHV-6, VZV dUTPase IgG Abs and participants’ self-reports of pain or fatigue. In the ME/CFS cohort, 25 patients reported severe fatigue and 13 moderate fatigue and there were no cases reported of mild or no fatigue (ME/CFS mean fatigue (range: 0–10) score = 7.053 ± 1.541). Most individuals in the control group reported no (n = 11, 69%) or mild (n = 4, 25%) fatigue with only 1-person self-reporting moderate fatigue and a mean fatigue score = 0.5 ± 1.033. Two ME/CFS patients, who had missing values for fatigue and pain, were excluded from the analyses, thus 38 ME/CFS patients and all 16 participants in the control group were included in pairwise analyses.

Group comparisons using non-parametric Mann-Whitney U tests revealed significant differences in fatigue severity between dUTPase IgG Ab-positive and Ab-negative individuals for EBV (p = 0.006), HHV-6 (p = 0.017), and VZV (p = 0.021) but not for HERV-K (p = 0.798). Pain also differed significantly in EBV (p = 0.011) and HHV-6 (p = 0.008) dUTPase IgG Ab positive individuals. There was not a statistically significant difference in pain between dUTPase IgG Ab-positive and Ab-negative individuals for VZV (p = 0.088) or HERV-K (p = 0.602). Linear regression correlation analyses indicated that higher serum levels of EBV (Figure 5A) and HHV-6 (Figure 5B) dUTPase IgG Abs were significantly associated with both fatigue (Spearman’s ρ = 0.38, p = 0.006; Spearman ρ = 0.37, p = 0.007, respectively) and pain severity (Figure 6A-B) (Spearman’s ρ = 0.35, p = 0.011; Spearman’s ρ = 0.34, p = 0.015, respectively). Whereas serum levels of VZV or HERV-K dUTPase IgG Abs did not show significant associations with either fatigue (Figure 5C-D) or pain (Figure 6C-D). Further group comparisons using Kruskal- Wallis tests to examine the distribution of virus dUTPase IgG Ab levels across fatigue (Figure 5A-D right graphs), or pain (Figure 6A-D right graphs) severity categories (No fatigue/pain, mild, moderate and severe) in ME/CFS and healthy control cohorts showed significant differences for EBV and HHV-6 dUTPase IgG serum levels across fatigue (p = 0.002 for both) and pain (p = 0.024 and p = 0.023, respectively) severity groups, but not for VZV or HERV-K (Figures 5C-D and 6C-D right graphs). Median-split analyses comparing fatigue and pain severity in individuals with low versus high virus dUTPase IgG Abs further confirmed higher fatigue in individuals with heightened EBV and HHV-6 dUTPase IgG Ab levels. We also observed that fatigue severity positively correlated with greater pain in ME/CFS patients (Spearman’s ρ = 0.63, p < 0.001) (Suppl. Figure 1). Altogether, these data indicate that elevated IgG Ab levels against EBV and HHV-6 dUTPases correlate with fatigue and pain symptom severity in this ME/CFS cohort.

FIGURE 5 |.

Correlation analysis of fatigue and virus dUTPase antibody levels in ME/CFS and healthy control cohorts. (A-D left graphs) Scatter plots depicting the relationship between fatigue severity scores and dUTPase IgG serum levels for (A) EBV, (B) HHV-6, (C) VZV and (D) HERV-K across all participants. The green linear regression line on each plot indicates the association trend. The Spearman’s rank correlation coefficient (ρ) and its corresponding p-value are displayed in the plot: (A) EBV: ρ = 0.38, p = 0.006; (B) HHV-6: ρ = 0.37, p = 0.007; (C) VZV: ρ = 0.18, p = 0.195; (D) HERV-K: ρ = 0.01, p = 0.920. (A-D right graphs) Scatter plots showing individual virus dUTPase Ab (OD) values across Fatigue Severity categories (No fatigue, Mild, Moderate, Severe), for EBV (A, left graph), HHV-6 (B, left graph), VZV (C, left graph) and HERV-K (D, left graph) differentiated by “Case” (ME/CFS, blue) and “Control” (orange) cohorts. Each point represents an individual. Black error bars indicate the mean ± SD for each severity-cohort subgroup. Kruskal-Wallis test comparing virus dUTPase Ab (OD) across Fatigue Severity groups yielded H = 15.197, p = 0.002 for EBV and HHV-6 H = 5.157, p = 0.161 for VZV and H = 3.961, p = 0.266 for HERV-K.

FIGURE 6 |.

Correlation analysis of pain and virus dUTPase antibody levels in ME/CFS and healthy control cohorts. Scatter plots depicting the relationship between pain severity scores and dUTPase IgG serum levels for (A) EBV, (B) HHV-6, (C) VZV and (D) HERV-K across all participants. The brown linear regression line on each plot shows the trend. The Spearman’s rank correlation coefficient (ρ) and its corresponding p-value are displayed in the plot: (A) EBV: ρ = 0.35, p = 0.011; (B) HHV-6: ρ = 0.34, p = 0.015; (C) VZV: ρ = 0.09, p = 0.54; (D) HERV-K: ρ = 0.06, p = 0.689. (A-D right graphs) Scatter plots showing individual virus dUTPase Ab (OD) values across Pain Severity categories (No pain, Mild, Moderate, Severe), for EBV (A, left graph), HHV-6 (B, left graph), VZV (C, left graph) and HERV-K (D, left graph) differentiated by “Case” (ME/CFS, blue) and “Control” (orange) cohorts. Each point represents an individual. Black error bars indicate the mean ± SD for each severity-cohort subgroup. Kruskal-Wallis test comparing virus dUTPase Ab (OD) across Pain Severity groups yielded H = 15.197, p = 0.002 for EBV and H = 9.495, p = 0.023 HHV-6, H = 6.874, p = 0.076 for VZV and H = 0.602, p = 0.896 for HERV-K.

3.4 |. Evaluation of Cytokines/Chemokines in Longitudinal Serum Samples of ME/CFS Patients

Previous studies by Cox et al [28] showed a significant elevation in serum concentrations of activin A and IL-21 but abnormally low CXCL13 in a cohort of ME/CFS patients. To determine whether these findings were a common phenomenon to ME/CFS, we examined longitudinal samples for fluctuations in the concentrations of activin A, IL-21 and CXCL13 over the 3-month study period. As shown in Table 2, all 873 serum samples examined from ME/CFS cases (n = 40) had normal levels of activin A ranging from 3.67 to 558.88 pg/mL [12, 17]. Similarly to the expression pattern of activin A, no significant changes were observed in the concentrations of CXCL13 or IL-21 over time. All ME/CFS cases expressed CXCL13 and IL-21 concentrations within normal levels (0–115 pg/mL), which ranged from 11.25 to 91.61 pg/ml and 0.009–4.18 pg/mL, respectively (Table 2).

TABLE 2 |.

Evaluation of cytokines levels in longitudinal serum of ME/CFS patients.^a

PID	Days	Activin A	IL-21	CXCL13
111	84	50.73 ± 60.35	ND	NDb
114	84	102.22 ± 60.36	ND	ND
115	80	348.32 ± 68.42	0.17 ± 0.41	29.42 ± 4.69
116	85	257.21 ± 51.42	ND	ND
118	83	281.26 ± 54.75	4.11 ± 7.53	40.85 ± 9.50
119	84	499.77 ± 98.65	ND	ND
120	84	144.86 ± 21.28	0.48 ± 1.10	80.49 ± 12.59
122	85	448.44 ± 93.17	0.14 ± 0.37	29.62 ± 6.40
123	82	262.32 ± 37.59	ND	ND
124	84	197.29 ± 51.51	ND	ND
125	84	230.84 ± 45.87	ND	ND
127	84	390.82 ± 115.47	ND	ND
128	81	220.25 ± 63.28	ND	ND
132	74	145.34 ± 54.02	1.76 ± 2.63	41.83 ± 14.70
134	84	147.93 ± 96.79	0.16 ± 0.35	19.54 ± 5.38
138	46	181.64 ± 63.32	ND	ND
141	83	235.60 ± 61.77	0.009 ± 0.002	70.94 ± 56.23
142	81	363.64 ± 76.89	0.50 ± 1.99	45.23 ± 5.30
143	83	305.40 ± 98.61	1.39 ± 2.54	34.04 ± 16.87
144	84	158.97 ± 40.04	ND	ND
151	84	126.81 ± 84.18	0.15 ± 0.26	11.25 ± 3.40
152	84	373.28 ± 140.30	ND	ND
157	84	228.15 ± 33.26	ND	ND
163	84	154.25 ± 84.52	ND	ND
165	85	444.11 ± 102.89	ND	ND
166	84	279.08 ± 60.60	1.54 ± 2.51	57.05 ± 33.72
169	84	289.56 ± 59.74	ND	ND
171	84	56.55 ± 58.78	0.35 ± 0.54	23.56 ± 8.83
172	84	3.67 ± 0.0001	0.11 ± 0.17	28.94 ± 15.15
173	81	425.50 ± 59.24	4.18 ± 2.72	61.50 ± 15.65
174	84	133.59 ± 242.63	0.06 ± 0.11	36.15 ± 6.29
176	84	318.09 ± 74.53	0.009 ± 0.001	28.94 ± 6.58
177	60	402.13 ± 54.59	ND	ND
178	84	174.37 ± 31.47	ND	ND
179	84	558.88 ± 75.88	0.84 ± 1.86	70.61 ± 12.93
181	84	257.72 ± 47.17	ND	ND
184	84	299.06 ± 80.21	ND	ND
186	46	76.52 ± 50.29	0.19 ± 0.41	89.38 ± 98.62
187	84	505.76 ± 87.25	ND	ND
189	49	136.43 ± 51.99	0.11 ± 0.26	91.61 ± 81.84
RANGE	46-85	3.67-558.88	0.009-4.18	11.25-91.61

^aLongitudinal serum samples from ME/CFS (11–25 serum samples/patient; n = 20–40 cases) were examined for cytokine/chemokine levels by ELISA as described in Methods. Data represent the mean ± SD of an n = 4 measurements/analyte. Cytokine/chemokine concentrations are expressed as pg/mL.

^bNot determined.

Longitudinal sera of ME/CFS patients were also tested for IFN-γ levels. Similarly to a recent study in military veterans diagnosed with chronic multi-symptom illness (CMI) [45], the ME/CFS cohort showed no significant changes in IFN-γ concentration over the study period compared to the healthy control group (p = 0.90) (Supplementary Figures 2–5). Of the 40 ME/CFS patients examined only 3 (PID: 119, PID: 179 Supplementary Figurs 3A–B and PID: 120, Supplementary Figure 4B) exhibited elevated serum levels of IFN-γ, which ranged from 27.91 to 91.35 pg/mL. Despite evidence suggesting reactivation of multiple herpesviruses, as determined by heightened levels of Abs to the lytic protein dUT-Pase. Interestingly, in the healthy control cohort three of the subjects who were negative for herpesvirus reactivation exhibited elevated IFN-γ levels indicating an ongoing infection at the time of blood draw (Supplementary Table 1 and suppl. Figure 2C, CTL: 57 and 3C, CTL: 68 and CTL: 76).

4 |. Discussion

There have been several case definitions used for diagnosis of patients with ME/CFS [1–4, 46], all of which focus on similar sets of symptoms, but differ markedly in the number of symptoms required and how those symptoms are defined. In 2015 the IOM proposed that the Canadian Consensus criteria be employed for research purposes [1]. Jason et al [47] indicated that subjects meeting the research criteria were significantly more impaired on a wide variety of symptoms and functional areas when compared to those meeting the clinical criteria. Thus, it is important to consider what criteria was used when extracting data findings from different ME/CFS studies.

This study aimed to elucidate previously unknown relationships between serum markers (dUTPase IgG Abs) of herpesvirus (EBV, HHV-6, VZV) exposure/reactivation, or HERV-K, cytokine/chemokines (IFN-γ, activin A, IL-21, CXCL13), and self-reported pain and fatigue severity in longitudinal sera from ME/CFS and healthy control women collected over a 1–3-month period. We confirmed the high prevalence of dUTPase IgG Abs to multiple herpesviruses in this cohort of ME/CFS compared to the control group (75% vs 31%). More striking was the observation that 12 ME/CFS patients (30%) were positive for Abs to all EBV, HHV-6, VZV and HERV-K dUTPases compared to only 1 person in healthy controls (6%).

Group comparisons revealed significant differences in fatigue severity between antibody-positive and antibody-negative individuals for EBV, HHV-6 and VZV but not HERV-K. Pain scores also differed significantly for EBV and HHV-6 positive individuals, whereas no differences were observed in VZV or HERV-K dUTPase Ab-positive individuals. Further statistical analyses using linear regression models indicated significant positive associations between EBV and HHV-6 dUTPase IgG Abs with both fatigue and pain severity scores. To our knowledge this is the first report identifying significant relationships between symptoms in ME/CFS and serum levels of EBV and HHV-6 dUTPase IgG Abs.

Activin A, IL-21 and CXCL13 serum levels did not change over the study period in ME/CFS patients and stayed within the normal range for these cytokines. This is in stark contrast to our previous “Good Day Bad Day ME/CFS” study in which patients exhibited increased activin A and IL-21 and abnormally low CXCL13 which was associated with aberrant follicular helper T cell differentiation [28]. The serum samples employed in our studies were obtained from ME/CFS patients diagnosed using two different case-definition criteria: the Fukuda plus Reeves modification in the previous study [21, 28] and Fukuda in the present study. While this could explain the differences in activin A and IL-21 concentrations between the two studies, it is also possible it may reflect regional differences in unknown environmental factors. However, the one consistent feature common to both studies was the significant increase in dUTPase Abs to multiple herpesviruses, which suggests reactivation of these viruses is occurring in a significant number of individuals diagnosed with ME/CFS by either case definition. Surprisingly, IFN-γ levels in ME/CFS patients remained unchanged and within the normal range throughout the study period. This is consistent with a recent study by Cox et al., conducted in military veterans diagnosed with CMI, which suggested a deficient anti-viral response to EBV, HHV-6 and VZV reactivation [45].

There has been increasing interest in determining the potential role of human endogenous retroviruses (HERV) in the pathophysiology of ME/CFS. However, conflicting data has been reported [36–39], in part due to the heterogeneity of the populations as well as differences in the methodology used ranging from transcriptomics analysis to serological approaches. A recent study by Gimenez-Orenga K et al [40] found that ME/CFS and fibromyalgia cases could be distinguished based upon HERV transcriptomic profiles. In the present study, 52% of the ME/CFS patients and 50% of the controls examined exhibited anti-HERV-K-dUTPase antibodies. Notably, there was an increase in the prevalence of ME/CFS patients simultaneously co-expressing Abs to HERV-K and multiple herpesviruses’ dUTPases compared to the healthy controls (72.5% and 31%, respectively). While interesting, care must be taken in attributing a significance or lack thereof to this finding since anti-HERV-K dUTPase antibodies could be directed to the gag and/or gag-pro dUTPase precursor polyproteins in addition to the dUTPase proteins itself.

A previous study by our group conducted in single sera of a ME/CFS cohort (n = 55) reported that 55% of the cases were positive for multiple herpesvirus dUTPase antibodies and 53% were positive for HHV-6 dUTPase antibodies while in the control cohort (n = 151) 17% were positive for multiple herpesvirus dUTPase antibodies and 30% were positive for HHV-6 [21]. In the same study using longitudinal sera samples (n = 4) collected over an 18-month period from a ME/CFS cohort (n = 74), 49% exhibited antibodies against EBV and HHV-6 dUTPases and 54% exhibited antibodies against the HHV-6 dUTPase. Expanding on these observations, the current study demonstrates that 72.5% of the ME/CFS cohort (n = 40) exhibited antibodies to multiple herpesvirus dUTPases and 67.5% for the HHV-6 dUTPase, while in the control cohort (n = 16) only 31% and 6.25% were positive for multiple herpesvirus dUTPases and HHV-6 dUTPase, respectively. These results demonstrate that the lytic protein dUTPase was produced in sufficient quantities in these ME/CFS patients to induce a humoral response and suggest multiple herpesviruses may have reactivated. Sequential reactivation of several herpesviruses has also been implicated in the pathogenesis of Drug-induced Hypersensitivity Syndrome/Drug Reaction with eosinophilia and Systemic Symptoms (DiHS/DReSS) [48]. While additional studies are needed, herpesvirus reactivation in ME/CFS is likely caused by a defective adaptive immune response of virus-specific memory CD8⁺ T cells. This is supported by recent studies showing an impaired EBV-specific T cell response and diminished IFN-γ production [49, 50] as well as CD8⁺ T cell dysfunction in ME/CFS patients [51] which may result in T cell exhaustion [52]. Notably, we have previously reported that EBV dUTPase also alters T cell function in vitro and in vitro [28, 53]. In light of these findings, we propose that the abortive lytic replication of a herpesvirus, probably HHV-6, results in the chronic production of the dUTPase as well as other virus proteins that induce T cell exhaustion due to the expression of cross-reactive epitopes on the viral proteins [54–56] and the sequential reactivation of other herpesviruses contribute to the pathophysiological symptoms of fatigue, PEM and cognitive impairment characteristic of ME/CFS. Furthermore, disease severity may be related not only to the number of herpesviruses reactivated but also which ones are reactivated during the illness. However, further studies are needed to define this later premise.

In summary, this study demonstrates heightened levels of dUT-Pase Abs to multiple herpesviruses and HERV-K, suggests a faulty IFN-γ anti-viral response and identifies previously unknown associations between serum levels of EBV and HHV-6 dUTPase Abs and increased pain and increased fatigue symptoms in ME/CFS patients. Potential limitations of the study include the absence of male participants in the study, in part justified by the higher prevalence of this illness in women, as well as the shorter time-frame for serum collection in the control cohort.

Supplementary Material

Supple Fig 5

Supporting Figure 5. IFN-γ fluctuations over time in individual ME/CFS patients and healthy controls.

Supple Fig 3

Supporting Figure 3. IFN-γ fluctuations over time in individual ME/CFS patients and healthy controls. IFN-γ fluctuations over time in individual ME/CFS patients and healthy controls.

Supple Fig 4

Supporting Figure 4. IFN-γ fluctuations over time in individual ME/CFS patients and healthy controls.

Supple Fig 1

Supporting Figure 1. Correlation analysis of pain versus fatigue severity in ME/CFS and healthy control cohorts.

Supple Fig 2

Supporting Figure 2. IFN-γ fluctuations over time in individual ME/CFS patients and healthy controls.

Additional supporting information can be found online in the Supporting Information section.

FIGURE 2 |.

Herpesvirus dUTPase IgG antibody expression patterns in ME/CFS patients. Longitudinal sera (11–25 samples/participant) collected over a 3-month period from 10 of 40 ME/CFS patients (PIDs: 178, 171, 119, 122, 134, 151, 111, 123, 179, 144) and 4 of 16 healthy controls (CTLs: 68, 64, 76, 55) were examined for antibodies to the dUTPase protein of EBV, HHV-6, VZV and HERV-K by ELISA, as described in Methods. (A-C) Graphs of virus anti-dUTPase IgG Abs changes over time in individual ME/CFS patients (A-B) and healthy controls (C). Data represent the mean ± SEM of n ≥ 3 measurements/sample for each recombinant dUTPase protein at each time point. A positive reaction was defined as a serum sample that led to a signal three times over the background OD of the control serum (EBV = 0.392, green dotted line; HHV-6 = 0.327, blue dotted line; VZV = 0.290, red dotted line; and HERV-K = 0.292, pink dotted line; 95% Confidence interval).

FIGURE 3 |.

Herpesvirus dUTPase IgG antibody expression patterns in ME/CFS patients. Longitudinal sera (11–25 samples/participant) collected over a 3-month period from 10 of 40 ME/CFS patients (PIDs: 169, 166, 132, 152, 143, 174, 120, 124, 141, 127) and 4 of 16 healthy controls (CTLs: 72, 60, 56, 54) were examined for antibodies to the dUTPase protein of EBV, HHV-6, VZV and HERV-K by ELISA, as described in Methods. (A-C) Graphs of virus anti-dUTPase IgG Abs changes over time in individual ME/CFS patients (A-B) and healthy controls (C). Data represent the mean ± SEM of – ≥ 3 measurements/sample for each recombinant dUTPase protein at each time point. A positive reaction was defined as a serum sample that led to a signal three times over the background OD of the control serum (EBV = 0.392, green dotted line; HHV-6 = 0.327, blue dotted line; VZV = 0.290, red dotted line; and HERV-K = 0.292, pink dotted line; 95% Confidence interval).

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.