HERV activation segregates ME/CFS from fibromyalgia while defining a novel nosologic entity

Abstract

Research of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) and fibromyalgia (FM), two acquired chronic illnesses affecting mainly females, has failed to ascertain their frequent co-appearance and etiology. Despite prior detection of human endogenous retrovirus (HERV) activation in these diseases, the potential biomarker value of HERV expression profiles for their diagnosis, and the relationship of HERV expression profiles with patient immune systems and symptoms had remained unexplored. By using HERV-V3 high-density microarrays (including over 350k HERV elements and more than 1500 immune-related genes) to interrogate the transcriptomes of peripheral blood mononuclear cells from female patients diagnosed with ME/CFS, FM, or both, and matched healthy controls ( n = 43), this study fills this gap of knowledge. Hierarchical clustering of HERV expression profiles strikingly allowed perfect participant assignment into four distinct groups: ME/CFS, FM, co-diagnosed, or healthy, pointing at a potent biomarker value of HERV expression profiles to differentiate between these hard-to-diagnose chronic syndromes. Differentially expressed HERV–immune–gene modules revealed unique profiles for each of the four study groups and highlighting decreased γδ T cells, and increased plasma and resting CD4 memory T cells, correlating with patient symptom severity in ME/CFS. Moreover, activation of HERV sequences coincided with enrichment of binding sequences targeted by transcription factors which recruit SETDB1 and TRIM28, two known epigenetic silencers of HERV, in ME/CFS, offering a mechanistic explanation for the findings. Unexpectedly, HERV expression profiles appeared minimally affected in co-diagnosed patients denoting a new nosological entity with low epigenetic impact, a seemingly relevant aspect for the diagnosis and treatment of this prevalent group of patients.

Full text

Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 1 of 24
HERV activation segregates ME/CFS
from fibromyalgia while defining a novel
nosologicentity
Karen Giménez- Orenga1, Eva Martín- Martínez2, Lubov Nathanson3, Elisa Oltra4*
1Escuela de Doctorado, Catholic University of Valencia, Valencia, Spain; 2National
Health Service, Manises Hospital, Valencia, Spain; 3Institute for Neuro- Immune
Medicine, Dr. Kiran C. Patel College of Osteopathic Medicine, Nova Southeastern
University, Fort Lauderdale, United States; 4Department of Pathology, School of
Health Sciences, Catholic University of Valencia, Valencia, Spain
eLife Assessment
This important study substantially expands observations of HERV expression in the clinical settings.
The evidence provided by the authors that HERV activity is an underlying etiological factor in ME/
CFS and fibromyalgia is compelling and suggests further investigation into mechanisms. This work
will be of broad interest to clinicians and researchers alike.
Abstract Research of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) and fibro-
myalgia (FM), two acquired chronic illnesses affecting mainly females, has failed to ascertain their
frequent co- appearance and etiology. Despite prior detection of human endogenous retrovirus
(HERV) activation in these diseases, the potential biomarker value of HERV expression profiles for
their diagnosis, and the relationship of HERV expression profiles with patient immune systems and
symptoms had remained unexplored. By using HERV- V3 high- density microarrays (including over
350k HERV elements and more than 1500 immune- related genes) to interrogate the transcriptomes
of peripheral blood mononuclear cells from female patients diagnosed with ME/CFS, FM, or both,
and matched healthy controls (n = 43), this study fills this gap of knowledge. Hierarchical clustering
of HERV expression profiles strikingly allowed perfect participant assignment into four distinct
groups: ME/CFS, FM, co- diagnosed, or healthy, pointing at a potent biomarker value of HERV
expression profiles to differentiate between these hard- to- diagnose chronic syndromes. Differentially
expressed HERV–immune–gene modules revealed unique profiles for each of the four study groups
and highlighting decreased γδ T cells, and increased plasma and resting CD4 memory T cells,
correlating with patient symptom severity in ME/CFS. Moreover, activation of HERV sequences coin-
cided with enrichment of binding sequences targeted by transcription factors which recruit SETDB1
and TRIM28, two known epigenetic silencers of HERV, in ME/CFS, offering a mechanistic explanation
for the findings. Unexpectedly, HERV expression profiles appeared minimally affected in co- diag-
nosed patients denoting a new nosological entity with low epigenetic impact, a seemingly relevant
aspect for the diagnosis and treatment of this prevalent group of patients.
Introduction
Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), classified by the WHO with the ICD- 11
8E49 code as a postviral fatigue syndrome, and fibromyalgia (FM) (ICD- 11 MG30.0 for chronic primary
pain) (Harrison et al., 2021), are chronic, disabling, acquired diseases, characterized by complex
RESEARCH ARTICLE
*For correspondence:
elisa.oltra@ucv.es
Competing interest: The authors
declare that no competing
interests exist.
Funding: See page 19
Preprint posted
09 December 2024
Sent for Review
10 December 2024
Reviewed preprint posted
05 February 2025
Reviewed preprint revised
15 April 2025
Version of Record published
08 May 2025
Reviewing Editor: Nicholas
Dopkins, Northwell Health,
United States
Copyright Giménez- Orenga
etal. This article is distributed
under the terms of the Creative
Commons Attribution License,
which permits unrestricted use
and redistribution provided that
the original author and source
are credited.

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 2 of 24
symptomatology that affects multiple organs (Bateman etal., 2021; Wolfe etal., 2010). Diagnosis
of ME/CFS and FM continues to be based on the clinical assessment of unspecific symptoms, such as
debilitating fatigue, generalized pain, cognitive impairment or intestinal, sleep, and immune distur-
bances (Carruthers etal., 2011; Carruthers etal., 2003; Wolfe etal., 2016; Wolfe etal., 2010;
Wolfe et al., 1990). Their frequent co- diagnosis drove the hypothesis of a single syndrome and
promoted the search for common or differentiating factors (Abbi and Natelson, 2013; Natelson,
2019; Wessely etal., 1999). However, despite molecular support for ME/CFS and FM constituting
different entities (Groven etal., 2021; Light etal., 2012; Nepotchatykh etal., 2023), their frequent
joint appearance (over 50% in females) (Castro- Marrero etal., 2017) remains enigmatic.
Extensive research into ME/CFS and FM has not yet been able to ascertain their origin and patho-
physiology. However, several environmental factors have been suggested as triggering agents (Chu
etal., 2019; Furness et al., 2018; Tschopp etal., 2023). Viral infections are, particularly, gaining
momentum after the emergence of a type of persistent post- viral syndrome with symptoms that
closely resemble those in ME/CFS (Komaroff and Lipkin, 2023) and FM (Clauw and Calabrese,
2024), which has been defined as post- COVID- 19 condition (ICD- 11 RA02) (Soriano etal., 2022). The
prevalence of post- COVID- 19 cases meeting ME/CFS criteria has been estimated at 58% (Jason and
Dorri, 2022) raising concerns of a health, social, and economic burden of unprecedented dimensions.
Viral infections trigger dramatic changes in host cells gene expression, affecting their metabo-
lism and epigenetic landscape (Liu etal., 2020). Although functionally not well understood, these
epigenetic changes involve de- repression of human endogenous retroviruses (HERVs) (Macchietto
et al., 2020) sequences that were incorporated from exogenous viral infections during evolution
and that currently represent about 8% of our genome (Giménez- Orenga and Oltra, 2021). In fact,
SARS- CoV- 2 infection alters peripheral blood mononuclear cell (PBMC) ERV expression in human,
monkey, and mice expressing human ACE2 receptors, associating with immune- response activation
and histone modification genes in severe COVID- 19 cases (Guo etal., 2024), and the presence of
HERV proteins in postmortem tissues of lungs, heart, gastrointestinal tract, brain olfactory bulb, and
nasal mucosa from COVID- 19 patients (Charvet etal., 2023).
Although still poorly understood, advances in sequencing technology have aided in the discovery
of the wide range of physiological processes in which HERV participate, including coordination of the
immune response (Kassiotis, 2023), interaction with microbiota (Dopkins etal., 2022; Lima- Junior
etal., 2021), or shaping neurological functions (Ferrari etal., 2021), among other. Derangement
of HERV expression associates with disease (e.g., multiple sclerosis; Gruchot etal., 2023), systemic
lupus erythematosus (Khadjinova etal., 2022), or post- COVID- 19 condition (Giménez- Orenga etal.,
2022). Particularly, the expression of HERV- encoded proteins has been shown to stimulate an immune
response present in autoimmunity (Gruchot etal., 2023; Khadjinova etal., 2022). In addition, HERV
influence on pathophysiology extends to their long terminal repeats (LTRs) regulatory regions, shaping
host gene expression (Ito etal., 2017).
Previous studies detected overexpression of some HERV families in immune cells of ME/CFS
(Rodrigues et al., 2019) and FM (Ovejero etal., 2020) at the transcript level. However, the use
of directional RT- qPCR (reverse transcription followed by quantitative polymerase chain reaction)-
based approaches using degenerated primers limited the detection of affected genomic loci. In this
study, we elevated these analyses to obtain genome- wide HERV profiles (HERV expression profiles) of
patients having received ME/CFS, FM diagnosis, or both, hypothesizing that HERV expression profiles
may reveal distinct underlying pathomechanisms that justify their classification as separate diseases,
and/or common fingerprints explaining their joint high prevalence. To test our hypothesis, we scru-
tinized HERV and gene transcriptomes by using high- density microarray technology in a selected
cohort of female patients carefully phenotyped by a single expert clinician into three defined groups
of patients: patients fulfilling Canadian and International ME/CFS diagnosis criteria (Carruthers etal.,
2011; Carruthers etal., 2003), patients fulfilling FM diagnosis ACR (American College of Rheuma-
tology) criteria (Wolfe etal., 2016; Wolfe etal., 2010; Wolfe etal., 1990), or patients complying
with both, ME/CFS and FM diagnosis criteria, being called co- diagnosed from now on. The results of
comparing HERV profiles within PBMCs across these three patient groups as well as to HERV profiles
of matched healthy subjects allowed, not only for the separation of patients from healthy individuals
but also for perfect discrimination of patients into three distinct disease groups, suggesting distinct
subjacent pathomechanisms. Unexpectedly, HERV expression profiles appeared minimally affected

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 3 of 24
in co- diagnosed patients denoting a new nosological entity with low epigenetic impact. In addition,
HERV profiling also exposed some commonalities between the FM and ME/CFS subgroups and
marked quantitative differences within the ME/CFS group that correlated with immune disturbances
and patient symptomatology, aspects that may well set new criteria for the differential diagnosis of FM
and ME/CFS as well as for patient subtyping with expected impact in precision medicine programs.
Results
Demographics and clinical characteristics of participants
The study compared a total of 43 female subjects: 8 ME/CFS cases, 10 FM, 16 co- diagnosed cases,
and 9 matched healthy controls (Figure1). The average age for participants was 54 ± 3years (range
50–58) for ME/CFS subjects, 50 ± 5 years (range 42–58) for FM, 47 ± 15years (range 22–70) for
co- diagnosed individuals, and 51 ± 6 (range 43–61) for controls (Figure1—figure supplement 1A,
Supplementary file 1A). Questionnaires used for patient symptom phenotyping included Fibromy-
algia Impact Questionnaire (FIQ) (Burckhardt etal., 1991), Multi Fatigue Inventory (MFI) (Smets
etal., 1995) for general fatigue, and Short- Form- 36 Health Survey (SF- 36) (Mchorney etal., 1993) for
quality- of- life assessment (Figure 1—figure supplement 1A). Average total and subdomain scores
are shown in Table 1 for each patient group, while itemized questionnaire scores are provided in
Supplementary file 1A. According to total FIQ scores, after considering that a score<39 indicates
mild affection, ≥39 through≤59 is assigned to moderately affected patients, and>59to severe affec-
tion (Rivera and González, 2004) around 88% of cases in the studied cohort correspond with severe
cases while 12% are moderate. No statistically significant difference between patient groups was
observed for any questionnaire domain except for MFI Physical Fatigue (p = 0.011) and Reduced
Motivation (p = 0.038), and only between the ME/CFS and co- diagnosed groups (Supplementary
file 1B).
HERV signatures discriminate ME/CFS, FM, and co-diagnosed groups
from healthy controls
Genome- wide HERV expression profiles for each of the four study groups (three disease groups: ME/
CFS, FM, and co- diagnosed, plus one control group corresponding to healthy participants), using
custom high- density Affymetrix HERV- V3 microarrays (Becker et al., 2017), showed that a set of
489 HERV (502 probesets) is differentially expressed (DE) between at least two of the groups (FDR
<0.1and |log2FC| >1) (Figure1; Supplementary file 1C), confirming dysregulation of particular sets
of HERV elements in the immune systems of ME/CFS, FM, and co- diagnosed cases as compared to
healthy controls. Volcano plots show the number of HERV elements found significantly over- or under-
expressed for each of the comparisons, evidencing an enhanced dysregulation of HERV in ME/CFS
as compared to FM, co- diagnosed, or healthy control groups (Figure1A). Consistently, hierarchical
clustering of samples based on DE HERV fingerprints clearly differentiate these pathologies from
healthy subjects, strikingly segregating ME/CFS from all other compared groups (Figure1B); while
hierarchical clustering of DE HERV sequences revealed four main clusters consisting of loci specifically
overexpressed (Cluster 1, 251 HERV, 293 probesets) or underexpressed (Cluster 2, 119 HERV, 163
probesets) in ME/CFS, loci specifically overexpressed in FM (Cluster 3, 68 HERV, 74 probesets), and
loci underexpressed in all three pathologies (Cluster 4, 54 HERV, 64 probesets). DE of the HERV loci
for each study group was corroborated by analysis of HERV expression levels encompassed in each
cluster, as illustrated with violin plots (Figure1B), with some of the changes being validated by the
alternative RT- qPCR approach, as shown in Figure1—figure supplement 1B.
Closer examination of Clusters 1 and 2 indicated that the ME/CFS group could be subdivided into
two subgroups showing differential levels of HERV expression profiles (Figure1C). While both ME/
CFS subgroups (subgroups 1 and 2) show clear upregulation (Cluster 1, p < 0.0001) or downregu-
lation (Cluster 2, p < 0.0001) of several HERV loci as compared to all the other study groups, they
show quantitative differences, with the ME/CFS subgroup 2 exhibiting a more pronounced dysregu-
lation than subgroup 1, particularly affecting the downregulated Cluster 2 (Figure1C). Interestingly,
subgroup 2 ME/CFS patients present more accused levels of fatigue (p < 0.05) and reduced motiva-
tion as compared to the co- diagnosed group and the ME/CFS subgroup 1 (Supplementary file 1D)

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 4 of 24
Figure 1. DE of HERV elements discriminates myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), bromyalgia (FM), and co- diagnosis. (A)
Volcano plots showing log2(fold changes) and the adjusted p- values for all HERV assessed with HERV- V3 microarray for each set of groups, as indicated.
Red dots indicate DE HERV (FDR <0.1and |log2FC| >1). Gray boxes show numbers of overexpressed or underexpressed HERV elements. (B) HERV
expression heatmap and cluster analysis of ME/CFS (n = 8, green), FM (n = 10, orange), co- diagnosed (n = 16, yellow), and healthy control (n = 9, blue)
samples. The heatmap includes all HERV probes displaying signicant DE between at least two of the compared groups (FDR <0.1and |log2FC| >1).
Clusters 1–4 correspond to groups of HERV probes displaying signicant over/under- expression in at least one group. Box and violin plots summarize
the distribution and expression levels of the DE probes in each cluster per study group. The scaled mean expression value (z- score) for each HERV
Figure 1 continued on next page

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 5 of 24
indicating that quantitative HERV fingerprint differences might help to assess the degree of ME/CFS
severity.
In line with our findings, unsupervised principal component analysis (PCA) of DE HERV loci
supports perfect discrimination of samples by study group and differentiates the two identified ME/
CFS subgroups (Figure1D). While principal component 1 (PC1) perfectly segregates the two ME/CFS
subgroups from FM, co- diagnosed, and control groups; principal component 2 (PC2) allows partition
of FM, co- diagnosed, and control samples, but not so much of ME/CFS from co- diagnosed cases. Thus,
PC1 explains better the variance observed across groups (29.1%) than PC2 (9.3%), suggesting a prom-
inent role of DE HERV loci in ME/CFS pathology. Top 10 contributing HERVs to principal components
probe is plotted. Box plots show the median z- score value and the rst and third quartiles. (C) Box and violin plots for Clusters 1 and 2 differentiating
ME/CFS subgroups 1 and 2. (D) Principal component analysis of DE HERV. (E) Bar plot of top 10 DE HERV with greater inuence/contribution to
principal component 1 (PC1) or principal component 2 (PC2). Statistical tests: unpaired two- sample Wilcoxon test with Benjamini–Hochberg p- value
correction (***p < 0.001, ****p < 0.0001).
The online version of this article includes the following gure supplement(s) for gure 1:
Figure supplement 1. Differences between patients and controls.
Figure 1 continued
Table 1. Patient health status assessment with FIQ, MFI, and SF- 36 (Burckhardt etal., 1991; Mchorney etal., 1993; Smets etal.,
1995) questionnaires.
Control (n = 9)
Mean ± SD [range]
ME/CFS (n = 8)
Mean ± SD [range]
FM (n = 10)
Mean ± SD [range]
Co- diagnosed (n = 16)
Mean ± SD [range]
Age 51 ± 2 [43–61] 54 ± 3 [50–58] 50 ± 5 [42–58] 47 ± 15.48 [22 – 58]
BMI 25.03 ± 2.07 [22.10–28.04] 26.29 ± 3.33 [21.34–31.22] 25.31 ± 1.73 [23.68–28.40] 24.80 ± 4.72 [18.71–30.43]
FIQ
Total FIQ 28.16 ± 13.08 [0–2.86] 74.4 ± 4.8 [56.3–85.6] 75.9 ± 3.4 [51.6–92.8] 74.6 ± 2.5 [47.8–96.3]
Function 2.38 ± 0.57 [1.98–3.63] 5.2 ± 2.2 [1.7–7.9] 6.8 ± 1.5 [4.3–8.9] 5.2 ± 1.8 [1.7–9.2]
Overall 10.01 ± 0 [10.01–10.01] 8.9 ± 3.0 [1.4–10.0] 8.2 ± 3.4 [0–10.0] 7.9 ± 2.7 [1.4–10]
Symptoms 0.32 ± 0.95 [0–2.86] 6.6 ± 4.8 [0–10.0] 6.0 ± 3.4 [0–10.0] 6.2 ± 2.5 [2.9–10]
MFI
General fatigue 11.56 ± 4.19 [5–17] 17.6 ± 3.1 [11–20] 16.3 ± 4.6 [10–20] 13.6 ± 3.2 [7–20]
Physical fatigue 10.11 ± 4.01 [6–17] 18.0 ± 1.9 [16–20] 17.0 ± 3.7 [12–20] 14.1 ± 2.9 [12–20]
Reduced activity 7.33 ± 2.74 [4–12] 15.9 ± 3.4 [12–20] 15.9 ± 4.4 [9–20] 12.9 ± 2.8 [11–20]
Reduced motivation 7.22 ± 3.31 [4–15] 15.4 ± 3.7 [9–19] 14.8 ± 4.1 [9–20] 11.6 ± 2.4 [8–16]
Mental fatigue 7.11 ± 2.93 [4–13] 15.5 ± 3.7 [11–20] 15.1 ± 3.8 13.3 ± 3.1 [9–19]
SF- 36
Physical functioning 86.67 ± 13.23 [65–100] 44.5 ± 14.5 [15–60] 31.5 ± 13.9 35.3 ± 14.8 [0–65]
Role physical 83.33 ± 17.95 [50–100] 3.2 ± 8.8 [0–25] 10.6 ± 17.7 [0–43.8] 14.5 ± 21.9 [0–75]
Bodily pain 58.61 ± 12.63 [45–80] 19.3 ± 11.7 [0–32.5] 19.3 ± 17.8 [0–45] 17.3 ± 14.2 [0–45]
General health 65.56 ± 15.09 [35–85] 18.9 ± 11.0 [3.8–35.0] 24.8 ± 16.1 [0–50] 23.4 ± 14.8 [0–45]
Vitality 56.94 ± 9.08 [43.75–75.00] 13.6 ± 10.1 [0–30] 11.1 ± 16.8 [0–43.8] 14.5 ± 11.6 [0–40]
Social functioning 84.72 ± 13.66 [62.50–100] 33.6 ± 21.9 [4.0–77.5] 25.0 ± 23.2 [0–62.1] 29.5 ± 25.6 [0–75]
Role emotional 79.63 ± 20.88 [50–100] 12.8 ± 35.3 [0–100] 43.3 ± 37.2 [0–100] 12.8 ± 38.9 [0–100]
Mental health 71.67 ± 18.54 [40–95] 34.9 ± 21.2 [3.4–72.0] 44.7 ± 19.9 [12–70] 45.2 ± 17.1 [15–80]
FIQ, Fibromyalgia Impact Questionnaire; MFI, Multi Fatigue Inventory; SF- 36, Short- Form 36 Health Survey; SD, standard deviation; SE, standard error.
Range refers to the possible values in the studied group.

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 6 of 24
PC1 and PC2 are shown (Figure 1E). In summary, these results show that ME/CFS, FM, and their
co- diagnosis present unique HERV expression profiles capable of discriminating patient subtypes
from healthy subjects, providing potential biomarkers for their differential diagnosis and stratification.
ME/CFS shows the greatest HERV dysregulation which, importantly, associates with patient symp-
tomatology, suggesting a prominent role of DE HERV in the pathomechanism of this disease, while
the co- diagnosed group presents the closest HERV expression profiles to those of healthy controls
indicating much less involvement of epigenetic changes in these patients, thus, defining a different
nosological entity from ME/CFS or FM.



















* ****
**** ****
*
**** ***
**
**
p=0.065
**



 








3
6
6
1
1
1
3
1
4
1
2
1
1
100
75
50
25
0
100
75
50
25
0
100
75
50
25
0
Figure 2. Diverse solitary long terminal repeat (LTR) families are deregulated in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS),
bromyalgia (FM), and co- diagnosed conditions. (A) Relative contribution of HERV families to each cluster, calculated as the proportion of HERV loci
assigned to each family relative to total HERV loci in each cluster. Families with a representation of at least 2.5% are shown. Main HERV families (B)
downregulated in ME/CFS, FM and co- diagnosed groups (Cluster 4), (C) upregulated in ME/CFS (Cluster 1), (D) downregulated in ME/CFS (Cluster 2),
and (E) upregulated in FM (Cluster 3). Box and violin plots summarize distributions and expression levels of the different HERV probes belonging to
the same family in the different study groups. The scaled mean expression value (z- score) for each HERV probe is plotted. Box plots show the median
z- score value and the rst and third quartiles. (F) Proportion of HERV subdomains expressed by cluster. (G) Genomic context of the DE HERV loci by
cluster. Statistical tests: Fisher’s exact test, t- test or Wilcoxon test with Benjamini–Hochberg p- value correction (*p < 0.05, **p < 0.01, ***p < 0.001, ****p
< 0.0001).
The online version of this article includes the following gure supplement(s) for gure 2:
Figure supplement 1. DE HERV families by study group.
Figure supplement 2. DE HERV structures by study group.
Figure supplement 3. DE HERV genomic location by study group.
Figure supplement 4. Intergenic solo long terminal repeat (LTR) DE across study groups.

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 7 of 24
Dysregulated HERV families are disease specific
Since preferential HERV- family derangements may indicate particularly altered functions in disease
(Ito et al., 2017), we next examined the family composition of the identified DE HERV clusters
(Figure2). Overall, ME/CFS (Clusters 1 and 2) outstands with the largest heterogeneity with up to 66
HERV dysregulated families, contrasting with only 22 families in FM (Cluster 3) and 16 in Cluster 4, the
latter representing DE HERV common to all patient groups (Supplementary file 1E, Figure2—figure
supplement 1).
Among DE HERV families 8 appear to be common across all clusters, and 6 of them represent the
most abundantly deregulated (>2.5%) (MLT1, MST, THE1, HERV16, ERVL, and HERVL33) (Figure2A,
Supplementary file 1E), suggesting an unspecific association with all disease groups. Within these
families, however, MLT1 appears as the only significantly downregulated in all patient groups (p <
0.01) (Figure2—figure supplement 1), accounting for more than a third of the DE HERV in each
cluster (Figure2A, Supplementary file 1E), a finding that may be, at least partially, due to the fact of
MLT1 being the most abundant HERV family in the human genome (Figure2A).
Interestingly, unique sets of DE HERV families were found for each disease (Supplementary file
1E). Up to 21 families, comprising HERV- H, HUERSP3, and RRHERVI among other, appeared specif-
ically upregulated in Cluster 1 (Figure 2B, Figure2—figure supplement 1, and Supplementary
file 1E), while 12 families, including HERV- P, HERV- Rb, and HERV- XA34, were specifically silenced
in Cluster 2, both strikingly DE in ME/CFS patients from subgroup 2 (Figure2C, Figure2—figure
supplement 1, Supplementary file 1E). Regarding FM, three families consisting of HERV- T, LTR84,
and PRIMA41, were found specifically upregulated (Cluster 3) (Figure2D, Figure2—figure supple-
ment 1, Supplementary file 1E). Five families were commonly downregulated across ME/CFS, FM,
and co- diagnosed groups, including HERV- W, HERVK14C, and MER67 (Figure2E, Figure2—figure
supplement 1, Supplementary file 1E). Contrary to the literature (Ovejero etal., 2020), reporting
upregulation of HERV- W family in FM, our results showed downregulation of HERV- W family in all
patient groups, more accentuated in FM and co- diagnosed groups than ME/CFS with respect to
healthy levels (Figure2E). In all, these results reveal family- specific derangement of HERV expression
in immune cells of FM and ME/CFS but not in co- diagnosed cases.
In light of the potential involvement of HERV- encoded proviral proteins in disease development
(Giménez- Orenga and Oltra, 2021) and to deepen the analysis of disease- specific HERV fingerprints,
we then determined the proportion of these HERV fingerprints corresponding to solitary and proviral
elements, as well as the different subdomains of the HERV structure according to probe detection in
the microarrays that is, LTR containing regulatory sequences versus gag, pol, and env protein- coding
genes. Although functional information of these sequences is quite limited, the nature of the detected
HERV motifs may inform of potential regulatory mechanisms leading to or perpetuating disease status
(e.g., toxic env proteins or non- coding regulatory sequences). In agreement with hg38 HERV genomic
data, it was found that the great majority of the dysregulated HERV loci corresponded to solitary
LTR elements, which are known to be the most abundant in our genome, with few exceptions corre-
sponding to proviral HERV (Figure2F, Figure2—figure supplement 2).
Lastly, given the potential of HERV LTRs to influence the expression of neighboring genes through
the regulatory elements they hold, we examined the genomic context of the DE HERV loci by cluster
(Figure 2G, Figure 2—figure supplement 3). Although not allocating significant association to
promoter regions, repressed loci in ME/CFS (Cluster 2) and overexpressed loci in FM (Cluster 3) were
significantly associated with gene 3′ endings, and silenced loci in the three pathologies (Cluster 4)
significantly associated with intronic regions. These results suggest that the molecular mechanisms
by which disease- specific deregulated HERV loci could be participating in the development or main-
tenance of disease likely involve changes in transcript turnover or availability of alternative splicing
events. In addition, DE analysis of intergenic solo LTRs show that they are underexpressed in ME/CFS
(Figure2—figure supplement 4), pointing at the participation of epigenetic regulatory mechanisms
in the disease.
HERV fingerprints associate with abnormal immune gene expression
Since disease- associated HERV derangement could indicate, even mediate, differences in global gene
expression, we proceeded to analyze potential associations between disease- associated HERV finger-
prints and a set of genes involved in eight pathways (immunity, inflammation, cancer, central nervous

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 8 of 24
system disorders, differentiation, telomere maintenance, chromatin structure, and gag- like genes)
detected by probes present in this same HERV- V3 high- density microarrays (Becker et al., 2017).
Differential transcript analysis revealed 368 genes (1037 probesets) DE between at least two study
groups (FDR <0.1and |log2FC| >1) (Supplementary file 1F). In line with DE HERV expression, ME/
CFS immune cells showed a more accused dysregulation in gene expression profiles compared to
FM’s and co- diagnosed groups (Figure3, Figure3—figure supplement 1A). Hierarchical clustering
of samples and PCA based on DE genes, however, did not cluster all samples by study group as HERV
signatures did, with the only exception of ME/CFS subgroup 2, which exhibited a markedly different
gene expression profile further supporting its distinction from the remaining study groups (Figure3—
figure supplement 1B).
To uncover potentially affected physiologic functions linked to DE HERV, we examined how DE
HERVs and DE genes with similar expression patterns grouped together in modules based on their
intrinsic relationships by their hierarchical co- clustering (Figure3). Then, the functional significance
of these modules was assessed by gene ontology (GO) analysis of the DE genes within each module.
The hierarchical clustering analysis resulted in the identification of eight distinct modules, each char-
acterized by unique combinations of DE HERV and DE gene patterns across all four study groups
(Figure3). In particular, 37 HERV (19 from Module 1 and 18 from Module 6) with decreased expression
levels across the three disease groups, correlated with several genes involved in the differentiation of
immune cells and regulation of cell–cell adhesion (Figure3A). Remarkably, ME/CFS exhibited upreg-
ulation of 258 HERV loci with expression levels that correlated with 34 genes involved in alpha- beta T
cell activation and T- helper 17cell commitment- related genes (Module 3). In addition, 78 HERV down-
regulated in ME/CFS correlated with pathogen detection- related genes (Figure 3B, GO:0016045,
GO:0016032, GO:0098543), suggesting a special involvement of the immune system in the disease.
Lastly, several HERV and genes exhibited upregulated expression in FM while appearing downreg-
ulated in ME/CFS. These FM- specific DE expressed genes are involved in telomere structure orga-
nization and maintenance (Modules 2 and 5), cellular response to glucocorticoids (Module 7), and
regulation of NF-κB signaling (Module 8) (Figure3C), indicating divergent immune states in ME/CFS
and FM with a differential impact in inflammatory processes.
Co-expression of DE HERV and immune-response genes predominantly
occur by mechanisms other than co-transcription
To further investigate the relationship between HERV dysregulation and immune abnormalities, we
mapped the genomic localization of HERV and genes within each module, looking for genomic over-
laps that could explain their correlation by co- expression. Surprisingly, the results showed no major
overlaps of DE HERV and their correlated DE genes (Figure4A), with distances that separate them at
least 100 kbp (Figure4B, Supplementary file 1G), thereby suggesting overall independent expres-
sion of DE HERV and DE genes. As an exception, it was noticed that some elements from the MLT1
family co- localized with DE genes within a 30 kbp window (Supplementary file 1G). For example,
the element MLT1_5q32, from Module 1, appears located 18,114bp upstream of the CD74 gene,
involved in MHC class II antigen processing, or the element MLT1_4q24, belonging to Module 6,
lies within NFκB1 first intron. Both HERV loci and their co- localized genes exhibited downregula-
tion in ME/CFS, FM, and co- diagnosed groups (p < 0.05). Likewise, an exception was noticed for
the MLT1_8q12.1 and MLT1_Xp22.3 HERV elements in Module 2, located within LYN ’s tenth intron
and TLR8- long isoform’s second exon, respectively. Their genomic co- localization may explain that
their specific lower expression levels in the ME/CFS subgroup 2 occur as result of co- transcription
(Figure4C).
Since common trans- regulatory mechanisms may explain correlation by co- expression in the
absence of co- transcription, we analyzed DE HERV in each module for potential enrichment in tran-
scription factor binding sites (TFBS) by comparing the annotated TFBS within dysregulated HERV
genomic regions against the full set of TFBS in the genome (based on publicly available human
ChIP- seq data provided by ReMap2022) (Figure4D). Then, we performed GO analysis for HERV- gene
sets enriched TFBS to infer their function (Figure4—figure supplement 1). Interestingly enough, the
results showed enrichment of binding sites for transcription factors involved in the immune response
in Modules 1 through 7, further supporting the role of immune disturbances in ME/CFS, FM, and
co- diagnosed cases, while no significant enrichment of TFBS was found in Module 8. In addition, HERV

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 9 of 24

Log2 (mRNA expression level)
Module 1
(19 HERV / 55 genes)
Module 6
(18 HERV / 167 genes)
HERVs Genes
ComorbidityFMControl ME subg.1 ME2 subg.2

Log2 (mRNA expression level)
Module 3
(258 HERV / 34 genes)
Module 4
(78 HERV / 144 genes)
Log2 (mRNA expression level)
ComorbidityFMControl ME subg.1 ME2 subg.2

Module 2
(35 HERV / 294 genes)
Module 5
(8 HERV / 214 genes)
HERVs Genes
(8
HERV
/
214
genes)
Log2 (mRNA expression level)
Module 7
(83 HERV / 36 genes)
Module 8
(3 HERV / 93 genes)
ComorbidityFMControl ME subg.1 ME2 subg.2
0 2 4 6 8 10
-Log10(p.value)
0 1 2 3 4 5 6
-Log10(p.value)
4 6 8 10
-Log10(p.value)
0 1 2 3 4 5 6 7
-Log10(p.value)
HERVsGenes
****
****
****
****
****
****
(258
HERV
/
34
genes)
HERVsGenes
ComorbidityFMControl ME subg.1 ME2 subg.2
(78
HERV
/
144
genes)
********
****
********
********
****
**** ********
****
****
****
********
****
****
********
****
****
****
Figure 3. HERV expression correlates with immune- response genes. Hierarchical clustering of DE HERV and genes (FDR <0.1and |log2FC| >1) providing
eight modules with highly correlated expression levels. Box and violin plots summarize the distribution and expression level of the different HERV (blue)
and gene (green) probes in Modules (A) 1 and 6, (B) 3 and 4, and (C) 2, 5, 7, and 8, per study group. The scaled mean expression value (z- score) for each
HERV and gene probe sets are plotted. Box plots show the median z- score values with rst and third quartiles. Statistical tests: Wilcoxon test (****p <
0.0001).
Figure 3 continued on next page

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 10 of 24
The online version of this article includes the following gure supplement(s) for gure 3:
Figure supplement 1. Differential gene expression for myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), bromyalgia (FM), and
comorbidity groups.
Figure 3 continued
Figure 4. DE HERV sequences and DE immune- response genes are majorly independently transcribed. (A) Genomic distribution of HERV, and genes
encompassed in Modules 1 through 8. (B) Genomic distance between DE HERV and their nearest DE genes per module, plotted as density curves
illustrating the distribution of distances from 0 to 100million base pair (Mbp). (C) Violin dot plots showing the distribution and expression levels of some
examples of co- localized HERV loci and genes in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) subgroup 1 (n = 3), ME/CFS subgroup
2 (n = 5), FM (n = 10), co- diagnosed (n = 16), and healthy control (n = 9) samples. Blue and orange highlighted headings indicates the pair HERV/gene
belonged to Modules 1 or 6, or Modules 2, 5, 7, or 8, respectively. (D) Heatmap of enriched transcription factor binding sites (TFBS) in HERV loci of
Modules 1 through 7. Module 8 did not show signicant enrichment of TFBS. Statistical tests: Wilcoxon test (*p < 0.05; **p < 0.01, ***p < 0.001, ****p <
0.0001).
The online version of this article includes the following gure supplement(s) for gure 4:
Figure supplement 1. Gene ontology analysis of transcription factors binding sites enriched in HERV loci of Modules 1–7.

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 11 of 24
downregulated across the three patient groups (Modules 1 and 6) were enriched in TFBS involved in
angiogenesis, while those upregulated in FM but downregulated in ME/CFS (Modules 2, 5, and 7),
were enriched in TFBS for miRNA transcription and glucocorticoid signaling. It is worth mentioning
that HERV upregulated in ME/CFS (Module 3) were markedly enriched with binding sites for chro-
matin remodeling factors, from which SETDB1 and TRIM28 stand out as key epigenetic repressors
of HERV expression and CTCF for its importance in the establishment of topologically associating
domains in chromosomes. These results therefore support the possibility that DE HERV expression
occurs independently, but in coordination with their correlated DE genes by yet to define mecha-
nisms, likely driven by and/or driving epigenetic changes in immune cells of ME/CFS patients.
Figure 5. HERV ngerprints correlate with immune cell proles. (A) Measurements of immune cell proportions
by CIBERSORTx per study group (n = 8 myalgic encephalomyelitis/chronic fatigue syndrome [ME/CFS], n =
10bromyalgia [FM], n = 16 co- diagnosed, n = 9 controls). (B) Association between modules and immune cell
proportions was evaluated by correlating eigengenes from each module with cell proportion values obtained
from CIBERSORTx analysis. Boxes show Pearson correlation values and associated p- values (*p < 0.05, **p <
0.01, ***p < 0.001, ****p<0.0001) between gene expression levels of each module and quantity of cells from each
specic type. A value of 1 (green) and –1 (blue) quantify strongest positive and negative correlations, respectively,
while 0 (white) shows no correlation. (C) Scatter plots between top HERV loci expression levels and immune cell
proportion as measured by CIBERSORTx. Pearson correlation coefcients and associated p- values are shown in
plots. Correlation 95% condence intervals are: 0.33–0.74 for LTR65_20p11.21 versus γδ T cells, –0.81 to –0.46 for
MER49_11q13.1 versus resting CD4 memory T cells, and –0.70 to –0.25 for MER72_4q34.3 versus plasma cells.

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 12 of 24
DE HERV strongly associates with plasma and T-cell levels
Given the detected relationships between HERV dysregulation and immunological disturbances, we
investigated differences in immune cell populations across the study groups by performing CIBER-
SORTx analysis on normalized microarray gene expression data. CIBERSORTx is an analytical tool
that provides an estimation of cell type abundance using gene expression data (Newman etal.,
2019). The results point at a decrease in γδ T cells as well as an increase in resting CD4 T memory
and plasma cells (p < 0.05) in ME/CFS patients from subgroup 2 (Figure5A). To further understand
the relationship between gene signatures and the potential differences in immune cell populations
with DE HERV, we analyzed correlations across HERV and genes included in Modules 1 through 8 and
immune cell proportions, finding that dysregulated HERV loci moderately to strongly correlate with
plasma cell and resting CD4 memory T cells increase, and γδ T cell population decrease (|R| > 0.4, p <
0.01) (Figure5B). Curiously, HERV and genes upregulated in ME/CFS subgroup 2 (Module 3), nega-
tively correlate with γδ T cell abundance (e.g., LTR65_20p11.21, R = 0.57, p = 6.3e−05) and positively
correlate with resting CD4 memory T cells, and plasma cell abundances (e.g., LTR65_20p11.21, R =
0.57, p = 6.3e−05; MER49_11q13.1, R = −0.67, p = 1e−06) (Figure5C), while the tendency of correla-
tion appears reversed for the other modules. These results suggest a potential relationship between
HERV expression and immune cell fractions in a subgroup of ME/CFS patients.
HERV and immune alterations associate with ME/CFS diagnosis
Lastly, in an effort to detect potential relationships between deregulated HERV and genes, and patient
diagnosis, we analyzed the association between DE of HERV and genes belonging to each module
Figure 6. HERV and gene deregulation associates with disease symptomatology. (A) Association between modules and disease determined by
correlations of eigengenes from each module with binarized disease traits. Boxes show Pearson correlation values and associated p- values (*p < 0.05,
**p < 0.01, ***p < 0.001) by module and each disease trait. A value of 1 (orange) and –1 (blue) quantify strongest positive and negative correlations,
respectively, while 0 (white) shows no correlation. (B) Barplots showing top 10 HERV loci (upper) and top 10 genes (lower) with highest correlations with
myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) disease. Bar height and color display the strength of each correlation, being high orange
bars more correlated than lower blue bars. (C) Association between top 10 HERV loci (left) and top 10 genes (right) with disease traits as determined by
their correlation with ME/CFS patient symptoms as assessed by questionnaire scores. Boxes show Pearson correlation values and associated p- values
(*p < 0.05, **p < 0.01) by DE HERV or DE gene as indicated for Fibromyalgia Impact Questionnaire (FIQ) subdomains (Function, Overall, and Symptoms)
or total FIQ; for Multi Fatigue Inventory (MFI) subdomains (GF: general fatigue, Pfa: physical fatigue, RA: reduced activity, RM: reduced motivation, and
MF: mental fatigue), and for Short- Form- 36 Health Survey (SF- 36) subdomains (PF: physical functioning, RP: role physical, BP: bodily pain, GH: general
health, VT: vitality, SF: social functioning, RE: role emotional, and MH: mental health).

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 13 of 24
and the diagnosis received. Interestingly, activation of HERV and genes involved in immune response
(Module 3) seem to strongly associate with diagnosis of ME/CFS (subgroup 2) (|R| > 0.89, p < 0.001)
(Figure6A). In fact, top correlated genes and HERV with ME/CFS diagnosis (Figure6B), showed the
strongest association with disease symptoms (Figure6C), particularly with some domains of the MFI
and SF- 36 questionnaires, including physical and mental fatigue (p < 0.05). These results suggest
the participation of these HERV and genes in ME/CFS disease while uncovering their potential as
biomarkers.
Discussion
Researchers have strived for decades to elucidate the etiology and pathophysiology of ME/CFS and
FM, with disease complexity and patient heterogeneity hindering this task. With no biomarkers avail-
able and unclear pathophysiology, ME/CFS and FM are often misdiagnosed, further complicating the
identification of effective treatments. In this study, we tried to overcome patient heterogeneity by
studying a very finely phenotyped cohort diagnosed by a single expert clinician, representing one of
the few studies (Nepotchatykh etal., 2023) including samples of patients diagnosed with either ME/
CFS, FM, or both (co- diagnosed group) by two different clinical criteria (Canadian (Carruthers etal.,
2003) and International Consensus (Carruthers et al., 2011) criteria, and the 1990 (Wolfe etal.,
1990) and 2011 (Wolfe etal., 2016; Wolfe etal., 2010) ACR criteria, respectively). Due to previously
reported sex- associated differences in these diseases, only female patients were included, limiting the
external validity of the results in the male population while minimizing potential sex- associated bias.
Triggering agents for ME/CFS and FM, such as infections, stress, or trauma (Chu et al., 2019;
Furness et al., 2018), seem capable of disrupting the epigenetic mechanisms constraining HERV
expression (Rangel etal., 2022). Dysregulation of HERV has been proposed to contain more distinc-
tive features than the host transcriptome itself in response to viral infections (Marston etal., 2021),
and it is becoming detected in additional complex diseases (Giménez- Orenga and Oltra, 2021),
offering a new perspective to study their physiopathology. In fact, HERV profiling could constitute
an indirect approach (surrogate marker) to evidence epigenetic derangements and 3D alterations of
the chromatin structure with obvious functional consequences. Previous research revealed aberrant
activation of HERV in ME/CFS (Rodrigues etal., 2019) and FM (Ovejero etal., 2020). However, these
two studies consisting of broad amplification of HERV by degenerate primer sets, did not allow for
specific HERV loci identification. In this study, we fill this gap of knowledge by genome- wide analyzing
annotated HERV sequences in the human genome and their relationship with immune gene expression
and patient symptoms. The results show for the first time a comparison of PBMCs HERV expression
profiles of ME/CFS, FM, and co- diagnosed patient groups as compared to HERV expression profiles
of healthy subjects. Despite the limited number of samples (8 ME/CFS cases, 10 FM, 16 comorbid
cases, and 9 matched healthy controls), our results show a specific HERV fingerprint for each disease
that allows perfect discrimination into three distinct patient groups and their separation from healthy
participants, opening the possibility of HERV expression profile- based diagnostic of these symptom-
related, diseases. Methodology that could be extended to the study of other chronic diseases such as
post- COVID- 19 condition. Particularly interesting will be to determine whether post- COVID- 19 cases
present different HERV fingerprints that justify commonalities to ME/CFS and/or FM, or if by contrast
show an evolution closer to the more prevalent co- diagnosed state. Worth noting are the results
obtained in the co- diagnosed group which, unexpectedly, presented a completely different HERV
profile to those identified for ME/CFS and for FM while strikingly closer to that of healthy controls.
Findings that support a distinct subjacent pathomechanism for the co- diagnosed group of patients
and, therefore, the identification of a novel nosologic entity.
In addition, our study reveals enhanced HERV dysregulation in the ME/CFS group when compared
to FM, co- diagnosed cases, or healthy controls and significant quantitative differences within a more
severely affected ME/CFS subgroup (subgroup 2) supporting inter- individual HERV expression vari-
ability toward patient subtyping, with potential relevance for patient tailored treatments, and poten-
tial for HERV expression levels to define degree of severity in ME/CFS. The study also defined a set of
DE HERV commonly downregulated in all three patient groups (ME/CFS, FM, and their co- diagnosis,
Cluster 4), a deeper study of which may help explain clinical commonalities between these conditions.
In fact, the lack of a specific association of Cluster 4 DE HERV with a patient- study group opens the
possibility of their derangement in additional complex and chronic diseases.

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 14 of 24
Analysis of DE HERV loci at family level within each cluster interestingly showed the involvement of
different HERV families in each of the study groups, with ME/CFS outstanding by the largest hetero-
geneity, including deregulation in families such as HERV- H or HERV- P. Deregulation of HERV- H and
HERV- P families has been previously linked to disease (Bergallo etal., 2019; Morris etal., 2019; Yi
etal., 2006) with potential involvement in immune function performance (Goodchild etal., 1993;
Kulski etal., 1999; Mager etal., 1999; Mangeney etal., 2001; Matsuzawa etal., 2021). Another
HERV family consistently linked to disease is the HERV- W (Dolei etal., 2014; Giménez- Orenga etal.,
2022; Kremer etal., 2013), whose upregulation has been reported in FM (Ovejero etal., 2020).
Contrarily, our results evidenced downregulation of HERV- W family in all patient groups, particularly
the HERV- W- LTRSU3_19q13.42_st element. Dissimilar results may be due to the use of different tech-
niques to evaluate family expression (degenerate primers in RT- qPCR vs. probes in microarray), or
accuracy in patient diagnosis. Our study further revealed that the majority of downregulated HERV
in ME/CFS, FM, and co- diagnosed groups belonged to a sole HERV family, the MLT1. This family was
found enriched near DE genes involved in the response to infection (Bogdan etal., 2020). MLT1
family members are among the most frequently overlapping with human lncRNAs (Ramsay etal.,
2017). However, the potential role of MLT1- derived lncRNAs in the diseases studied here remains
speculative at this stage.
In accordance with genomic data (Lander etal., 2001), most deregulated HERV loci consisted of
solitary LTRs, known for their ability to influence gene expression through their regulatory sequences
(Ito etal., 2017). For example, a HERV- H solitary LTR located at the HHLA2 locus serves as its main
polyadenylation signal (Goodchild etal., 1993; Mager etal., 1999), allowing the encoding of the
HHLA2 protein, which regulates CD4 and CD8 T cell functions along with antigen- presenting cell
proliferation (Zhao etal., 2013). Furthermore, it is known that non- allelic homologous recombination
events leading to solitary LTRs create inter- individual variation, not only by removing coding potential
but also by altering the cis- regulatory or transcriptional activity of the affected particular genomic
regions (Rebollo etal., 2012; Thomas etal., 2018) even causing microdeletions (Hermetz etal.,
2012; Sanchez- Valle etal., 2010; Shuvarikov et al., 2013), which offer possibilities for exploring
familial risk factors of these diseases.
Analysis of differential gene expression in the exact same samples provided additional evidence
of HERV- associated immune changes in the ME/CFS group as compared to FM, co- diagnosed, or
healthy control groups, further supporting the existence of two ME/CFS subgroups differing, once
more, at the quantitative level only.
Although genomic localization of correlated HERV and genes suggested major mechanisms other
than co- transcription, it should be noted that gene probes included in the microarray do not fully
picture the complete mRNA transcriptome, and therefore the possibility that other DE genes locate
nearer to deregulated HERV cannot be ruled out. Furthermore, long- range interactions between gene
promoters and LTR regulatory elements have been described for some HERV families, including MLT1,
MER21, MER41, or LTR54 (Ito etal., 2017), reported as DE in this study. In this regard, it has been
shown that the HERV- H family, upregulated in ME/CFS subgroup 2, can influence gene expression
over long genomic distances, for example by creating topologically associated domains (Zhang etal.,
2019), perhaps explaining at least some of our findings of DE immune genes being co- expressed with
DE HERV. Under the hypothesis that independently transcribed DE HERV could be regulated by the
same trans- acting mechanisms as DE genes, we interestingly found enrichment for TFBS regulating
the immune response within DE HERV, suggesting trans- mediated coordination of deranged HERV
and gene expression. Furthermore, HERV upregulated in ME/CFS were markedly enriched in chro-
matin remodeling factor binding sites, including SETDB1 and TRIM28, both being key epigenetic
repressors of HERV expression, pointing to the impairment of TRIM28/SETDB1 axis as the under-
lying mechanism of HERV upregulation in ME/CFS. In support of this possibility, influenza A virus, a
suspected trigger of ME/CFS (Magnus etal., 2015), has been shown to activate HERV expression by
impairing the repressor TRIM28/SETDB1 axis (Schmidt etal., 2019).
As the literature endorses changes in immune cell subsets in ME/CFS (Brenu etal., 2014; Brenu
etal., 2011; Curriu etal., 2013), we aimed to indirectly measure potentially skewed immune cell type
ratios in our study groups with the use of CIBERSORTx software (Newman etal., 2019). The detected
increase in plasma cells and resting CD4 T memory cells in ME/CFS subgroup 2, suggests the pres-
ence or the history of previous infections stimulating the immune system in these subjects, coinciding

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 15 of 24
with greater dysregulation in immune- related genes and HERV upregulation and increased severity of
symptoms. On another hand, this same subgroup of patients exhibited lower levels of γδ T cells, an
event reported to our knowledge in other autoimmune diseases like systemic lupus erythematosus
and psoriasis (Fay etal., 2016; Wang etal., 2012). It has been described that γδ T cells mediate the
immune response against viral infections like EBV, HHV- 6, or CMV (Zhao etal., 2018), which together
with natural killer cell dysfunction reported in the mentioned autoimmune diseases (Henriques etal.,
2013; Kim etal., 2019) but also in ME/CFS (Eaton- Fitch etal., 2019), suggests an impaired cytotoxic
function rendering patients with lower counts of this cell type, becoming more susceptible to viral
reactivation. Interestingly, anomalies in immune cell abundance highly correlated with observed HERV
expression patterns, indicating that HERV deregulation may either reflect expression derangement in
those cell subpopulations or somehow influence immune cell ratios. The analysis of PBMC, rather than
isolated subpopulations, is limited at distinguishing between these non- exclusive possibilities. In addi-
tion, we found strong associations between the HERV elements MER31_3p26.1 and HERV16_3p21.31
expression correlating with physical and mental fatigue according to health status assessment
instruments MFI (Smets etal., 1995) and SF- 36 (Mchorney etal., 1993), where higher scores indi-
cate a worse or better health status, respectively. In this regard, both HERV elements, belonging
to Module 3, that is, upregulated in ME/CFS subgroup 2, positively and negatively correlated with
MFI and SF- 36 subdomains, respectively, indicating that increased expression of MER31_3p26.1 and
HERV16_3p21.31 elements may affect patient health status. Interestingly enough, previous correla-
tions between HERV expression and physical fatigue are reported in the post- viral syndrome post-
COVID- 19 condition (Giménez- Orenga et al., 2022), which shares symptomatology with ME/CFS
(Komaroff and Lipkin, 2023). Further study of MER31_3p26.1 and HERV16_3p21.31 elements and
their potential pathogenic function seem granted.
Overall, this study not only shows the potential of HERV expression profiles as biomarkers for ME/
CFS, FM, and for the definition of ME/CFS and FM co- diagnosis but also exposes their potential
involvement in immune anomalies and connection with patient symptoms in these diseases, with
etiological potential. Our results reveal family- specific HERV deregulation in ME/CFS and FM immune
cells and their association with changes in pathogen- detection genes and immune cell subsets. Recent
findings by Marston etal., 2021 demonstrated specific disruption of HERV families in response to
specific viral infection agents. Perhaps a database enabling the comparison of disease HERV expres-
sion profiles with those induced by exogenous infections could help reveal the triggering agents
behind diseases of suspected viral etiology, such as ME/CFS and FM.
In summary, this study pioneers the comparison of female ME/CFS, FM, and co- diagnosed HERV
expression profiles showing the superior diagnostic potential of HERV fingerprints over immune tran-
scriptomes, with clinical implications, importantly providing an effective tool for the objective differ-
ential diagnosis of these hard- to- diagnose co- occurring diseases. Given the limited sample size of
our cohort, validation of the findings in extended cohorts is a must. The potential mechanisms point
at underlying epigenetic disturbances strikingly derepressing HERV silencing, particularly in ME/CFS.
Whether these disturbances primarily affect certain immune subpopulations, estimated decreased for
γδ T cells, or increased for plasma and resting CD4 memory T cells, correlating with patient symptom
severity in ME/CFS, or reflect the presence of DEHERV profiles in immune cell subpopulations biased
in these patient groups is an aspect that awaits further work.
Materials and methods
Study design
This cross- sectional observational study was approved by the Public Health Research Ethics Committee
DGSP- CSISP of Valencia, núm. 20190301/12, Valencia, Spain. The study included a total of 43 female
patients invited to participate from local patient associations who were clinically diagnosed with ME/
CFS (n = 8), FM (n = 10), or both (n = 16) (National Biobank Registry Ref 0006024) and 9 healthy
control individuals’ population- matched for age and BMI (National Biobank Registry Ref 0006034).
Patients were diagnosed by an FM and ME/CFS specialized clinician (Hospital de Manises, Valencia,
Spain) using the 1990 (Wolfe etal., 1990) and 2011 (Wolfe etal., 2016; Wolfe etal., 2010) Amer-
ican College of Rheumatology (ACR) criteria for FM and the Canadian (Carruthers etal., 2003) and
International Consensus (Carruthers etal., 2011) criteria for ME/CFS diagnosis. Patients with health

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 16 of 24
problems other than FM and ME/CFS were excluded from the study. Individuals with any similar o
related pathology, including a medical history of chronic pain and/or fatigue, or serious health compli-
cations, were excluded from control group, as well as medicated healthy controls. Written informed
consent was obtained from all study participants and patient health status was also evaluated with
the use of standardized questionnaires, including the FIQ case report form (Burckhardt etal., 1991;
Rivera and González, 2004), the MFI questionnaire (Smets etal., 1995), and the quality- of- life SF- 36
instrument (Mchorney etal., 1993). Participating patients agreed to withdraw medication at least
12hr prior to blood draw. Microarray analysis of RNA extracted from PBMCs was performed along
with complementary bioinformatic analysis to evaluate differences in HERV profiles among patients
with potential application as biomarkers or indicators of the underlying pathomechanisms.
Isolation of PBMCs and total RNA extraction
Up to 10ml of whole blood were collected via venipuncture after a 12- hr overnight fasting in K2EDTA
tubes (Becton Dickinson, Franklin Lakes, NJ, USA) and processed within 2hr by dilution at 1:1 (vol/
vol) ratio in phosphate- buffered saline solution (PBS) with layering on top of 1volume of Ficoll- Paque
Premium (GE Healthcare, Chicago, IL, USA) and separation by density centrifugation at 500 × g for
30min (20°C, brakes off). The PBMC layer was isolated and washed with PBS and resuspended in
red blood cell lysis buffer (155mM NH4Cl, 10mM NaHCO3, 0.1mM EDTA, and pH 7.4), kept on ice
for 5min, and centrifuged (20°C at 500 × g for 10min), to remove contaminating erythrocytes. The
washed pellets were adjusted to a final concentration of 107cells/ml in freezing medium (90% FBS,
10% DMSO), aliquoted, and deeply frozen in liquid nitrogen until use. Total RNA was extracted using
RNeasy Mini Kit (QIAGEN, MD, USA) according to the manufacturer’s instructions. RNA quality was
assessed using Agilent TapeStation 4200 (Agilent). All RNA samples had an RNA integrity number
above 7.
Transcriptome analysis by microarray
To minimize batch effects the samples were anonymized (blinded to the operator) and scrambled
across groups. HERV transcriptome was scrutinized using custom high- density HERV- V3 microarrays
(Becker et al., 2017), capable to discriminate 174,852 HERV elements, 179,142 MaLR elements,
and putative active 1072 LINE- 1 elements at the locus level, in addition to detecting a set of 1559
genes involved in eight potentially relevant cellular pathways (immunity, inflammation, cancer, central
nervous system affections, differentiation, telomere maintenance, chromatin structure, and gag- like
genes). These custom HERV- V3 arrays also allow to discern the different subdomains within proviral
sequences: 3′ or 5′ LTRs and gag/pol/env regions of proviral HERV and solitary LTRs originated after
non- allelic recombination events between the two LTRs causing complete proviral elimination (Thomas
etal., 2018). To analyze HERV transcriptomes of PBMC- derived RNA samples, cDNA was synthesized
and amplified from 45ng of RNA using the Ovation Pico WTA System V2 kit (Nugen) according to the
manufacturer’s instructions. The resulting amplified ssDNA was purified using the QIAquick purifica-
tion kit (QIAGEN, MD, USA). Total DNA concentration was measured with NanoDrop 2000 spectro-
photometer (Thermo Scientific) and quality was assessed on the Bioanalyzer 2100. Five micrograms
of purified ssDNA were enzymatically fragmented into 50–100bp fragments and biotin- labeled with
the Encore Biotin Module kit from Nugen according to the manufacturer’s instructions. The resulting
target was mixed with standard hybridization controls and B2 oligonucleotides following supplier
recommendations. The hybridization cocktail was heat- denatured at 95°C for 2min, incubated at 50°C
for 5min, and centrifuged at 16,000 × g for 5min to pellet the residual salts. HERV- V3 microarrays
were pre- hybridized with 200μl of hybridization buffer and placed under stirring (60rpm) in an oven at
50°C for 10min. Hybridization buffer was then replaced by denatured hybridization cocktail. Hybrid-
ization was performed at 50°C for 18hr in the oven under constant stirring (60 rpm). Washing and
staining were carried out according to the manufacturer’s protocol, using a fluidic station (GeneChip
fluidic station 450, Affymetrix). Arrays were finally scanned using a fluorometric scanner (GeneChip
scanner 3000 7G, Affymetrix). Biotin purification, hybridization, and reading steps were performed by
Sampled (Piscataway, NJ 08854, USA). Overall, 1,397,352 probes were detected, the vast majority
of them corresponding to HERV (1,290,800 probesets), followed by genes (103,724 probesets), and
LINE- 1 (2828 probesets).

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 17 of 24
Identification of DE HERV and genes
All bioinformatic analyses were performed with RStudio software version 4.2.1. Microarray CEL files
were processed and analyzed using R oligo package (Carvalho and Irizarry, 2010). Data were normal-
ized, adjusted for background noise, and summarized using the RMA (Robust Multi- Array) algorithm.
DE analysis was performed using limma R package (Ritchie etal., 2015), considering DE those probes
with a Benjamini–Hochberg adjusted p- value (FDR) <0.1and an absolute log2 fold- change >1. The
analysis results were presented in volcano plots using EnhancedVolcano R package (Blighe et al.,
2022).
Clustering analysis
DE HERV or genes (FDR <0.1and |log2FC|>1) were used to perform clustering analysis. PCA was
performed to analyze the behavior of samples under study based on DE HERV expression profiles
by using factoextra R package (Kassambara and Mundt, 2025). Samples were plotted according
to the first two PCA principal components and the contribution of variables to each of the principal
components was represented in barplots. Scaled expression data of DE HERV or gene probesets were
used to cluster samples and probes. Data were represented in heatmap plots by using pheatmap R
package (Kolde etal., 2018). Rows and columns were clustered using Euclidean and Pearson correla-
tion distances, respectively. HERV clusters were visually identified based on heatmap results and
extracted from dendrograms using the pheatmap implemented function. Mean probe expression by
group was calculated for all probes encompassed in each cluster and their distribution and expression
level were visualized by box and violin plots. Average- linkage agglomerative hierarchical clustering
with Pearson correlation distance was performed on normalized expression data to identify modules
of HERV and genes with correlated expression patterns across samples. Modules of HERV–genes were
produced by applying a dynamic tree cut of dendrograms using a minimum cluster size of 110 HERV
or genes, as described by Marston etal., 2021.
Enrichment analysis
Microarray probes were annotated based on the HERV- V3 annotation file kindly provided by
bioMérieux, allowing retrieval of HERV subdomain (gag, pol, env, or LTR), family, and genomic loca-
tion of each probe. Family group information for solitary LTR probes was added based on Dfam and
EnHERV databases. Original hg37 HERV- V3 coordinates were converted into hg38 genomic coor-
dinates by using UCSC LiftOver tool (Kuhn etal., 2013) and the corresponding chromosome band
information was added. The genomic context of the HERV loci was assessed with Goldmine R package
(Bhasin and Ting, 2016). Enrichment analysis for HERV family, HERV subdomain, and genomic context
was performed for the HERV encompassed in each cluster by comparing their frequency in the cluster
to their corresponding on hg38 genomic data according to HERV- V3 microarray annotation (Becker
etal., 2017). The frequency of families, HERV subdomains, and genomic context (intron, exon, inter-
genic, promoter, or 3′ end) for the DE HERV encompassed in each cluster and in the overall hg38
genomic data were represented as stacked bar plots with ggplot2 R package (Wickham, 2016). To
assess for TFBS enrichment, the ReMapEnrich Shiny 1.4 web interface was used providing DE HERV
loci genomic coordinates as input. Enriched TFBS were represented as a heatmap using pheatmap R
package (Kolde etal., 2018), with colors defining modules in which the enrichment was found, and
no enrichments were shown as blank.
Overrepresentation functional analysis
Genes encompassed in modules identified by hierarchical clustering as well as TF binding to identified
enriched TFBS in HERV of each module were functionally analyzed by GO using the R package cluster-
Profiler (Wu etal., 2021) with an adjusted p- value cutoff of 0.05. For each module, the top 5 GO terms as
ranked by increasing adjusted p- value were visualized in a heatmap using pheatmap R package (Kolde
etal., 2018). Statistical significance of each GO term as adjusted p- values was represented by a color
gradient, with darker colors representing more significant terms in the module. Terms sharing significant
overrepresentation in other modules (adjusted p < 0.05) were colored on heatmaps accordingly.
Evaluation of overlaps between HERV and gene sequences
Genomic coordinates of HERV and genes encompassed in each module were transformed to genomic
range objects (Granges) using the GenomicRanges R package (Lawrence etal., 2013). They were

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 18 of 24
subsequently plotted over a linear chromosomal representation with karyoploteR package (Gel
and Serra, 2017), which displays the density of sequences throughout chromosomes. The genomic
distance between HERV and genes in the same module was determined with a specific function
included in GenomicRanges package and displayed as density plots.
Measurement of immune cell relative amounts
Digital cytometry analysis was performed with CIBERSORTx (Newman et al., 2019) software on
normalized non- logarithmic gene expression data for all samples. The default LM22 leukocyte gene
signature matrix (Newman etal., 2015) was used as a reference. LM22 was generated using Affy-
metrix HGU133A microarray data distinguishing 22 human hematopoietic cell phenotypes, including
different types of T cells, B cells (naïve and memory), plasma cells, NK cells, monocytes, dendritic cells,
and some myeloid subsets (Newman etal., 2015). Output results were visually inspected for poten-
tial differences in immune cell counts and significant differences were assessed by Welch’s t- test with
adjusted p- value <0.05 from specific immune cell types. Results were plotted as violin plots.
Trait-measure and symptom association
For each module extracted by hierarchical clustering the eigengenes were calculated with the use
of the function moduleEigengenes implemented in WGCNA package (Langfelder and Horvath,
2008). A matrix of normalized expression data and a list assigning each sequence to each module
was provided as input. Calculated module eigengenes were then correlated to external microarray
data, including immune cell proportion values provided by CIBERSORTx, questionnaire scores for FIQ,
MFI, and SF- 36, or the clinical diagnosis. This last variable was binarized with the use of binarizeCat-
egorialColumns. Results were plotted by using pheatmap package (Kolde etal., 2018) with boxes
showing Pearson correlation values and associated p- values between gene expression levels of each
module and the disease trait. A correlation value of 1 (orange or green) and –1 (blue) were used to
quantify the strongest positive and negative correlations, respectively, while 0 (white) was assigned
for no correlation.
Real-time quantitative polymerase chain reaction
DE of specific HERV sequences was validated using the same RNA samples analyzed by microarray (n
= 8 controls, n = 8 ME/CFS, n = 7FM, and n = 14 co- diagnosed). Reverse transcription was performed
using High- Capacity cDNA reverse Transcription kit (Applied Biosystems, Waltham, MA, USA, cat.
4308228), with 1µg of total RNA according to the manufacturer’s guidelines. cDNAs were used for
real- time PCR using PowerUP Sybr Green Master Mix (Applied Biosystems, cat. 100029283) and a
Lightcycler LC480 instrument (Roche, Penzberg, Germany). Standard amplification conditions were
applied, including a single hotstart polymerase preactivation cycle at 94°C for 15min, up to 45 ampli-
fication cycles, each one consisting of three steps: denaturation at 95°C for 15s, annealing at 60°C for
30s, and extension at 70°C for 30s. Sequences of specific primers used are detailed in Supplemen-
tary file 1H. GAPDH levels were used for relative quantification of the RNAs amplified, and 2−∆Ct
analysis to calculate expression was applied.
Statistical analysis
All statistical analyses were done in R v4.2.1. Data distributions were tested for normality. Normally
distributed data were tested using two- tailed unpaired Student’s t- tests; non- normal data were
analyzed with non- parametric statistical test, as detailed. For enrichment analyses, we used a Fish-
er’s exact test to calculate p- value, considering enriched in the provided list if an adjusted p- value
(FDR) was less than 0.05. We chose to use Fisher’s exact test in the analysis of contingency tables to
compare ME/CFS and healthy group, as it is more appropriate for small sample sizes in comparison to
the chi- square test or G- test of independence.
Acknowledgements
We particularly thank all the patients who participated in the study and Dr. Vicente Serra (Umivale,
Valencia, Spain) for his help in the recruitment of healthy volunteers. We also acknowledge bioMérieux
(Ain, France) for allowing the use of their custom- made HERV- V3 high- density microarrays and Sampled
LTD (Piscataway, NJ, USA) for assisting us in the analysis by microarray. This study was funded by an ME

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 19 of 24
Research UK (SCIO charity number SC036942) grant, and by Generalitat Valenciana CIAICO/2021/103
grant to EO. KG- O is supported by the Generalitat Valenciana ACIF2021/179 grant. Funders were not
involved in any of the research stages.
Additional information
Funding
Funder Grant reference number Author
ME Research UK SC036942 Elisa Oltra
Generalitat Valenciana CIAICO/2021/103 Elisa Oltra
Generalitat Valenciana ACIF2021/179 Karen Giménez-Orenga
The funders had no role in study design, data collection, and interpretation, or the
decision to submit the work for publication.
Author contributions
Karen Giménez- Orenga, Data curation, Formal analysis, Investigation, Writing – original draft, Writing
– review and editing; Eva Martín- Martínez, Investigation, Methodology, Writing – review and editing;
Lubov Nathanson, Investigation, Writing – review and editing; Elisa Oltra, Conceptualization, Data
curation, Supervision, Funding acquisition, Writing – original draft, Project administration, Writing –
review and editing
Author ORCIDs
Karen Giménez- Orenga
https://orcid.org/0000-0001-9790-7327
Eva Martín- Martínez
https://orcid.org/0000-0003-3261-494X
Lubov Nathanson
https://orcid.org/0000-0003-1038-9083
Elisa Oltra
https://orcid.org/0000-0003-0598-2907
Ethics
This study has been performed in accordance with the Declaration of Helsinki. It was approved by
the Public Health Research Ethics Committee DGSP- CSISP of Valencia, núm. 20190301/12, Valencia,
Spain. All participants signed an informed consent to participate in this study.
Peer review material
Reviewer #1 (Public review): https://doi.org/10.7554/eLife.104441.3.sa1
Reviewer #2 (Public review): https://doi.org/10.7554/eLife.104441.3.sa2
Reviewer #3 (Public review): https://doi.org/10.7554/eLife.104441.3.sa3
Author response https://doi.org/10.7554/eLife.104441.3.sa4
Additional files
Supplementary files
MDAR checklist
Supplementary file 1. Raw, processed, and complementary data. (A) Itemized participant
demographics and health status assessment with FIQ, MFI and SF- 36 questionnaires. (B) Statistical
analysis of participant demographics and health status assessment with FIQ, MFI and SF- 36
questionnaires. (C) Differentially expressed HERV. (D) Patient health status assessment with FIQ, MFI,
and SF- 36 questionnaires separating ME/CFS into two subgroups. (E) Relative contribution of HERV
families to each cluster. (F) Differentially expressed genes. (G) Genomic distance between correlated
HERVs and genes in each module. (H) Primers for RT- qPCR validation.
Data availability
All data associated with the study are available in the paper or supplementary materials, except
for HERV- V3 microarray annotation provided by bioMérieux under material transfer agreement.
Microarray data is available at GSE269047 from the NCBI GEO database.

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 20 of 24
The following dataset was generated:
Author(s) Year Dataset title Dataset URL Database and Identifier
Giménez- Orenga K,
Martín- Martínez E,
Nathanson L, Oltra E
2025 HERV activation segregates
ME/CFS from bromyalgia
while dening a novel
nosologic entity
https://www. ncbi.
nlm. nih. gov/ geo/
query/ acc. cgi? acc=
GSE269047
NCBI Gene Expression
Omnibus, GSE269047
References
Abbi B, Natelson BH. 2013. Is chronic fatigue syndrome the same illness as fibromyalgia: evaluating the “single
syndrome” hypothesis. QJM 106:3–9. DOI: https://doi.org/10.1093/qjmed/hcs156, PMID: 22927538
Bateman L, Bested AC, Bonilla HF, Chheda BV, Chu L, Curtin JM, Dempsey TT, Dimmock ME, Dowell TG,
Felsenstein D, Kaufman DL, Klimas NG, Komaroff AL, Lapp CW, Levine SM, Montoya JG, Natelson BH,
Peterson DL, Podell RN, Rey IR, etal. 2021. Myalgic encephalomyelitis/chronic fatigue syndrome: Essentials of
diagnosis and management. Mayo Clinic Proceedings 96:2861–2878. DOI: https://doi.org/10.1016/j.mayocp.
2021.07.004, PMID: 34454716
Becker J, Pérot P, Cheynet V, Oriol G, Mugnier N, Mommert M, Tabone O, Textoris J, Veyrieras JB, Mallet F.
2017. A comprehensive hybridization model allows whole HERV transcriptome profiling using high density
microarray. BMC Genomics 18:286. DOI: https://doi.org/10.1186/s12864-017-3669-7, PMID: 28390408
Bergallo M, Mareschi K, Montanari P, Calvi C, Berger M, Bini I, Daprà V, Galliano I, Fagioli F. 2019. Human
endogenous retrovirus, HERV- P and HERV- R in pediatric leukemia patients. Journal of Hematopathology
12:51–56. DOI: https://doi.org/10.1007/s12308-019-00352-0
Bhasin JM, Ting AH. 2016. Goldmine integrates information placing genomic ranges into meaningful biological
contexts. Nucleic Acids Research 44:5550–5556. DOI: https://doi.org/10.1093/nar/gkw477, PMID: 27257071
Blighe K, Rana S, Lewis M. 2022. EnhancedVolcano: publication- ready volcano plots with enhanced colouring
and labeling. ab0428e. Github. https://github.com/kevinblighe/EnhancedVolcano
Bogdan L, Barreiro L, Bourque G. 2020. Transposable elements have contributed human regulatory regions that
are activated upon bacterial infection. Philosophical Transactions of the Royal Society B 375:20190332. DOI:
https://doi.org/10.1098/rstb.2019.0332
Brenu EW, van Driel ML, Staines DR, Ashton KJ, Ramos SB, Keane J, Klimas NG, Marshall- Gradisnik SM. 2011.
Immunological abnormalities as potential biomarkers in Chronic Fatigue Syndrome/Myalgic Encephalomyelitis.
Journal of Translational Medicine 9:81. DOI: https://doi.org/10.1186/1479-5876-9-81, PMID: 21619669
Brenu EW, Huth TK, Hardcastle SL, Fuller K, Kaur M, Johnston S, Ramos SB, Staines DR, Marshall- Gradisnik SM.
2014. Role of adaptive and innate immune cells in chronic fatigue syndrome/myalgic encephalomyelitis.
International Immunology 26:233–242. DOI: https://doi.org/10.1093/intimm/dxt068, PMID: 24343819
Burckhardt CS, Clark SR, Bennett RM. 1991. The fibromyalgia impact questionnaire: development and
validation. The Journal of Rheumatology 18:728–733 PMID: 1865419.
Carruthers BM, Jain AK, De Meirleir KL, Peterson DL, Klimas NG, Lerner AM, Bested AC, Flor- Henry P, Joshi P,
Powles ACP, Sherkey JA, van de Sande MI. 2003. Myalgic Encephalomyelitis/Chronic Fatigue Syndrome.
Journal of Chronic Fatigue Syndrome 11:7–115. DOI: https://doi.org/10.1300/J092v11n01_02
Carruthers BM, van de Sande MI, De Meirleir KL, Klimas NG, Broderick G, Mitchell T, Staines D, Powles ACP,
Speight N, Vallings R, Bateman L, Baumgarten- Austrheim B, Bell DS, Carlo- Stella N, Chia J, Darragh A, Jo D,
Lewis D, Light AR, Marshall- Gradisnik S, etal. 2011. Myalgic encephalomyelitis: International consensus criteria.
Journal of Internal Medicine 270:327–338. DOI: https://doi.org/10.1111/j.1365-2796.2011.02428.x, PMID:
21777306
Carvalho BS, Irizarry RA. 2010. A framework for oligonucleotide microarray preprocessing. Bioinformatics
26:2363–2367. DOI: https://doi.org/10.1093/bioinformatics/btq431
Castro- Marrero J, Faro M, Aliste L, Sáez- Francàs N, Calvo N, Martínez- Martínez A, de Sevilla TF, Alegre J. 2017.
Comorbidity in chronic fatigue syndrome/myalgic encephalomyelitis: A nationwide population- based cohort
study. Psychosomatics 58:533–543. DOI: https://doi.org/10.1016/j.psym.2017.04.010, PMID: 28596045
Charvet B, Brunel J, Pierquin J, Iampietro M, Decimo D, Queruel N, Lucas A, del M Encabo- Berzosa M, Arenaz I,
Marmolejo TP, Gonzalez AI, Maldonado AC, Mathieu C, Küry P, Flores- Rivera J, Torres- Ruiz F, Avila- Rios S,
Salgado Montes de Oca G, Schoorlemmer J, Perron H, etal. 2023. SARS- CoV- 2 awakens ancient retroviral
genes and the expression of proinflammatory HERV- W envelope protein in COVID- 19 patients. iScience
26:106604. DOI: https://doi.org/10.1016/j.isci.2023.106604
Chu L, Valencia IJ, Garvert DW, Montoya JG. 2019. Onset patterns and course of myalgic encephalomyelitis/
chronic fatigue syndrome. Frontiers in Pediatrics 7:12. DOI: https://doi.org/10.3389/fped.2019.00012, PMID:
30805319
Clauw DJ, Calabrese L. 2024. Rheumatology and Long COVID: lessons from the study of fibromyalgia. Annals of
the Rheumatic Diseases 83:136–138. DOI: https://doi.org/10.1136/ard-2023-224250
Curriu M, Carrillo J, Massanella M, Rigau J, Alegre J, Puig J, Garcia- Quintana AM, Castro- Marrero J, Negredo E,
Clotet B, Cabrera C, Blanco J. 2013. Screening NK-, B- and T- cell phenotype and function in patients suffering
from Chronic Fatigue Syndrome. Journal of Translational Medicine 11:68. DOI: https://doi.org/10.1186/1479-
5876-11-68, PMID: 23514202

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 21 of 24
Dolei A, Garson JA, Arru G, Clerici M, Germi R, Marche PN, Perron H. 2014. Multiple sclerosis- associated
retrovirus and related human endogenous retrovirus- W in patients with multiple sclerosis. Journal of
Neuroimmunology 266:87–88. DOI: https://doi.org/10.1016/j.jneuroim.2013.11.009, PMID: 24355750
Dopkins N, O’Mara MM, Singh B, Marston JL, Bendall ML, Nixon DF. 2022. How human endogenous retroviruses
interact with the microbiota in health and disease. Trends in Microbiology 30:812–815. DOI: https://doi.org/10.
1016/j.tim.2022.05.011, PMID: 35672223
Eaton- Fitch N, du Preez S, Cabanas H, Staines D, Marshall- Gradisnik S. 2019. A systematic review of natural
killer cells profile and cytotoxic function in myalgic encephalomyelitis/chronic fatigue syndrome. Systematic
Reviews 8:279. DOI: https://doi.org/10.1186/s13643-019-1202-6, PMID: 31727160
Fay NS, Larson EC, Jameson JM. 2016. Chronic inflammation and γδ T cells. Frontiers in Immunology 7:210.
DOI: https://doi.org/10.3389/fimmu.2016.00210, PMID: 27303404
Ferrari R, Grandi N, Tramontano E, Dieci G. 2021. Retrotransposons as drivers of mammalian brain evolution.
Life 11:376. DOI: https://doi.org/10.3390/life11050376, PMID: 33922141
Furness PJ, Vogt K, Ashe S, Taylor S, Haywood- Small S, Lawson K. 2018. What causes fibromyalgia? An online
survey of patient perspectives. Health Psychology Open 5:2055102918802683. DOI: https://doi.org/10.1177/
2055102918802683, PMID: 30275965
Gel B, Serra E. 2017. karyoploteR: an R/Bioconductor package to plot customizable genomes displaying arbitrary
data. Bioinformatics 33:3088–3090. DOI: https://doi.org/10.1093/bioinformatics/btx346, PMID: 28575171
Giménez- Orenga K, Oltra E. 2021. Human endogenous retrovirus as therapeutic targets in neurologic disease.
Pharmaceuticals 14:495. DOI: https://doi.org/10.3390/ph14060495, PMID: 34073730
Giménez- Orenga K, Pierquin J, Brunel J, Charvet B, Martín- Martínez E, Perron H, Oltra E. 2022. HERV- W ENV
antigenemia and correlation of increased anti- SARS- CoV- 2 immunoglobulin levels with post- COVID- 19
symptoms. Frontiers in Immunology 13:1020064. DOI: https://doi.org/10.3389/fimmu.2022.1020064, PMID:
36389746
Goodchild NL, Wilkinson DA, Mager DL. 1993. Recent evolutionary expansion of a subfamily of RTVL- H human
endogenous retrovirus- like elements. Virology 196:778–788. DOI: https://doi.org/10.1006/viro.1993.1535,
PMID: 8372448
Groven N, Reitan SK, Fors EA, Guzey IC. 2021. Kynurenine metabolites and ratios differ between Chronic
Fatigue Syndrome, Fibromyalgia, and healthy controls. Psychoneuroendocrinology 131:105287. DOI: https://
doi.org/10.1016/j.psyneuen.2021.105287, PMID: 34090138
Gruchot J, Lewen I, Dietrich M, Reiche L, Sindi M, Hecker C, Herrero F, Charvet B, Weber- Stadlbauer U,
Hartung HP, Albrecht P, Perron H, Meyer U, Küry P. 2023. Transgenic expression of the HERV- W envelope
protein leads to polarized glial cell populations and a neurodegenerative environment. PNAS
120:e2308187120. DOI: https://doi.org/10.1073/pnas.2308187120, PMID: 37695891
Guo X, Zhao Y, You F. 2024. Identification and characterization of endogenous retroviruses upon SARS- CoV- 2
infection. Frontiers in Immunology 15:1294020. DOI: https://doi.org/10.3389/fimmu.2024.1294020, PMID:
38646531
Harrison JE, Weber S, Jakob R, Chute CG. 2021. ICD- 11: an international classification of diseases for the
twenty- first century. BMC Medical Informatics and Decision Making 21:206. DOI: https://doi.org/10.1186/
s12911-021-01534-6, PMID: 34753471
Henriques A, Teixeira L, Inês L, Carvalheiro T, Gonçalves A, Martinho A, Pais ML, da Silva JAP, Paiva A. 2013. NK
cells dysfunction in systemic lupus erythematosus: relation to disease activity. Clinical Rheumatology 32:805–
813. DOI: https://doi.org/10.1007/s10067-013-2176-8, PMID: 23377197
Hermetz KE, Surti U, Cody JD, Rudd MK. 2012. A recurrent translocation is mediated by homologous
recombination between HERV- H elements. Molecular Cytogenetics 5:6. DOI: https://doi.org/10.1186/1755-
8166-5-6, PMID: 22260357
Ito J, Sugimoto R, Nakaoka H, Yamada S, Kimura T, Hayano T, Inoue I. 2017. Systematic identification and
characterization of regulatory elements derived from human endogenous retroviruses. PLOS Genetics
13:e1006883. DOI: https://doi.org/10.1371/journal.pgen.1006883, PMID: 28700586
Jason LA, Dorri JA. 2022. ME/CFS and post- exertional malaise among patients with long COVID. Neurology
International 15:1–11. DOI: https://doi.org/10.3390/neurolint15010001, PMID: 36648965
Kassambara A, Mundt F. 2025. Factoextra: extract and visualize the results of multivariate data analyses. 1.0.6.
CRAN_Status_Badge. https://rpkgs.datanovia.com/factoextra/index.html
Kassiotis G. 2023. The immunological conundrum of endogenous retroelements. Annual Review of Immunology
41:99–125. DOI: https://doi.org/10.1146/annurev-immunol-101721-033341, PMID: 36630597
Khadjinova AI, Wang X, Laine A, Ukadike K, Eckert M, Stevens A, Bengtsson AA, Lood C, Mustelin T. 2022.
Autoantibodies against the envelope proteins of endogenous retroviruses K102 and K108 in patients with
systemic lupus erythematosus correlate with active disease. Clinical and Experimental Rheumatology 40:1306–
1312. DOI: https://doi.org/10.55563/clinexprheumatol/2kg1d8, PMID: 34665695
Kim JH, Park K, Lee SB, Kang S, Park JS, Ahn CW, Nam JS. 2019. Relationship between natural killer cell activity
and glucose control in patients with type 2 diabetes and prediabetes. Journal of Diabetes Investigation
10:1223–1228. DOI: https://doi.org/10.1111/jdi.13002, PMID: 30618112
Kolde R, Lizee A, Metsalu T. 2018. Pheatmap: pretty heatmaps. b333453. Github. https://github.com/raivokolde/
pheatmap
Komaroff AL, Lipkin WI. 2023. ME/CFS and Long COVID share similar symptoms and biological abnormalities:
road map to the literature. Frontiers in Medicine 10:1187163. DOI: https://doi.org/10.3389/fmed.2023.
1187163

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 22 of 24
Kremer D, Schichel T, Förster M, Tzekova N, Bernard C, van der Valk P, van Horssen J, Hartung H- P, Perron H,
Küry P. 2013. Human endogenous retrovirus type W envelope protein inhibits oligodendroglial precursor cell
differentiation. Annals of Neurology 74:721–732. DOI: https://doi.org/10.1002/ana.23970, PMID: 23836485
Kuhn RM, Haussler D, Kent WJ. 2013. The UCSC genome browser and associated tools. Briefings in
Bioinformatics 14:144–161. DOI: https://doi.org/10.1093/bib/bbs038, PMID: 22908213
Kulski JK, Gaudieri S, Inoko H, Dawkins RL. 1999. Comparison between two human endogenous retrovirus
(HERV)- Rich regions within the major histocompatibility complex. Journal of Molecular Evolution 48:675–683.
DOI: https://doi.org/10.1007/PL00006511
Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W,
Funke R, Gage D, Harris K, Heaford A, Howland J, Kann L, Lehoczky J, LeVine R, McEwan P, McKernan K, etal.
2001. Initial sequencing and analysis of the human genome. Nature 409:860–921. DOI: https://doi.org/10.
1038/35057062
Langfelder P, Horvath S. 2008. WGCNA: an R package for weighted correlation network analysis. BMC
Bioinformatics 9:559. DOI: https://doi.org/10.1186/1471-2105-9-559, PMID: 19114008
Lawrence M, Huber W, Pagès H, Aboyoun P, Carlson M, Gentleman R, Morgan MT, Carey VJ. 2013. Software for
computing and annotating genomic ranges. PLOS Computational Biology 9:e1003118. DOI: https://doi.org/
10.1371/journal.pcbi.1003118, PMID: 23950696
Light AR, Bateman L, Jo D, Hughen RW, Vanhaitsma TA, White AT, Light KC. 2012. Gene expression alterations
at baseline and following moderate exercise in patients with Chronic Fatigue Syndrome and Fibromyalgia
Syndrome. Journal of Internal Medicine 271:64–81. DOI: https://doi.org/10.1111/j.1365-2796.2011.02405.x,
PMID: 21615807
Lima- Junior DS, Krishnamurthy SR, Bouladoux N, Collins N, Han S- J, Chen EY, Constantinides MG, Link VM,
Lim AI, Enamorado M, Cataisson C, Gil L, Rao I, Farley TK, Koroleva G, Attig J, Yuspa SH, Fischbach MA,
Kassiotis G, Belkaid Y. 2021. Endogenous retroviruses promote homeostatic and inflammatory responses to the
microbiota. Cell 184:3794–3811.. DOI: https://doi.org/10.1016/j.cell.2021.05.020, PMID: 34166614
Liu X, Hong T, Parameswaran S, Ernst K, Marazzi I, Weirauch MT, Fuxman Bass JI. 2020. Human Virus
Transcriptional Regulators. Cell 182:24–37. DOI: https://doi.org/10.1016/j.cell.2020.06.023, PMID: 32649876
Macchietto MG, Langlois RA, Shen SS. 2020. Virus- induced transposable element expression up- regulation in
human and mouse host cells. Life Science Alliance 3:e201900536. DOI: https://doi.org/10.26508/lsa.
201900536, PMID: 31964680
Mager DL, Hunter DG, Schertzer M, Freeman JD. 1999. Endogenous retroviruses provide the primary
polyadenylation signal for two new human genes (HHLA2 and HHLA3). Genomics 59:255–263. DOI: https://
doi.org/10.1006/geno.1999.5877
Magnus P, Gunnes N, Tveito K, Bakken IJ, Ghaderi S, Stoltenberg C, Hornig M, Lipkin WI, Trogstad L,
Håberg SE. 2015. Chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME) is associated with pandemic
influenza infection, but not with an adjuvanted pandemic influenza vaccine. Vaccine 33:6173–6177. DOI:
https://doi.org/10.1016/j.vaccine.2015.10.018
Mangeney M, de Parseval N, Thomas G, Heidmann T. 2001. The full- length envelope of an HERV- H human
endogenous retrovirus has immunosuppressive properties. The Journal of General Virology 82:2515–2518.
DOI: https://doi.org/10.1099/0022-1317-82-10-2515, PMID: 11562544
Marston JL, Greenig M, Singh M, Bendall ML, Duarte RRR, Feschotte C, Iñiguez LP, Nixon DF. 2021. SARS- CoV- 2
infection mediates differential expression of human endogenous retroviruses and long interspersed nuclear
elements. JCI Insight 6:e147170. DOI: https://doi.org/10.1172/jci.insight.147170, PMID: 34731091
Matsuzawa A, Lee J, Nakagawa S, Itoh J, Takahashi Ueda M, Mitsuhashi S, Kochi Y, Kaneko- Ishino T, Ishino F.
2021. HERV- Derived Ervpb1 Is conserved in simiiformes, exhibiting expression in hematopoietic cell lineages
including macrophages. International Journal of Molecular Sciences 22:4504. DOI: https://doi.org/10.3390/
ijms22094504, PMID: 33925887
Mchorney CA, Johne W, Anastasiae R. 1993. The MOS 36- item short- form health survey (SF- 36). Medical Care
31:247–263. DOI: https://doi.org/10.1097/00005650-199303000-00006
Morris G, Maes M, Murdjeva M, Puri BK. 2019. Do human endogenous retroviruses contribute to multiple
sclerosis, and if so, how? Molecular Neurobiology 56:2590–2605. DOI: https://doi.org/10.1007/s12035-018-
1255-x, PMID: 30047100
Natelson BH. 2019. Myalgic encephalomyelitis/chronic fatigue syndrome and fibromyalgia: Definitions,
similarities, and differences. Clinical Therapeutics 41:612–618. DOI: https://doi.org/10.1016/j.clinthera.2018.
12.016, PMID: 30795933
Nepotchatykh E, Caraus I, Elremaly W, Leveau C, Elbakry M, Godbout C, Rostami- Afshari B, Petre D, Khatami N,
Franco A, Moreau A. 2023. Circulating microRNA expression signatures accurately discriminate myalgic
encephalomyelitis from fibromyalgia and comorbid conditions. Scientific Reports 13:1896. DOI: https://doi.org/
10.1038/s41598-023-28955-9
Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA. 2015. Robust
enumeration of cell subsets from tissue expression profiles. Nature Methods 12:453–457. DOI: https://doi.org/
10.1038/nmeth.3337
Newman AM, Steen CB, Liu CL, Gentles AJ, Chaudhuri AA, Scherer F, Khodadoust MS, Esfahani MS, Luca BA,
Steiner D, Diehn M, Alizadeh AA. 2019. Determining cell type abundance and expression from bulk tissues
with digital cytometry. Nature Biotechnology 37:773–782. DOI: https://doi.org/10.1038/s41587-019-0114-2,
PMID: 31061481

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 23 of 24
Ovejero T, Sadones O, Sánchez- Fito T, Almenar- Pérez E, Espejo JA, Martín- Martínez E, Nathanson L, Oltra E.
2020. Activation of transposable elements in immune cells of fibromyalgia patients. International Journal of
Molecular Sciences 21:1366. DOI: https://doi.org/10.3390/ijms21041366
Ramsay L, Marchetto MC, Caron M, Chen SH, Busche S, Kwan T, Pastinen T, Gage FH, Bourque G. 2017.
Conserved expression of transposon- derived non- coding transcripts in primate stem cells. BMC Genomics
18:214. DOI: https://doi.org/10.1186/s12864-017-3568-y, PMID: 28245871
Rangel SC, da Silva MD, da Silva AL, de MB Dos Santos J, Neves LM, Pedrosa A, Rodrigues FM, Trettel CDS,
Furtado GE, de Barros MP, Bachi ALL, Romano CM, Nali LHDS. 2022. Human endogenous retroviruses and the
inflammatory response: A vicious circle associated with health and illness. Frontiers in Immunology 13:1057791.
DOI: https://doi.org/10.3389/fimmu.2022.1057791, PMID: 36518758
Rebollo R, Miceli- Royer K, Zhang Y, Farivar S, Gagnier L, Mager DL. 2012. Epigenetic interplay between mouse
endogenous retroviruses and host genes. Genome Biology 13:R89. DOI: https://doi.org/10.1186/gb-2012-13-
10-r89, PMID: 23034137
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK. 2015. limma powers differential expression
analyses for RNA- sequencing and microarray studies. Nucleic Acids Research 43:e47. DOI: https://doi.org/10.
1093/nar/gkv007, PMID: 25605792
Rivera J, González T. 2004. The Fibromyalgia Impact Questionnaire: a validated Spanish version to assess the
health status in women with fibromyalgia. Clinical and Experimental Rheumatology 22:554–560 PMID:
15485007.
Rodrigues LS, da Silva Nali LH, Leal COD, Sabino EC, Lacerda EM, Kingdon CC, Nacul L, Romano CM. 2019.
HERV- K and HERV- W transcriptional activity in myalgic encephalomyelitis/chronic fatigue syndrome. Auto-
Immunity Highlights 10:12. DOI: https://doi.org/10.1186/s13317-019-0122-8, PMID: 32257068
Sanchez- Valle A, Wang X, Potocki L, Xia Z, Kang S- HL, Carlin ME, Michel D, Williams P, Cabrera- Meza G,
Brundage EK, Eifert AL, Stankiewicz P, Cheung SW, Lalani SR. 2010. HERV- mediated genomic rearrangement of
EYA1 in an individual with branchio- oto- renal syndrome. American Journal of Medical Genetics. Part A
152A:2854–2860. DOI: https://doi.org/10.1002/ajmg.a.33686, PMID: 20979191
Schmidt N, Domingues P, Golebiowski F, Patzina C, Tatham MH, Hay RT, Hale BG. 2019. An influenza virus-
triggered SUMO switch orchestrates co- opted endogenous retroviruses to stimulate host antiviral immunity.
PNAS 116:17399–17408. DOI: https://doi.org/10.1073/pnas.1907031116, PMID: 31391303
Shuvarikov A, Campbell IM, Dittwald P, Neill NJ, Bialer MG, Moore C, Wheeler PG, Wallace SE, Hannibal MC,
Murray MF, Giovanni MA, Terespolsky D, Sodhi S, Cassina M, Viskochil D, Moghaddam B, Herman K,
Brown CW, Beck CR, Gambin A, etal. 2013. Recurrent HERV- H- mediated 3q13.2- q13.31 deletions cause a
syndrome of hypotonia and motor, language, and cognitive delays. Human Mutation 34:1415–1423. DOI:
https://doi.org/10.1002/humu.22384, PMID: 23878096
Smets EM, Garssen B, Bonke B, De Haes JC. 1995. The Multidimensional Fatigue Inventory (MFI) psychometric
qualities of an instrument to assess fatigue. Journal of Psychosomatic Research 39:315–325. DOI: https://doi.
org/10.1016/0022-3999(94)00125-o, PMID: 7636775
Soriano JB, Murthy S, Marshall JC, Relan P, Diaz JV, WHO Clinical Case Definition Working Group on Post-
COVID- 19 Condition. 2022. A clinical case definition of post- COVID- 19 condition by A Delphi consensus. The
Lancet. Infectious Diseases 22:e102–e107. DOI: https://doi.org/10.1016/S1473-3099(21)00703-9, PMID:
34951953
Thomas J, Perron H, Feschotte C. 2018. Variation in proviral content among human genomes mediated by LTR
recombination. Mobile DNA 9:36. DOI: https://doi.org/10.1186/s13100-018-0142-3, PMID: 30568734
Tschopp R, König RS, Rejmer P, Paris DH. 2023. Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS):
A preliminary survey among patients in Switzerland. Heliyon 9:e15595. DOI: https://doi.org/10.1016/j.heliyon.
2023.e15595, PMID: 37131449
Wang L, Kang N, Zhou J, Guo Y, Zhang X, Cui L, Ba D, He W. 2012. Downregulation of CD94/NKG2A inhibitory
receptor on decreased γδ T cells in patients with systemic lupus erythematosus. Scandinavian Journal of
Immunology 76:62–69. DOI: https://doi.org/10.1111/j.1365-3083.2012.02705.x, PMID: 22486170
Wessely S, Nimnuan C, Sharpe M. 1999. Functional somatic syndromes: one or many? The Lancet 354:936–939.
DOI: https://doi.org/10.1016/S0140-6736(98)08320-2
Wickham H. 2016. Ggplot2, Elegant Graphics for Data Analysis. R. Springer. DOI: https://doi.org/10.1007/978-3-
319-24277-4
Wolfe F, Smythe HA, Yunus MB, Bennett RM, Bombardier C, Goldenberg DL, Tugwell P, Campbell SM,
Abeles M, Clark P, Fam AG, Farber SJ, Fiechtner JJ, Michael Franklin C, Gatter RA, Hamaty D, Lessard J,
Lichtbroun AS, Masi AT, Mccain GA, etal. 1990. The american college of rheumatology 1990 criteria for the
classification of fibromyalgia. Arthritis & Rheumatism 33:160–172. DOI: https://doi.org/10.1002/art.
1780330203
Wolfe F, Clauw DJ, Fitzcharles M, Goldenberg DL, Katz RS, Mease P, Russell AS, Russell IJ, Winfield JB,
Yunus MB. 2010. The American college of rheumatology preliminary diagnostic criteria for fibromyalgia and
measurement of symptom severity. Arthritis Care & Research 62:600–610. DOI: https://doi.org/10.1002/acr.
20140
Wolfe F, Clauw DJ, Fitzcharles MA, Goldenberg DL, Häuser W, Katz RL, Mease PJ, Russell AS, Russell IJ,
Walitt B. 2016. Revisions to the 2010/2011 fibromyalgia diagnostic criteria. Seminars in Arthritis and
Rheumatism 46:319–329. DOI: https://doi.org/10.1016/j.semarthrit.2016.08.012

Research article Medicine | Microbiology and Infectious Disease
Giménez- Orenga etal. eLife 2025;14:RP104441. DOI: https://doi.org/10.7554/eLife.104441 24 of 24
Wu T, Hu E, Xu S, Chen M, Guo P, Dai Z, Feng T, Zhou L, Tang W, Zhan L, Fu X, Liu S, Bo X, Yu G. 2021.
clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation 2:100141. DOI: https://
doi.org/10.1016/j.xinn.2021.100141, PMID: 34557778
Yi JM, Kim HM, Kim HS. 2006. Human endogenous retrovirus HERV- H family in human tissues and cancer cells:
expression, identification, and phylogeny. Cancer Letters 231:228–239. DOI: https://doi.org/10.1016/j.canlet.
2005.02.001
Zhang Y, Li T, Preissl S, Amaral ML, Grinstein JD, Farah EN, Destici E, Qiu Y, Hu R, Lee AY, Chee S, Ma K, Ye Z,
Zhu Q, Huang H, Fang R, Yu L, Izpisua Belmonte JC, Wu J, Evans SM, etal. 2019. Transcriptionally active
HERV- H retrotransposons demarcate topologically associating domains in human pluripotent stem cells. Nature
Genetics 51:1380–1388. DOI: https://doi.org/10.1038/s41588-019-0479-7, PMID: 31427791
Zhao R, Chinai JM, Buhl S, Scandiuzzi L, Ray A, Jeon H, Ohaegbulam KC, Ghosh K, Zhao A, Scharff MD, Zang X.
2013. HHLA2 is a member of the B7 family and inhibits human CD4 and CD8 T- cell function. PNAS 110:9879–
9884. DOI: https://doi.org/10.1073/pnas.1303524110, PMID: 23716685
Zhao Y, Lin L, Xiao Z, Li M, Wu X, Li W, Li X, Zhao Q, Wu Y, Zhang H, Yin J, Zhang L, Cho CH, Shen J. 2018.
Protective role of γδ T cells in different pathogen infections and its potential clinical application. Journal of
Immunology Research 2018:5081634. DOI: https://doi.org/10.1155/2018/5081634, PMID: 30116753