Genome-Epigenome Interactions Associated with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, Herrera et al, 2018

[Indigophoton]

This is a preprint - made available by the authors on bioRxiv before publication while the paper is still in peer review.

Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is an example of a complex disease of unknown etiology. Multiple studies point to disruptions in immune functioning in ME/CFS patients as well as with specific genetic polymorphisms and alterations of the DNA methylome in lymphocytes. However, the association between DNA methylation and genetic background in relation to the ME/CFS is currently unknown. In this study we explored this association by characterizing the genomic (~4.3 million SNPs) and epigenomic (~480 thousand CpG loci) variability between populations of ME/CFS patients and healthy controls. We found significant associations of methylation states in T-lymphocytes at several CpG loci and regions with ME/CFS phenotype. These methylation anomalies are in close proximity to genes involved with immune function and cellular metabolism. Finally, we found significant correlations of genotypes with methylation phenotypes associated with ME/CFS. The findings from this study highlight the role of epigenetic and genetic interactions in complex diseases, and suggest several genetic and epigenetic elements potentially involved in the mechanisms of disease in ME/CFS.

T-cell lymphocytes appear to be a primary cell type underlying immune and neuroendocrine abnormalities observed in ME/CFS patients. Functional impairment in T-cell glucocorticoid receptor and increased dexamethasone sensitivity are characteristic of some
ME/CFS patients (14,20). Furthermore, genetic polymorphisms within non-coding regions of T-cell receptor loci (15), as well as differential methylation in CD4+ 51 T helper lymphocyte cells (Brenu et al., 2014), have been associated with the disease. The possible interactions between genomic and T-cell epigenomic variation in ME/CFS remain unknown.

In this study, we aimed to explore the association between DNA methylation profiles of T-cells and single nucleotide polymorphisms (SNPs) in ME/CFS patients. We quantified lymphocyte proportions and isolated CD3+ T-cells (including both CD4+ T helper cells and CD8+ T killer cells) via fluorescence activated cell sorting. We characterized the variation in genomic(~4.3 million SNPs) and epigenomic (~480 thousand CpG loci) variability among ME/CFS patients and healthy controls.

Using this approach, we: 1) tested the association of genome-wide SNP genotypes with ME/CFS disease status; 2) tested the association of differentially methylated CpG loci and regions in CD3+ T-cells with ME/CFS disease status; 3) performed a methylation quantitative trait analysis to investigate the possible interactions between genetic background and methylation phenotypes of CD3+ T-cells associated with ME/CFS disease status.

Conclusions
We identified over one hundred differentially methylated CpG loci associated with ME/CFS in T lymphocytes. Approximately half of these were clustered in differentially methylated regions of 500bp in size or less. Our data and analyses suggest that there is an indirect role of genotype influencing DNA methylation patterns associated with ME/CFS. We found no substantial large-effect direct associations of specific genotypes with ME/CFS disease phenotype. Larger scale genome wide association studies are necessary to test for potential small-effect associations between genotype and ME/CFS phenotype.

All of the methylation values at differentially methylated loci in T lymphocytes had significant correlations with specific genotypes at neighboring SNPs (within a window of 1 Mbp), indicating that particular genetic backgrounds may influence methylation levels differently in ME/CFS patients than in controls. The genomic elements associated with genetic and epigenetic variants characteristic of ME/CFS patients in this study constitute targets for future research. Understanding the molecular mechanisms of genetic-epigenetic interactions of these targets will be key to develop new treatments for ME/CFS, and can serve as a model to understand the molecular basis of related complex diseases.

 

[Trish]

The preprint link is here:
https://www.biorxiv.org/content/early/2017/12/22/237958

Abstract
Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is an example of a complex disease of unknown etiology.

Multiple studies point to disruptions in immune functioning in ME/CFS patients as well as with specific genetic polymorphisms and alterations of the DNA methylome in lymphocytes. However, the association between DNA methylation and genetic background in relation to the ME/CFS is currently unknown.

In this study we explored this association by characterizing the genomic (~4.3 million SNPs) and epigenomic (~480 thousand CpG loci) variability between populations of ME/CFS patients and healthy controls.

We found significant associations of methylation states in T-lymphocytes at several CpG loci and regions with ME/CFS phenotype. These methylation anomalies are in close proximity to genes involved with immune function and cellular metabolism.

Finally, we found significant correlations of genotypes with methylation phenotypes associated with ME/CFS.

The findings from this study highlight the role of epigenetic and genetic interactions in complex diseases, and suggest several genetic and epigenetic elements potentially involved in the mechanisms of disease in ME/CFS.

 

[Indigophoton]

Thanks @Trish, forgot to include the link!

 

[forestglip]

Final paper published December 2018:

---

Genome-epigenome interactions associated with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome

Santiago Herrera, Wilfred C. de Vega, David Ashbrook, Suzanne D. Vernon, Patrick O. McGowan

[Line breaks added]

Abstract
Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a complex disease of unknown etiology. Multiple studies point to disruptions in immune functioning in ME/CFS patients as well as specific genetic polymorphisms and alterations of the DNA methylome in lymphocytes. However, potential interactions between DNA methylation and genetic background in relation to ME/CFS have not been examined.

In this study we explored this association by characterizing the epigenetic (~480 thousand CpG loci) and genetic (~4.3 million SNPs) variation between cohorts of ME/CFS patients and healthy controls.

We found significant associations of DNA methylation states in T-lymphocytes at several CpG loci and regions with ME/CFS phenotype. These methylation anomalies are in close proximity to genes involved with immune function and cellular metabolism.

Finally, we found significant correlations of genotypes with methylation modifications associated with ME/CFS. The findings from this study highlight the role of epigenetic and genetic interactions in complex diseases, and suggest several genetic and epigenetic elements potentially involved in the mechanisms of disease in ME/CFS.

Web | PDF | Epigenetics | Open Access

 

[forestglip]

I'm interested in this study because the following review found that, out of the six GWAS/TGAS that had been performed prior to 2020, only this one performed all the recommended quality control checks for a GWAS:

Review of the Quality Control Checks Performed by Current Genome-Wide and Targeted-Genome Association Studies on ME/CFS, 2020, Sepulveda et al [Article] [S4ME]

---

Excluding overlappers

In addition, we excluded individuals with health-related quality of life RAND-36 measurements that overlapped between cases and controls. This exclusion of intermediate illness phenotypes was aimed at increasing the power to detect possible associations between disease status and (epi)genotypes by decreasing the heterogeneity in phenotype symptomatology within each group.

An interesting technique I hadn't seen before. They excluded cases and controls that overlapped on subjective symptom scores to make sure the two groups are very different.

---

Initial GWAS

None of the more than 2 million variable SNP loci targeted in this study with the Human Omni 5–4 Array (Illumina Inc.) had a significant association (α = 0.05) with ME/CFS after p-value corrections with Bonferroni, Holm, Benjamini and Hochberg, or permutation methods when data from both sexes were analysed together (healthy control, n = 48 vs. ME/CFS, n = 61).

No significant SNPs in the initial testing.

This contrasts with three older studies that found dozens to hundreds of significant SNPs:

The earliest study by Smith et al. evaluated 116,204 SNPs (n = 40 CFS, n = 40 non-ME/CFS) using the Affymetrix GeneChip Mapping 100K array, and found 65 SNPs associated with ME/CFS (p < 0.001).

Rajeevan et al. used the Affymetrix Immune and Inflammation Chip to focus on ~11,000 SNPs located in genes involved in immune and inflammation pathways (n = 121 ME/CFS, n = 50 non-ME/CFS). Of these, 32 were associated with ME/CFS (p < 0.05).

Most recently, Schlauch et al. evaluated 906,600 SNPs with the Affymetrix Genome-Wide SNP Array 6.0 (n = 42 ME/CFS, n = 38 non-ME/CFS) and found 442 SNPs that were associated with ME/CFS (P < 3.3 × 10−5).

---

Females only: rs11712777

They followed up by only testing the females, and this time they found one significant SNP with a p-value of .0374 after multiple test correction:

Because of the known increased prevalence of ME/CFS in females, we performed independent analyses of data from females only (healthy control, n = 27 vs. ME/CFS, n = 34). These analyses revealed a significant association (χ2 genotypic test, permutation-corrected p-value = 0.0374, OR = 0.1845, 1/OR = 5.42) of one SNP (rs11712777, chr3:42347678) with the ME/CFS disease phenotype.

This specific SNP wasn't tested in the three other studies. One SNP tested in Schlauch et al (though they don't say it was significant) is in linkage disequilibrium (LD, meaning in an individual that has one of these SNPs, they are likely to have the other) with the above SNP:

The SNP that we found in significant association with ME/CFS in females, rs11712777, was not included in any of these datasets. One SNP in the Schlauch et al. data, rs1468604, is in linkage disequilibrium (LD) with rs11712777 (r2 = 0.8716; European population).

The Genotype-Tissue-Expression database shows that this SNP is significantly associated with the expression of the gene CCK (the amount of the CCK gene that the body creates).

The Genotype-Tissue-Expression database (GTEx) indicates that rs11712777, and the genes in LD with it, form an expression quantitative trail loci (eQTL) altering the expression of the CCK (cholecystokinin peptide hormone) gene.

---

CCK

CCK has a number of active forms, expressed in a variety of tissues, including the intestine and blood, and plays a role in appetite, body weight and immune system function.

Rats without the gene CCK experience less sickness behavior when administed bacterial endotoxins. (Source)

A rat knockout (KO) of the cholecystokinin B receptor (CCKBR) shows attenuated sickness behaviour. This sickness behaviour in rats has remarkable similarity to some of the symptoms of ME/CFS, including fatigue, malaise, hyperalgesia, sleepiness, anhedonia, weight loss and diminished activity.

CCK is also co-localized with sleep-promoting preoptic neurons in the hypothalamus, which may relate to fatigue and unrefreshing sleep symptoms in ME/CFS.

I think this means the CCK gene is expressed in these neurons that are involved in sleep.

Finally, recent evidence suggests that CCK has a role regulating the differentiation of CD4+ T-cells, and that CCK-expressing neurons are a critical cellular component of the hypothalamic–pituitary–adrenal axis.

---

Related SNPs
A SNP in linkage disequilibrium with the significant SNP:

In addition to rs11712777, SNP rs17223780 (R2 = 0.8799) binds DNase in CD14+ monocytes (http://www.regulomedb.org/snp/chr3/42363368), indicating a possible regulatory role in the immune system.

A nearby SNP:

Another SNP in the vicinity of rs11712777 (D’ = 0.7211, R2 = 0.0126), rs33449 (chr3:42400801), is associated with increased daytime resting duration (http://www.ebi.ac.uk/gwas/search?query=3:42347678-42372207;).

 

[forestglip]

Epigenetics/DNA methylation

They continued by testing methylation differences between ME/CFS and controls. DNA methylation is the biological process where methyl group molecules are attached to the DNA at specific locations, which affects the expression of genes, but the DNA itself doesn't change.

They found 133 significantly differentially methylated loci, meaning in ME/CFS at these specific DNA locations, methyl groups were attached more or less frequently compared to healthy controls, likely altering the gene activity associated with these locations. (As far as I can tell, these were significant with permutation analysis but without multiple test correction, though I may be missing something.)

Of the 467,971 CpG loci targeted with the Human Methylation 450K Array (Illumina Inc.), 141 had significant associations with the ME/CFS phenotype (raw p-value < 0.05) and a mean percentage methylation difference between cases and controls greater than 5% when data from both sexes were analysed together ()). None of these differentially methylated loci were significant after FDR corrections, however 133 had significant empirical p-values < 0.05 calculated through permutation analyses (these are referred to as differentially methylated probes – DMPs; Supplementary Table S1).

---

Comparison to other studies

Interestingly, this group's previous study on ME/CFS found tens of thousands of differentially methylated loci, versus 133 here. This may be because the other study used PBMC's in general, while this study only used T cells:

These results are in contrast with previous findings by our group, which revealed thousands to tens of thousands of differentially methylated CpG loci associated with ME/CFS in PBMCs, using the same 450K array [Citation11,Citation23]. It is possible that due to the relatively small methylation differences observed in DMPs (less than 15%, see Supplementary Table S1) in these studies may not be directly comparable when examined in the context of greater differences in ME/CFS symptom profiles [Citation40]. Alternatively, the differences between the targeted cell populations (i.e. PBMCs vs. isolated T-cells) may have contributed to the differences in the number of differentially methylated CpGs. The number of cell types within PBMCs may broaden the spectrum of epigenetic marks and thus increase the number of possible associations with the ME/CFS disease phenotype.

Supporting this idea about PBMC vs T cells, another study that only looked at CD4+ T cells found a similar number (120) of loci to this T cell study (133):

Consistent with this idea, Brenu et al. [Citation24] found 120 differentially methylated CpGs associated with ME/CFS in CD4+ T-cells (p < 0.001) using the 450K array (n = 25 ME/CFS, n = 18 non-ME/CFS). This number of differentially methylated CpGs is similar to the 133 DMPs we found in this study, which targeted a broader T-cell population (including CD4+ and CD4- T-cells). However, the only overlap between our study and the study from Brenu et al. [Citation24] corresponded to the HLA-DQB1 (major histocompatibility complex, class II, DQ beta 1) gene. HLA-DQB1 encodes a protein that is part of the DQ heterodimer, a cell surface receptor that is essential in immune signalling.

These two studies had differentially methylated loci in the HLA-DQB1 gene in common.

Recent studies focusing on CD4+ T-cells of patients affected by immune disorders such as rheumatoid arthritis [Citation41] and multiple sclerosis [Citation42] have found differential methylation in HLA-DQB1. This result is thus consistent with a potential immune dysregulation in ME/CFS.

In comparison to their own previous study on PBMCs, there were differentially methylated loci common to 31 genes:

We found 31 genes associated with DMPs in T-cells that were common to this study and a previous study by our group [Citation11]. These genes, which include PAX8 (paired box 8), and ATP4B (ATPase H+/K+ transporting beta subunit) (Supplementary Table S1), are involved in the regulation of cellular processes and cell signaling.

---

Genome - Epigenome Associations
They wanted to see if mutations in the DNA were associated with the significant methylation sites above, and found that there were significant associations of SNPs with all of the methylation loci:

To identify associations between SNP genotypes and DNA methylation levels at significantly differentially methylated CpG loci, we performed methylation quantitative trait loci (mQTLs) analysis using linear additive regression models. All the DMPs identified according to empirical p-values had significant associations (FDR corrected p-values < 0.05) with SNP genotypes (independent of disease phenotype). In total there were 13,060 significant cis-mQTLs (Supplementary Table S5). shows the strongest SNP-DMPs cis-mQTLs associations (according to correlation coefficient R2) in each of the 7 DMP-containing DMRs that were common in analyses of methylation data from both sexes and from females only.

The two genes that had SNPs associated with the most differentially methylated loci between ME/CFS and controls were SPATC1L and DUSP22.

SPATC1L (spermatogenesis and centriole associated 1 like) and DUSP22 (dual specificity phosphatase 22) were the two genes containing cis-mQTLs with the largest differentially methylated regions: 11 DMPs (7 hypermethylated probes in 5' UTR and 4 hypomethylated probes in 3' UTR region) in SPATC1L and 10 hypermethylated probes in the 5' UTR of DUSP22 (Supplementary Table S3 and ). While the exact function of SPATC1L is not well understood, it has been previously associated with xenobiotic response and differential methylation in the promoter of this gene is characteristic of certain ethnic groups in human populations [Citation46]. DUSP22 hypermethylation has also been observed in the 5' UTR region in T-cells of rheumatoid arthritis patients [Citation47]. In T-cells, DUSP22 is known to inhibit proliferation and autoimmunity through inactivating Lck and preventing the activation of the T-cell receptor [Citation48]. However, it remains to be confirmed how hypermethylation in the 5' UTR region affects the overall activity of DUSP22 in T-cells of ME/CFS patients.

So the overall idea is that mutations in these genes significantly affect how much methylation happens nearby (with methylation affecting how much of a gene is made). And methylation at these specific locations happens to be significantly associated with having ME/CFS in this study.

The mutations in these genes aren't directly associated with having ME/CFS though, so the mutations might not be the main thing causing the differential methylation between groups.

 

[forestglip]

So the only significant finding in terms of genetic mutations was a SNP which affects the expression of CCK (aka cholecystokinin). The database they used seems to say that this SNP decreases expression of CCK in cultured fibroblasts but increases expression in colon cells. I couldn't figure out from the paper which allele (C or T) was the risk allele for ME/CFS.

This bit was very intriguing in the context of ME/CFS:

Rats without the gene CCK experience less sickness behavior when administed bacterial endotoxins. (Source)

Could it be that increased CCK plays a part in ME/CFS symptoms?

I tried looking through published proteomics papers to see which ones tested this protein. I might make a thread for compiling such studies for easy reference, but for now here are the ones I checked:

Spoiler: Proteomics studies with supplementary data

- Deep phenotyping of post-infectious myalgic encephalomyelitis/chronic fatigue syndrome
- Supplementary Table 17: https://www.nature.com/articles/s41467-024-45107-3#Sec77

- Charting the Circulating Proteome in ME/CFS: Cross System Profiling and Mechanistic insights
- Supplementary Tables 2 and 9: https://www.medrxiv.org/content/10.1101/2025.05.28.25328245v1.supplementary-material

- In-Depth Analysis of the Plasma Proteome in ME/CFS Exposes Disrupted Ephrin-Eph and Immune System Signaling
- Supplementary Table 1: https://www.mdpi.com/2227-7382/9/1/6

- Proteomics and cytokine analyses distinguish myalgic encephalomyelitis/chronic fatigue syndrome cases from controls
- Additional File 2: https://link.springer.com/article/10.1186/s12967-023-04179-3#Sec20
- UniProt Ids

- Plasma proteomic profiling suggests an association between antigen driven clonal B cell expansion and ME/CFS
- Supplementary Tables 1-4: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0236148#sec016

- A SWATH-MS analysis of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome peripheral blood mononuclear cell proteomes reveals mitochondrial dysfunction
- Tables in the paper and Table S4: https://link.springer.com/article/10.1186/s12967-020-02533-3#Sec17

- Data-independent LC-MS/MS analysis of ME/CFS plasma reveals a dysregulated coagulation system, endothelial dysfunction, downregulation of complement machinery
- Supplementary Material 1: https://link.springer.com/article/10.1186/s12933-024-02315-x#Sec17
- UniProt IDs

- Bioenergetic and Proteomic Profiling of Immune Cells in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome Patients: An Exploratory Study
- Table S6: https://www.mdpi.com/2218-273X/11/7/961?trk=public_post_comment-text

- Replicated blood-based biomarkers for myalgic encephalomyelitis not explicable by inactivity
- Dataset EV6 and others: https://www.embopress.org/doi/full/10.1038/s44321-025-00258-8#supplementary-materials

- Immune cell proteomes of Long COVID patients have functional changes similar to those in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome
- Supplementary file: https://www.researchsquare.com/article/rs-3335919/v1
- Only significant proteins shown

- Dysregulation of extracellular vesicle protein cargo in female myalgic encephalomyelitis/chronic fatigue syndrome cases and sedentary controls in response to maximal exercise
- Supplementary files 2 and 3: https://isevjournals.onlinelibrary.wiley.com/doi/full/10.1002/jev2.12403#support-information-section

Of these, I could only find that CCK was tested in two:

Year	Study Title	Study DOI	Fold Change	Log FC	P-value	Adjusted P-value	Notes
2021	In-Depth Analysis of the Plasma Proteome in ME/CFS Exposes Disrupted Ephrin-Eph and Immune System Signaling​	https://doi.org/10.3390/proteomes9010006	1.161	0.216	0.142	0.745
2025	Charting the Circulating Proteome in ME/CFS: Cross System Profiling and Mechanistic insights	https://doi.org/10.1101/2025.05.28.25328245	1.046	0.065	0.174	0.432	Specifically in “Metabotype 3”, the unadjusted p-value is 0.003

So no really big differences, but it was (non-significantly) increased in ME/CFS in both studies. (Low unadjusted p-value in one of the machine-learning based subgroups in the second study.)

Information about CCK from Wikipedia (slightly edited for clarity):

Digestion

peptide hormone of the gastrointestinal system responsible for stimulating the digestion of fat and protein.

synthesized and secreted by enteroendocrine cells in the duodenum, the first segment of the small intestine.

Its presence causes the release of pancreatic juice from the pancreas and bile from the gallbladder.

It is released rapidly into the circulation in response to a meal. The greatest stimulator of CCK release is the presence of fatty acids and/or certain amino acids in the chyme entering the duodenum.

inhibits gastric emptying

stimulates the contraction of the gall bladder and the relaxation of the sphincter of Oddi (Glisson's sphincter), resulting in the delivery of bile into the duodenal part of the small intestine.

mediates satiety by acting on the CCK receptors distributed widely throughout the central nervous system

Nervous system

It is the most abundant neuropeptide in the central nervous system.

In addition to its role in digestion, CCK is involved in regulating various behavioral phenomena, including satiety, appetite, anxiety, thermoregulation, sexual behavior, memory, and the response to drugs of abuse, particularly within the cortex and limbic regions of the brain.

CCK also has stimulatory effects on the vagus nerve.

CCK is found extensively throughout the central nervous system, with high concentrations found in the limbic system.

In both humans and rodents, studies clearly indicate that elevated CCK levels causes increased anxiety. Reliably causes anxiety and panic attacks (panicogenic effect) when administered to humans and is commonly used in scientific research for this purpose of in order to test new anxiolytic drugs.

Several studies have implicated CCK as a cause of visual hallucinations in Parkinson's disease.

shown to interact with orexin neurons, which control appetite and wakefulness (sleep).

indirect effects on sleep regulation.

CCK in the body cannot cross the blood–brain barrier, but certain parts of the hypothalamus and brainstem are not protected by the barrier.

Edit: Removed bolding

 

[ME/CFS Science Blog]

Could it be that increased CCK plays a part in ME/CFS symptoms?

The link to sickness behaviour is interesting, thanks for highlighting this!

If I understand correctly this was only significant in the female-only comparison which had only 34 ME/CFS patients and 27 controls.

These analyses revealed a significant association (χ2 genotypic test, permutation-correctedp-value = 0.0374, OR = 0.1845, 1/OR = 5.42) of oneSNP (rs11712777, chr3:42347678) with the ME/CFSdisease phenotype

I think it means that most common allele (C, which is present in 62% of the population) occurs even more often in ME/CFS patients.

It would be good to know how many of the patients and controls had the C allele (Couldn't find this in the supplementary data). Because of the small sample size and subgroup analysis there's a risk that it was simply the controls who had a low prevalence of this allele (rather than ME/CFS patients having more of it).

If the prevalence in controls was 62% and the odds ratio was 5.42, then the prevalence in ME/CFS patients would have been around 90% (if I hadn't made a calculation error).

 

[forestglip]

I think it means that most common allele (C, which is present in 62% of the population) occurs even more often in ME/CFS patients.

Ok, I wasn't positive if the OR is always based on the minor allele frequency, but that's probably right.

If the prevalence in controls was 62% and the odds ratio was 5.42, then the prevalence in ME/CFS patients would have been around 90% (if I hadn't made a calculation error).

I got the same.

Edit: So I think that means the gene expression database is saying that if ME/CFS has the T allele less often, expression of CCK would be higher in cultured fibroblasts and lower in colon cells in ME/CFS. Interesting that it has opposite effects in different cells.

 

[hotblack]

Interesting that it has opposite effects in different cells.

It also sounds as if different receptors are involved according to this Wikipedia page on cholecystokinin antagonists

The CCKA receptor is mainly expressed in the small intestine, and is involved in the regulation of enzyme secretion by the pancreas, secretion of gastric acid in the stomach, intestinal motility and signaling of satiety (fullness). The CCKB receptor is expressed mainly in the central nervous system, and has functions relating to anxiety and the perception of pain.

Which sites this article on cholecystokinin (only a summary of each section is available but it looks interesting)