Evaluating the Causal Role of Genetically Inferred Immune Cells and Inflammatory Cytokines on ME/CFS, 2025, Duan et al

John Mac

Senior Member (Voting Rights)
https://www.mdpi.com/2227-9059/13/5/1200

Abstract
Background/Objectives: Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a multifaceted and diverse disorder with an ambiguous etiology. Recent evidence indicates that immune system impairment and inflammatory mechanisms are pivotal to the initiation and advancement of ME/CFS. Nonetheless, the causal relationships among these factors remain inadequately comprehended.

Methods: This study investigated the causative contributions of immunological dysfunction and inflammatory variables in ME/CFS utilizing genome-wide association study (GWAS) data. We employed Mendelian randomization (MR) to investigate associations between 91 inflammatory cytokines, 731 immune cell characteristics, and the risk of ME/CFS. Summary statistics for immune cell traits and inflammatory cytokines were sourced from European GWAS cohorts (n = 3757 and n = 14,824, respectively), while ME/CFS data were obtained from the UK Biobank (n = 462,933, including 2076 cases). We predominantly employed the inverse variance weighted (IVW) approach, complemented by MR-Egger, weighted median, BWMR, and MR-RAPS tests to guarantee robust and precise outcomes.

Results: The study revealed significant causal links between various inflammatory factors, immune cell characteristics, and the risk of ME/CFS. Increased CXCL5 and CCL20 levels were significantly linked to a higher risk of ME/CFS, while elevated TNF levels were inversely related to ME/CFS risk. Furthermore, 13 immune cell characteristics were identified as having substantial causal associations with the likelihood of ME/CFS. These data are supportive of the causality that immune system dysfunction and inflammatory variables play a pivotal role in the development of ME/CFS.

Conclusions: This study provides new insights into the causal role of immune system dysfunction in the development of ME/CFS, contributing to a deeper understanding of its underlying mechanisms. These results offer a foundation for identifying diagnostic biomarkers and developing targeted therapeutic strategies. Future research should validate these findings using multi-center cohort studies and further investigate the mechanisms behind key factors to enable the development of personalized treatment approaches.
 
Is «causal links» appropriate here?
Probably not.
For a couple reasons but the one that stands out to me is “Horizontal Pleiotropy”. Ie. a single SNP may influence many different things and we might only know about some. So assuming the cytokines play a causal role in ME/CFS according to some SNP variations linked to cytokines and ME/CFS is probably not fully supported. Since those SNPs might be doing something else as well and that could be what links it to ME/CFS.
 
Is «causal links» appropriate here?
The structure of a mendelian randomization study enables you to claim causation, but only if certain criteria are adequately met. There was some discussion of MR studies in this thread: https://www.s4me.info/threads/evidence-for-a-causal-association-between-human-cmv-infection-and-chronic-back-pain-a-one‐sample-mendelian-randomization-study-2025-naeini-et-al.43598/

I'm no expert in MR, but from my brief skimming, this paper does seem to do a good job of addressing those criteria [edit: though it does have the limitation of being a two sample rather than one-sample study]. If there does happen to be a genetics expert on the forum I'd definitely appreciate hearing their assessment.
 
This looks very similar to another recent study that was on PVFS:

Causal relationship between immune cells and post-viral fatigue syndrome: a Mendelian randomization study, 2025, Wang et al

They both used the same 731 immune traits from a separate GWAS, although this thread's study also looked at cytokine traits. As with the other one, I don't see any indication that they adjusted for multiple tests.

Good chance to compare the two studies to see if any immune cell traits were significant in both. I think the only exact match is "CD8+ Natural Killer T %lymphocyte" in Duan et al with "CD8br NKT %lymphocyte" in Wang et al. Though I can't figure out what "br" means. Though I assume they're the same since Wang et al used "br" on many traits that match in Duan et al except for they used "+". It looks like in both studies, a higher proportion of NKT cells is a risk factor.

These two are close to matching, but one is on CD45RA- and one is on CD45RA+:
Duan: CD3 on CD28+ CD45RA- CD8+ T cell
Wang: CD3 on CD28+ CD45RA+ CD8br

Significant in Duan et al (ME/CFS)
CD8+ Natural Killer T %T cell
TCRgd T cell %T cell
CD80 on CD62L+ myeloid Dendritic Cell
Monocyte Absolute Count
CD80 on myeloid Dendritic Cell
CD19 on IgD+ CD38- B cell
CD3 on naive CD8+ T cell
CD3 on CD28+ CD45RA- CD8+ T cell
Activated & secreting CD4 regulatory T cell %CD4 regulatory T cell
TCRgd T cell %lymphocyte
CD8+ Natural Killer T %lymphocyte
CD19 on memory B cell
FSC-A on CD8+ T cell

Significant in Wang et al (PVFS)
CD20 on IgD+ CD38-
CD20 on memory B cell
CD20 on unsw mem
CD24 on CD24+ CD27+
CD25 on IgD+ CD24+
CD25 on memory B cell
CD27 on IgD- CD38br
CD38 on IgD+
IgD- CD24- AC
Plasmacytoid DC AC
HVEM on EM CD8br
CD11b on CD33dim HLA DR-
CD33- HLA DR+ AC
CD33dim HLA DR+ CD11b+ AC
CD3- lymphocyte AC
CD4/CD8br
CD8br %T cell
CD8br NKT %lymphocyte
CD8dim %leukocyte
CD8dim %T cell
HLA DR+ CD8br %lymphocyte
SSC-A on granulocyte
T cell AC
CD127 on CD28+ CD45RA- CD8br
CD25hi CD45RA- CD4 not Treg %CD4+
CD28 on resting Treg
CD3 on CD28+ CD45RA+ CD8br
CD39+ CD8br AC
 
Interesting that elevated TNF levels were inversely related to ME /CFS risk.

The effect size, however, were small and as @forestglip mentioned it doesn't look like they correct for multiple comparisons. So the significant results the paper highlights may simply be false positives.

I wonder if we could do our own MR using the DecodeME summary data.
 
The source data about immune cells can be downloaded here:

But it seems like quite a lot of work to download and filter all of these (suspect the authors had automated scripts for this).
 
It looks like in both studies, a higher proportion of NKT cells is a risk factor.

Studies from the first few pages of results on Google Scholar (and one found as a citation) related to NKT cells in people with ME/CFS:

- Abnormal T-Cell Activation And Cytotoxic T-Cell Frequency Discriminates Symptom Severity In Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, 2025, Preprint: MedRxiv | S4ME | Article
Firstly, we compared the frequencies of [...] CD3+CD56+ NKT-like cells) [...] in ex vivo PBMCs from people with ME-MM [mild/moderate ME/CFS] or ME-SA [severe ME/CFS], and found there were no differences between the two clinical groups

- Association of T and NK cell phenotype with the diagnosis of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), 2018, Frontiers in Immunology | S4ME | Article
ME/CFS patients showed significantly [...] higher NKT-like cells (CD3+CD16+/−CD56+) than the healthy individuals.

- Regulatory T, natural killer T and γδ T cells in multiple sclerosis and chronic fatigue syndrome/myalgic encephalomyelitis: a comparison, 2016, Asian Pacific Journal of Allergy and Immunology | Article
The percentage of CD4+iNKT cells was significantly elevated in the CFS/ME group compared with healthy controls (3B).

The percentage of double negative iNKT cells (CD4-CD8- ) was significantly decreased in the CFS/ME compared with the healthy controls group (4C).

The percentage of CD56+CD16+ iNKT cells was significantly decreased in the CFS/ME group compared with the healthy controls group (3D).

- Examination of Innate and Adaptive Immune Cells in Chronic Fatigue Syndrome/Myalgic Encephalomyelitis Patients with Varying Degrees of Symptom Severity, 2015, Thesis | Article
In summary, NKT cell fatty acid oxidation is increased in ME/CFS samples compared to healthy controls, but CD36, the fatty acid transporter on the cell surface, is not contributing to this increase in ME/CFS cells – in fact, CD36 had lower levels in activated ME/CFS cell samples.

- Longitudinal analysis of immune abnormalities in varying severities of Chronic Fatigue Syndrome/Myalgic Encephalomyelitis patients, 2015, Journal of Translational Medicine | Article
Over time, iNKT CD62L expression significantly increased in moderate CFS/ME patients [...]. [...] CD56−CD16− iNKT phenotypes, [...] were significantly increased in severe CFS/ME patients at 6 months.

- Analysis of the Relationship between Immune Dysfunction and Symptom Severity in Patients with Chronic Fatigue Syndrome/Myalgic Encephalomyelitis (CFS/ME), 2014, Journal of Clinical & Cellular Immunology | Article
Significant increases in [...], iNKT [invariant natural killer T cell] phenotypes, memory and naive B cells were also shown in CFS/ME participants.

Severe CFS/ME patients demonstrated increased [...], total iNKT, iNKT cell and NK phenotypes compared to moderate CFS/ME patients.
 
Last edited:
Back
Top Bottom