"Autoimmunity-related risk variants in PTPN22 and CTLA4 are associated with ME/CFS with infectious onset"

Steiner et al (2020). Front Immunol. 11:578.

Paper presented by João Malato on 29/01/2021.

On the topic

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a complex multisystemic disease with a prevalence estimated between 0.2 and 0.3%. This is a disease with numerous research challenges including the lack of a disease biomarker. In theory, the discovery of a biomarker could improve accuracy of disease diagnosis and the understanding of the underlying pathological mechanisms. It could also help discerning differences between subgroups of patients, including those whose disease appeared to have been triggered by an infection.

Brief summary

The paper describes a genetic association study comparing two groups of ME/CFS patients according to their disease trigger (infection versus non-infection) with healthy controls. Five candidate single nucleotide polymorphisms (SNPs) were genotyped in peripheral blood mononuclear cells. These SNPs were located in genes previously associated with autoimmune diseases (PTPN22, CTLA4, TNF, and IRF5).

The analysis showed an association between the allelic and genotypic distributions of the PTPN22 and CTLA4 SNPs and the infection-triggered ME/CFS group. When comparing infection-triggered patients with patients without such trigger there was evidence for decreased B-cell frequency in the former group and higher CrP levels in the later.

In summary, the study suggested a genetic predisposition associated with autoimmunity in patients whose disease was triggered by infection.

Our comment

Our group discussion identified 5 issues with the paper. First, ME/CFS patients and the healthy controls were not well matched in terms of age and gender distributions. Second, the use of broadly defined Caucasian study participants did not ensure that patients and healthy controls came from the same genetic population. This is particularly problematic when the healthy controls were recruited from staff. Therefore, we could not rule out that the findings could be due to the existence of some cryptic population structure in the study participants. Third, it was not clearly stated which allele from each SNP was indeed considered the risk allele. This brought some confusion into our discussion given that there is a discrepancy between Tables 2 and 3 about the PTPN22 SNP (A>G versus G>A, respectively). Fourth, there is an overstatement in the interpretation of Figure 2A. It is stated that there was “trend toward lower C3 levels”. However, the Mann-Whitney test applied to these data did not reject the null hypothesis of equal medians across the groups. Fifth, when looking at the PTPN22 SNP results on infection-triggered ME/CFS (Figure 2B), it is understandable that the comparison across all genotypes was not possible because there was only one individual with the genotype AA. However, we should remark that the frequency of individuals with the GG genotype was three times the frequency of individuals with the genotypes GA and GG. This fact could potentially bias the interpretation of the results.

Tables 2 and 3

There is a discrepancy in the allelic information about rs2476601 (highlighted in red).

Closing remarks

As more studies on ME/CFS are performed, there is a growing evidence for the increased role of autoimmunity on disease aetiology, specifically, in patients whose disease was triggered by an infection. Therefore, studies like this one are important not only to identify the inherent immunological mechanisms of the disease, but also to try to distinguish between subgroups of patients that ultimately can refer to different clinical entities.

More paper discussions can be found in the Archive.