A MAP1B Variant Tracks With Progressive MS in Women

Finding genes that shape how severe multiple sclerosis (MS) becomes has proved much harder than finding genes that determine who develops it. Two recent genome-wide studies of MS severity returned only a few variants clearing the significance threshold, and some of those failed to hold up in the studies that followed. Djuric and colleagues took a targeted route instead of scanning the genome again. They began from their own earlier work, a sequencing panel of 138 ferroptosis-related genes measured in immune cells from patients at the two extremes of severity, mild relapsing-remitting and severe secondary progressive disease. From the three genes most strongly differentially expressed in that panel, CDKN1A, EGLN2 and MAP1B, they selected variants likely to regulate expression, then added two variants already reported by the severity genome-wide studies. The study covers 845 patients from Serbia, 604 relapsing-remitting and 241 progressive, and every one of them was genotyped for HLA-DRB1*15:01 so that its effect could be adjusted for throughout.

How the Variants Were Chosen
The selection step deserves attention because it is fully documented rather than asserted. Most variants capable of regulating a gene sit in non-coding introns or flanking regions, which make up roughly 80% of the genome, so choosing among them is the real work. Five criteria were combined: a minor allele frequency above 0.1 in the European non-Finnish population; status as an expression quantitative trait locus in the tissues that matter here, meaning blood, peripheral blood mononuclear cells, lymphocytes and brain; a RegulomeDB score no lower than 1f, marking a high probability of affecting protein binding; existing trait associations in the GWAS Catalogue; and linkage and regulatory motif information from HaploReg. The paper tabulates each criterion for each of the seven variants, which means another group could apply the same filter and arrive at the same shortlist.

The New Finding: a MAP1B Variant in Women
MAP1B rs62363242 was the one variant showing a genotype distribution that differed between relapsing and progressive patients, and it did so in females only. A-allele-containing genotypes were more frequent among progressive women than relapsing women (p = 0.03), as was the A allele itself (p = 0.03), with a genotype-by-sex effect at p = 0.026. The adjusted odds ratio was 1.56 (95% CI 1.06–2.29, p = 0.02), independent of HLA-DRB1*15:01, and the allele alone gave 1.35 (1.02–1.79, p = 0.03). The authors report the statistical power for this association as 62%, a number most papers leave out. The biology, as they present it, cuts both ways: mice deficient in MAP1B have thinner myelin sheaths and fewer large myelinated axons, while accumulation of the protein leads to progressive cell death, and they state plainly that no straightforward conclusion is available about whether it protects or harms the adult brain.

The Same Variant Also Tracks With Iron
In progressive patients, carriers of the rs62363242 A allele had lower serum free iron (13.66 ± 4.78 against 15.91 ± 4.14 µmol/L, p = 0.03) and higher transferrin (2.53 ± 0.44 against 2.27 ± 0.37 g/L, p = 0.03), with a trend toward lower ferritin (p = 0.08). That combination, iron lower or unchanged alongside raised transferrin, matches what has been reported for MS patients in general, and here it appears specifically by genotype and specifically in progressive disease. The authors connect it to work on iron-deficient embryos, where expression of MAP1B rose markedly and was proposed to drive increased apoptosis. Having the same gene turn up in both the clinical association and the circulating iron measurements, in the same patient group, is what makes this worth following further.

CDKN1A Haplotypes Carry Expression and Disability Together
Haplotypes built from rs3176326 and rs3176336 were associated with CDKN1A messenger RNA levels in immune cells, independently of sex. In relapsing patients the GT haplotype gave higher levels than the reference (p = 0.03), and in secondary progressive patients the AT haplotype, carrying both rare alleles, did the same (p = 0.01). In both groups it is the rs3176336 T allele that sits in the higher-expressing haplotype, and that variant has been proposed as a splicing quantitative trait locus for CDKN1A, with the T allele showing the highest intron excision ratio. The GT haplotype also tracked with higher disability in relapsing patients, a mean EDSS of 2.78 (2.49–3.07) against 2.35 (2.17–2.52) for the reference haplotype (p = 0.02). Genotype, gene expression and clinical score lining up in the same direction is more than most candidate-gene studies manage to show.

Where the Severity Variants Landed
RAB4B-EGLN2 rs111833532 predicted all three disability measures in relapsing patients, with the D-allele dominant model giving p = 0.028 for EDSS, 0.05 for the severity score and 0.018 for the age-related severity score, in models that also included sex, disease duration, age at onset and HLA-DRB1*15:01. The same genotypes carried higher plasma 4-hydroxynonenal in progressive patients (p = 0.04), a product of lipid peroxidation that rises during ferroptosis, which ties the clinical signal to the process the panel was built around. The DYSF–ZNF638 variant rs10191329 predicted EDSS (p = 0.03) and the age-related score (p = 0.04), but only in male progressive patients. MTSS1 rs9643199 showed an effect in progressive females in the opposite direction to the original genome-wide study, and the authors explain why rather than leaving it: they analysed cross-sectional severity scores while the earlier work used longitudinal ones, and they say validation in a larger group is needed before drawing any conclusion.

What the Authors Say Needs Doing Next
The limitations are stated specifically enough to act on. The relapsing and progressive groups are unequal in size, which the authors describe as a constant difficulty in studies of this design. The subgroup in which messenger RNA was measured comprised only 48 people, and they suggest this is why no associations emerged beyond the CDKN1A haplotypes. Biological relevance for several of the variants still needs experimental work or animal models. And the previously unstudied associations, the MAP1B finding above all, need validating in larger groups and in different populations. One further result they flag for confirmation is the association between the HLA-DRB1*15:01 A allele and higher serum hexanoyl-lys adduct, an early-stage marker of lipid peroxidation, which they believe has not been reported before. Taken together, what this study offers is a small set of mechanistically anchored candidates, each with the measurement that would test it already named.

Disclaimer: This blog post is based on the cited research article and is intended for informational purposes only. It is not intended to provide medical advice. Please consult with a healthcare professional for any health concerns.

Reference:
Djuric, T., Djordjevic, A., Kuveljic, J., Stefanovic, M., Dincic, E., Kolakovic, A., & Zivkovic, M. (2025). Association of functional gene variants in DYSF–ZNF638, MTSS1 and ferroptosis-related genes with multiple sclerosis severity and target gene expression. International Journal of Molecular Sciences, 26, 4986. https://doi.org/10.3390/ijms26114986