Twelve Genes, Twenty Years, No Answer: Vitamin D and Genetic Susceptibility to MS
Scazzone and colleagues set out to review every genetic study linking the vitamin D pathway to multiple sclerosis (MS) risk, covering single-nucleotide polymorphisms (SNPs) in 12 vitamin D pathway genes. Their conclusion is negative: the reviewed studies cannot clarify whether, or to what extent, vitamin D-related gene variants influence MS risk. That result deserves attention precisely because the mechanistic prior is so strong. Ramagopalan and colleagues reported in 2011 that more than 80% of MS-associated genes carry vitamin D responsive elements (VDREs) in their promoter regions, and Disanto and colleagues showed in 2012 that these VDREs sit preferentially in open chromatin in immune cells. Vitamin D therefore has direct transcriptional access to most of the known MS risk architecture. Two decades of candidate-gene work on the genes that set vitamin D status still cannot show that variation in those genes changes who develops MS. The gap between those two statements is what this review documents.
Twelve Genes, Because the Pathway Offers Twelve Points to Intervene
Vitamin D3 forms in skin when UVB radiation converts 7-dehydrocholesterol to pre-vitamin D3, which rearranges into vitamin D3 by a temperature-dependent step. Activation then requires two hydroxylations: the first in the liver, producing 25-hydroxyvitamin D [25(OH)D], catalysed mainly by CYP2R1 with minor contributions from CYP27A1 and CYP3A4; the second mostly in kidney proximal tubule cells, where CYP27B1 produces the active hormone 1,25-dihydroxyvitamin D. CYP24A1 runs the opposite direction, hydroxylating both metabolites at C-24 and C-23 into inactive forms bound for excretion. Transport and delivery add more control points: vitamin D binding protein (VDBP, encoded by GC) carries 85 to 90% of circulating metabolites, albumin binds another 10 to 15%, and under 1% circulates free. The megalin-DAB2-cubilin system internalises the 25(OH)D-VDBP complex. At the target cell, VDR heterodimerises with retinoid-X-receptor alpha to drive genomic effects, while MARRS handles the rapid non-genomic response. Every one of these steps generated a candidate gene, which is how the review's SNP table reached its length.
VDR Is the Most Studied Gene and the Least Settled
Four VDR variants dominate the literature, still named after the restriction enzymes once used to genotype them: ApaI (rs7975232), BsmI (rs1544410), TaqI (rs731236), and FokI (rs2228570). The distinction between them is functional. ApaI, BsmI, and TaqI regulate VDR expression without altering the protein, while FokI changes both the protein structure and its transcriptional activity. Fukazawa and colleagues opened the field in 1999 by reporting a BsmI association with MS in a Japanese population. The review then lists more than 25 case-control studies across ethnically distinct populations reaching contradictory results, and names the reasons: most had small samples and correspondingly low statistical power, none accounted for interactions with other genetic or environmental factors, and the MS phenotype itself is clinically heterogeneous. A gene studied for 20 years by that many groups without convergence is telling you something about study design rather than about biology.
CYP27B1 Produced a Rare Loss-of-Function Variant, Then Lost It
CYP27B1 carries the strongest single-variant story in the review, and also the sharpest rebuttal. The Australia and New Zealand consortium GWAS in 2009, on 1,618 cases and 3,413 controls, flagged CYP27B1 as its most likely MS candidate, and Sundqvist and colleagues confirmed rs703842 in a large Swedish case-control study in 2010. Sawcer and colleagues then found rs12368653 in a GWAS of more than 9,000 MS patients in 2011. In the same year, Ramagopalan and colleagues sequenced 43 families with at least four affected members and found rs118204009 in one family, carried by every affected individual; genotyping in 3,046 parent-affected child trios supported the association, and in vitro work showed the variant causes complete loss of CYP27B1 function and reduced 1,25-dihydroxyvitamin D. Two years later Ban and colleagues published a study in Annals of Neurology finding no evidence of association between mutant CYP27B1 alleles and MS. As of this review, the rare variant remains unconfirmed by any independent group.
CYP24A1 and GC: Functional Evidence That Nobody Reproduced
CYP24A1 followed the same pattern. The Sawcer GWAS of 9,772 cases from 15 European countries detected rs2248359 among 29 novel susceptibility loci, and Ramasamy and colleagues supplied functional support in 2014 by showing that rs2248359 regulates CYP24A1 expression in human brain. Then Orton and colleagues genotyped 1,360 MS patients and 1,677 unaffected family members and found no statistically significant result, with matching nulls reported in Han Chinese, Danish, and Tasmanian populations. GC is the clearest case of a gene that predicts the biomarker but not the disease. Two non-synonymous SNPs in exon 11, rs7041 and rs4588, produce the Gc1f, Gc1s, and Gc2 isoforms, whose combinations yield six VDBP phenotypes with different binding affinities, from Gc1f/1f at the highest to Gc2/2 at the lowest. Two independent GWAS confirmed these variants as genetic determinants of 25(OH)D levels. Every study of VDBP SNPs and MS risk since Lindblom and colleagues first looked in 1988 has failed to find an association. The review reads this as VDBP acting on MS only through vitamin D status, if at all.
Four Pathway Components Nobody Has Properly Tested
The review is at its most useful where it maps what is missing. No author has investigated whether FGF-23 or Klotho variants influence MS risk, despite both regulating vitamin D through a defined feedback loop in which 1,25-dihydroxyvitamin D induces FGF-23 secretion and FGF-23 in turn downregulates CYP27B1 and upregulates CYP24A1. MARRS has been evaluated exactly once, by Pytel and colleagues in 2019, with unsatisfying results, even though the mechanistic case is specific: impaired MARRS activity dysregulates STAT3 signalling, which drives naïve T cells toward the pro-inflammatory Th17 phenotype central to MS. RXR-alpha has been examined by two research groups, which disagreed. The megalin-DAB2-cubilin genes were first investigated in 2019, also inconsistently. The pathway's non-genomic and transport arms are close to untested while VDR accumulated its 25 conflicting papers.
The Negative Result Is a Design Problem, and Two Parts Are Fixable
MS risk is polygenic and mediated by variants common in the general population, so any single variant carries only a modest effect. Candidate-gene studies with a few hundred patients cannot detect modest effects, and most entries in the review's table are candidate studies; only a handful come from GWAS. The review's own account of the Mendelian randomization work sharpens the point. Mokry, Rhead, and Gianfrancesco and their colleagues each found evidence that low vitamin D raises MS risk, and the finding survived adjustment for confounders including sun exposure and diet. Those analyses pool many vitamin D variants into a single instrument. Pooling recovers a signal that testing one SNP at a time does not. The second fixable problem is measurement. The review notes that the lack of standardised serum 25(OH)D measurement blocked consensus guidelines, and states that studies evaluating 25(OH)D should be taken with a grain of salt unless internationally recognised reference materials and procedures were used. Standardised assays exist now. Applying them, and testing the vitamin D pathway as a variant burden rather than one SNP at a time, would tell us whether the 20-year null reflects biology or just underpowered arithmetic.
Disclaimer: This blog post is based on the cited review 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.
References:
Scazzone, C., Agnello, L., Bivona, G., Lo Sasso, B., & Ciaccio, M. (2021). Vitamin D and genetic susceptibility to multiple sclerosis. Biochemical Genetics, 59(1), 1-30. https://doi.org/10.1007/s10528-020-10010-1
