An Iron-Responsive MS Subtype, Proposed From Two Children

Two boys with paediatric relapsing-remitting multiple sclerosis (MS), both with severe recurrent iron deficiency, both followed for a decade. Van Rensburg and colleagues sequenced their exomes to ask whether an inherited defect in iron handling could explain the iron deficiency and, through it, the disease. It is worth being clear at the outset what this paper is: a report on two patients, not a cohort study, offering a hypothesis with an unusually long observation attached rather than evidence that the hypothesis holds. Both children were diagnosed under the paediatric MS consensus criteria, with lesions separated anatomically and over time confirmed on serial MRI. Case 1 became symptomatic at 3 years 8 months with relapses at 4 years 4 months and 5 years 1 month; Case 2 had onset at 5 years with relapses at 5 years 8 months and 11 years 8 months.

The Gene List, and What Selecting on Iron Genes Means
Sequencing returned 24,916 variants in Case 1 and 26,550 in Case 2. These were annotated and prioritised through an in-house pipeline built around genes with a role in iron deficiency, drawn from roughly 20,000 human genes. Variants emerged in TMPRSS6, which encodes a serine protease that suppresses hepcidin and so governs iron absorption; TF, encoding transferrin; CUBN, encoding cubilin; SLC25A37, encoding mitoferrin-1; CD163, a scavenger receptor for haemoglobin after haemorrhage; and COQ3, in coenzyme Q synthesis. All reported variants were verified by Sanger sequencing, which is proper practice. The structural problem sits in the method description itself: two children selected for iron deficiency, screened through a pipeline selected for iron genes, will return iron gene variants. Without allele frequencies in a matched population and without a comparison group, carrying heterozygous common polymorphisms in these genes carries limited weight on its own.

What the Variants Would Do If They Matter
Taken on its own terms, the proposed mechanism is coherent and specific rather than vague. CUBN is the most interesting node, because cubilin sits at the junction of three deficiencies the paper cares about at once. Compromised cubilin function means enhanced urinary excretion of transferrin-bound iron, impaired reabsorption of vitamin D in the kidney, and vitamin B12 malabsorption in the ileum, and all three appear on the paper's own list of demyelination risk factors. SLC25A37 encodes mitoferrin-1, the importer required to build mitochondrial haem and the iron-sulfur clusters of the electron transport chain. The authors note that the rs3736032 AA genotype has a frequency below 0.1 in all natural population groups, which they read as compatible with a severe impact on survival, and both children are heterozygous for it. Neither child had symptoms of megaloblastic anaemia, so the CUBN variants are not clinically expressed in the way high-penetrance mutations in that gene are.

Neither Child Carried the Main MS Risk Allele
The most useful genetic result here is a negative one. Taqman tagging with rs9271366 showed that neither child carries HLA-DRB1*15:01, the allele with the largest single effect on MS risk. That observation does the real work in the paper's argument, since it removes the most common genetic explanation and leaves room to look elsewhere. The base rate is worth keeping in view alongside it: a substantial share of people with MS do not carry this allele, so two consecutive non-carriers is an ordinary occurrence rather than a finding in itself. What it does do is make the question the authors are asking a reasonable one to ask.

Ten Years Without Relapse, and What That Can Carry
Iron supplementation at 4 mg/kg/day began after a third relapse in both children, and their iron levels dropped whenever it was stopped. Case 1 has since gone 11 years without a relapse and Case 2 nine years. Case 2 had repeat MRI at 12 years 3 months and 14 years 9 months following epileptic seizures, which were attributed to healed scar tissue from earlier demyelination, with no active lesions found; Case 1 never required further imaging. The length of that follow-up is the strongest thing in the paper. Three limits constrain what it establishes. The intervention was not iron alone but a broader restorative programme of nutrients, so several variables changed at once. There was no control, no blinding, and no comparison group. And relapse frequency in relapsing-remitting disease declines over time without any intervention at all. The abstract describes the follow-up as confirming the role of iron deficiency in these patients, which is a stronger word than an uncontrolled observation in two children will bear.

The Theory of MS This Sits Inside
The paper is arguing against the standard autoimmune framing, and it names its source rather than gesturing at one. It cites a 2018 histopathology study by Prineas and Parratt that found no evidence for a diagnostic circulating anti-myelin or anti-oligodendrocyte autoantibody, and concluded that a better account involves phagocytosis of oligodendrocytes and damaged myelin by activated microglia and monocyte-derived macrophages, secondary to the death of the oligodendrocytes. From there the authors build a three-step model: a shortage of substrates for ATP synthesis produces mitochondrial oxidative stress and oligodendrocyte apoptosis; variants in inflammatory pathways such as CD163 prolong the immune response that follows; and nutritional supplementation relieves the oxidative stress and allows remyelination. Both children were exposed to second-hand smoke from their parents, which the authors include as a contributor to reactive oxygen species.

What the Paper Concedes, and What It Leaves Out
Their stated limitation is specific and honest: urinary iron was never measured, because the possibility of renal iron loss was not considered until the CUBN variants turned up. They also note that the functional significance of most variants in their table was not demonstrated in vitro, with clinical relevance for iron metabolism reported only for TMPRSS6 rs855791 and TF rs1880669. Two further points belong in a reader's assessment. There is no control group, so nothing here separates these variant patterns from what two children of similar ancestry drawn at random would carry. And the disclosure shapes how the recommendations should be read: two authors are named inventors on a patent filed by their university, and one is a director and shareholder of a company spun out of the South African Medical Research Council, in the area of genetic testing the paper argues for expanding. Their own closing recommendation meets the scale problem directly and is the right next step, calling for these variants to be investigated in a larger group of MS patients alongside measurements of renal excretion of iron, vitamin B12, and vitamin D.

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:
van Rensburg, S. J., Peeters, A. V., van Toorn, R., Schoeman, J., Moremi, K. E., van Heerden, C. J., & Kotze, M. J. (2019). Identification of an iron-responsive subtype in two children diagnosed with relapsing-remitting multiple sclerosis using whole exome sequencing. Molecular Genetics and Metabolism Reports, 19, 100465. https://doi.org/10.1016/j.ymgmr.2019.100465