Part IV
Genetics
HLA-DRB1*15:01 dominates a polygenic landscape of more than 230 immune-system loci. The familial, twin and GWAS data fit a model of common-variant, immune-driven susceptibility — with environment (EBV, vitamin D, smoking) converting genetic risk into disease.
1. Family & Twin Studies — Heritability of MS
MS aggregates in families and behaves as a complex trait of moderate heritability:
| Relationship | Lifetime risk | Recurrence vs population |
|---|---|---|
| General population (NW Europe) | ~0.3% | 1× |
| First-degree relative | ~3% | ~10× |
| Sibling | ~3–5% | ~10–20× |
| Dizygotic twin | ~5% | ~20× |
| Monozygotic twin | ~25–30% | ~80–100× |
| Child of two MS parents | ~30% | ~100× |
The discordance between MZ (~30%) and DZ (~5%) twin concordance gives a narrow-sense heritability estimate of \(h^2 \approx 0.5\) on the liability scale. The fact that MZ concordance is well below 100% establishes that environment is essential — consistent with the EBV picture from Part III.
2. HLA-DRB1*15:01 — the Major Susceptibility Allele
The MHC association with MS, first described by Jersild (1972), is the strongest in any neurological disease. Fine-mapping has placed the dominant signal on the HLA-DRB1*15:01 allele (within the DR2 haplotype DRB1*15:01-DQB1*06:02):
- One copy — OR ~3 for MS in northern European populations.
- Two copies (homozygote) — OR ~6 (additive on the log-OR scale).
- HLA-DRB1*15:01 carriage explains roughly 10–20% of MS heritability on the liability scale.
- Mechanistically, DR2 presents myelin peptides (notably MBP85-99) with high affinity to autoreactive CD4+ T cells — the structural basis is captured in PDB 1IEA (shown in Part III).
The same allele is also strongly associated with other autoimmune diseases (narcolepsy, type-1 diabetes), arguing for a generic effect of altered antigen presentation rather than an MS-specific mechanism.
3. Other HLA Class-II and Class-I Effects
Beyond DRB1*15:01, the MHC harbours independent susceptibility and protective signals:
- HLA-DRB1*03:01, *13:03, *08:01 — secondary risk alleles.
- HLA-A*02:01 — a class-I protective allele (OR ~0.6); presents EBV-derived peptides and may aid CD8+ control of EBV.
- HLA-B*44:02 — class-I protective.
- Epistatic interaction between DRB1*15:01 and A*02:01 has been described — the protective effect of A*02:01 is greatest in DRB1*15:01 carriers.
The net MHC effect on MS risk requires modelling all class-I and class-II alleles jointly. Modern imputation panels (HLA*IMP:02, SNP2HLA) make this routine in GWAS.
4. The IMSGC GWAS — 233 Non-MHC Loci
The International Multiple Sclerosis Genetics Consortium (IMSGC) has progressively scaled MS GWAS:
- 2007 — first MS GWAS (931 trios), confirms IL-7R, IL-2RA.
- 2011 — IMSGC + WTCCC2 (Nature 2011): 9772 cases, 17,376 controls; 57 non-MHC loci.
- 2013 — ImmunoChip analysis (47 638 cases + ctrls): 110 non-MHC loci.
- 2019 — IMSGC (Science 2019): 47,429 cases, 68,374 controls, 233 non-MHC genome-wide-significant loci; 32 independent MHC effects.
- 2023 — IMSGC (Nature 2023): identification of a non-coding variant near DYSF/ZNF638 associated with severity (independent of susceptibility).
The architecture is profoundly immune-system-centric: enrichment analyses show non-MHC associations cluster in CD4+ T cells (Th1/Th17), B cells, and dendritic cells. No locus implicates an oligodendrocyte- or neuron-intrinsic gene at genome-wide significance — consistent with MS being primarily an immune disease, the CNS being the target rather than the source of the defect.
5. Cytokine-Receptor and Co-Stimulation Genes
Several non-MHC loci have well-mapped functional consequences:
| Gene | Pathway | Functional effect |
|---|---|---|
| IL-7R (CD127) | T-cell homeostasis | Risk allele alters splicing → ↑ soluble IL-7R, ↑ inflammatory T-cell signalling |
| IL-2RA (CD25) | Treg / IL-2 axis | Risk allele ↓ Treg function (target of daclizumab) |
| TNFRSF1A | TNF receptor 1 | R92Q variant produces decoy receptor; explains paradoxical worsening of MS by anti-TNF agents |
| CD58 | T-cell co-stim | Risk allele lowers CD58 → reduced Treg suppression |
| TYK2 | JAK/STAT | Loss-of-function variant protective; implicates IL-12/23 pathway |
| CD40 | B-T crosstalk | Risk allele alters B-cell activation |
| EOMES, RGS1, STAT4 | Th1/Th17 differentiation | Multiple risk alleles tune effector T-cell programmes |
| CYP27B1, CYP24A1 | Vitamin D metabolism | Mendelian randomisation supports causal vitamin D effect |
Several loci overlap with other autoimmune diseases (rheumatoid arthritis, type-1 diabetes, IBD); shared pathways include IL-2/IL-7 signalling, TYK2/JAK-STAT, and antigen presentation. MS-specific effects concentrate in T- and B-cell programmes rather than innate-immunity genes.
6. Polygenic Risk Score
An MS polygenic risk score (PRS), summing effect-weighted genotypes across >200 loci, has area-under-curve of ~0.65–0.70 for case-control discrimination. Top decile of PRS confers ~5× increased risk vs lowest decile. PRS does not currently warrant clinical screening, but is increasingly used:
- To enrich populations for prevention trials (EBV vaccine).
- To stratify CIS patients for conversion risk.
- In pleiotropy analyses across autoimmune diseases.
Formally, on the liability scale, the proportion of variance explained by the PRS is \(R^2 \approx 0.20\); HLA-DRB1*15:01 alone contributes ~10%; the rest is split among 233+ loci of individually small effect (typical OR 1.05–1.20).
7. The Heritability Gap
Twin studies place MS heritability at \(h^2 \approx 0.5\). The cumulative variance explained by all GWAS-significant loci, even when the MHC is fully accounted for, is closer to ~0.20. The missing heritability may reflect:
- Rare variants of larger effect not captured by SNP arrays (whole-exome / whole-genome sequencing initiatives are underway).
- Structural variants in the MHC (DRB1 copy-number, DR/DQ haplotype effects).
- Gene-gene interactions (epistasis), particularly within MHC.
- Gene-environment interactions absorbing apparent heritability under standard models.
- Twin concordance estimates that may overstate \(h^2\) (shared environments).
The gap is comparable to that observed in other complex autoimmune diseases and is not specific to MS.
8. Gene–Environment Interactions
The most replicated gene-environment effect in MS is between HLA-DRB1*15:01 and smoking (and infectious mononucleosis):
- HLA-DRB1*15:01 carriage alone — OR ~3.
- HLA-DRB1*15:01 + smoking — OR ~14 (multiplicative interaction; Hedstrom et al., Brain 2011).
- HLA-DRB1*15:01 + IM history — OR ~9.
- HLA-DRB1*15:01 + adolescent obesity — OR ~6 in women.
These interactions reframe MS prevention as a gene-targeted environmental intervention: the same exposure (smoking, EBV, low vitamin D) yields very different population effects depending on HLA background.
9. Pharmacogenomics & Severity Genetics
MS pharmacogenomics is younger than its susceptibility genetics, but several signals are clinically actionable:
- JCV serology — risk of progressive multifocal leukoencephalopathy on natalizumab is stratified primarily by JC virus antibody status (clinical, not genetic).
- Glatiramer acetate response — HLA-DRB1*15:01 carriers may respond differently; data conflicting.
- Interferon-β neutralising antibodies — HLA-DRB1*15:01 and *07:01 affect rates.
- DYSF/ZNF638 locus — the IMSGC 2023 paper identified a non-coding variant associated with disease severity (time to EDSS 6) but not susceptibility — the first severity locus.
- CYP27B1, CYP24A1, GC — vitamin D pathway genotypes modulate response to vitamin D supplementation.
The dissociation of susceptibility (immune) and severity (potentially CNS-intrinsic) genetics is a key 2020s finding: genes that determine whether someone develops MS may differ from those that determine how badly they progress.