Quantifying Genetic Burden in Multiple Sclerosis: Insights from Multicase and Sporadic Families
Gourraud and colleagues (2011) address a central problem in multiple sclerosis (MS) genetics: MS is moderately heritable yet highly polygenic, with many common variants of modest effect identified through genome-wide association studies (GWAS). The authors ask whether these known susceptibility variants aggregate differently in families with multiple affected individuals (“multicase” families) versus families with a single affected individual (“sporadic” families). To do so, they introduce a weighted, log-additive composite metric—the MS genetic burden (MSGB)—that integrates risk contributions from a set of established loci (including the major histocompatibility complex [MHC] via an HLA-DRB1*15:01 tagging marker) and, optionally, sex as a covariate reflecting the known female predominance in MS.
Study design, cohorts, and genotyping strategy
The analysis leverages a substantial family-based resource: 1,213 independent MS families, partitioned into 403 multicase families (≥1 first-degree relative also affected) and 810 sporadic families (no known MS history in relatives). The study focuses on probands, their parents, and siblings when available, and uses spouses of MS patients as genetically unrelated controls. Importantly, ascertainment and diagnostic criteria were harmonized across datasets, and key clinical descriptors (sex ratio, age at onset, disease duration, proportion relapsing–remitting versus progressive forms) were broadly comparable between the multicase probands, affected relatives, and sporadic probands (Table 1). Genotyping employed TaqMan assays with a high overall success rate (~99%), supporting reliability of the downstream MSGB calculations.
Constructing the MS Genetic Burden (MSGB)
MSGB is computed as a weighted sum of risk alleles using effect sizes (odds ratios) derived from prior GWAS and meta-analyses, thereby minimizing in-sample “overfitting” of weights. The model includes one representative SNP per associated region (Table 2), capturing loci such as IL2RA, IL7R, CD58, TYK2, TNFRSF1A, IRF8, and others; the MHC contribution is represented by rs3135388 (tagging HLA-DRB1*15:01) with a comparatively large effect size. The authors also incorporate gender as a fixed odds ratio (1.6) in some MSGB formulations to reflect epidemiologic sex differences, and they compute alternative MSGB variants excluding gender and/or excluding MHC to isolate component contributions. This multi-component construction enables explicit tests of whether non-MHC common variants “compensate” when HLA risk or female sex is absent.
Key finding: higher aggregation of known risk variants in multicase families
Across probands and parents, MSGB demonstrates a consistent gradient of genetic loading. Probands from multicase families show higher MSGB than sporadic probands (Figure 1), and both mothers and fathers in multicase families also carry higher burdens than their sporadic-family counterparts. Moreover, all family member categories (probands and parents) have higher MSGB than spouse-controls, supporting enrichment of known risk alleles within MS pedigrees. The authors emphasize the large within-group variability: despite statistically robust group shifts, MSGB values overlap substantially across categories, underscoring heterogeneity in how common risk variants distribute among families. The ordered trend—from multicase probands toward controls through sporadic probands and parents—remains significant even when excluding gender and when restricting the score to non-MHC SNPs.
Dissecting contributions: HLA and sex dominate discriminatory signal
A mechanistically important observation emerges when MSGB is stratified by HLA-DRB1*15:01 dose and gender (Figure 2). When the full score is used (gender + MHC + non-MHC), the subgroup differences among probands are strong and track expected HLA dose and sex effects. However, when the score is restricted to non-MHC SNPs only, the between-subgroup contrasts largely disappear, indicating that most of the separability across strata is driven by HLA and sex, rather than the currently known non-MHC common variants. Critically, the authors do not observe a compensatory increase in non-MHC burden among “low-risk” groups (e.g., HLA-negative males). In fact, HLA-negative male probands can trend toward lower scores relative to controls under certain parameterizations, reinforcing the interpretation that non-MHC common variants, as modeled here, do not offset absence of major effects from HLA and sex.
Within-family risk and the limits of prediction
Using sibling data in multicase families, the study probes whether MSGB can distinguish affected from unaffected relatives under relatively shared environments. Unaffected siblings exhibit lower MSGB than probands but still higher than spouse-controls (Figure 3A; Table 5), consistent with intermediate genetic loading. When comparing siblings within the same sibship, having an MSGB greater than or equal to the proband is associated with increased odds of being affected (odds ratio ~2.1). Nevertheless, predictive performance is weak: receiver operating characteristic analyses of sib–proband contrasts yield areas under the curve near chance (approximately 0.53–0.57 depending on score components), and the non-MHC component adds little incremental discrimination beyond HLA and sex (Figure 3B). This is a scientifically consequential result: even with multiple validated loci aggregated in a weighted score, case–control status prediction remains infeasible at clinically meaningful accuracy using the then-available common-variant architecture.
Interpretation, limitations, and implications for MS genetics
The authors position MSGB primarily as an integrative research instrument rather than a diagnostic tool: it quantifies cumulative known genetic susceptibility, allows comparison of “genetic loading” across study designs, and may serve as a covariate analogous to population-structure principal components in association models. They also underscore why multigenerational, multiaffected families remain valuable even in the GWAS era: families may enrich for risk architectures that improve power and may help prioritize individuals at the low end of MSGB for resequencing or linkage efforts (where rare, more penetrant variants or strong environmental triggers might be more salient). The discussion also candidly notes methodological constraints: modeling the MHC with a single tagging SNP is reductive; effect sizes may vary across contexts; alternative inheritance models and gene–gene or gene–sex interactions are not fully captured; and the weighting does not incorporate uncertainty of published odds ratios. Collectively, the work reframes a key message for complex-trait genetics: aggregating common variants can illuminate population and family-level burden gradients, yet translation to individual-level prediction requires substantially more genetic and environmental resolution than provided by early GWAS signals.
Disclaimer: This blog post is based on the provided 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.
References:
Gourraud, P. A., McElroy, J. P., Caillier, S. J., Johnson, B. A., Santaniello, A., Hauser, S. L., & Oksenberg, J. R. (2011). Aggregation of multiple sclerosis genetic risk variants in multiple and single case families. Annals of neurology, 69(1), 65-74.
