Background Due to the lack of prospective studies with longitudinal data on relapse, past genetic studies have not attempted to identify genetic factors that predict relapse risk (the primary endpoint of many pivotal clinical trials testing the efficacy of multiple sclerosis (MS) disease-modifying drugs) at a genome-wide scale.

Methods We conducted a genome-wide association analysis (GWAS) to identify genetic variants that predict MS relapse risk, using a three-stage approach. First, GWAS was conducted using the southern Tasmania MS Longitudinal Study with 141 cases followed prospectively for a mean of 2.3 years. Second, GWAS was conducted using the Ausimmune Longitudinal Study with 127 cases having a classic first demyelinating event followed for 5 years from onset. Third, the top hits with p<5.0×10⁻⁶ from the first two stages were combined with a longitudinal US paediatric MS cohort with 181 cases followed for 5 years after onset. Predictors of time to relapse were evaluated by a mixed effects Cox model. An inverse variance fixed effects model was then used to undertake a meta-analysis.

Results In the pooled results, using these three unique longitudinal MS cohorts, we discovered one novel locus (LRP2; most significant single nucleotide polymorphism rs12988804) that reached genome-wide significance in predicting relapse risk (HR=2.18, p=3.30×10⁻⁸). LRP2 is expressed on the surface of many central nervous system cells including neurons and oligodendrocytes and is a critical receptor in axonal guidance.

Conclusions The finding of a genetic locus that has extensive effects on neuronal development and repair is of interest as a potential modulator of MS disease course.