The major histocompatibility complex human leukocyte antigen (HLA)-DRB1*15 (DR2) haplotype is strongly associated with risk of multiple sclerosis (MS). The primary susceptibility has been localized to only approximately 200 kb encompassing the HLA-DR and -DQ loci. Further dissection of disease association with this region is demanding because of the high levels of linkage disequilibrium (LD). Recently, evidence was obtained for the involvement of a gene, potentially encoding an immune co-receptor, in another DR2-associated inflammatory condition, sarcoidosis. The implicated gene, BTNL2, is adjacent to DR and is in strong LD with HLA-DRB1. This fact, combined with a sequence relationship between BTNL2 and myelin oligodendrocyte glycoprotein, an autoantigen associated with MS, makes the gene an attractive candidate. To determine whether BTNL2 contributes to MS, we genotyped 1136 well-characterized MS families from the UK and the USA, as well as an African-American case-control data set, making this among the largest genetic studies in MS. Family-based and case-control association studies were performed for the BTNL2 and HLA-DRB1 loci. In all family data sets, the protein-truncating allele of BTNL2, implicated in sarcoidosis, was significantly over-transmitted to cases (combined data sets: global P=2.4x10(-11)). Given that the protein-truncating allele of BTNL2 virtually always occurred with DRB1*15, an effect could only be tested in DRB1*15-negative individuals or pedigrees. However, despite adequate power to detect an independent association, no difference in transmission of BTNL2 alleles or genotypes was observed in DRB1*15-negative individuals with MS. Conditional logistic regression modeling also strongly supported the conclusion that BTNL2 does not confer additional disease risk. The association of BTNL2 with MS observed in the African-American data set was also secondary to the primary DRB1*15 association.