HLA associations in schizophrenia: are we re-discovering the wheel?

Associations between human leukocyte antigen (HLA) polymorphisms on chromosome 6p and schizophrenia (SZ) risk have been evaluated for over five decades. Numerous case-control studies from the candidate gene era analyzed moderately sized samples and reported nominally significant associations with several loci in the HLA region (sample sizes, n = 100-400). The risk conferred by individual alleles was modest (odds ratios < 2.0). The basis for the associations could not be determined, though connections with known immune and auto-immune abnormalities in SZ were postulated. Interest in the HLA associations has re-emerged following several recent genome-wide association studies (GWAS); which utilized 10- to 100-fold larger samples and also identified associations on the short arm of chromosome 6. Unlike the earlier candidate gene studies, the associations are statistically significant following correction for multiple comparisons. Like the earlier studies; they have modest effect sizes, raising questions about their utility in risk prediction or pathogenesis research. In this review, we summarize the GWAS and reflect on possible bases for the associations. Suggestions for future research are discussed. We favor, in particular; efforts to evaluate local population sub-structure as well as further evaluation of immune-related variables in future studies.