Previous studies in narcolepsy, an autoimmune disorder affecting hypocretin (orexin) neurons and recently associated with H1N1 influenza, have demonstrated significant associations with five loci. Using a well-characterized Chinese cohort, we refined known associations in TRA@ and P2RY11-DNMT1 and identified new associations in the TCR beta (TRB@; rs9648789 max P = 3.7×10 −9 OR 0.77), ZNF365 (rs10995245 max P = 1.2×10 −11 OR 1.23), and IL10RB-IFNAR1 loci (rs2252931 max P = 2.2×10 −9 OR 0.75). Variants in the Human Leukocyte Antigen (HLA)- DQ region were associated with age of onset (rs7744020 P = 7.9×10 −9 beta −1.9 years) and varied significantly among cases with onset after the 2009 H1N1 influenza pandemic compared to previous years (rs9271117 P = 7.8×10 −10 OR 0.57). These reflected an association of DQB1*03:01 with earlier onset and decreased DQB1*06:02 homozygosity following 2009. Our results illustrate how genetic association can change in the presence of new environmental challenges and suggest that the monitoring of genetic architecture over time may help reveal the appearance of novel triggers for autoimmune diseases.
Narcolepsy-hypocretin deficiency results from a highly specific autoimmune attack on hypocretin cells. Recent studies have established antigen presentation by specific class II proteins encoded by (HLA DQB1*06:02 and DQA1*01:02) to the cognate T cell receptor as the main disease pathway, with a role for H1N1 influenza in the triggering process. Here, we have used a large and well-characterized cohort of Chinese narcolepsy cases to examine genetic architecture not observed in European samples. We confirmed previously implicated susceptibility genes (T cell receptor alpha, P2RY11), and identify new loci (ZNF365, IL10RB-IFNAR1), most notably, variants at the beta chain of the T cell receptor. We found that one HLA variant, (DQB1*03:01), is associated with dramatically earlier disease onset (nearly 2 years). We also identified differences in HLA haplotype frequencies among cases with onset following the 2009 H1N1 influenza pandemic as compared to before the outbreak, with fewer HLA DQB1*06:02 homozygotes. This may be the first demonstration of such an effect, and suggests that the study of changes in GWAS signals over time could help identify environmental factors in other autoimmune diseases.