The primary cause of neurological syndromes with antibodies against glutamic acid decarboxylase 65 (GAD65-Ab) is unknown, but genetic predisposition may exist as it is suggested by the co-occurrence in patients and their relatives of other organ-specific autoimmune diseases, notably type 1 diabetes mellitus (T1DM), and by the reports of a few familial cases. We analyzed the human leukocyte antigen (HLA) in 32 unrelated patients and compared them to an ethnically matched sample of 137 healthy controls. Four-digit resolution HLA alleles were imputed from available Genome Wide Association data, and full HLA next-generation sequencing-based typing was also performed. HLA DQA1*05:01–DQB1*02:01–DRB1*03:01 was the most frequent class II haplotype in patients (13/32, 41%). DQB1*02:01 was the only allele found to be significantly more common in patients than in controls (20/137, 15%, corrected p = 0.03, OR 3.96, 95% CI [1.54–10.09]). There was also a trend towards more frequent DQA1*05:01 among patients compared to controls (22/137, 16%; corrected p = 0.05, OR 3.54, 95% CI [1.40–8.91]) and towards a protective effect of DQB1*03:01 (2/32, 6% in patients vs. 42/137, 31% in control group; corrected p = 0.05, OR 0.15, 95% CI [0.02–0.65]). There was no significant demographic or clinical difference between DQ2 and non-DQ2 carriers (p > 0.05). Taken together, these findings suggest a primary DQ effect on GAD65-Ab neurological diseases, partially shared with other systemic organ-specific autoimmune diseases such as T1DM. However, it is likely that other non-HLA loci are involved in the genetic predisposition of GAD65-Ab neurological syndromes.
