Abstract
BACKGROUND AND OBJECTIVES: Neuromyelitis optica spectrum disorders (NMOSDs) comprise inflammatory processes of the CNS. Most patients with NMOSD have serum immunoglobulin (Ig) G autoantibodies directed against the astrocytic water channel aquaporin-4 (AQP4-IgG). In this study, we analyzed HLA allelic frequencies in a large cohort of patients with NMOSD, stratified by ethnicity and AQP4-IgG status, compared with healthy controls. METHODS: Next-generation sequencing-based HLA class I and II genotyping was performed in 174 White, 45 Black, and 41 Hispanic AQP4-IgG-positive (AQP4-IgG+) patients with NMOSD; 49 White patients with AQP4-IgG-negative (AQP4-IgG-) NMOSD; and 2,427 White, 244 Black, and 155 Hispanic controls. Correction for multiple testing was performed using the Bonferroni method. RESULTS: In White AQP4-IgG+ patients with NMOSD, the most significantly associated alleles were HLA-DQA1*05:01:01 (30.1% vs 11.1%, odds ratio 3.43 [95% CI 2.65-4.42], corrected p = 8.95E-17), HLA-DQB1*02:01:01 (29.9% vs 11.3%, 3.34 [2.58-4.31], corrected p = 4.33E-16), and HLA-DRB1*03:01:01 (29.2% vs 11.6%, 3.15 [2.43-4.06], corrected p = 3.66E-14), followed by HLA-B*08:01:01 (26% vs 10.4%, 3.02 [2.3-3.94] corrected p = 8.79E-12), HLA-C*07:01:01 (27.8% vs 14%, 2.36 [1.81-3.04], corrected p = 2.44E-07), and HLA-DRB3*01:01:02 (28% vs 14.7%, 2.27 [1.73-2.95], corrected p = 2.19E-06). The frequency of HLA-DRB1*08:04:01 was higher in Black AQP4-IgG+ patients with NMOSD than in Black controls but did not achieve statistical significance (19.3% vs 5.7%, 3.92 [1.91-7.86], corrected p = 0.08). Nevertheless, when compared with a larger cohort of Black controls (n = 16,178), the frequency of HLA-DRB1*08:04 (19.3% vs 5.1%, 4.46 [2.45-7.66], corrected p = 1.88E-03) was significantly higher in Black AQP4-IgG+ patients with NMOSD. No significant HLA associations were detected in AQP4-IgG+ Hispanic patients or White AQP4-IgG- patients with NMOSD. DISCUSSION: This study confirms the previously recognized association of HLA-DRB1*03:01 with AQP4-IgG+ NMOSD in White patients and extends this association to the HLA-DRB1*03:01:01∼HLA-DQA1*05:01:01∼HLA-DQB1*02:01:01 haplotype. Furthermore, it identifies an association of HLA-DRB1*08:04 with AQP4-IgG+ NMOSD in Black patients. However, no HLA associations were detected in White AQP4-IgG- patients with NMOSD. The immunogenetic differences between AQP4-IgG+ and AQP4-IgG- NMOSD support pathophysiologic distinctions between these entities.