High-resolution MHC sequencing reveals novel genetic risk variants underlying antithyroid drug-induced agranulocytosis

Abstract

Background: Antithyroid drug-induced agranulocytosis (TIA) is a rare and severe complication Graves disease (GD) treatment. Previous genome-wide association studies (GWAS) have identified ethnicity-specific human leukocyte antigen (HLA) risk alleles for TIA, but limitations persist in major histocompatibility complex (MHC) region coverage, variant resolution, and population generalizability. Methods: We conducted the first comprehensive MHC probe-capture sequencing (138× depth, covering 4.97 Mb) using a 2-stage study (Stage 1: 40 TIA cases, 93 GD controls; Stage 2: 20 TIA cases, 49 GD controls), with 1200 healthy controls as background references. High-resolution HLA typing, single-nucleotide polymorphism (SNP) association analysis, meta-analysis, and logistic regression were employed to identify TIA-related genetic variants, and a clinical nomogram was constructed for TIA risk prediction. Results: A total of 5661 MHC variants were detected, among which 133 TIA-risk SNPs were identified and grouped into 12 linkage disequilibrium blocks. The lead SNP rs9270003 (HLA-DRB1) was the most strongly associated variant (odds ratio, 10.19; Padj = 5.99 × 10−7). The C4A missense variant rs201016130 (Ser1286Ala) was correlated with milder TIA and lower serum C4 levels, revealing a novel link between complement system variation and TIA severity. Four human leukocyte antigen (HLA) alleles (HLA-B*38:02:01:02, HLA-B*39:01:01:02L, HLA-C*07:02:01:01, DRB1*08:03:02:01) were associated with TIA and validated in the Stage 2. Notably, homozygosity of HLA-C*07:02:01:01 was linked to more severe TIA phenotypes. A preliminary risk-stratification nomogram was developed (C-index = 0.792), necessitating prospective validation for further confirmation. Conclusion: Targeted MHC sequencing with high-resolution HLA typing identified novel genetic markers associated with TIA susceptibility and severity. Our findings may help lay a foundation for further investigation into individualized risk assessment in GD management.