Genome-Wide Association Study of Late-Onset Myasthenia Gravis: Confirmation of TNFRSF11A and Identification of ZBTB10 and Three Distinct HLA Associations

• Published in: Molecular medicine (Cambridge, Mass.) Vol. 21; no. 1; pp. 769 - 781
• Main Authors: Seldin, Michael F, Alkhairy, Omar K, Lee, Annette T, Lamb, Janine A, Sussman, Jon, Pirskanen-Matell, Ritva, Piehl, Fredrik, Verschuuren, Jan J G M, Kostera-Pruszczyk, Anna, Szczudlik, Piotr, McKee, David, Maniaol, Angelina H, Harbo, Hanne F, Lie, Benedicte A, Melms, Arthur, Garchon, Henri-Jean, Willcox, Nicholas, Gregersen, Peter K, Hammarstrom, Lennart
• Format: Journal Article
• Language: English
• Published: England BioMed Central 01.01.2015; The Feinstein Institute Press
• Subjects: Age; Datasets; Eigenvalues; Females; Genomes; Histology; Laboratories; Males; Medicin och hälsovetenskap; Myasthenia gravis; Principal components analysis; Quality control
• ISSN: 1076-1551; 1528-3658; 1528-3658
• DOI: 10.2119/molmed.2015.00232

To investigate the genetics of late-onset myasthenia gravis (LOMG), we conducted a genome-wide association study imputation of>6 million single nucleotide polymorphisms (SNPs) in 532 LOMG cases (anti-acetylcholine receptor [AChR] antibody positive; onset age≥50 years) and 2,128 controls matched for sex and population substructure. The data confirm reported associations (rs4574025, = 3.9 × 10 , odds ratio [OR] 1.42) and identify a novel candidate gene, , achieving genome-wide significance (rs6998967, = 8.9 × 10 , OR 0.53). Several other SNPs showed suggestive significance including rs2476601 ( = 6.5 × 10 , OR 1.62) encoding the PTPN22 R620W variant noted in early-onset myasthenia gravis (EOMG) and other autoimmune diseases. In contrast, EOMG-associated SNPs in showed no association in LOMG, nor did other loci suggested for EOMG. Many SNPs within the major histocompatibility complex (MHC) region showed strong associations in LOMG, but with smaller effect sizes than in EOMG (highest OR ~2 versus ~6 in EOMG). Moreover, the strongest associations were in opposite directions from EOMG, including an OR of 0.54 for in LOMG ( = 5.9 × 10 ) versus 2.82 in EOMG ( = 3.86 × 10 ). Association and conditioning studies for the MHC region showed three distinct and largely independent association peaks for LOMG corresponding to (a) MHC class II (highest attenuation when conditioning on ), (b) and (c) MHC class III SNPs. Conditioning studies of human leukocyte antigen (HLA) amino acid residues also suggest potential functional correlates. Together, these findings emphasize the value of subgrouping myasthenia gravis patients for clinical and basic investigations and imply distinct predisposing mechanisms in LOMG.