Risk for myasthenia gravis maps to a (151) Pro→Ala change in TNIP1 and to human leukocyte antigen-B*08

Abstract

Objective: The objective of this study is to comprehensively define the genetic basis of early onset myasthenia gravis (EOMG).

Methods: We have carried out a 2-stage genome-wide association study on a total of 649 North European EOMG patients. Cases were matched 1:4 with controls of European ancestry. We performed imputation and conditional analyses across the major histocompatibility complex, as well as in the top regions of association outside the human leukocyte antigen (HLA) region.

Results: We observed the strongest association in the HLA class I region at rs7750641 (p = 1.2 × 10(-92) ; odds ratio [OR], 6.25). By imputation and conditional analyses, HLA-B*08 proves to be the major associated allele (p = 2.87 × 10(-113) ; OR, 6.41). In addition to the expected association with PTPN22 (rs2476601; OR, 1.71; p = 8.2 × 10(-10) ), an imputed coding variant (rs2233290) at position 151 (Pro→Ala) in the TNFAIP3-interacting protein 1, TNIP1, confers even stronger risk than PTPN22 (OR, 1.91; p = 3.2 × 10(-10) ).

Interpretation: The association at TNIP1 in EOMG implies disease mechanisms involving ubiquitin-dependent dysregulation of NF-κB signaling. The localization of the major HLA signal to the HLA-B*08 allele suggests that CD8(+) T cells may play a key role in disease initiation or pathogenesis.

Figure 1

Manhattan plot of association signals in early-onset myasthenia gravis (EOMG). The strength of association (-log₁₀ p values; ordinate) is plotted against the position on each chromosome in base pairs (abscissa). Details of markers achieving p <5 × 10⁻⁴ are provided in Supplementary Table 2.

Figure 2

TNIP1 association signals in EOMG. Association signals were examined after inclusion of imputation results using 1000 Genome sequencing data and the genotyped SNPs (diamond symbols) and imputed SNPs (circles) are shown. The ordinate shows the strength of the association signals, with the position on chromosome 5 shown in Mb (HG18 map) along the abscissa. The p value for each SNP is shown before (a) and after conditioning with rs2233290 (b), a SNP that encodes a ¹⁵¹Pro→Ala change. The color-coded symbols correspond to the strength of linkage disequilibrium (r²), showing the most significant associated SNP (rs1559127) as a blue circle. See Supplementary Materials for imputation methods and details for imputed SNPs and results (Supplementary Table 3).

Figure 3

Analysis of the HLA region association signals in EOMG. In each panel, the symbols show the strength of the association signal (ordinate) for this region of chromosome 6 shown in Mb (HG18) along the abscissa; the SNPs and HLA determinants are given color-coded symbols as described in Figure 2. For a, the −Log₁₀ p value before conditioning is shown. For b–d the −Log₁₀ p values are shown after conditioning on the indicated HLA determinant(s). See Supplementary Materials for methods and additional conditioning studies (Supplementary Table 4, Fig. S1).