HLA-DRB1 and DQB1 alleles in Japanese type 1 autoimmune hepatitis: The predisposing role of the DR4/DR8 heterozygous genotype

Objective Autoimmune hepatitis (AIH) is a chronic progressive liver disease. AIH is composed predominantly of type 1 in Japanese populations. The genetic and environmental factors are associated with the pathogenesis of AIH. HLA-DRB1*03:01 and *04:01 are associated with type 1 AIH in European and *04:05 in Japanese populations. Here, we conducted an HLA association study in order to find HLA alleles or haplotypes predisposing or protective for Japanese AIH. Methods HLA-DRB1 and DQB1 genotyping of 360 type 1 AIH patients and 1026 healthy controls was performed. Results The predisposing association of DRB1*04:01 (P = 0.0006, corrected P [Pc] = 0.0193, odds ratio [OR] 2.97, 95% confidence interval [CI] 1.62–5.43), DRB1*04:05 (P = 1.89×10−21, Pc = 5.86×10−20, OR 3.41, 95% CI 2.65–4.38), and DQB1*04:01 (P = 4.66×10−18, Pc = 6.99×10−17, OR 3.89, 95% CI 2.84–5.33) and the protective association of DRB1*13:02 (P = 0.0003, Pc = 0.0080, OR 0.48, 95% CI 0.32–0.72) with Japanese type 1 AIH were observed. An association of the DR4/DR8 heterozygous genotype with Japanese AIH was identified for the first time (P = 3.12×10−9, OR 3.52, 95% CI 2.34–5.29). Susceptible diplotypes were DRB1*04:05-DQB1*04:01/DRB1*08:02-DQB1*03:02 (P = 0.0004, OR 24.77, 95% CI 1.45–424.31) and DRB1*04:05-DQB1*04:01/DRB1*08:03-DQB1*06:01 (P = 1.18×10−6, OR 10.64, 95% CI 3.19–35.46). Serum levels of Immunoglobulin G and Immunoglobulin M, International Autoimmune Hepatitis Group score, positive rate of anti-smooth muscle antibodies, and the rate of definite AIH were higher in AIH patients with DRB1*04:05 than without. Conclusions The important roles of specific combinations of DRB1 and DQB1 alleles or haplotypes in the pathogenesis of type 1 AIH were suggested. The association of DR4/DR8 heterozygous genotype suggested the pathologic importance of trans-complementing DQα-β heterodimer molecules encoded by DQA1 allele of one haplotype and the DQB1 allele of the other haplotype, as it was proposed in the HLA association studies of Type 1 diabetes.


Introduction
Autoimmune hepatitis (AIH) is a very rare chronic progressive liver disease with autoimmune features [1,2,3]. Type 1 AIH is characterized by the presence of serum anti-nuclear antibodies (ANA) or anti-smooth muscle antibodies (ASMA) and type 2 AIH by type 1 liver-kidney microsomal antibodies. AIH is composed predominantly of type 1 in Japanese populations. Although the disease etiology is uncertain, it is considered that the genetic and environmental factors are associated with the pathogenesis of AIH. Many studies including a recent genomewide association study [4] showed the genetic association of AIH with genes located within human leukocyte antigen (HLA) region. HLA-DRB1 Ã 03:01 and Ã 04:01 are associated with AIH in European populations [5]; Ã 04:05 is associated in Japanese and Korean populations [6,7,8,9]. In addition, several studies have shown that DRB1 Ã 04:04, Ã 04:05, and Ã 13:01 are associated with AIH in Latin America [10,11,12,13]. DRB1 Ã 08 alleles are also reported to be associated with AIH in Indian and Iranian, but not in Pakistani populations [14,15,16]. On the other hands, DRB1 Ã 15:01 is protective for the susceptibility of AIH in European and Japanese populations [5,6]. DRB1 Ã 13:02, which differ by one amino acid residue from DRB1 Ã 13:01, is protectively associated with AIH in Latin America [11,13,17] and in Japan [8].
It was reported in the genome-wide association study that HLA is the sole strong genetic factor for the susceptibility of type 1 AIH [4]. The HLA region was scanned and the most important loci for the susceptibility of type 1 AIH was reported to be DRB1 [18]. It was suggested that no other genes in the HLA region are associated with type 1 AIH. However, DRB1 is in strong linkage disequilibrium with DQB1 and it is difficult to differentiate the role of DRB1 and DQB1 in the pathogenesis of type 1 AIH. Although HLA alleles are known to confer the risk for various autoimmune diseases, the precise mechanisms have not sufficiently been revealed. The risk alleles are different in these autoimmune diseases [19]. It was considered that different auto-antigens are presented by different disease-specific risk alleles; the presented auto-antigens are restricted by HLA alleles and are influenced by non-HLA genes, environmental factors, or precipitating events. The complex of auto-antigens and risk alleles stimulate self-reactive T cells, resulting in the eliciting of diseases [20]. In this study, we conducted an HLA association study in order to search HLA alleles or haplotypes predisposing or protective for Japanese AIH.

Patients and healthy controls
Three hundred sixty type 1 AIH patients were enrolled from the register of Japanese National Hospital Organization Liver Registry [21]. The AIH patients without any other types of liver diseases satisfied the criteria of International Autoimmune Hepatitis Group (IAIHG) for diagnosis of type I AIH [22]. The healthy controls (n = 1026; mean age ± SD, 37.7 ± 11.7 years, 303 male [29.8%]) were recruited at Sagamihara Hospital, the University of Tokyo, Teikyo University, and Kanazawa University [23,24] or by the Pharma SNP Consortium (Tokyo, Japan) [25]. All the patients and the healthy individuals were native Japanese living in Japan. The study was reviewed and approved by University of Tsukuba Research Ethics Committee, Nagasaki University Research Ethics Committee, and the NHO central Institutional Review Board. Informed consents in writing were obtained from all the participants. The study was performed in accordance with the principles expressed in the Declaration of Helsinki.

Statistical analysis
Differences of AIH characteristics were analyzed by Mann-Whitney's U test or Fisher's exact test using 2x2 contingency tables. Association of allele carrier frequencies, haplotype carrier frequencies, or amino acid residue carrier frequencies was analyzed by Fisher's exact test using 2x2 contingency tables under the dominant model. Differences of genotype frequencies or diplotype (the specific combination of DRB1-DQB1 haplotypes) frequencies were analyzed by Fisher's exact test using 2x2 contingency tables. Adjustment for multiple comparisons was conducted with Bonferroni method; corrected P (Pc) values were calculated by multiplying the P value by the number of alleles or amino acid residues tested.

Clinical features of type I AIH patients
Characteristics of the type I AIH patients are shown in Table 1. Among 360 AIH patients, 314 (87.2%) were positive for ANA, 118 (38.6%) were positive for ASMA. Of overall AIH, 227 (63.1%) were definite AIH.

HLA-DRB1 in type I AIH
To compare HLA-DRB1 allele carrier frequency of the AIH patients and the healthy controls, we performed HLA-DRB1 genotyping (  (Table 3). Serum levels of Immunoglobulin G (IgG), Immunoglobulin M (IgM), and IAIHG score were higher in AIH patients with DRB1 Ã 04:05 than without. Positive rate of ASMA and the rate of definite AIH were higher in AIH patients with DRB1 Ã 04:05 than without. The complication rate of cirrhosis tended to be higher in AIH patients with DRB1 Ã 13:02 than without. Thus, specific clinical features of AIH patients possessing DRB1 Ã 04:05 were observed.

HLA-DRB1 genotype in type I AIH
We investigated the genotype frequency in the AIH patients (   Certain amino acid residues in HLA-DRβ chains were associated with AIH The association with AIH with respect to each amino acid residue in the HLA-DRβ chain was analyzed. The amino acid residues of 11V, 13H, 33H, 57S, and 96Y in the DRβ chain showed associations with AIH (Fig 1). Thus, this association analysis suggested roles for specific amino acid residues in the HLA-DRβ chain.

Certain amino acid residues in HLA-DQβ chains were associated with AIH
The association with AIH with respect to each amino acid residue in the HLA-DQβ chain was analyzed in the comparison with the 413 healthy controls. The amino acid residues of 23L, 56L, 70E, and 71D in the DQβ chain showed associations with AIH (Fig 2). When each amino acid residue frequency in the DRβ chain for the 413 healthy controls was compared with that of the AIH patients, similar tendencies were observed (S1 Fig). Thus, this association analysis suggested roles for specific amino acid residues in the HLA-DQβ chains.

Discussion
Several studies have reported that type 1 AIH is associated with HLA-DRB1 Ã 03:01 and DRB1 Ã 04:01 in European [5] and DRB1 Ã 04:05 in Japanese populations (Fig 3) [6,7,8]. In the present study, we showed an association of Japanese AIH with DRB1 Ã 04:01 and Ã 04:05, indicating the common predisposing DRB1 Ã 04:01 allele for AIH between European and Japanese populations. DRB1 Ã 04:05 is also common predisposing allele for AIH between Latin America [13] and Japan. In previous studies, DRB1 Ã 13:02 was protectively associated with type 1 AIH in Latin America [11,13,17]. We also confirmed a protective association of DRB1 Ã 13:02 with Japanese AIH [8], but could not replicate the protective effects of DRB1 Ã 15:01 [5,6]. These data indicated that the common protective DRB1 Ã 13:02 allele for AIH between Latin America and Japan is also the protective allele shared by multiple autoimmune diseases [19]. Specific demographic features of Japanese AIH patients with DRB1 Ã 04:05 were observed (Table 3). Elevated serum levels of IgG and IgM were detected in AIH patients with DRB1 Ã 04:05, as it was previously described [7]. In the present study, the IAIHG score, the positive rate of ASMA, and the rate of definite AIH were newly found to be higher in Japanese AIH patients  [26,27] and DRB1 Ã 04:05 in Japanese populations [23,28] are associated with the susceptibility for rheumatoid arthritis (RA) and type 1 diabetes, in an analogous fashion to AIH. RA is a systemic autoimmune disease that affects synovial joints. RA-susceptible DRB1 alleles shared a conserved amino acid sequence at position 70-74 (QKRAA, RRRAA, or QRRAA) in HLA-DRβ chain and were designated as shared epitope alleles [23,26]. The shared epitope alleles include DRB1 Ã 01:01, Ã 04:01, Ã 04:04, Ã 04:05, Ã 04:10, Ã 10:01, Ã 14:02, and Ã 14:06. However, neither DRB1 Ã 01:01 nor Ã 04:10 seems to be a risk allele for AIH (Table 2). In the associations of DRB1 alleles with susceptibility to RA, a gene dosage effect was reported; homozygosity for predisposing DRB1 alleles confers higher OR than heterozygosity. However, we could not find any gene dosage effects of predisposing alleles or haplotypes in AIH. These data suggested the differential roles of DRB1 in the pathogenesis between AIH and RA.