HLA and CTLA4 polymorphisms may confer a synergistic risk in the susceptibility to Graves’ disease

Abstract

Graves’ disease (GD) is an autoimmune disease characterized by hyperthyroidism due to the presence of autoantibodies against thyroid-stimulating hormone receptor, which is measured as thyroid-stimulating hormone-binding inhibitory immunoglobulin (TBII). Most of the GD patients are TBII-positive, but TBII is undetectable in a proportion of GD patients. We previously reported the association of HLA-A^*02 and -DPB1^*0501 with TBII-positive GD, whereas TBII-negative GD showed association with HLA-A^*02 and DPB1^*0202. Recently, polymorphisms of cytotoxic T-lymphocyte-associated antigen-4 (CTLA4) gene are reported to be associated with GD. In this study, we investigated 329 (240 TBII-positive and 89 TBII-negative) GD patients and 378 controls for the polymorphisms in HLA-A, -DPB1 and CTLA4 (CT60, rs3087243, A/G) to investigate the contribution of these factors in the susceptibility to GD. A significant association with CTLA4 was found for the TBII-positive GD (G carriers in patients vs controls, 97.1 vs 91.8%; odds ratio (OR)=2.97, 95% confidence interval=1.29–6.87, P=0.008), but the association was weak and not significant for the TBII-negative GD (94.4 vs 91.8%; OR=1.50, 95% confidence interval=0.57–3.98, P=0.41). Stratification analyses suggested a possible synergistic interaction of CTLA4 with HLA-A^*02 and -DPB1^*0501 in the susceptibility to TBII-positive GD.

Introduction

Graves’ disease (GD) is a typical organ-specific autoimmune disease affecting thyroid gland, in which autoantibodies against the receptors for thyroid-stimulating hormone induce hyperthyroidism. The autoantibodies has been measured as thyroid-stimulating antibodies (TSAb) by a biological assay or as thyroid-stimulating hormone-binding inhibitory immunoglobulins (TBII) by a solid phase assay.¹ Although most of the GD patients are positive for both TBII and TSAb, a proportion of the patients carry low or undetectable level of TBII (TBII-negative GD).² However, TSAb can be detected in the TBII-negative GD, and it was reported that the TBII-negative patients exhibited mild goiter and followed benign prognosis, implying that the TBII-negative GD might be a clinical subtype different from the ordinary TBII-positive GD.³

Etiological mechanisms of GD are not fully understood, but certain genetic factors should be involved in the pathogenesis because of familial aggregation of the disease.⁴ To decipher the genetic factors candidate gene approaches have been used, and it is well documented that human leukocyte antigen (HLA) polymorphisms are associated with GD, such that HLA-B^*08-DRB1^*03-DQA1^*05-DQB1^*02 haplotype showed strong association in European populations.^{5, 6} However, the distribution of HLA alleles and haplotypes are quite different in different ethnic groups, and the European GD-associated HLA haplotype is virtually absent in Japanese.⁷ We previously reported that HLA-A^*02 and DPB1^*0501 are associated with Japanese GD.⁸ In addition, we demonstrated that HLA-A^*02 and DPB1^*0202 showed association with TBII-negative GD, indicating that the TBII-negative GD was different in HLA-linked genetic factors from the TBII-positive GD.⁹

There are several other genetic factors associated with GD, including polymorphisms in the gene for cytotoxic T-lymphocyte-associated antigen-4 (CTLA4).¹⁰ CTLA4 encodes a co-stimulatory molecule of T cells, which is involved in the regulation of T-cell activation,^{11, 12} and it has been reported that the polymorphisms in exon 1 and 3′-UTR region are associated with GD as well as Hashimoto thyroiditis, another autoimmune thyroid disease, in European and Asian populations.^{10, 13, 14, 15} However, it is not clear how the CTLA4 and HLA polymorphisms would confer the risk to GD; that is, they operate independently or synergistically in determining the genetic risk. To clarify the issue, we analyzed a CTLA4 polymorphism (rs3087243, also called as CT60) in addition to HLA-A^*02, DPB1^*0202 and DPB1^*0501 in association with GD.

Materials and methods

Subjects

A total of 329 Japanese GD patients (240 TBII-positives and 89 TBII-negatives) and 378 randomly selected Japanese healthy controls were the subjects. The TBII-negative patients were randomly selected from the previously reported 97 patients,⁹ whereas the TBII-positive patients were those analyzed previously⁸ (n=142) and newly recruited patients (n=98). The diagnosis of GD and assay for TBII were performed as described previously.^{2, 3, 9} Blood sample to prepare genomic DNA was obtained from each subject under a given informed consent. The study protocol was approved by the Ethics Reviewing Committee of Medical Research Institute, Tokyo Medical and Dental University.

Genotyping

Detection of HLA-A^*02 and HLA-DPB1 genotyping was carried out as described previously.⁹ The CTLA4 polymorphism (CT60, rs3087243) was analyzed by the PCR-restriction-fragment length polymorphism method using HpyCh4IV.^{10, 16}

Statistical analysis

Frequencies of alleles and genotypes were compared between the patients and controls using χ²-test. Strength of the association was expressed by odds ratio (OR). A stratification analysis was carried as described by Svejgaard and Ryder¹⁷ to test the independency or synergistic effects of two different alleles. P-values were corrected for multiple testing by the number of tested markers, HLA-A^*02, DPB1^*0501, DPB1^*0202 and CTLA4-CT60-G (n=4). The association was considered to be significant when the corrected P-value was less than 0.05.

Results and Discussion

Table 1 shows the carrier frequencies of HLA-A^*02, -DPB1^*0501 and-DPB1^*0202 in the studied populations, demonstrating that the frequencies of disease-associated HLA carriers were significantly high in the patients. When the GD patients were divided into TBII-positive and -negative groups, HLA-A^*02 and DPB1^*0501 were associated with TBII-positive GD, although the association of HLA-A^*02 was weak and marginal because it lost the significance after the Bonferroni’s correction for multiple testing. In contrast, HLA-A^*02 and DPB1^*0202 were significantly associated with TBII-negative GD.

Table 1 Frequencies of HLA-A^*02, -DPB1^*0501, -DPB1^*0202 and CTLA4-CT60-G in the Japanese populations

The analysis of a CTLA4 polymorphism CT60 demonstrated that the association with G allele was strong and significant for TBII-positive GD (OR=2.97, 95% confidence interval=1.29–6.87, P=0.031), but it was weak and not significant in TBII-negative GD (OR=1.50, 95% confidence interval=0.57–3.98, P=nonsignificant) (Table 1). Because TSAb can be detected in almost all of TBII-positive GD patients, while only a part of TBII-negative GD patients were positive for TSAb, the observation that the association with CT60-G allele was weak and not significant for the TBII-negative GD might be due to that CT60-G allele was associated with the presence of TSAb. Among the TBII-negative GD patients, 31 (34.8%) were TSAb-positive, whereas TSAb was not detected in the other 58 patients. Frequency of CTLA4-CT60-G carrier was 93.5% in the TSAb-positive patients, whereas it was 94.8% in the TSAb-negative patients. These findings suggested that the CTLA4-CT60-G allele was not associated with the presence of TSAb. We previously reported that the absence of TBII might be a predictor of good prognosis, because the efficacy of medication and the disease-associated HLA-DPB1 allele was different from the typical GD positive for TBII.⁸ In this study, we also demonstrated that the association with CT60 was weak for TBII-negative GD, implying that the contribution of CTLA4 in the pathogenesis was relatively small in the TBII-negative GD, further supporting that the absence of TBII was a good predictor for prognosis of GD.

We next performed a stratification analysis of three different GD-associated alleles, HLA-A^*02, DPB1^*0501 and CTLA4-CT60-G, in the TBII-positive GD (Table 2). It was found that the ORs of HLA-DPB1^*0501 were 3.28 and 3.30 in the presence and absence of CTLA4-CT60-G, respectively, suggesting that CTLA4-linked genetic factor would increase the HLA-DPB1-linked risk (Table 2b-1, test {3} and {4}). Similarly, ORs of CTLA4-CT60-G were 3.03 and 3.05 in the presence and absence of HLA-DPB1^*0501, respectively, showing that HLA-DPB1-linked genetic factor would increase the CTLA4-linked risk (Table 2b-1, test {5} and {6}). In addition, the presence of both HLA-DPB1^*0501 and CTLA4-CT60-G conferred OR of 9.99 (Table 2b-1, test {8}), which was much higher than HLA-DPB1^*0501 or CTLA4-CT60-G alone (OR=−3.30 or 3.05, Table 2b-1, test {4} or {6}, respectively), although it was not significant because the 95% confidence interval of ORs for each category showed a considerable overlapping as shown in Table 2b-1. These observations suggested a possible synergistic role of HLA-DPB1^*0501 and CTLA4-CT60-G in the susceptibility to TBII-positive GD. It was also found that HLA-A^*02 and CTLA4-CT60-G conferred a possible synergistic risk, albeit to a less extent, because the presence of both HLA-A^*02 and CTLA4-CT60-G conferred OR of 3.75 (Table 2b-2, test {8}), which was much higher than HLA-A^*02 or CTLA4-CT60-G alone (OR=1.04 or 2.50, Table 2b-2, test {4} or {6}, respectively), although it was not statistically significant because of the overlapping of 95% confidence interval of ORs (Table 2b-2). Because there is no linkage disequilibrium between HLA-A^*02 and DPB1^*0501,⁷ the findings in this study implied a synergistic contribution of HLA-linked factors and CTLA4-linked factor in the susceptibility.

Table 2 Stratification analysis of HLA and CTLA4 in GD