HLA-A and DPB 1 Loci Confer Susceptibility to Graves' Disease Rui-Ping Dong, Akinori Kimura, Ryoko Okubo, Hirotoshi Shinagawa, Hajime Tamai, Yasuharu Nishimura, and Takehiko Sasazuki
ABSTRACT: To investigate HLA-linked genetic factors
involved in the pathogenesis of Graves' disease, 76 patients and 317 healthy controls in the Japanese population were examined for HLA-A, B, C, DR, and DQ specificities by serologic typing and for HLA-DPB1 alleles by DNA typing by using the PCR-SSOP method. The frequencies of HLA-A2, B46, Cw11, and DPBI*0501 were increased and those of HLA-A24, B7, Bw52, and DR1 were decreased in the patients. The increased frequencies of HLA-A2 and DPB 1"0501 in the patients were statistically
significant when the corrected p value (Pc) was applied (Pc ~ 0.02 and p, ~ 0.002, respectively). ORs for a risk to develop the disease were calculated among individuals positive for DPB 1"0501 and/or HLA-A2, and the highest OR (10.5) was observed in individuals possessed both DPBI*0501 and HLA-A2. This observation suggests a synergic involvement of a HLA class II allele (DPBI*0501) and an HLA class I allele (HLA-A2) in the pathogenesis of Graves' disease. Human Immunology 35, 165-172 (1992)
ABBREVIATIONS
AHVR IDDM OR PCR
allelic hypervariable region insulin-dependent diabetes mellitus odds ratio polymerase chain reaction
RA RR RFLP SSOP
rheumatoid arthritis relative risk restriction fragment length polymorphism sequence-specific oligonucleotide probe
INTRODUCTION Graves' disease is a typical organ-specific autoimmune disease and genetic factors are suggested to be involved in its pathogenesis, based on the existence of multiplex families [ 1 - 3 ] and the difference in concordant rate of the disease between monozygotic and dizygotic twins [4]. The major histocompatibility complex has been shown to control both immune responsiveness to natural antigens and susceptibility to some diseases in humans and in experimental animals [ 5 - 7 ] , and the association between the disease and HLAs has been reported in various ethnic groups, e.g., HLA-A1, B8, and DR3 in
From the Department of Genetics (R.P.D., A.K., R.O., H.S., Y.N., T.S.), Medical Institute of Bioregulation; and the Department of Psychosomatic Medicine (H. T.), Faculty of Medicine, Kyushu University, Fukuoka, Japan. Address reprint requests to Dr. T. Sasazuki, Department of Genetics, Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812, Japan. ReceivedJuly 14, 1992," accepted September 18, 1992. Human Immunology 35, 165-172 (1992) © American Society for Histocompatibility and Immunogenetics, 1992
Caucasians [8], B46 and DR9 in Chinese [9], and A10, B8, and DQw2 in Asian Indians [10]. In the Japanese population, HLA-B35, B5 [11], B46 [12], and A l l , B46, B48, DR8, DQ4, and DPw2 [13] were reported to be associated with the disease. On the other hand, we have previously reported that the susceptibility to Graves' disease is controlled by at least two major genes linked to HLA and Gm, respectively [2, 14, 15]. These observations indicate the involvement of HLAlinked genetic factor(s) in the pathogenesis of Graves' disease. The HLA genes have been analyzed by serologic typing and it is known that the typing of class II HLAs, especially HLA-DP, is sometimes difficult due to the limited availability of well-defined serologic or cellular reagents, the presence of autoantibodies, and medications. This difficulty can now be overcome by HLA typing at the D N A level by using the polymerase chain reaction (PCR)[16-18]. Because PCR-sequence-specific oligonucleotide probe (SSOP) analysis reveals not 165 0198-8859/92/$5.00
R.-P. Dong et al.
166
only that two alleles are different, but how they differ, specific amino acid residues of the HLAs, which are critical for the susceptibility to Graves' disease, will be identified by this method. We report here the analysis of 76 Japanese patients with Graves' disease for HLA-A, B, C, DR, and DQ by serologic typing and for the HLA-DPB1 gene by the PCR-SSOP typing to investigate further the HLAlinked genetic factor(s). MATERIALS A N D M E T H O D S
Patients. A total of 76 unrelated Japanese patients with Graves' disease (11 men and 65 women; mean age, 32.4 years; range, 11-58 years) were the subjects of this study. Graves' disease was diagnosed on the basis of clinical symptoms, thyroid function tests, and I23I uptake. As the control population, 317 healthy unrelated Japanese were used. Serologic typing. A panel of highly selected alloantisera in the Ninth Japan HLA Workshop or the Tenth International Histocompatibility Workshop [19] were used to identify HLA-A, B, C, DR, and DQ specificities by using a standard complement-dependent microcytotoxicity assay [20]. D N A typing of DPBI alleles by PCR-SSOP analysis. DNA was extracted from peripheral granulocytes of each subject as described previously [21]. Genomic DNA was subjected to 30 cycles of PCR in a thermal cycler (Perkin Elmer Cetus, Norwalk, CT) to amplify the second exon of DPB1 gene, using thermostable DNA
polymerase (Ampli Taq, Perkin Elmer Cetus, obtained via Takara, Kyoto). The primers and SSOPs were those used in the DNA component of the 11th International HLA Workshop [22]. Conditions of PCR and procedures of hybridization with SSOPs were described previously [21]. The allele designations are those from WHO HLA Nomenclature Committee [23].
Statistical analysis. The frequencies of HLA-A, B, C, DR, DQ, and DPB1 alleles in 76 patients with Graves' disease were compared with those in 317 healthy unrelated Japanese controls. Strength of the statistical association between Graves' disease and a genetic marker was expressed by the relative risk (RR) as given by Woolf [24] or by the odds ratio (OR) [25], and the statistical significance was examined by the chi-squared test with Yates' correction. The corrected p value (Pc) was calculated by multiplying the p value by 108, the number of alleles tested at class I and class II loci. Linkage disequilibrium between given HLA alleles was tested by the method of Mittal [26]. RESULTS
Serologic HLA typing. Among the HLA-A, B, C, DR, and DQ specificities tested, the frequencies of A2(67.1% vs 41.6%, phenotype frequency in the patients versus that in the controls; X2 = 15.98, p < 0.0001, RR = 2.86), B46 (21.1% vs 10.7%; X 2 -- 5.89, p < 0.02, RR = 2.22), and C w l l (15.8% vs 7.3%; X2 -- 5.50, p < 0.03, RR = 2.94) showed significant positive association with the disease (Table 1). The increased frequency of HLAB46 is consistent with the previous report [12, 13],
TABLE 1 Association between Graves' disease and HLA determined by serologic typing % Patients (n = 76)
% Controls (n = 317)
Relative risk
X2
pb
A2 A24
67.1 47.4
41.6 61.5
2.86 0.56
15.98 5.06
ABSTRACT: To investigate HLA-linked genetic factors
involved in the pathogenesis of Graves' disease, 76 patients and 317 healthy controls in the Japanese population were examined for HLA-A, B, C, DR, and DQ specificities by serologic typing and for HLA-DPB1 alleles by DNA typing by using the PCR-SSOP method. The frequencies of HLA-A2, B46, Cw11, and DPBI*0501 were increased and those of HLA-A24, B7, Bw52, and DR1 were decreased in the patients. The increased frequencies of HLA-A2 and DPB 1"0501 in the patients were statistically
significant when the corrected p value (Pc) was applied (Pc ~ 0.02 and p, ~ 0.002, respectively). ORs for a risk to develop the disease were calculated among individuals positive for DPB 1"0501 and/or HLA-A2, and the highest OR (10.5) was observed in individuals possessed both DPBI*0501 and HLA-A2. This observation suggests a synergic involvement of a HLA class II allele (DPBI*0501) and an HLA class I allele (HLA-A2) in the pathogenesis of Graves' disease. Human Immunology 35, 165-172 (1992)
ABBREVIATIONS
AHVR IDDM OR PCR
allelic hypervariable region insulin-dependent diabetes mellitus odds ratio polymerase chain reaction
RA RR RFLP SSOP
rheumatoid arthritis relative risk restriction fragment length polymorphism sequence-specific oligonucleotide probe
INTRODUCTION Graves' disease is a typical organ-specific autoimmune disease and genetic factors are suggested to be involved in its pathogenesis, based on the existence of multiplex families [ 1 - 3 ] and the difference in concordant rate of the disease between monozygotic and dizygotic twins [4]. The major histocompatibility complex has been shown to control both immune responsiveness to natural antigens and susceptibility to some diseases in humans and in experimental animals [ 5 - 7 ] , and the association between the disease and HLAs has been reported in various ethnic groups, e.g., HLA-A1, B8, and DR3 in
From the Department of Genetics (R.P.D., A.K., R.O., H.S., Y.N., T.S.), Medical Institute of Bioregulation; and the Department of Psychosomatic Medicine (H. T.), Faculty of Medicine, Kyushu University, Fukuoka, Japan. Address reprint requests to Dr. T. Sasazuki, Department of Genetics, Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812, Japan. ReceivedJuly 14, 1992," accepted September 18, 1992. Human Immunology 35, 165-172 (1992) © American Society for Histocompatibility and Immunogenetics, 1992
Caucasians [8], B46 and DR9 in Chinese [9], and A10, B8, and DQw2 in Asian Indians [10]. In the Japanese population, HLA-B35, B5 [11], B46 [12], and A l l , B46, B48, DR8, DQ4, and DPw2 [13] were reported to be associated with the disease. On the other hand, we have previously reported that the susceptibility to Graves' disease is controlled by at least two major genes linked to HLA and Gm, respectively [2, 14, 15]. These observations indicate the involvement of HLAlinked genetic factor(s) in the pathogenesis of Graves' disease. The HLA genes have been analyzed by serologic typing and it is known that the typing of class II HLAs, especially HLA-DP, is sometimes difficult due to the limited availability of well-defined serologic or cellular reagents, the presence of autoantibodies, and medications. This difficulty can now be overcome by HLA typing at the D N A level by using the polymerase chain reaction (PCR)[16-18]. Because PCR-sequence-specific oligonucleotide probe (SSOP) analysis reveals not 165 0198-8859/92/$5.00
R.-P. Dong et al.
166
only that two alleles are different, but how they differ, specific amino acid residues of the HLAs, which are critical for the susceptibility to Graves' disease, will be identified by this method. We report here the analysis of 76 Japanese patients with Graves' disease for HLA-A, B, C, DR, and DQ by serologic typing and for the HLA-DPB1 gene by the PCR-SSOP typing to investigate further the HLAlinked genetic factor(s). MATERIALS A N D M E T H O D S
Patients. A total of 76 unrelated Japanese patients with Graves' disease (11 men and 65 women; mean age, 32.4 years; range, 11-58 years) were the subjects of this study. Graves' disease was diagnosed on the basis of clinical symptoms, thyroid function tests, and I23I uptake. As the control population, 317 healthy unrelated Japanese were used. Serologic typing. A panel of highly selected alloantisera in the Ninth Japan HLA Workshop or the Tenth International Histocompatibility Workshop [19] were used to identify HLA-A, B, C, DR, and DQ specificities by using a standard complement-dependent microcytotoxicity assay [20]. D N A typing of DPBI alleles by PCR-SSOP analysis. DNA was extracted from peripheral granulocytes of each subject as described previously [21]. Genomic DNA was subjected to 30 cycles of PCR in a thermal cycler (Perkin Elmer Cetus, Norwalk, CT) to amplify the second exon of DPB1 gene, using thermostable DNA
polymerase (Ampli Taq, Perkin Elmer Cetus, obtained via Takara, Kyoto). The primers and SSOPs were those used in the DNA component of the 11th International HLA Workshop [22]. Conditions of PCR and procedures of hybridization with SSOPs were described previously [21]. The allele designations are those from WHO HLA Nomenclature Committee [23].
Statistical analysis. The frequencies of HLA-A, B, C, DR, DQ, and DPB1 alleles in 76 patients with Graves' disease were compared with those in 317 healthy unrelated Japanese controls. Strength of the statistical association between Graves' disease and a genetic marker was expressed by the relative risk (RR) as given by Woolf [24] or by the odds ratio (OR) [25], and the statistical significance was examined by the chi-squared test with Yates' correction. The corrected p value (Pc) was calculated by multiplying the p value by 108, the number of alleles tested at class I and class II loci. Linkage disequilibrium between given HLA alleles was tested by the method of Mittal [26]. RESULTS
Serologic HLA typing. Among the HLA-A, B, C, DR, and DQ specificities tested, the frequencies of A2(67.1% vs 41.6%, phenotype frequency in the patients versus that in the controls; X2 = 15.98, p < 0.0001, RR = 2.86), B46 (21.1% vs 10.7%; X 2 -- 5.89, p < 0.02, RR = 2.22), and C w l l (15.8% vs 7.3%; X2 -- 5.50, p < 0.03, RR = 2.94) showed significant positive association with the disease (Table 1). The increased frequency of HLAB46 is consistent with the previous report [12, 13],
TABLE 1 Association between Graves' disease and HLA determined by serologic typing % Patients (n = 76)
% Controls (n = 317)
Relative risk
X2
pb
A2 A24
67.1 47.4
41.6 61.5
2.86 0.56
15.98 5.06
