The Association of CTLA4 Polymorphism with Type 1 Diabetes Is ...

0021-972X/06/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 91(3):1087–1092 Copyright © 2006 by The Endocrine Society doi: 10.1210/jc.2005-1407

The Association of CTLA4 Polymorphism with Type 1 Diabetes Is Concentrated in Patients Complicated with Autoimmune Thyroid Disease: A Multicenter Collaborative Study in Japan Hiroshi Ikegami, Takuya Awata, Eiji Kawasaki, Tetsuro Kobayashi, Taro Maruyama, Koji Nakanishi, Akira Shimada, Shin Amemiya, Yumiko Kawabata, Susumu Kurihara, Shoichiro Tanaka, Yasuhiko Kanazawa, Mie Mochizuki, and Toshio Ogihara, on behalf of the Japanese Study Group on Type 1 Diabetes Genetics* Department of Geriatric Medicine (H.I., Y.Kaw., T.O.), Osaka University Graduate School of Medicine, Yamadaoka, Suita, Osaka 565-0871, Japan; Division of Endocrinology and Diabetes, Department of Medicine (T.A., S.K.), Saitama Medical School, Saitama 350-0495, Japan; Department of Metabolism/Diabetes and Clinical Nutrition (E.K.), Nagasaki University Hospital of Medicine and Dentistry, Nagasaki 852-8501, Japan; Third Department of Internal Medicine (T.K., S.T.) and Department of Pediatrics (S.A., M.M.), Interdisciplinary Graduate School of Medical and Engineering, University of Yamanashi, Yamanashi 409-3898, Japan; Department of Internal Medicine (T.M.), Saitama Social Insurance Hospital, Saitama 330-0074, Japan; Department of Endocrinology and Metabolism (K.N.), Toranomon Hospital, Tokyo 105-8470, Japan; and Department of Internal Medicine (A.S., Y.Kan.), Keio University School of Medicine, Tokyo 160-8582, Japan Context: Transracial studies are a powerful tool for genetic association studies of multifactorial diseases, such as type 1 diabetes. The low incidence of type 1 diabetes in Asian countries, however, makes it difficult to perform large-scale studies in Asia. Objective: To overcome this, we have assembled a multicenter study group in Japan and studied the association of CTLA4 polymorphisms with type 1 diabetes relative to autoimmune thyroid disease (AITD) phenotypes. Subjects: Subjects included a total of 1837 samples, including 1114 cases (769 with type 1 diabetes and 345 with AITD) and 723 control subjects. Methods: The ⫹6230G⬎A and ⫹49G⬎A polymorphisms of CTLA4 as well as HLA-DRB1 and -DQB1 were genotyped.

T

YPE 1 DIABETES IS caused by autoimmune destruction of insulin-producing ␤-cells of the pancreas in genetically susceptible individuals (1). Susceptibility to type 1 diabetes is determined by multiple genes, with IDDM1 in the human leukocyte antigen (HLA) region showing the strongest effect (1). Although HLA is a major component in type 1 diabetes susceptibility, non-HLA genes are also necessary. More than 10 susceptibility loci have been mapped outside the HLA region by genome scan (2, 3) and/or candidate gene

First Published Online December 13, 2005 * For names of members of the Japanese Study Group on Type 1 Diabetes Genetics, see Acknowledgments. Abbreviations: AITD, Autoimmune thyroid disease; CI, confidence interval; HLA, human leukocyte antigen; NS, not significant; SNP, single nucleotide polymorphism; Tg, thyroglobulin; TPO, thyroid peroxidase. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

Results: The ⫹6230G⬎A polymorphism was significantly associated with type 1 diabetes complicated with AITD (odds ratio, 1.54; P ⫽ 0.027) and with AITD alone (odds ratio, 1.31; P ⫽ 0.045) but not with type 1 diabetes without AITD. The association with type 1 diabetes positive for autoantibodies to both pancreatic islets and thyroid was particularly strong (odds ratio, 1.87; P ⫽ 0.001). Type 1 diabetic patients with the disease-associated GG genotype were characterized by a significantly higher frequency of AITD (P ⫽ 0.013), of positivity for both AITD and antiislet autoantibody (P ⫽ 0.00086), and of highrisk HLA genotypes (P ⫽ 0.034). Conclusions: Given the high frequency of AITD in patients with type 1 diabetes, these data suggest the possibility that the association of CTLA4 with type 1 diabetes in previous studies may have been secondary to AITD, suggesting the importance of subclassification of type 1 diabetes relative to AITD in genetic studies. (J Clin Endocrinol Metab 91: 1087–1092, 2006)

approaches (4 – 6). Most of these loci, however, showed only a modest effect (2, 3), making it difficult to fine-map and identify disease-causing variants in small-scale studies. To fine-map and identify disease-causing variants with a modest effect, two tools are required: a large number of samples from affected and unaffected subjects and a panel of haplotypes consisting of different combinations of alleles at nearby loci. Recent large-scale studies have clearly indicated the importance of a large number of samples in the genetic dissection of type 1 diabetes with sufficient statistical power (3). Almost all large-scale studies to date, however, have been performed in Caucasian populations, making it sometimes difficult to genetically dissect type 1 diabetes even with largescale studies because of the limited number of haplotypes within Caucasian populations. Clearly, large-scale studies in populations other than Caucasians, such as Japanese, are important for genetic dissection of multifactorial diseases in general and for identification of non-HLA genes in type 1

1087

1088

J Clin Endocrinol Metab, March 2006, 91(3):1087–1092

Ikegami et al. • CTLA4 Is Associated with Type 1 DM plus AITD

diabetes in particular. Despite the importance of large-scale studies in Japanese, the number of Japanese samples in previous studies has been limited because of the very low incidence (less than 1/10 of that in Caucasians) of type 1 diabetes in Japanese (7). To overcome this limitation, we have assembled a multicenter study group in which each member had previously collected moderate sized samples (200 –300) and had experience in genetic association studies on type 1 diabetes. Seven such leading groups in the field of genetics of type 1 diabetes in Japan agreed for the first time to collaborate to perform a large-scale study on susceptibility to type 1 diabetes, and a total of up to 1500 type 1 diabetic patients and control subjects were combined for the collaborative effort. As the first step in using these valuable resources, we studied the association of CTLA4, which has recently been reported to be responsible for IDDM12, a non-HLA susceptibility gene for type 1 diabetes on chromosome 2q33 (6). The results demonstrated that the association of CTLA4 with type 1 diabetes is limited to a subgroup of patients complicated with autoimmune thyroid disease (AITD). Subjects and Methods Subjects A total of 1837 subjects, including 1114 cases and 723 control subjects, were studied. The study was approved by the ethical committee of each institute, and informed consent was obtained from subjects. Cases consisted of 769 patients with type 1 diabetes, including 172 patients with both type 1 diabetes and AITD and 345 patients with AITD alone (AITD without type 1 diabetes). The clinical characteristics of the subjects are shown in Table 1. The diagnosis of type 1 diabetes was made by endocrinologists, based on both clinical features and laboratory data. All the patients were ketosis-prone, required insulin injections to sustain their life, and lacked endogenous insulin secretion as judged by urinary C-peptide levels of less than 3.3 nmol/d and/or were positive for antiislet autoantibodies. AITD was defined as Graves’ disease, Hashimoto’s thyroiditis, or positivity for antibodies against thyroid peroxidase (TPO) and/or thyroglobulin (Tg). Graves’ disease and Hashimoto’s thyroiditis were diagnosed by endocrinologists clinically and confirmed by abnormal levels of thyroid hormones and positivity for autoantibodies to TPO, Tg, and/or TSH receptor.

Genotyping of CTLA4 polymorphisms Two single nucleotide polymorphisms (SNPs) in the CTLA4 gene, ⫹6230G⬎A (CT60 in Ref. 9, rs3087243) and ⫹49G⬎A (rs231775), were TABLE 1. Clinical characteristics of cases and control subjects

n Sex (male/female) Age at onset (yr) Islet autoantibodies (%)b AITD (total) Graves’ disease Hashimoto’s thyroiditis Autoantibodiesc

Type 1 diabetes

AITD only

Control subjects

769 343/426 27.3 ⫾ 17.3 75.6 172 44 54 74

345 58/287 40.8 ⫾ 14.1 ND 345 213 132 0

723 308/415 36.7 ⫾ 13.5a ND 0 0 0 ND

Data are mean ⫾ SD. ND, Not determined. a Age at the time of this study for control subjects. b Information on islet autoantibodies was available for 525 patients with type 1 diabetes, and 397 were positive. c Positive for antibodies against TPO and/or thyroglobulin (Tg) without clinical manifestation of thyroid dysfunction

genotyped in all samples. ⫹6230G⬎A was chosen because a very recent study with a large number of Caucasian samples showed that ⫹6230G⬎A is most strongly associated with type 1 diabetes as well as AITD and that it is also associated with variation in CTLA4 gene splicing (6). ⫹49G⬎A was chosen because it was reported to be associated with type 1 diabetes in previous small-scale studies in Japanese (8 –12), and the SNP is the only polymorphism leading to amino acid substitution in the CTLA4 molecule (6). In addition to ⫹6230G⬎A and ⫹49G⬎A, three SNPs, MH30, Jo30, and rs1863800, which were also reported to be associated with type 1 diabetes and AITD in a recent large-scale study in Caucasian populations (6), were genotyped in a subset of samples. Genotyping was performed as reported previously (6).

Genotyping of HLA Because of the well-known strong effect of IDDM1 in the HLA region on susceptibility to type 1 diabetes, DRB1 and DQB1 were genotyped in a subset of samples; 302 patients with type 1 diabetes and 239 control subjects. Samples randomly selected from all centers were used. Genotyping was performed as reported previously (13, 14). HLA haplotypes (DRB1-DQB1) were estimated based on the known HLA haplotypes in Japanese (15).

Antiislet autoantibody assays Anti-GAD antibody was detected by means of a commercially available RIA kit using 125I-labeled recombinant human GAD65 as a tracer reagent (Cosmic, Tokyo, Japan). Autoantibodies to IA-2 and ICA were determined, as described previously (16, 17). Subjects positive for autoantibodies to GAD, IA-2, and/or ICA were defined as positive for antiislet autoantibody. The timing of autoantibody testing varied from patient to patient.

Statistical methods Data are given as mean ⫾ sd. Allele and genotype frequencies were compared between patients and control subjects. Statistical analysis of the differences between groups was performed by ␹2 test. Odds ratio and its 95% confidence interval (CI) were also calculated. Haplotypes were estimated using the EM algorithm, and D’ value was calculated using Haploview version 2.03 software to evaluate linkage disequilibrium. P ⬍ 0.05 was considered statistically significant.

Results ⫹6230G⬎A polymorphism and type 1 diabetes

The ⫹6230G⬎A polymorphism of CTLA4 was not associated with type 1 diabetes as a whole [odds ratio for GG genotype, 1.08; not significant (NS)] (Table 2). Because the CTLA4 polymorphism was reported to be more strongly associated with AITD than with type 1 diabetes in a previous study (6) and because AITD is the most commonly associated autoimmune disease in patients with type 1 diabetes in Japanese (18, 19), type 1 diabetic patients were divided into two groups, those with and without AITD. The frequency of the GG genotype was significantly higher in patients with type 1 diabetes complicated with AITD than in control subjects [65.5 vs. 55.2%; odds ratio, 1.54 (95% CI, 1.09 –2.18); P ⫽ 0.027] (Table 2). The association was even stronger in type 1 diabetic patients with AITD who were also positive for antiislet autoantibodies [69.7 vs. 55.2%; odds ratio, 1.87 (1.27–2.74); P ⫽ 0.001]. The ⫹6230G⬎A polymorphism was not associated with type 1 diabetes with antiislet autoantibody when patients complicated with AITD were excluded [odds ratio, 1.05 (0.76 –1.44); NS]. To minimize the effect of the difference in the prevalence of AITD relative to the duration of type 1 diabetes, type 1 diabetic patients with duration more than 10 yr were analyzed separately. The results were essentially the same in that the frequency of the GG genotype was high in



1089

TABLE 2. Frequency of ⫹6230G⬎A polymorphism of CTLA4 in cases and controls

Type 1 diabetes All With AITD With AITD and islet autoantibodiesc AITD alone All Graves’ disease Hashimoto’s Controls

GG (%)

GA (%)

AA (%)

Oddsa (95% CI)

P

G allele

P

439 (57.2) 112 (65.5) 99 (69.7)

285 (37.2) 52 (30.4) 37 (26.1)

43 (5.6) 7 (4.1) 6 (4.2)

1.08 (0.88 –1.33) 1.54 (1.09 –2.18) 1.87 (1.27–2.75)

NSb 0.027 0.001

1163 (75.8%) 276 (80.7%) 235 (82.7%)

NS 0.027 0.005

213 (61.7) 130 (61.0) 83 (62.9) 395 (55.2)

116 (33.6) 71 (33.3) 45 (34.1) 283 (39.6)

16 (4.6) 12 (5.6) 4 (3.0) 37 (5.2)

1.31 (1.01–1.70) 1.27 (0.93–1.73) 1.37 (0.94 –2.01)

0.045 NS NS

542 (78.6%) 331 (77.7%) 211 (79.9%) 1073 (75.0%)

NS NS NS

a

For GG genotype. P ⬎ 0.05. c Islet autoantibodies (anti-GAD, IA-2, and/or ICA). b

type 1 diabetics with AITD (64.3%) and in those with AITD and islet autoantibodies (72.2%) as compared with type 1 diabetics as a whole (53.3%). The ⫹6230G⬎A polymorphism was significantly associated with AITD alone (without type 1 diabetes) [odds ratio for GG, 1.31 (1.01–1.70); P ⫽ 0.045] (Table 2). ⫹49G⬎A polymorphism and type 1 diabetes

The ⫹49G⬎A polymorphism was not significantly associated with type 1 diabetes [odds ratio, 1.08 (0.87–1.33)], even when present with antiislet autoantibody and/or AITD. The tendency for an association, however, was similar to that for ⫹6230G⬎A, in that the odds ratio for the GG genotype was higher in type 1 diabetic patients with AITD [odds ratio, 1.25 (0.89 –1.76)] and highest in type 1 diabetic patients with AITD who were positive for antiislet autoantibody [odds ratio, 1.35 (0.94 –1.95)] as compared with type 1 diabetic patients as a whole [odds ratio, 1.08 (0.87–1.33)]. The GG genotype of the ⫹49G⬎A polymorphism was significantly associated with AITD [odds ratio, 1.36; P ⫽ 0.018 (1.05–1.75)]. Interaction with HLA

Because of the well-known contribution of IDDM1 in the HLA to susceptibility to type 1 diabetes, a subset of patients (n ⫽ 302) and control subjects (n ⫽ 239) randomly selected from the subjects were typed for HLA. As reported previously, the DRB1*0405-DQB1*0401 (odds ratio, 2.14; 27.5 vs. 15.1%; P ⬍ 1 ⫻ 10⫺6) and DRB1*0901-DQB1*0303 (odds ratio, 1.84; 26.2 vs. 16.1%; P ⬍ 1 ⫻ 10⫺4) haplotypes were positively associated with type 1 diabetes in the Japanese population. The DRB1*1501-DQB1*0602 (odds ratio, 0.22; P ⬍ 1 ⫻ 10⫺4) and DRB1*1502-DQB1*0601 (odds ratio, 0.21; P ⬍ 1 ⫻ 10⫺8) haplotypes were negatively associated with type 1 diabetes, indicating that the type 1 diabetic patients in the present study had typical type 1 diabetes in Japanese in terms of the HLA association. The frequency of the GG genotype of the ⫹6230G⬎A polymorphism in CTLA4 was significantly higher in type 1 diabetic patients with highest risk HLA genotypes, i.e. possessing two doses of high-risk haplotypes in Asians, DRB1*0405 and DRB1*0901, than in those without these genotypes (64.1 vs. 51.0%; odds ratio, 1.72; P ⫽ 0.034). The frequency of the G allele was significantly higher in patients with high-risk HLA genotypes than in those without (79.3 vs. 71.1%; P ⫽ 0.036).

In patients with AITD alone, no significant association of DRB1*0405 (odds ratio, 1.09; NS) and DRB1*0901 (odds ratio, 1.06; NS) with the disease was observed. In patients with type 1 diabetes complicated with AITD, in contrast, the frequencies of the DRB1*0405 (odds ratio, 2.75) and DRB1*0901 (odds ratio, 1.98) haplotypes were higher than those in control subjects. Clinical characteristics of patients with different CTLA4 genotypes

Type 1 diabetic patients with the GG genotype at ⫹6230G⬎A as compared with other genotypes were characterized by a significantly higher frequency of AITD (25.5 vs. 18.0%; P ⫽ 0.013), a significantly higher frequency of AITD with positivity for antiislet antibody (22.6 vs. 13.1%; P ⫽ 0.00086), and a significantly higher frequency of highrisk HLA genotypes, i.e. DRB1*0405/0405, DRB1*0901/0901, or DRB1*0405/0901 (35.5 vs. 24.3%; P ⫽ 0.034). No significant differences were observed in the frequency of antiislet autoantibody (75.8 vs. 75.4%) and the age at onset of type 1 diabetes (27.0 ⫾ 17.1 vs. 27.6 ⫾ 17.7 yr) between the patients with the GG genotype and those with other genotypes. Differences between Japanese and Caucasian populations

The frequencies of alleles in control subjects were significantly different between Japanese and Caucasian populations for both the ⫹6230G⬎A and ⫹49G⬎A polymorphisms (⫹6230G⬎A/G allele, 75.0 vs. 52.3%; ⫹49G⬎A/A allele, 63.1 vs. 35.8%; P ⬍ 1 ⫻ 10⫺10 for both) (Table 3), indicating a marked difference in gene frequencies between the two populations. The ⫹49G⬎A polymorphism was in linkage disequilibrium with the ⫹6230G⬎A polymorphism (D⬘ ⫽ 0.94; 95% CI, 0.90 – 0.97). Based on the distribution of the haplotypes consisting of ⫹49G⬎A:⫹6230G⬎A in cases and controls, three common haplotypes (A:A, G:G, and A:G) were observed in the Japanese population, as in the case of the Caucasian population (Table 3), suggesting that linkage disequilibrium between these loci and haplotypes consisting of these polymorphisms are similar in the two populations. Haplotype frequencies, however, were markedly different between the Japanese and Caucasian populations, with a much higher frequency of the G:G haplotype and a lower frequency of the A:A haplotype in the Japanese than in the Caucasian population (Table 3).

1090



TABLE 3. Frequencies of alleles and haplotypes of CTLA4 polymorphisms in cases and control subjects in Japanese and Caucasian populations Caucasiansa

Japanese

SNP/allele ⫹6230/G ⫹49/G Haplotypes (⫹49G⬎A:⫹6230G⬎A) G:G A:A A:G

Cases (n ⫽ 1538) (%)

Controls (n ⫽ 1446) (%)

Odds ratio (95% CI)

Controls (n ⫽ 1304)

75.8 64.0

75.0 63.1

1.04 (0.88 –1.23) 1.04 (0.89 –1.21)

52.3 35.8

62.4 22.7 13.5

62.2 23.7 13.1

1.00 (0.87–1.16) 0.94 (0.79 –1.11) 1.04 (0.84 –1.28)

36.7 46.6 16.3

Haplotypes were estimated by an accelerated EM algorithm using Haploview (version 2.03). Ref. 6. n ⫽ 844 for haplotypes.

a

To compare alleles at other polymorphic loci in CTLA4 and the haplotypes of CTLA4 between the Japanese and Caucasian populations, a subset of subjects was typed for additional markers in CTLA4 that were reported to be associated with autoimmune diseases in a recent large-scale study in Caucasian populations (6). The frequencies of alleles at all the polymorphic loci typed in CTLA4 were significantly different between Japanese and Caucasian populations in that the frequencies of alleles reported to be associated with type 1 diabetes and AITD (6) were significantly higher in the Japanese than in the Caucasian population (MH30/G, 73.7 vs. 55.0%; Jo30/G, 72.9 vs. 50.3%; rs1863800/C, 74.4 vs. 54.8%; P ⬍ 1 ⫻ 10⫺7 for all). Despite the marked difference in frequencies of alleles, the relative frequencies of alleles at these polymorphic loci in each population were similar in that there was a relatively lower frequency of the diseaseassociated G allele at ⫹49G⬎A than the frequencies of disease-associated alleles at other loci, rs1863800, Jo31, MH30, and ⫹6230G⬎A (66.8 vs. 72.9 –74.4% in Japanese and 35.8 vs. 50.3–54.8% in Caucasians). Discussion

In the present study with the largest number of samples to date from Japanese patients with type 1 diabetes, polymorphisms in CTLA4 were significantly associated with type 1 diabetes complicated with AITD. Because CTLA4 is involved in the regulation of immune function in general, but not the immune response to specific organs, as evidenced by the wide range of autoimmune phenomena in mice with targeted disruption of the gene for CTLA4 (20, 21), the stronger association of the CTLA4 polymorphism with autoimmune reaction to both pancreatic islets and thyroid gland than with type 1 diabetes alone observed in the present study is reasonable. It is also consistent with the observation in the nonobese diabetic mouse, an animal model of type 1 diabetes, because a recent study demonstrated that the gene for CTLA4 is responsible for Idd5.1, a susceptibility gene for type 1 diabetes (6, 22), and nonobese diabetic mice develop not only antiislet autoimmunity, but also autoimmunity to other organs, including the thyroid gland (23). In both Caucasian and Japanese populations, the frequency of antithyroid autoimmunity in patients with type 1 diabetes was reported to be as high as 50% (18, 19, 24 –28). Thyroid autoimmunity as evaluated by the presence of antiTPO antibody and/or anti-Tg antibody was reported to be present in 21.6% of 7097 patients with type 1 diabetes in

Germany (25), anti-TPO antibody was reported to be positive in 27.9% of 111 newly diagnosed patients with type 1 diabetes in Spain (26) and 17% of 157 patients in Belgium (27), and thyroid microsomal autoantibodies were positive in 20.1% of 643 white patients in the United States (28), indicating that latent antithyroid autoimmunity without clinical presentation of the disease is present at high frequency. The stronger association of the CTLA4 polymorphism with AITD than with type 1 diabetes in recent large-scale studies in Caucasian populations (6) together with the high frequency of AITD in patients with type 1 diabetes (24 –28) raises the possibility that the association of CTLA4 with type 1 diabetes could be due to the association of the polymorphism with coexisting AITD. The stronger association of CTLA4 with type 1 diabetes complicated with AITD than with type 1 diabetes without AITD in the present study suggests this possibility. The stronger association of the CTLA4 polymorphism with both type 1 diabetes and AITD than with type 1 diabetes alone may not be limited to Japanese because a similar tendency was previously suggested by a French group using the ⫹49G⬎A polymorphism with a smaller number of subjects (29). The odds ratio of approximately 1.5 for AITD (6) together with the 20% frequency of AITD in type 1 diabetic patients is expected to give rise to an odds ratio of 1.1 for type 1 diabetes even if the odds ratio for type 1 diabetes itself is 1.0. The observed odds ratios for type 1 diabetes in Japanese (1.08 in this study) and Caucasian (1.14, Ref. 6) populations are consistent with this hypothesis. Although the number of subjects was limited, when type 1 diabetic patients with AITD were subgrouped by the associated AITD, the association of the polymorphism appeared to be particularly stronger for latent AITD, i.e. antithyroid autoimmunity without clinical manifestation of thyroid dysfunction, in type 1 diabetes in the present study [odds ratio, 1.88 (1.12–3.16)] than for Graves’ disease [odds ratio, 1.42 (0.75–2.67)] and Hashimoto’s thyroiditis [odds ratio, 1.27 (0.72–2.24)], suggesting the possibility that the CTLA4 polymorphism may not be associated with type 1 diabetes alone and that the modest association (odds ratio of 1.1–1.2) of the polymorphism with type 1 diabetes in previous studies may have been due to the association of the polymorphism with type 1 diabetes complicated with antithyroid autoimmunity, most of which was latent and therefore was not clinically appreciated. To clarify this possibility, reevaluation of previous studies showing a positive association of CTLA4 with type 1 diabetes by phenotyping the subjects relative to AITD


as well as further studies in Caucasian populations with a large number of well-characterized patients with type 1 diabetes, whose AITD status and/or positivity for antithyroid antibodies are characterized, are necessary. Although a large number of samples are required for identification of non-HLA genes for type 1 diabetes, because of the much weaker effect of each non-MHC locus on disease susceptibility as compared with the extremely strong effect of HLA, it is difficult to collect a large number of Asian samples in each research group because of the very low incidence (less than 1/10 of that in Caucasians) of type 1 diabetes in most Asian countries, including Japan (7). We therefore assembled a multicenter study group with a total number of samples of more than 1800, including more than 750 cases with type 1 diabetes. Given the very low frequency (less than 1/10 of that in Caucasians) of type 1 diabetes in Japan, the time and effort taken to collect the number of cases in the present study may correspond to more than 7500 cases of type 1 diabetes in Caucasians. Because of the well-known contribution of IDDM1 in the HLA region to susceptibility to type 1 diabetes, it is important to study the effect of CTLA4 on type 1 diabetes susceptibility relative to HLA genotypes. The present study demonstrated that the association of CTLA4 with type 1 diabetes is affected by the HLA genotypes, with a stronger association in patients with very high-risk genotypes consisting of two doses of the high risk haplotypes, DRB1*0405-DQB1*0401 and DRB1*0901-DQB1*0303, than in those with other genotypes. A recent study in a large pedigree in Sweden with multiple cases of type 1 diabetes and AITD suggested a strong genetic interaction of HLA and the CTLA4 region in conferring susceptibility to autoimmune diseases (30). Interaction of CTLA4 with HLA was also suggested in previous studies with smaller numbers of patients in Japanese (12) as well as Caucasian (31) populations. The findings of the present study with a much larger number of samples are consistent with these observations. Given the function of CTLA4 as a negative regulatory molecule of the immune system in general and that of class II HLA in the presentation of a specific antigen to initiate an immune response to the antigen, it is reasonable to speculate that a susceptible allele at CTLA4 leads to autoimmunity, and specific HLA haplotypes target autoimmune attack to pancreatic islets. A similar mechanism has recently been proposed for a rat model of type 1 diabetes, the Komeda diabetes-prone rat, in which the major nonMHC susceptibility locus, Iddm/kdp1, together with MHC accounts for most of the genetic predisposition to diabetes (32). Iddm/kdp1 has recently been positionally cloned and identified as a nonsense mutation in Cblb, which acts as a negative regulator of the immune system, as in the case of CTLA4 (32). A nonsense mutation of Cblb in the Komeda diabetes-prone rat is associated with autoimmunity against multiple organs including the thyroid gland, and type 1 diabetes develops only when combined with a diseasesusceptible MHC class II u haplotype (32). Although these data suggest the interaction of CTLA4 with HLA in conferring susceptibility to type 1 diabetes, it is difficult to obtain a definite conclusion with case-control studies subgrouped by HLA as in the present study. Further studies with a large


1091

number of families with well-defined genotypes for CTLA4 and HLA are necessary. The effect of non-HLA genes on type 1 diabetes susceptibility is modest, as compared with the extremely strong effect of HLA-linked genes (2, 3). Although the effect of non-HLA genes is weaker than that of the HLA, identification of these genes will provide important information on biological pathways in the development of type 1 diabetes, leading to the development of effective methods for prediction, prevention, and intervention in the disease. Given the strong protection in a mouse model of type 1 diabetes by a small modification of the environment and/or each step of the etiological pathway (33, 34), information on genes predisposing to the disease is important even if the effect of each gene is small. A large number of samples in both Japanese and Caucasian populations will facilitate genetic dissection and identification of susceptibility genes for type 1 diabetes, leading to better understanding of the pathogenesis of type 1 diabetes as well as other autoimmune diseases. Acknowledgments We thank Miss Miho Uga and Miyuki Moritani for technical assistance. Members of the Japanese Study Group on Type 1 Diabetes Genetics are: Takuya Awata, Hiroshi Ikegami, Eiji Kawasaki, Tetsuro Kobayashi, Taro Maruyama, Koji Nakanishi, and Akira Shimada. Received June 24, 2005. Accepted December 7, 2005. Address all correspondence and requests for reprints to: Hiroshi Ikegami, M.D., Ph.D., Department of Geriatric Medicine, Osaka University Graduate School of Medicine, Yamadaoka, Suita, Osaka 5650871, Japan. E-mail: [email protected]. This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas, a Grant-in-Aid for Scientific Research, and a Grant-in-Aid for Exploratory Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to H.I.).

References 1. Atkinson MA, Eisenbarth GS 2001 Type 1 diabetes: new perspectives on disease pathogenesis and treatment. Lancet 358:221–229 2. Davies JL, Kawaguchi Y, Bennett ST, Copeman JB, Cordell HJ, Pritchard LE, Reed PW, Gough SCL, Jenkins SC, Palmer SM, Balfour KM, Rowe BR, Farrall M, Barnett AH, Bain SC, Todd JA 1994 A genome-wide search for human type 1 diabetes susceptibility genes. Nature 371:130 –136 3. Cox NJ, Wapelhorst B, Morrison VA, Johnson L, Pinchuk L, Spielman RS, Todd JA, Concannon P 2001 Seven regions of the genome show evidence of linkage to type 1 diabetes in a consensus analysis of 7657 multiplex families. Am J Hum Genet 69:820 – 830 4. Julier C, Hyer RN, Davies J, Merlin F, Soularue P, Briant L, Cathelineau G, Deschamps I, Rotter JI, Froguel P, Boitard C, Bell JI, Lathrop GM 1991 Insulin-IGF2 region on chromosome 11p encodes a gene implicated in HLADR4-dependent diabetes susceptibility. Nature 354:155–159 5. Meloni GF, Lucarelli P, Pellecchia M, Eisenbarth GS, Comings D, Mustelin T 2004 A functional variant of lymphoid tyrosine phosphatase is associated with type 1 diabetes. Nat Genet 36:337–338 6. Ueda H, Howson JMM, Esposito L, Heward J, Snook H, Chamberlain G, Rainbow DB, Hunter KMD, Smith AN, Di Genova G, Herr MH, Dahlman I, Payne F, Smyth D, Lowe C, Twells RCJ, Howlett S, Healy B, Nutland S, Rance HE, Everett V, Smink LJ, Lam AC, Cordell HJ, Walker NM, Bordin C, Hulme J, Motzo C, Cucca F, Hess JF, Metzker ML, Rogers J, Gregory S, Allahabadia A, Nithiyananthan R, Tuomilehto-Wolf E, Tuomilehto J, Bingley P, Gillespie KM, Undlien DE, Ronningen KS, Guja C, Ionescu-Tirgoviste C, Savage DA, Maxwell AP, Carson DJ, Patterson CC, Franklyn JA, Clayton DG, Peterson LB, Wicker LS, Todd JA, Gough SCL 2003 Association of the T-cell regulatory gene CTLA4 with susceptibility to autoimmune disease. Nature 423:506 –511 7. Karvonen M, Viik-Kajander M, Moltchanova E, Libman I, LaPorte R, Tuomilehto J 2000 Incidence of childhood type 1 diabetes worldwide. Diabetes Mondiale (DiaMond) Project Group. Diabetes Care 23:1516 –1526 8. Awata T, Kurihara S, Iitaka M, Takei S, Inoue I, Ishii C, Negishi K, Izumida

1092

9.

10.

11.

12.

13. 14.

15.

16.

17.

18.

19.


T, Yoshida Y, Hagura R, Kuzuya N, Kanazawa Y, Katayama S 1998 Association of CTLA-4 gene A-G polymorphism (IDDM12 locus) with acute-onset and insulin-depleted IDDM as well as autoimmune thyroid disease (Graves’ disease and Hashimoto thyroiditis) in the Japanese population. Diabetes 47: 128 –129 Yanagawa T, Maruyama T, Gomi K, Taniyama M, Kasuga A, Ozawa Y, Terauchi M, Hiroshi H, Maruyama H, Saruta T 1999 Lack of association between CTLA-4 gene polymorphism and IDDM in Japanese subjects. Autoimmunity 29:53–56 Takara M, Komiya I, Kinjo Y, Tomoyose T, Yamashiro S, Akamine H, Masuda M, Takasu N 2000 Association of CTLA-4 gene A/G polymorphism in Japanese type 1 diabetic patients with younger age of onset and autoimmune thyroid disease. Diabetes Care 23:975–978 Abe T, Yamaguchi Y, Takino H, Fujita N, Yamaguchi-Degawa M, Ozaki M, Yamakawa K, Sera Y, Sakamachi H, Uotani S, Kawasaki E, Awata T, Yamasaki H, Eguchi K 2001 CTLA-4 gene polymorphism contributes to the mode of onset of diabetes with antiglutamic acid decarboxylase antibody in Japanese patients: genetic analysis of diabetic patients with antiglutamic acid decarboxylase antibody. Diabet Med 18:726 –731 Mochizuki M, Amemiya S, Kobayashi K, Kobayashi K, Shimura Y, Ishihara T, Nakagomi Y, Onigata K, Tamai S, Kasuga A, Nanazawa S 2003 Association of the CTLA-4 gene 49 a/g polymorphism with type 1 diabetes and autoimmune thyroid disease in Japanese children. Diabetes Care 26:843– 847 Awata T, Kuzuya T, Matsuda A, Iwamoto Y, Kanazawa Y 1992 Genetic analysis of HLA class II alleles and susceptibility to type 1 (insulin-dependent) diabetes mellitus in Japanese subjects. Diabetologia 35:419 – 424 Kawabata Y, Ikegami H, Kawaguchi Y, Fujisawa T, Shintani M, Ono M, Nishino M, Uchigata Y, Lee I, Ogihara T 2002 Asian-specific HLA haplotypes reveal heterogeneity of the contribution of HLA-DR and -DQ haplotypes to susceptibility to type 1 diabetes. Diabetes 51:545–551 Tokunaga K, Ishikawa Y, Ogawa A, Wang H, Mitsunaga S, Moriyama S, Lin L, Bannai M, Watanabe Y, Kashiwase K, Tanaka H, Akaza T, Tadokoro K, Juji T 1997 Sequence-based association analysis of HLA class I and II alleles in Japanese supports conservation of common haplotypes. Immunogenetics 46:199 –205 Kawasaki E, Sera Y, Fujita N, Yamauchi M, Ozaki M, Abe T, Yamakawa K, Uotani S, Takino H, Yamasaki H, Yamaguchi Y, Uchigata Y, Matsuura N, Eguchi K 2001 Association between IA-2 autoantibody epitope specificities and age of onset in Japanese patients with autoimmune diabetes. J Autoimmun 17:323–331 Sera Y, Kawasaki E, Abiru N, Ozaki M, Abe T, Takino H, Kondo H, Yamasaki H, Yamaguchi Y, Akazawa S, Nagataki S, Uchigata Y, Matsuura N, Eguchi K 1999 Autoantibodies to multiple islet autoantigens in patients with abrupt onset type 1 diabetes and diabetes diagnosed with urinary glucose screening. J Autoimmun 13:257–265 Kida K, Mimura G, Kobayashi T Nakamura K, Sconoda S, Inouye H, Tsuji K 1989 Immunogenetic heterogeneity in type 1 (insulin-dependent) diabetes among Japanese: HLA antigens and organ-specific autoantibodies. Diabetologia 32:34 –39 Chikuba N, Akazawa S, Yamaguchi Y, Kawasaki E, Takino H, Takao Y, Maeda Y, Okuno S, Yamamoto H, Yokota A, Yoshimoto M, Nagataki S 1992


20. 21.

22. 23. 24. 25. 26. 27.

28. 29.

30.

31.

32. 33. 34.

Type 1 (insulin-dependent) diabetes mellitus with co-existing autoimmune thyroid disease in Japan. Intern Med 31:1076 –1080 Waterhouse P, Penninger JM, Timms E, Wakeham A, Shahinian A, Lee KP, Thompson CB, Griesser H, Mak TW 1995 Lymphoproliferative disorders with early lethality in mice deficient in. CTLA-4. Science 270:985–988 Tivol EA, Borriello F, Schweitzer AN, Lynch WP, Bluestone JA, Sharpe AH 1995 Loss of CTLA-4 leads to massive lymphoproliferation and fatal multiorgan tissue destruction, revealing a critical negative regulatory role of CTLA-4. Immunity 3:541–547 Hill NJ, Lyons PA, Armitage N, Todd JA, Wicker LS, Peterson LB 2000 NOD Idd5 locus controls insulitis and diabetes and overlaps the orthologous CTLA4/IDDM12 and NRAMP loci in humans. Diabetes 49:1744 –1747 Bernard NF, Ertug F, Margolese H 1991 High incidence of thyroiditis and anti-thyroid autoantibodies in NOD mice. Diabetes 41:40 – 46 Burek CL, Rose NR, Guire KE, Hoffmann WH 1990 Thyroid autoantibodies in black and white children and adolescents with type 1 diabetes mellitus and their first-degree relatives. Autoimmunity 4:157–167 Fernandez-Castaner M, Molina A, Lopez-Jimenez L, Gomez JM, Soler J 1999 Clinical presentation and early course of type 1 diabetes in patients with and without thyroid autoimmunity. Diabetes Care 22:377–381 Kordonouri O, Klinghammer A, Lang EB, Gruters-Kieslich A, Grabert M, Holl RW 2002 Thyroid autoimmunity in children and adolescents with type 1 diabetes. Diabetes Care 25:1346 –1350 Abrams P, De Leeuw I, Vertommen J, Bergian Diabetes Registry 1996 In new-onset insulin-dependent diabetic patients the presence of anti-thyroid peroxidase antibodies is associated with islet cell autoimmunity and the high risk haplotype HLA DQA1*0301-DQB1*0302. Diabet Med 13:415– 419 Riley WJ, Maclaren NK, Lezotte DC, Spillar RP, Rosenbloom AL 1981 Thyroid autoimmunity in insulin-dependent diabetes mellitus: the case for routine screening. J Pediatr 99:350 –354 Djilali-Saiah I, Larger E, Harfouch-Hammoud E, Timsit J, Clerc J, Bertin E, Assan R, Boitard C, Bach JF, Caillat-Zucman S 1998 No major role for the CTLA-4 gene in the association of autoimmune thyroid disease with IDDM. Diabetes 47:125–127 Einarsdottir E, Soderstrom I, Lofgren-Burstrom A, Haraldsson S, NilssonArdnor S, Penha-Goncalves C, Lind L, Holmgren G, Holmberg M, Asplund K, Holmberg D 2003 The CTLA4 region as a general autoimmunity factor: an extended pedigree provides evidence for synergy with the HLA locus in the etiology of type 1 diabetes mellitus, Hashimoto’s thyroiditis and Graves’ disease. Eur J Hum Genet 11:81– 84 Vaidya B, Imrie H, Petros P, Young ET, Kelly WF, Carr D, Large DM, Toft AD, McCarthy MI, Kendall-Taylor P, Pearce SH 1999 The cytotoxic T lymphocyte antigen-4 is a major Graves’ disease locus. Hum Mol Genet 8:1195– 1199 Yokoi N, Komeda K, Wang H-Y, Yano H, Kuma K, Kitada K, Saitoh Y, Seino Y, Yasuda K, Serikawa T, Seino S 2002 Cblb is a major susceptibility gene for rat type 1 diabetes mellitus. Nat Genet 31:391–394 Todd JA 1991 A protective role of the environment in the development of type 1 diabetes? Diabet Med 8:906 –910 Bowman MA, Leiter EH, Atkinson MA 1994 Prevention of autoimmune diabetes in the NOD mouse: implication for therapeutic intervention in human disease. Immunol Today 15:115–120

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.