IT IS WELL known that type 1 diabetes mellitus (T1DM) is

Transcription

1 /01/$03.00/0 The Journal of Clinical Endocrinology & Metabolism 86(8): Printed in U.S.A. Copyright 2001 by The Endocrine Society Two Distinct MICA Gene Markers Discriminate Major Autoimmune Diabetes Types GIOVANNI GAMBELUNGHE, MEHRAN GHADERI, CRISTINA TORTOIOLI, ADRIANO FALORNI, FAUSTO SANTEUSANIO, PAOLO BRUNETTI, CARANI B. SANJEEVI, AND ALBERTO FALORNI ON BEHALF OF THE UMBRIA TYPE 1DIABETES REGISTRY Departments of Internal Medicine and Endocrine and Metabolic Sciences (G.G., C.T., F.S., P.B., Al.F.) and Gynecological, Obstetric, and Pediatric Sciences (Ad.F.), University of Perugia, Perugia, Italy; and Department of Molecular Medicine, Karolinska Institute (G.G., M.G., C.B.S.), S Stockholm, Sweden The polymorphism of the major histocompatibility complex class I chain-related A gene is associated with type 1 diabetes mellitus. The major histocompatibility complex class I chainrelated A gene 5 allele is significantly more frequent in Caucasian type 1 diabetes mellitus children than in healthy subjects, but no information is available on the association with adult-onset type 1 diabetes mellitus or with the so-called slowly progressive latent autoimmune diabetes of the adult in the same ethnic group. In this study we estimated the frequency of major histocompatibility complex class I chainrelated A gene alleles and human leukocyte antigen-drb1*03- DQA1*0501-DQB1*0201 and human leukocyte antigen-drb1*04- DQA1*0301-DQB1*0302 in 195 type 1 diabetes mellitus subjects, in 80 latent autoimmune diabetes of the adult subjects, and in 158 healthy subjects from central Italy. Major histocompatibility complex class I chain-related A gene 5 was significantly associated with type 1 diabetes mellitus only in the 1 25 yr age group at diagnosis, and the odds ratio of the simultaneous presence of both major histocompatibility complex class I chain-related A gene 5 and human leukocyte antigen-drb1*03- IT IS WELL known that type 1 diabetes mellitus () is a complex genetic disease resulting from the autoimmune destruction of pancreatic -cells (1). Several genes have been associated with susceptibility and/or protection for (2 7), but the disease risk is mostly influenced by genes located in the class II region of the major histocompatibility complex [human leukocyte antigen (HLA)]. is positively associated with HLA-DRB1*03-DQA1*0501- DQB1*0201 (DR3-DQ2) and HLA-DRB1*04-DQA1*0301- DQB1*0302 (DR4-DQ8), but is negatively associated with HLA-DRB1*1501-DQA1*0102-DQB1*0602 (DR2-DQ6) (2). However, several lines of evidence support the hypothesis that other HLA genes contribute to the risk for autoimmune diabetes. We recently demonstrated that the major histocompatibility complex (MHC) class I chain-related A (MICA) gene, located telomeric to the TNFa gene between the B- associated transcript and the HLA-B genes, influences the risk for (8). In addition, other studies (9 11) have observed an association between MICA gene polymorphism and in other ethnic groups. Abbreviations: GAD65Ab, glutamic acid decarboxylase 65 autoantibodies; HLA, human leukocyte antigen; LADA, latent autoimmune diabetes of the adult; MHC, major histocompatibility complex; MICA, MHC class I chain-related A gene; OR, odds ratio;, type 1 diabetes mellitus. DQA1*0501-DQB1*0201 and/or human leukocyte antigen- DRB1*04-DQA1*0301-DQB1*0302 was as high as 54 and higher than 388 when compared with double negative individuals. Adult-onset type 1 diabetes mellitus (age at diagnosis, >25 yr) and latent autoimmune diabetes of the adult were significantly associated with major histocompatibility complex class I chainrelated A gene 5.1, which was not significantly increased among diabetic children. Only the combination of major histocompatibility complex class I chain-related A gene 5.1 and human leukocyte antigen-drb1*03-dqa1*0501-dqb1*0201 and/or human leukocyte antigen-drb1*04-dqa1*0301-dqb1*0302 conferred increased risk for adult-onset type 1 diabetes mellitus or for latent autoimmune diabetes of the adult. Our study provides demonstration of the existence of distinct genetic markers for childhood/young-onset type 1 diabetes mellitus and for adultonset type 1 diabetes mellitus/latent autoimmune diabetes of the adult, namely major histocompatibility complex class I chainrelated A gene 5 and major histocompatibility complex class I chain-related A gene 5.1, respectively. (J Clin Endocrinol Metab 86: , 2001) The MICA gene polymorphism consists of a variable number of repetitions of GCT in the exon 5 that identifies five different alleles (denominated 4, 5, 5.1, 6, and 9) (12). The frequency of the MICA5 allele is significantly increased in Caucasian subjects compared with that in age-, sexand geographically-matched healthy controls (8). The association of MICA5 with seems to be little influenced by any of the known HLA markers previously shown to confer susceptibility for the disease, including HLA-DRB1*03- DQA1*0501-DQB1*0201, -DRB1*04-DQA1*0301-DQB1*0302, TNFa-2, HLA-B8, and HLA-B15 (8). The heterogeneity of the clinical and immunological features of in relation to age at clinical onset is a complicating factor. Childhood is characterized by an abrupt onset and ketosis and is associated with HLA- DRB1*04-DQA1*0301-DQB1*0302 and a high frequency of insulin and IA-2 autoantibodies (13, 14). On the other hand, the so-called latent autoimmune diabetes of the adult (LADA) is a slowly progressive form of adult-onset autoimmune diabetes that is noninsulin dependent at the time of clinical diagnosis and is characterized by the presence of glutamic acid decarboxylase 65 (GAD65) autoantibodies (GAD65Ab) and/or islet cell antibodies (15 17). Although the prevalence of immune and genetic markers of islet autoimmunity differs significantly between childhood and adult- 3754

2 Gambelunghe et al. MICA Polymorphism in Type 1 Diabetes The Journal of Clinical Endocrinology & Metabolism, August 2001, 86(8): onset (13, 18, 19), no exclusive marker of LADA has been identified to date. GAD65Ab are present in 75% diabetic children, and insulin autoantibodies and IA-2 autoantibodies as well as HLA-DRB1*04-DQA1*0301-DQB1*0302 are more frequent in adult subjects than in age-matched healthy controls. It is still unclear whether LADA is under the control of distinct genetic markers or is only the consequence of the action of protective environmental factors on a genetic background predisposing for autoimmune diabetes. MICA5 does not confer increased risk for adult-onset (8), and it is still unknown whether the MICA gene polymorphism has any role in the pathogenesis of LADA. It has been shown that gene markers located inside or in proximity to the HLA class I region are associated with the age at onset of Caucasian (20). More recently, an age-dependent association of the MICA gene polymorphism with was suggested in the Japanese population (10). However, no information is currently available on the association of MICA gene polymorphism with adult-onset Caucasian autoimmune diabetes. In the present study we tested the association of MICA gene polymorphism with Caucasian in relation to age at clinical onset of the disease, and we provide the demonstration that two distinct genetic markers discriminate major types, namely childhood/young-onset and adult-onset /LADA. Subjects Subjects and Methods Genomic DNA was obtained by phenol-chloroform purification and using standard procedures from EDTA-treated peripheral blood samples from 275 unrelated subjects with autoimmune diabetes mellitus, including 195 cases of (age at diagnosis: median, 23 yr; range, 1 42; male/female ratio, 108/87) and 80 cases of LADA (age at diagnosis: median, 51 yr; range, 25 68; male/female ratio, 22/58), recruited by the Umbria Type 1 Diabetes Registry (central Italy) between January 1, 1993, and March 31, The diagnosis of was made according to the National Diabetes Data Group (21). A total of 94% subjects was positive for at least one of the three major islet autoantibodies: islet cell antibodies, GAD65 antibodies, and tyrosine phosphatase-like molecule IA-2 antibodies. The incidence of in Umbria between 1993 and 2000 was 9.1 new cases/100,000 yr in the age group 0 29 yr, and the degree of case ascertainment of our registry was 99%. By screening 725 diabetic subjects initially classified as type 2 diabetes, we identified 80 LADA subjects (age at diagnosis: median, 51 yr; range, 25 68; male/ female ratio, 22/58) on the basis of the presence of GAD65Ab, the maintenance of a good metabolic control with diet and/or hypoglycemic agents for at least 1 yr after the diagnosis of diabetes mellitus, and an age at diagnosis over 25 yr, according to previously described criteria (15 17). A total of 66 of 80 (82%) LADA subjects were converted to insulin treatment within 6 yr after the diagnosis. The results for LADA subjects were compared with those for typical subjects of similar age (diagnosis after the age of 25 yr) and in younger subjects (age 25 yr at diagnosis). Blood samples collected during routine analyses were available from 158 unrelated healthy control subjects (schoolchildren and employees of the University of Perugia; age: median, 33 yr; range, 6 62; male/female ratio, 83/75), with no family history of diabetes and geographically matched with the diabetic patients. All patients and healthy individuals gave their informed consent for the study. MICA and HLA-DR, -DQ genotyping were performed in all patients and healthy controls for whom a genomic DNA sample was available. MICA genotyping MICA genotyping was performed according to previously described procedures (12) using 5 -CCTTTTTTTCAGGGAAAGTGC-3 as forward and 5 -CCTTACCATCTCCAGAAACTGC-3 as reverse primer. The reverse primer was labeled at the 5 -end with fluorescent reagent 6-HEX (Amersham Pharmacia Biotech, Uppsala, Sweden). After amplification, the number of the GCT triplet repeat units was determined using an ABI PRISM automated DNA sequencer (PE Applied Biosystems, Foster City, CA). HLA-DR and -DQ genotyping PCR-amplified products of the polymorphic second exon of the DQA1, DQB1, and DRB1 genes were manually dotted onto nylon membranes (Amersham Pharmacia Biotech, Arlington Heights, IL), under denaturing conditions. The membranes were hybridized with sequencespecific oligonucleotides (SSOs), 3 -end labeled with [ 32 P]deoxy-CTP, and washed in stringency conditions before exposure to x-ray film, as previously described (22). Statistical analysis The odds ratio (OR) was calculated according to Woolf and Miettinen (23, 24). Differences in allele/haplotype frequencies between the diabetic subjects and the control group were tested by the 2 method. Yates correction or Fisher s exact test was used when necessary. The strongest HLA association was tested using the method of Svejgaard and Ryder (25). In this analysis the association of autoimmune diabetes with MICA gene polymorphism (factor A) was compared with the presence of HLA-DR3-DQA1*0501-DQB1*0201 and/or DR4-DQA1*0301-DQB1*0302 (factor B). The probability values were corrected (Pc) for the number of comparisons, according to the number of alleles or haplotypes observed among diabetic subjects and the number of subgroups of subjects. The dependency of the frequency of MICA alleles or HLA-DR-DQ haplotypes on sex and age at onset was assessed by logistic regression analysis. Allele or haplotype frequencies for less than 25-yr-old and more than 25-yr-old subjects are shown in Tables 1 5 because of the age effect revealed by the logistic regression analysis in the entire population and for comparison with the LADA population, which was initially identified from a group of type 2 diabetes subjects with a clinical diagnosis after the age of 25 yr. A P (or Pc) 0.05 was considered significant. Results In the present study we observed that MICA5 was significantly more frequent in 195 subjects than in 158 healthy controls (OR, 5.1; corrected P ; Tables 1 and 2). The logistic regression analysis revealed that the association of MICA5 with was negatively dependent on age at clinical onset of the disease (P ). The frequency of MICA5 was not associated with age in the healthy control group. MICA5 was present in only 22% of subjects older than 25 yr and in 11% of LADA subjects, which is not statistically different from the frequency observed in healthy controls (15%; Table 1). On the other hand, the analysis of 195 subjects showed that the frequency of MICA5.1 was positively dependent on age at onset of (logistic regression analysis, P ), and this allele was present in 62% of subjects older than 25 yr and in 77% of LADA subjects compared with 33% healthy subjects (Table 1). The frequency of MICA5.1 was not dependent on age in the healthy control group. Hence, the frequency of MICA5.1 in adult-onset autoimmune diabetes was significantly higher than that observed in healthy subjects (OR 3.4 and 7.0; corrected P 0.001; Tables 1 and 2). No statistically significant relationship between the conversion to insulin treatment and the presence of MICA5.1 was observed in LADA subjects. None of the other MICA alleles, namely alleles 4, 6, and 9,

3 3756 The Journal of Clinical Endocrinology & Metabolism, August 2001, 86(8): Gambelunghe et al. MICA Polymorphism in Type 1 Diabetes TABLE 1. Association between MICA alleles and HLA-DR, -DQ haplotypes and autoimmune diabetes in relation to age at onset and clinical presentation (n 107) (n 88) LADA (n 80) Healthy controls (n 158) Age (yr) a 23 (1 42) 17 (1 24) 31 (25 42) 51 (25 68) 33 (6 62) Males/females 108/87 54/53 54/34 22/58 83/75 MICA5 b 93 c (48%) 74 (69%) 19 (22%) 9 (11%) 24 (15%) Pc d Pc e Pc NS e Pc NS d MICA5.1 b 78 f (40%) 23 (21%) 55 (62%) 62 (77%) 52 (33%) Pc NS d Pc NS e Pc e Pc d DRB1*03/DQA1*0501/DQB1*0201 b (47%) (51%) (43%) (41%) (13%) Pc g Pc h Pc h Pc g DRB1*04/DQA1*0301/DQB1*0302 b (28%) (32%) (22%) (19%) (8%) Pc g Pc h Pc NS h Pc NS g a Age at diagnosis in diabetic subjects or age at blood sampling for genetic analysis in healthy controls. b Corrected P values (Pc) vs. healthy controls are shown. c The frequency of MICA5 in subjects was negatively dependent on age at disease onset (logistic regression analysis, P ). d Correction factor 5 (number of MICA alleles). e Correction factor 10 (number of MICA alleles number of subgroups of subjects). f The frequency of MICA5.1 in subjects was positively dependent on age at disease onset (logistic regression analysis, P ). g Correction factor 26 (number of HLA-DR-DQ haplotypes). h Correction factor 52 (number of HLA-DR-DQ haplotypes number of subgroups of subjects). TABLE 2. Odds ratios of MICA alleles and HLA-DR, -DQ haplotypes for autoimmune diabetes in relation to age at onset and clinical presentation (n 107) (n 88) LADA (n 80) MICA ( ) ( ) ( ) ( ) Pc Pc Pc NS Pc NS MICA ( ) ( ) ( ) ( ) Pc NS Pc NS Pc Pc DRB1*03/DQA1*0501/DQB1* ( ) ( ) ( ) ( ) Pc Pc Pc Pc DRB1*04/DQA1*0301/DQB1* ( ) ( ) ( ) ( ) Pc Pc Pc NS Pc NS OR (95% CI) of MICA alleles and HLA-DR, -DQ haplotypes are shown. Pc, Corrected P value (see Table 1 for correction factors). was significantly associated, positively or negatively, with autoimmune diabetes regardless of age at clinical diagnosis. Although decreasing with increasing age at onset, the frequency of both HLA-DRB1*03-DQA1*0501-DQB1*0201 and DRB1*04-DQA1*0301-DQB1*0302 in diabetic subjects did not vary significantly among the different age groups (Table 1). Similarly, the frequencies of the studied HLA haplotypes were not dependent on age in the healthy control group. Moreover, only HLA-DRB1*03-DQA1*0501-DQB1*0201 was statistically associated with adult-onset and with LADA even though the frequency of DRB1*04-DQA1*0301- DQB1*0302 was lower in healthy subjects than in adult diabetic subjects. In subjects, the presence of MICA5 or MICA5.1 was not significantly associated with the presence of any of the major islet autoantibodies. To perform the Svejgaard and Ryder s test (25), we calculated the frequencies of the combination of MICA5 (Table 3) or MICA5.1 (Table 4) with at risk class II haplotypes, represented by HLA-DRB1*03-DQA1*0501-DQB1*0201 and/or DRB1*04-DQA1*0301-DQB1*0302. This analysis revealed the independent associations of MICA5 and class II gene polymorphism with childhood/young-onset (age at onset, 25 yr; Table 5, lines 3 6). The interaction between the MICA gene and the HLA class II polymorphism was revealed by the increase in the OR in the presence of both MICA5 and HLA-DRB1*03-DQA1*0501-DQB1*0201 and/or DRB1*04-DQA1*0301-DQB1*0302 in young-onset (OR 54; OR 388 when compared with double-negative individuals; Tables 3 and 5, line 8). Thus, 44 of 107 (41%) subjects with childhood/young-onset were simultaneously positive for MICA5 and at risk class II haplotypes compared with only 2 of 158 (1.3%) healthy individuals (Pc ; Table 3). MICA5 was significantly associated with young-onset even in the absence of at risk class II

4 Gambelunghe et al. MICA Polymorphism in Type 1 Diabetes The Journal of Clinical Endocrinology & Metabolism, August 2001, 86(8): TABLE 3. Combination of MICA5 and at risk HLA-DR, -DQ haplotypes in less than 25 yr-old type 1 diabetic subjects (n 107) Healthy controls (n 158) MICA At risk HLA class II haplotypes a,b (29%) (41%) (1.3%) Pc c Pc d MICA At risk HLA class II haplotypes a,b (19%) (28%) (14%) Pc NS c Pc 0.05 d MICA At risk HLA class II haplotypes a,b (33%) (25%) (18%) Pc c Pc NS d MICA At risk HLA class II haplotypes a,b (19%) (6%) (67%) Pc c Pc d a At risk HLA class II haplotypes are DRB1*03-DQA1*0501-DQB1*0201 or DRB1*04-DQA1*0301-DQB1*0302. b Corrected P values (Pc) vs. healthy controls are shown. c Correction factor 4 (number of MICA5/at risk HLA-DR-DQ subgroups). d Correction factor 8 (number of MICA5/at risk HLA-DR-DQ subgroups number of age subgroups of subjects). TABLE 4. Combination of MICA5.1 and at risk HLA-DR, -DQ haplotypes in more than 25-yr-old type 1 diabetic subjects and in LADA subjects (n 88) LADA (n 80) Healthy controls (n 158) MICA At risk HLA class II haplotypes a,b (26%) (34%) (42%) (9%) Pc c Pc d Pc c MICA At risk HLA class II haplotypes a,b (14%) (28%) (35%) (24%) Pc NS Pc NS Pc NS MICA At risk HLA class II haplotypes a,b (40%) (26%) (15%) (13%) Pc c Pc NS Pc NS MICA At risk HLA class II haplotypes a,b (20%) (12%) (8%) (54%) Pc c Pc d Pc c a At risk HLA class II haplotypes are DRB1*03-DQA1*0501-DQB1*0201 or DRB1*04-DQA1*0301-DQB1*0302. b Corrected P values (Pc) vs. healthy controls are shown. c Correction factor 4 (number of MICA5.1/at risk HLA-DR-DQ subgroups). d Correction factor 8 (number of MICA5.1/at risk HLA-DR-DQ subgroups number of age subgroups of subjects). haplotypes (Table 3). No significant linkage between MICA5 and HLA-DRB1*03-DQA1*0501-DQB1*0201 or DRB1*04- DQA1*0301-DQB1*0302 was detected in either diabetic patients (Table 5, line 9) or healthy subjects (Table 5, line 10). MICA5.1 contributed significantly to the risk for adultonset (age at onset, 25 yr) in the absence of both HLA-DRB1*03-DQA1*0501-DQB1*0201 and DRB1*04- DQA1*0301-DQB1*0302 (Table 5, line 4), but its contribution did not reach statistical significance after correction of the P value in subjects positive for at risk HLA class II haplotypes (Table 5, line 3), probably because of the number of subjects studied. In LADA subjects, the Sveigaard and Ryder test of the strongest HLA association showed the independent association of MICA5.1 and class II gene polymorphism (Table 5, lines 3 6). The strong interaction between MICA5.1 and class II haplotypes in conferring risk for adult-onset / LADA was revealed by the high OR of the combined association of the two markers (Table 5, line 8). No significant linkage between MICA5.1 and HLA-DRB1*03-DQA1*0501- DQB1*0201 or DRB1*04-DQA1*0301-DQB1*0302 was detected in either diabetic patients (Table 5, line 9) or healthy controls (Table 5, line 10). Taken together, 64 of 168 (38%) adult /LADA subjects were simultaneously positive for both MICA5.1 and at risk class II haplotypes, compared with only 14 of 158 (9%) healthy individuals (Pc ; Table 4). Discussion The MICA gene product is mainly expressed in the gastrointestinal epithelium and other epithelial cells and in keratinocytes, endothelial cells, and monocytes (26 28). Although its role is still unclear, it seems that MICA mediates the activation of natural killer cells and T cells (29, 30), thus being involved in stress-mediated responses during infections or autoimmune processes. The possible pathogenetic

5 3758 The Journal of Clinical Endocrinology & Metabolism, August 2001, 86(8): Gambelunghe et al. MICA Polymorphism in Type 1 Diabetes TABLE 5. Test of the strongest HLA association in subjects with autoimmune diabetes comparing MICA gene polymorphism with DR3-DQ2 and/or DR4-DQ8 25 yr 25 yr LADA OR (95% CI) P Pc OR (95% CI) P Pc OR (95% CI) P Pc Comparison a Test 1. A vs. non-a Individual assoc ( ) ( ) ( ) B vs. non-b Individual assoc. 8.4 ( ) ( ) ( ) A B vs. A B Indiv. MICA assoc ( ) ( ) NS NS 4.2 ( ) A B vs. A B Indiv. MICA assoc ( ) ( ) ( ) A B vs. A B Indiv. DR3-DQ2 DR4-DQ8 assoc ( ) ( ) NS 3.3 ( ) A B vs.a B Indiv. DR3-DQ2 DR4-DQ8 assoc. 17 ( ) (4 21.7) (2.8 23) A B vs. A B Diff. between assoc. 1.3 ( ) NS NS 0.6 ( ) NS NS 1.3 ( ) NS NS 8. A B vs. A B Combined assoc. 388 ( ) (8 41) ( ) Linkage in pts. 0.3 ( ) NS NS 0.5 ( ) NS NS 0.6 (0.1 3) NS NS 9. Assoc A and B in patients Linkage in contr. 0.3 ( ) NS NS 1.5 (0.6 4) NS NS 1.5 (0.6 4) NS NS 10. Assoc. A and B in controls a A, MICA5 in 25 yr and MICA5.1 in 25 yr /LADA; B, DRB1*03-DQA1*0501-DQB1*0201 and/or DRB1*04-DQA1*0301-DQB1*0302. Entries in the 2 2 contingency tables have been calculated on the basis of the frequencies reported in Tables 3 and 4. 95% CI, 95% confidence interval; Pc, corrected P value: correction factor was 20 for type 1 diabetes (10 comparisons 2 patients subgroups) and 10 for LADA subjects. Assoc., Association; Indiv., individual; diff., difference; contr., healthy controls; pts., patients. role of MICA in human diseases is controversial. Nevertheless, its location in the HLA system and its polymorphism warrant studies finalized at testing the association between the MICA gene and or other immune-mediated human diseases. This is also supported by the results of several studies suggesting the existence of -associated genes in the chromosomic region containing the MICA gene (20, 31). It has been shown that MICA gene polymorphism influences the risk for Behçet s disease, anterior uveitis, Takayasu s arteritis, Buerger s disease, and psoriasis (32 35). In addition, we demonstrated that MICA5.1 is significantly increased in Italian patients with autoimmune Addison s disease (36). The association of MICA5.1 with Addison s disease has also been confirmed in the U.S. population (37). The results of our previous study of 95 Italian subjects (8), showing the significant association with MICA5, provide the best rationale for our present analysis of the influence of MICA gene polymorphism on the age at onset of autoimmune diabetes. In our study of 195 subjects, we confirm the significant association with MICA5 and the inverse relationship between the presence of this allele and the age at onset of. In addition, we provide new evidence of the significant association between MICA5.1 and adult-onset. This association was observed also in subjects with the so-called LADA, characterized by the presence of GAD65Ab in adult diabetic patients not requiring insulin therapy at diagnosis, suggesting a common genetic background for LADA and adult-onset. To the best of our knowledge, our study provides the first evidence of the association between LADA and MICA gene polymorphism, thus confirming indirectly the autoimmune origin of LADA. The demonstration of the association between MICA gene polymorphism and in different ethnic groups strengthens the importance of MICA as a genetic marker of the disease. On the other hand, different MICA alleles have been found in association with in different ethnic groups. Our study shows that in Caucasians the risk for childhood is increased in the presence of MICA5, whereas in Chinese the positively associated allele is MICA9 (9), and in Japanese (10) and Koreans (11) it is MICA4. Furthermore, our results on adult-onset are at variance with those of another study performed in the Japanese population (10). In Japanese subjects, adult-onset is negatively associated with MICA5.1 (10), the same allele we found to be positively associated with adult-onset Italian and Addison s disease. It is very well known that Japanese and Caucasian patients have different HLA- DR-DQ associations (38, 39) in part because of a different distribution of the HLA-DR and DQ haplotypes, and different linkage disequilibrium in the general population. Thus, a different distribution of MICA alleles in the general population in relation to ethnic group may also be responsible for the different results observed. However, we cannot rule out the possibility that the association between MICA and autoimmune diabetes is due to linkage disequilibrium with a still unidentified gene. If this is the case, a different linkage disequilibrium between MICA and the unknown gene in different ethnic groups may be the cause of the different results observed. HLA class II association with or LADA is depen-

6 Gambelunghe et al. MICA Polymorphism in Type 1 Diabetes The Journal of Clinical Endocrinology & Metabolism, August 2001, 86(8): dent upon age at disease onset (40 43). Accordingly, the hypothesis that the age-dependent association between MICA and autoimmune diabetes is secondary to a linkage disequilibrium with HLA-DR or DQ genes must be tested. Indeed, a strong disequilibrium exists across the HLA system. However, several lines of evidence support the hypothesis that the association of MICA with autoimmune diabetes is not exclusively determined by linkage disequilibrium with HLA class II genes. Firstly, the Svejgaard and Ryder analysis showed the independent associations of MICA and class II gene polymorphism with autoimmune diabetes, especially in the age group younger than 25 yr at the time of clinical diagnosis. More specifically, MICA5 was associated with childhood also in the absence of at risk HLA class II haplotypes. On the other hand, both MICA5.1 and HLA- DR3-DQ2 and/or DR4-DQ8 were required to significantly increase the risk of adult-onset /LADA. Accordingly, the association of class II gene polymorphism with is significantly reduced in the absence of MICA5 or MICA5.1 depending on the patients age. Taken together these results suggest an interaction between HLA class II genes and MICA in conferring an increased risk for autoimmune diabetes. Secondly, the linkage disequilibrium between the MICA and MICB genes is significantly weaker than that between MICA and HLA-B (44, 45), consistent with a recombination hot spot between the two MIC genes (46), which suggests an even weaker linkage between MICA and class II genes. Thirdly, we found the same class II haplotypes, namely HLA- DRB1*03-DQA1*0501-DQB1*0201 and DRB1*04-DQA1*0301- DQB1*0302, but two different MICA alleles, associated with childhood and adult-onset /LADA. Thus, it is unlikely that the different age-dependent MICA associations are mainly due to a linkage disequilibrium with class II haplotypes. In conclusion, our study demonstrates the age-dependent association of MICA gene polymorphism with autoimmune diabetes. Although we cannot exclude that MICA is associated with because of proximity with an as yet unidentified gene, the MICA gene polymorphism may be a useful genetic marker to estimate disease risk in the general population. The simultaneous presence of positively associated MICA and HLA class II haplotypes makes it possible to identify subjects at high risk for in the general population. Furthermore, the availability of distinct genetic markers for either rapid or slowly progressive may prove important for understanding the molecular mechanisms of genetic predisposition for this autoimmune disease as well as for the organization of clinical trials of primary and secondary prevention. Acknowledgments The following members of the Umbria Type 1 Diabetes Registry have collected clinical data and blood samples from diabetic subjects: A. Angeli (Gubbio), E. Baiocchi (Assisi), D. Belladonna (Todi), R. Bellanti (Città di Castello), M. Bracaccia (Orvieto), C. Campanelli (Città di Castello), G. Campolo (Todi), C. Cicioni (Terni), S. Coaccioli (Terni), A. Coletti (Gualdo Tadino), M. Cozzari (Cascia), G. De Giorgi (Perugia), G. Di Matteo (Perugia), G. Divizia (Spoleto), G. Fracassi (Marsciano), A. Fragasso (Foligno), A. Frascarelli (Assisi), S. Gagliardo (Narni), G. Giannico (Marsciano), A. Lilli (Gubbio), E. Madeo (Città della Pieve), G. Mancini (Orvieto), R. Marcacci (Castiglione del Lago), M. Napolitano (Foligno), G. Pennoni (Gualdo Tadino), E. Picchio (Perugia), S. Pocciati (Foligno), M. Scattoni (Narni), and E. Vignai (Città della Pieve). Received November 14, Accepted April 23, Address all correspondence and requests for reprints to: Alberto Falorni, M.D., Ph.D., Immunology and Immunogenetics Laboratory, Department of Internal Medicine and Endocrine and Metabolic Sciences, Via E. Dal Pozzo, Perugia, Italy. falorni@dimisem.med. unipg.it. This work was supported by the Juvenile Diabetes Foundation International (to A.F.), the Swedish Medical Research Council (K P C), the Karolinska Institute, the Swedish Diabetes Association, the Barndiabetes Fund, the Novo Nordisk Fund, the Åke Wiberg Fund, the Swedish Physicians Association (to C.B.S.), Telethon (Grant E.C787), and the Italian Ministry of University Research and Scientific Technology (MURST Project ). References 1. Lernmark Å, Falorni A 1997 Immunology of insulin-dependent diabetes mellitus. In: Pickup J, Williams C, eds. Textbook of diabetes. Oxford: Blackwell; Sanjeevi CB, Kockum I, Lernmark Å 1995 The role of major histocompatibility complex in insulin-dependent diabetes mellitus. Curr Opin Endocrinol Diabetes 2: Davies JL, Kawaguchi Y, Bennett ST, et al A genome-wide search for human type 1 diabetes susceptibility genes. Nature 371: Hashimoto L, Habita C, Beressi J, et al Genetic mapping of a susceptibility locus for insulin-dependent diabetes mellitus on chromosome 11q. Nature 371: Field LL, Tobias R, Magnus T 1994 A locus on chromosome 15q26 (IDDM3) produces susceptibility to insulin-dependent diabetes mellitus. Nat Genet 8: Owerbach D, Gabbay KH 1995 The HOXD8 locus (2q31) is linked to type 1 diabetes; interaction with chromosome 6 and 11 disease susceptibility genes. Diabetes 44: Nisticò L, Buzzetti R, Pritchard LE, et al The CTLA-4 gene region of chromosome 2q33 is linked to, and associated with, type 1 diabetes. Hum Mol Genet 5: Gambelunghe G, Ghaderi M, Cosentino A, Falorni Ad, Brunetti P, Falorni Al, Sanjeevi CB 2000 Association of MHC Class I chain-related A (MIC-A) gene polymorphism with type I diabetes. Diabetologia 43: Lee YJ, Huang FY, Wang CH, et al Polymorphism in the transmembrane region of the MICA gene and type 1 diabetes. J Pediatr Endocrinol Metab 13: Kawabata Y, Ikegami H, Kawaguchi Y, et al Age-related association of MHC class I chain-related gene A (MICA) with type 1 (insulin-dependent) diabetes mellitus. Hum Immunol 61: Park Y, Lee H, Sanjeevi CB, Eisenbarth GS 2001 MICA polymorphism is associated with type 1 diabetes in the Korean population. Diabetes Care 24: Ota M, Katsuyama Y, Mizuki N, et al Trinucleotide repeat polymorphism within exon 5 of the MICA gene (MHC class I chain-related gene A): allele frequency data in the nine population groups Japanese, Northern Han, Hui, Uygur, Kazakhstan, Iranian, Saudi Arabian, Greek and Italian. Tissue Antigens 49: Vandewalle CL, Decraene T, De Schuit FC, Leeuw IH, Pipeleers DG, Gorus FK, Belgian Diabetes Registry 1993 Insulin autoantibodies and high titre islet cell antibodies are preferentially associated with the HLA-DQA1*0301- DQB1*0302 haplotype at clinical onset of type I (insulin-dependent) diabetes before age 10 years but not at onset between age 10 and 40 years. Diabetologia 36: Seissler J, de Sonnaville JJ, Morgenthaler NG, et al. Immunological heterogeneity in type I diabetes: presence of distinct autoantibody patterns in patients with acute onset and slowly progressive disease. Diabetologia 41: Turner R, Stratton I, Horton V, et al UKPDS 25: autoantibodies to islet-cell cytoplasm and glutamic acid decarboxylase for prediction of insulin requirement in type 2 diabetes. Lancet 350: Tuomi T, Carlsson A, Li H, et al 1999 Clinical and genetic characteristics of type 2 diabetes with and without GAD antibodies. Diabetes 48: Falorni A, Gambelunghe G, Forini F, et al 2000 Autoantibody recognition of COOH-terminal epitopes of GAD65 marks the risk for insulin requirement in adult-onset diabetes mellitus. J Clin Endocrinol Metab 85: Vandewalle CL, Falorni A, Svanholm S, Lernmark Å, Pipeleers DG, Gorus FK, Belgian Diabetes Registry 1995 High diagnostic sensitivity of glutamate decarboxylase autoantibodies in insulin-dependent diabetes mellitus with clinical onset between age 20 and 40 years. J Clin Endocrinol Metab 80: Vandewalle CL, Falorni A, Lernmark Å, et al GAD65-autoantibodies are

7 3760 The Journal of Clinical Endocrinology & Metabolism, August 2001, 86(8): Gambelunghe et al. MICA Polymorphism in Type 1 Diabetes associated with HLA-DQ and 5 insulin gene-linked risk markers in new onset IDDM patients, aged years. Diabetes Care 20: Demaine AG, Hibberd ML, Mangles D, Millward BA 1995 A new marker in the HLA class I region is associated with the age at onset of IDDM. Diabetologia 38: National Diabetes Data Group 1979 Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 28: Sanjeevi CB, Seshiah V, Möller E, Olerup O 1992 Different genetic background for type I diabetes and malnutrition related diabetes in South Indians. Diabetologia 35: Woolf B 1995 On estimating the relation between blood group and disease. Ann Hum Genet 19: Miettinen OS 1976 The estimability and estimation in case-referent studies. Am J Epidemiol 103: Svejgaard A, Ryder L 1994 HLA and disease associations: detecting the strongest association. Tissue Antigens 43: Bahram S, Bresnahan M, Geraghty DE, Spies T 1994 A second lineage of mammalian major histocompatibility complex class I genes. Proc Natl Acad Sci USA 91: Bahram S, Mizuki N, Inoko H, Spies T 1996 Nucleotide sequence of the human MHC class I MICA gene. Immunogenetics 44: Zwirner NW, Fernandez-Vina MA, Statsny P 1998 MICA, a new polymorphic HLA-related antigen, is expressed mainly by keratinocytes, endothelial cells and monocytes. Immunogenetics 47: Bauer S, Groh V, Wu J, et al Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science 285: Wu J, Song Y, Bakker ABH, et al An activating immunoreceptor complex formed by NKG2D and DAP10. Science 285: Degli-Esposti M, McCann V, Abraham L, Spies T, Christiansen F, Dawkins R 1992 Ancestral haplotypes reveal the role of the central MHC in the immunogenetics of IDDM. Immunogenetics 36: Mizuki N, Ota M, Kimura M, Ohno S, Ando H, et al Triplet repeat polymorphism in the transmembrane region of the MICA gene: a strong association of six GCT repetitions with Behçet disease. Proc Natl Acad Sci USA 94: Goto K, Ota M, Maksymowych WP, et al Association between MICA gene A4 allele and acute anterior uveitis in white patients with and without HLA-B27. Am J Ophthalmol 126: Kimura A, Kobayashi Y, Takahashi M, et al MICA gene polymorphism in Takayasu s arteritis and Buerger s disease. Int J Cardiol 66(Suppl 1):S107 S Cheng L, Zhang SX, Xiao CY, et al The A5.1 allele of the major histocompatibility complex class I chain-related gene A is associated with psoriasis vulgaris in Chinese. Br J Dematol 143: Gambelunghe G, Falorni A, Ghaderi M, et al Microsatellite polymorphism of the MHC class I chain-related (MIC-A and MIC-B) genes marks the risk for autoimmune Addison s disease. J Clin Endocrinol Metab 84: Park YS, Sanjeevi CB, Yu L, et al. Genetic determinants of Addison s disease in patients with and without type 1 diabetes mellitus. Proc of the 4th Immunol of Diabetes Soc Congr. 1999; Aparicio JMR, Wakisaka A, Takada A, Matura N, Aizawa M 1988 HLA-DQ system and insulin-dependent diabetes mellitus in Japanese: does it contribute to the development of IDDM as it does in Caucasians? Immunogenetics 28: Awata T, Kuzuya T, Matsuda A, et al High frequency of aspartic acid at position 57 of the HLA-DQ chain in Japanese IDDM patients and nondiabetic subjects. Diabetes 39: Ludvigsson J, Samuelson U, Beauforts C, et al HLA-DR3 is associated with a more slowly progressive form of type I (insulin-dependent) diabetes. Diabetologia 29: Knip M, Ilonen J, Mustonen A, Akerblom HK 1986 Evidence of an accelerated -cell destruction in HLA-Dw3/Dw4 heterozygous children with type I (insulin-dependent) diabetes. Diabetologia 29: Caillat-Zucman S, Garchon H-J, Timsit J, et al Age-dependent HLA genetic heterogeneity of type I insulin-dependent diabetes mellitus. J Clin Invest 90: Horton V, Stratton I, Bottazzo GF, et al Genetic heterogeneity of autoimmune diabetes: age of presentation in adults is influenced by HLA DRB1 and DQB1 genotypes (UKPDS 43). Diabetologia. 42: Ando H, Mizuki N, Ota M, et al Allelic variants of the human MHC class I chain-related B gene (MICB). Immunogenetics 46: Kimura T, Goto K, Yabuki K, et al Microsatellite polymorphism within the MICB gene among Japanese patients with Behçet s disease. Hum Immunol 59: Shiina T, Tamiya G, Oka A, et al Nucleotide sequencing analysis of the 146-kilobase segment around the I BL and MICA genes at the centromeric end of the HLA class I region. Genomics 47: