HLA Genetic Discrepancy Between Latent Autoimmune Diabetes in Adults and Type 1 Diabetes: LADA China Study No. 6

Transcription

1 ORIGINAL ARTICLE HLA Genetic Discrepancy Between Latent Autoimmune Diabetes in Adults and Type 1 Diabetes: LADA China Study No. 6 Shuoming Luo,* Jian Lin,* Zhiguo Xie,* Yufei Xiang, Peilin Zheng, Gan Huang, Xia Li, Yu Liao, William A. Hagopian, Cong-Yi Wang, and Zhiguang Zhou Department of Metabolism and Endocrinology (S.L., J.L., Z.X., Y.X., P.Z., G.H., X.L., Y.L., C.-Y.W., Z.Z.), Second Xiangya Hospital and Diabetes Center, Institute of Metabolism and Endocrinology, Central South University, and Key Laboratory of Diabetes Immunology, Ministry of Education, National Clinical Research Center for Metabolic Diseases, Changsha, Hunan , China; Pacific Northwest Diabetes Research Institute and University of Washington (W.A.H.), Seattle, Washington 98122; and The Center for Biomedical Research (C.-Y.W.), Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei , China Context: The discrepancies in terms of human leukocyte antigen (HLA)-DRB1-DQA1-DQB1 conferred risks between latent autoimmune diabetes in adults (LADA) and type 1 diabetes (T1D) patients remained almost completely unknown. The goal of the current study is to determine and compare HLA-conferred risks between LADA and T1D. Design: A case-control study was conducted in a representative Chinese data set containing 520 T1D patients, 562 LADA patients, and 1065 controls. The frequencies and odds ratios for HLA susceptible haplotypes and genotypes and for arginine at residue 52 in the DQ- chain or aspartic acid at residue 57 in the DQ- chain were analyzed. Results: DRB1*0405-DQB1*0401 and DRB1*0901-DQB1*0303 are the major LADA susceptible haplotypes, which also confer comparable risks for T1D (odds ratio 2.02 vs 2.20 and 1.61 vs 2.30, respectively). The strongly associated T1D haplotype DRB1*0301- DQA1*05-DQB1*0201 is also associated with LADA but confers only half of the T1D risk (odds ratio 2.65 vs 4.84). Interestingly, the most susceptible T1D haplotypes, DRB1*0901-DQA1*05- DQB1*0201, DRB1*0301-DQB1*0201, and DRB1*0301-DQB1*0303, are not associated with LADA. Genotypes for DR3/DR3, DR3/DR9, and DR9/DR9 are highly associated with T1D susceptibility, whereas only DR9/DR9 confers risk for LADA. DR3/DR3 is the high-risk genotype in Chinese T1D patients, which manifests similar risk as the DR3/DR4 genotype in Caucasians but with a lower frequency. DR9/DR9 is the high risk LADA genotype in Chinese. Alleles with DQ- arginine at residue 52-positive, DQ- aspartic acid at residue 57- negative, and their combination formed in cis or trans confer susceptibility to T1D but not to LADA. Conclusion: Our results suggest that LADA risk conferred by HLA-DRB1-DQA1-DQB1 loci in Chinese differs significantly from that of T1D risk. This information would be useful for classifying Asian LADA patients, which should provides novel insight into the understanding of its pathoetiology as well. (J Clin Endocrinol Metab 101: , 2016) ISSN Print X ISSN Online Printed in USA Copyright 2016 by the Endocrine Society Received October 24, Accepted February 5, First Published Online February 11, 2016 * S.L., J.L., and Z.X. contributed equally to this work. Abbreviations: Arg52, arginine at residue 52; Asp57, lacking an aspartic acid at residue 57; CI, confidence interval; HLA, human leukocyte antigen; LADA, latent autoimmune diabetes in adults; OR, odds ratio; T1D, type 1 diabetes. doi: /jc J Clin Endocrinol Metab, April 2016, 101(4): press.endocrine.org/journal/jcem 1693

2 1694 Luo et al HLA Genetic Discrepancy Between LADA and T1D J Clin Endocrinol Metab, April 2016, 101(4): The prevalence of diabetes in China has increased dramatically over past several decades, and to date more than 100 million people are estimated to be affected by this devastating disorder (1). Type 1 diabetes (T1D) results from a cellular-mediated autoimmune destruction of the pancreatic -cells, in which -cell-specific autoantibodies can be served as useful markers for disease prediction and diagnosis (2). In general, the prevalence of T1D varies among ethnic groups substantially as manifested by that it is quite common in subjects originated from Finland and Sardinia, whereas it is relatively rare in the Asian populations such as in Chinese and Japanese (3). Recent studies indicated an increasing trend for T1D incidence in European countries (4, 5), and a similar trend was also noted in China (6). Latent autoimmune diabetes in adults (LADA) is a slowly progressive form of autoimmune-associated diabetes (7). It encloses a group of patients with features of type 2 diabetes but manifesting autoimmune components. They are characterized as the production of autoantibodies against islet cell antigens such as glutamic acid decarboxylase 65, and noninsulin-dependent for at least 6 months after diagnosis. The LADA China study showed that around 5.9 of diabetic patients were classified as LADA (8). Despite past extensive studies, the disease etiology underlying LADA remains poorly understood. It has been well recognized that the human leukocyte antigen (HLA) class II genes serve as a major genetic factor for T1D susceptibility. Particularly, the HLA-DRB1- DQA1-DQB1 genotypes and haplotypes confer the highest risk for T1D (9). The DQ chain with an arginine at residue 52 (Arg52 ) and the DQ- chain lacking an aspartic acid at residue 57 (Asp57 ) also confer susceptibility to T1D (10). However, significant variations in terms of the risks conferred by the HLA genes are noted in different ethnic groups (11). Given that LADA shares similar clinical features as T1D, it is plausible that HLA genes are also associated with LADA susceptibility. Indeed, several studies demonstrated feasible evidence supporting the association between HLA genes and LADA risk (12 15). It is noteworthy that most of these studies were conducted with focus to address HLA-conferred risks either in T1D or LADA susceptibility, whereas the discrepancies in terms of HLA-conferred risks between LADA and T1D patients remained completely unknown. The lack of this information not only hampered the development of approach for classifying this category of patients but also impacted the understanding of its disease pathoetiology. We thus in the current report assessed the contribution of HLA-DRB1-DQA1-DQB1 haplotypes and genotypes in LADA and T1D susceptibility in patients with Chinese Han origin. The impact of Arg52 polymorphism in DQ- chain and the effect of Asp57 in DQ- chain on LADA and T1D risk were also analyzed. Patients and Methods T1D patients T1D defined in this study is specific for classical type 1 diabetes (namely, type 1A diabetes). A total of 520 classical T1D patients with Chinese Han origin were recruited from June 1999 to December Of these, the vast majority of them were collected from Hunan province of China. The age at onset of disease (274 males and 246 females) ranged from 1 to 67 years (median onset 21 y old). However, most of the patients were diagnosed as children or juveniles. T1D patients were considered for inclusion by the following criteria: 1) diagnosis of diabetes (World Health Organization criteria of 1999); 2) acute onset and presence of diabetic ketosis or ketoacidosis; 3) positive for at least one of the islet autoantibodies (glutamic acid decarboxylase antibody, protein tyrosine phosphatase antibody, and zinc transporter 8 antibody); and 4) insulin dependency at the time of diagnosis. Exclusion criteria were as follows: patients with other forms of diabetes including LADA, maturity onset diabetes of the young, secondary diabetes, or type 2 diabetes. LADA patients A total of 562 LADA patients with Chinese Han origin (316 males and 246 females) within the same time period were enrolled. The age for onset of LADA ranged from 30 to 82 years (median 50 y old). Most of the LADA patients were enrolled in Hunan province with a few cases collected from other cities by the LADA China Study (8). Inclusion criteria for LADA patients were as follows: islet autoantibody-positive patients, initially noninsulin requirement for at least 6 months with disease onset at 30 years old or older. Patients with manifestations of acuteonset T1D, secondary diabetes, diabetes in pregnancy, gestational diabetes mellitus, and malignancy were excluded from the study. In general, LADA is distinguished from T1D by the absence of insulin requirement for a minimum of 6 months upon diagnosis and absence of ketoacidosis and weight loss, whereas LADA differs from type 2 diabetes by the manifestation of the presence for glutamic acid decarboxylase autoantibodies. Control subjects A group of 1065 control subjects were recruited from two cross-sectional surveys (590 males and 475 females; aged y, ranging from 1 to 80 y old) (16). All subjects had normal glucose tolerance and no family history of diabetes. Control subjects of all ages were included and matched to the cases ethnically and geographically. Ethics Informed consent was obtained from all study subjects or their parents. The study was approved by the Ethical Committee of Second Xiangya Hospital of Central South University and was carried out in agreement with the guidelines in the Declaration of Helsinki.

3 doi: /jc press.endocrine.org/journal/jcem 1695 Genotyping and construction of haplotypes All subjects were genotyped for HLA-DRB1, HLA-DQA1, and HLA-DQB1 by directly DNA sequencing as described previously (17). Sequencing exon 2 of each gene yielded two- or four-digit resolution for HLA-DQA1 and four-digit resolution for HLA-DQB1 and HLA-DRB1. Each PCR contained 3.5 L deionized water, 12.5 L2 Taq PCR master mix (Fermentas), 1 L primer eight primers (10 M) (18), and 1 L DNA to a final volume of 25 L. Amplification was carried out in a Biometra T gradient PCR instrument using the following program: a preheating step at 98 C for 3 minutes, 40 cycles of 95 C for 20 seconds, annealing at 56 C for 10 seconds, extension at 72 C for 30 seconds, and finally a hold at 4 C. All samples were sequenced on an ABI Prism 3730XL DNA analyzer (Applied Biosystems). Alleles for HLA-DRB1, DQA1, and DQB1 were assigned by comparing with sequence library from IMGT/HLA sequence database (Release IMGT_HLA_ _350.1) using Assign-SBT version (Conexio Genomics) (19). Because this typing system could not target all polymorphic sites, some alleles were not distinguished. For example, alleles for DQA1*0101 and DQA1*0104 cannot be distinguished, and therefore, DQA1*0101g was used to designate this ambiguity. Similarly, DQA1*0601g was used to designate the ambiguity of DQA1*0601 and DQA1*0602. The DR-DQ haplotypes were constructed by using the PHASE program (20, 21). Because of the strong linkage disequilibrium between DRB1, DQA1, and DQB1, the haplotypes can be deduced with high probability in individuals without family data. Analysis of polymorphisms with amino acid changes Based on published data and cdna and genomic sequences in the public databases, we defined five susceptible DQA1 alleles coding for DQ (Arg52 ) (DQA1*: 03, 0302, 0401, 05, and 0601g) and five Arg52 alleles as protective (DQA1*: 0101g, 0102, 0103, 0201, and 0202) (10, 22). Similarly, 11 DQB1 genes coding for DQ Asp57 were considered susceptible (DQB1*: 0201, 0302, 0304, 0305, 0501, 0502, 0505, 0604, 0605, 0610, and 0613), and nine DQB1 genes for Asp57 were considered protective (DQB1*: 0301,0303, 0401, 0402, 0403, 0503, 0601, 0602, and 0603) (10, 23). Statistical analysis Statistical analysis was conducted by SPSS statistical software (version 13; SPSS Inc). Frequency for alleles, genotypes, and haplotypes was obtained by simple counting. Differences between groups were analyzed by the 2 or Fisher s exact test when appropriate. P values were corrected for the number of different alleles, genotypes, or haplotypes tested with Bonferroni correction (denoted as P c ). Data were expressed as P c values, odds ratio (OR), and 95 confidence intervals (CIs) between patients and controls. Statistical significance was set at a value of P c.05. Results T1D associated HLA-DRB1-DQA1-DQB1 haplotypes Twenty-four haplotypes were detected to be significantly associated (positive or negative) with T1D susceptibility. Once a highly conservative Bonferroni correction for multiple testing was applied; 12 of these haplotypes still showed significant association. Table 1 shows the frequencies for HLA-DRB1-DQA1-DQB1 susceptible or protective haplotypes of patients with T1D and LADA. Some of these haplotypes are well-established risk factors for T1D in Caucasian populations, including the susceptible DRB1* 0301-DQA1*05-DQB1*0201(DR3) and DRB1*0405- DQB1*0302(DR4) haplotypes, and the protective DRB1*1101-DQA1*05-DQB1*0301 haplotype. Importantly, some disease-related haplotypes appeared to be unique in our Chinese data set, which were absent in the published Caucasian data sets (9, 24). As shown in Table 1, notable examples of unique protective haplotypes in our data set include DRB1*1401-DQA1*0101g-DQB1* 0502, DRB1*0803-DQA1*0103-DQB1*0601, DRB1* 1202-DQA1*0601g-DQB1*0301, and DRB1*0803- DQB1*0303. Unique susceptible haplotypes include DRB1*0301-DQB1*0201, DRB1* 0301-DQB1*0303, DRB1*0901-DQA1*05- DQB1*0201, DRB1*0901-DQB1*0303, and DRB1* 0405-DQB1*0401. Risk genotypes in T1D The frequencies for HLA-DRB1-DQA1-DQB1 genotypes of T1D patients and control subjects are shown in Table 2. It was noted that the frequencies for DR3/DR3 (3.65), DR3/DR9 (5.58), and DR9/DR9 (16.35) genotypes in T1D patients were significantly higher than that in control subjects (P c 0.001). They were high-risk genotypes and displayed a significant hierarchy of disease risk. The most susceptible genotype in our data set was the DR3/ DR3 homozygous genotype (OR 40.35), followed by the DR3/DR9 heterozygote (OR 12.52), and the DR9/DR9 homozygote was relatively the least one (OR 5.75). Comparison of HLA-conferred risks between LADA and T1D The frequencies for haplotypes DRB1*0301-DQA1* 05-DQB1*0201(DR3), DRB1*0405-DQB1* 0401, and DRB1*0901-DQB1*0303(DR9) were significantly higher both in LADA and T1D patients as compared with that of control subjects (Table 1). However, the frequency for the DR3 haplotype in LADA patients was noted to be significantly lower than that in T1D patients (8.27 vs 14.13, P c ). Haplotype DRB1*0405-DQB1*0401 showed similar frequencies between the LADA patients and T1D patients along with similar odds ratios (OR 2.02 vs 2.20). Of note, haplotype DRB1*0901-DQB1*0303 manifested the highest frequency in LADA patients (22.33), whereas its OR was the lowest (OR 1.61). Haplotypes

4 1696 Luo et al HLA Genetic Discrepancy Between LADA and T1D J Clin Endocrinol Metab, April 2016, 101(4): Table 1. HLA-DRB1-DQA1-DQB1 Haplotype Frequencies Among Chinese T1D, LADA and Controls DR-DQ Haplotypes T1D, LADA, Controls, T1D vs Controls LADA vs Controls LADA vs T1D OR (95 CI) P c OR (95 CI) P c OR (95 CI) P c n DRB1* ( ) NS 0.1 ( ) DQB1*0201 a DRB1* ( ) NS NS DQB1*0303 a DRB1* ( ) ( ) ( ) DQA1*05- DQB1*0201 DRB1* ( ) NS NS DQB1*0302 DRB1* ( ) ( ) NS DQB1*0401 a DRB1* ( ) NS NS DQA1*0103- DQB1*0601 b DRB1* ( ) ( ) NS DQB1*0303 a DRB1* ( ) NS NS DQA1*05- DQB1*0201 a DRB1* ( ) NS NS DQA1*05- DQB1*0301 DRB1* ( ) NS 2.79 ( ) DQA1*0601g- DQB1*0301 b DRB1* ( ) NS 8.19 ( ) DQA1*0101g- DQB1*0502 b DRB1*0803- DQB1*0303 b ( ) ( ) Abbreviations: DR3, DRB1*0301-DQA1*05-DQB1*0201; DR4, DRB1*0405-DQB1*0302; DR9, DRB1*0901-DQB1*0303; NS, no significance. a Unique protective haplotypes in Chinese. b Unique predisposing haplotypes in Chinese. DRB1*0301-DQB1*0303, DRB1*0405-DQA1* 03-DQB1*0302, and DRB1*0901-DQA1*05-DQB1*0201 werefoundconferringriskfort1dbutnotforlada. Itislikely that haplotype DRB1*0803-DQB1*0303 acts as a weakladaprotectivehaplotype(or0.99). Thefrequencyfor total susceptible haplotypes in LADA patients was found to be significantly lower than that in T1D patients (38.17 vs 59.23,P ).Onthecontrary,thefrequencyfor T1D-associated protective haplotypes was significantly higher in LADA as compared with that in T1D (13.08 vs 4.52, P ). The frequency for the DR9/DR9 homozygous genotype was markedly higher both in LADA and T1D than that in controls (Table 2). However, this genotype seemed to confer much higher risk for T1D, given that its frequency and OR in LADA were only half of that in Table 2. HLA-DRB1-DQA1-DQB1 Genotype Frequencies Among Chinese T1D, LADA and Controls DR-DQ Genotypes T1D, LADA, Controls, T1D vs Controls LADA vs Controls LADA vs T1D OR(95) P c OR(95) P c OR(95) P c n DR3/DR ( ) NS NS DR3/DR ( ) NS NS DR9/DR ( ) ( ) ( ) Others ( ) ( ) ( ) Abbreviations: DR3, DRB1*0301-DQA1*05-DQB1*0201; DR9, DRB1*0901-DQB1*0303; NS, no significance.

5 doi: /jc press.endocrine.org/journal/jcem 1697 T1D patients. It is noteworthy that the DR3/DR3 and DR3/DR9 genotypes confer genetic susceptibility only to T1D, not to LADA after Bonferroni correction. The impact of Arg52 and Asp57 polymorphisms on disease risk In line with our expectation, the Arg52 /Arg52 homozygous genotype in the DQ- and the Asp57 / Asp57 homozygous genotype in the DQ- chain were highly associated with T1D susceptibility (both P c.01, Supplemental Figure 1). In contrast, we failed to obtain evidence supporting the association between Arg52 and Asp57 genotypes and LADA susceptibility. Next, we analyzed the impact of combination of DQ- Arg52 and DQ- Asp57 formed in cis or trans on T1D susceptibility. A highly significant association to T1D was detected in individuals homozygous for two susceptibility heterodimers (S-S/S-S, the highest relative risk, OR 4.84), followed by S-S/S-P (OR 2.67), and the degree of susceptibility appears to be proportional to the number of dimmers (ORs S-S/S-S S-S/S-P S-S/P-P S-P/P-P). No significant association was observed for these combinations between LADA patients and controls (Supplemental Figure 2). The cis- and trans-encoded DQ- / heterodimers in disease risk To further assess whether all DQ- and susceptible heterodimers account for a similar risk to T1D, we compared the frequencies for different DQA1-DQB1 combinations formed in cis and/or in trans between patients and control subjects. The DQB1*0201 combination, formed mainly in trans, showed the highest OR (OR 7.71), whereas the DQA1*05- DQB1*0201 combination, predominantly formed in cis, had a lower OR (OR 6.32). Both DQB1*0303 and DQB1*0401 (both formed mainly in cis) manifested similar ORs (Supplemental Table 1). Interestingly, the DQB1*0201 combination and DQA1*05-DQB1*0201 combination displayed a similar trend of action as T1D susceptibility in LADA patients but with a relatively lower risk. Discussion Over the past decades, knowledge relevant to genetic factors in T1D susceptibility has grown substantially, particularly for the characterization of HLA. The HLA region contributes approximately 50 of the T1D genetic risk and mostly conferred by the HLA-DR and DQ genes (24). Although the impact of HLA genes on T1D susceptibility has been extensively studied, most of the related data, however, were from Caucasian populations (24) and notably from the T1DGC studies (9). Particularly, specific combinations of alleles or haplotypes of the DRB1, DQA1, and DQB1 genes (in cis and trans) determine the extent of risk and the distribution of DR-DQ haplotypes and genotypes ranges from highly susceptible to highly protective (25). Studies in Caucasians revealed that the highest risk HLA- DR and DQ haplotypes for T1D are DRB1*03- DQA1*0501-DQB1*0201 (DR3) and DRB1*04- DQA1*0301-DQB1*0302 (DR4), whereas the highest risk genotype is DR3/DR4 (OR 40). In contrast, DRB1*15-DQA1*0102-DQB1*0602 and DRB1*14- DQA1*0101-DQB1*0503 confer T1D protection predominantly in Caucasians (24). Unfortunately, the detailed impact of HLA genes on T1D susceptibility in Chinese population is yet to be addressed. A previous pilot study revealed that the three haplotypes, DR3, DR4, and DR4-DQA1*0301-DQB1*04, confer T1D susceptibility for Chinese patients in Taiwan (26). Data resulted from this study suggest that DR3 and DR4 haplotypes manifest a similar susceptible effect on Chinese patients as that on Caucasians, whereas the third haplotype has been reported only in the Japanese population (26), indicating a particular haplotype or genotype from these genes may manifest a distinctive impact on disease susceptibility in different populations. Therefore, population-specific HLA haplotypes and genotypes need to be established for the identification of high-risk Chinese patients, and our current report in fact fulfilled this mission. Indeed, some unique protective and susceptible haplotypes were identified from this study, which would be useful for classifying Chinese T1D patients. Similarities and differences in the architecture of HLA-conferred susceptibility to T1D between Chinese and Caucasians are summarized in Table 3. Our study revealed that the frequency for DR3 haplotype in Chinese T1D patients was significantly lower than that in Caucasian patients. In particular, the frequency for high-risk haplotypes (OR 5.0) was 6.0, and the frequency for low- to moderate-risk haplotypes (5.0 OR 1.0) was 53.3 in Chinese T1D patients, whereas the frequencies for high-risk haplotypes and low- to moderate-risk haplotypes in Caucasian patients were 30.6 and 39.1, respectively (9). These observations indicate that the frequency for high-risk haplotypes in Chinese patients is lower than the corresponding counterparts, whereas the frequency for low- to moderate-risk haplotypes is higher in Chinese patients than that in Caucasian patients. The protective haplotypes in T1D risk assessment for Chinese also differed greatly from that in Caucasians (Table 3).

6 1698 Luo et al HLA Genetic Discrepancy Between LADA and T1D J Clin Endocrinol Metab, April 2016, 101(4): Table 3. Similarities and Differences in the Architecture of HLA-Conferred Susceptibility to T1D and LADA Between Chinese and Caucasians Chinese Caucasians a T1D DR3 haplotype 14.1 (OR 4.74) 34.1 (OR 3.63) DR4 haplotype 1.8 (OR 3.83) 2.5 (OR 11.37) DRB1*1501-DQA1*0102-DQB1*0602 Protective trend Strong protection (OR 0.02) DRB1*1401-DQA1*0101g-DQB1*0503 Protective trend Strong protection (OR 0.03) DRB1*0701-DQA1*0201-DQB1*0303 Neutral Strong protection (OR 0.02) DR3/DR3 genotype 3.65 (OR 35.22) 30 (OR 6.69) DR3/DR4 genotype 1.90 (OR 1.00) 40 (OR 39.0) DR3/DR9 genotype 5.58 (OR 12.52) Rare DR9/DR9 genotype (OR 5.75) Rare LADA DR3 haplotype 8.27 (OR 2.65) (OR ) b,c DR4 haplotype 7.56 (OR 2.02) (OR ) b,c High-risk genotype DR9/DR9 (8.54, OR 2.75) DR3/DR4 (17.6, OR 12.0) c a Reference 9. b Reference 12. c Reference 29. The DR3/DR9 and DR9/DR9 genotypes were rare in Caucasians, but they were frequent in Chinese, and both were Chinese-specific T1D risk genotypes. In contrast, the DR3/DR4 genotypes were the major risk factor in Caucasians (frequencies as high as 40, OR 37 41) (24), whereas the frequency for DR3/DR4 genotypes in Chinese patients was as low as 1.9 and manifested a susceptible trend only for T1D risk. These significant discrepancies in terms of HLA conferred T1D risks between Chinese and Caucasian populations can be explained by the involvement of DR-DQ linkage disequilibrium, DQ- transcomplementation and probably other HLA-linked genes (27). Previous studies suggested that most T1D-associated HLA haplotypes also confer LADA susceptibility (28). In line with this observation, three of the six T1D-associated haplotypes in our data set are also noted to be associated with LADA susceptibility. Specifically, LADA-associated DRB1*0901-DQB1*0301 haplotype characterized in the Norway study is also associated with LADA susceptibility in our data set (28), whereas haplotypes DRB1*0405-DQB1*0401 and DRB1*0301-DQA1*05-DQB1*0201 are noted to confer susceptibility for both LADA and T1D in our data set, suggesting that both diseases share some common genetic factors. However, the related frequency and OR differ significantly, demonstrating the existence of discrepancy in terms of genetic susceptibility. Indeed, three T1D associated haplotypes failed to show evidence for conferring LADA susceptibility in our data set. In general, the results originated from our data set are consistent with previous reported data in Caucasians (15). These findings suggest that LADA shares some common HLA factors with T1D, whereas a significant discrepancy does exist between T1D and LADA in terms of disease susceptibility conferred by HLA genes. It is worthy of note that the susceptible haplotypes or genotypes are less frequent, whereas the protective haplotypes are more frequent in LADA patients as compared with that of T1D patients. This discrepancy may contribute to the manifestation of the late onset of LADA as compared with T1D, which also suggest that LADA is likely to be a mild form of T1D with a distinct etiological entity. This information could be also useful for classifying Asian LADA patients and should provide novel insight into the understanding of pathoetiology underlying LADA. Studies in Caucasian LADA patients revealed that DR3 and DR4 are common susceptible genotypes, and DRB1*1501-DQB1*0602 is a protective haplotype (12, 29). In contrast, data resulted from our study indicated that DR3/DR4 is the high-risk LADA genotype (Table 3). Furthermore, the major LADA-susceptible determinants in the Chinese are markedly different from that of Caucasians. These discrepancies suggest that LADA risk assessment in the Chinese by HLA genes could also differ greatly from that in Caucasians. Susceptibility to T1D in Caucasians has been shown to be highly correlated with Arg52 in the DQ- chain or Asp57 in the DQ- chain (30, 31), and their combination also confers T1D susceptibility (10). In our data set, we demonstrated evidence indicating that the DQ- Arg52, DQ- Asp57 and their combination confer to T1D risk but not to LADA susceptibility. Although the results were negative in LADA, this was the first study to dissect the relationship between amino acid polymorphism and LADA susceptibility. We further noted that the degree of

7 doi: /jc press.endocrine.org/journal/jcem 1699 T1D susceptibility depends on the number of DQ heterodimers, which can be formed in a particular subject according to his or her DQA1-DQB1 genotype. As expected, the highest OR was observed in T1D patients with four susceptible DQ heterodimers. This observation raised the question whether all susceptible DQ heterodimers confer similar risk of T1D. Analysis of DQ combinations formed in cis and/or trans showed that a restricted number of DQ combinations were significantly associated with T1D, but the conferred risk varied greatly. These data support the implication of additional amino acid polymorphisms in T1D susceptibility other than the recognized Asp57 and Arg52 polymorphisms. The strongest association was detected in the DQB1*0201 combination mainly formed by genes in trans position. We found that four different susceptible heterodimers are correlated significantly with T1D, and these four heterodimers are also correlated significantly with LADA. In general, genetic association studies assess the relationship between genetic variants and disease state on a population scale, and inconsistent results could be produced due to small sample size, particularly for those genes with weak risk. Although the data set presented in the current report is thus far the largest LADA and T1D data set in Chinese populations, we still cannot exclude the possibility that our data are completely absent of inconsistency due to small sample size. However, this situation is unlikely the case for our study because HLA genes are not low-risk T1D genes, and they confer greater than 50 of total genetic risk for T1D. Another critical point that may render data from the current study with limitation is the classification of T1D and LADA patients. T1D in this report was defined as the classical type 1A diabetes. Its main clinical features include acute onset, presence of diabetic ketosis or ketoacidosis, positive for at least one of the islet autoantibodies, disease onset usually at the stage of childhood or juvenile, lack of endogenous insulin secretion, and requirement for insulin injection at the time of diagnosis. Unlike T1D, LADA is absent in children or juveniles, and LADA patients do not require insulin therapy for at least a period of 6 months upon diagnosis, whereas LADA can be distinguished from type 2 diabetes by the presence of islet autoantibodies. It also should be noted that subjects included in this study were mostly recruited from Hunan province, in central China, and therefore, the statements discussed above may be related only to this particular area of China. Another critical issue that needs to be pointed out is the age of T1D onset. The vast majority of T1D patients included in this data set were manifested T1D onset at the age of childhood or juvenile stage, but a few T1D cases were actually with disease onset at the adult stage. To demonstrate whether age for T1D onset would affect our conclusion, we conducted studies by exclusion of those patients with manifestation of T1D onset at adult stage, and we failed to observe any discrepancy in terms of disease risk for a particular genotype or haplotype by either inclusion or exclusion of those cases with disease onset at adult stage. Nevertheless, additional nationwide and multicenter studies would be necessary to further confirm our conclusions. In summary, we demonstrated that the contribution of HLA-DRB1-DQA1-DQB1 loci to LADA susceptibility in the Chinese differs significantly from that to T1D and also differs obviously from that to LADA patients with Caucasian origin, which could be useful for the classification of Asian LADA patients. It is worthy of note that some strongly associated T1D HLA-DRB1-DQA1-DQB1-risk genes are also found to confer LADA susceptibility, but the conferred risk varies greatly. Our data should provide the feasibility to predict the risk for developing LADA and T1D according to the Chinese-specific haplotypes and genotypes, which could also be useful for prevention trials. Collectively these findings should provide novel insight into the understanding of disease pathoetiology underlying LADA and T1D. Acknowledgments Address all correspondence and requests for reprints to: Zhiguang Zhou, MD, PhD, Department of Metabolism and Endocrinology, Second Xiangya Hospital, Central South University, and Key Laboratory of Diabetes Immunology, Ministry of Education, and National Clinical Research Center for Metabolic Diseases, Changsha, Hunan , China. zhouzg@hotmail.com; or Cong-Yi Wang, MD, PhD, The Center for Biomedical Research, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei , China. wangcy@tjh.tjmu.edu.cn. Author contributions included the following: S.L. designed the study, performed the experiments and data analysis, and wrote the manuscript. J.L. and Z.X. designed the study, performed the experiments, and contributed to the discussion of the manuscript. Y.X., P.Z., G.H., and X.L. contributed to the discussion of the manuscript. Y.L. performed the experiments. W.A.H. reviewed and edited the manuscript. Z.Z. and C.-Y.W. designed the study, contributed to the data analysis, the discussion, and reviewed and edited the manuscript. This work was supported by Grants , , and from the National Natural Science Foundation of China; Grant 2015JC3020 from the Key Research and Development Program of Hunan Province Science and Technology Project; Grant IRT1195 from the Program for Changjiang Scholars and Innovative Research Team in University; Grants 2012BAI02B04 and 2013BAI09B12 from the National Key Technology Research and Development Program; and Grant

8 1700 Luo et al HLA Genetic Discrepancy Between LADA and T1D J Clin Endocrinol Metab, April 2016, 101(4): from the Research Fund for the Doctoral Program of Higher Education of China. Disclosure Summary: The authors have nothing to disclose. References 1. Xu Y, Wang L, He J, et al. Prevalence and control of diabetes in Chinese adults. JAMA. 2013;310: Canivell S, Gomis R. Diagnosis and classification of autoimmune diabetes mellitus. Autoimmun Rev. 2014;13: Atkinson MA, Eisenbarth GS, Michels AW. Type 1 diabetes. The Lancet. 2014;383: Patterson CC, Dahlquist GG, Gyurus E, Green A, Soltesz G. Incidence trends for childhood type 1 diabetes in Europe during and predicted new cases : a multicentre prospective registration study. Lancet. 2009;373: Bruno G, Maule M, Biggeri A, et al. More than 20 years of registration of type 1 diabetes in Sardinian children: temporal variations of incidence with age, period of diagnosis, and year of birth. Diabetes. 2013;62: Zhao Z, Sun C, Wang C, et al. Rapidly rising incidence of childhood type 1 diabetes in Chinese population: epidemiology in Shanghai during Acta Diabetol. 2014;51: Fourlanos S, Dotta F, Greenbaum CJ, et al. Latent autoimmune diabetes in adults (LADA) should be less latent. Diabetologia. 2005; 48: Zhou Z, Xiang Y, Ji L, et al. Frequency, immunogenetics, and clinical characteristics of Latent Autoimmune Diabetes in China (LADA China Study): a nationwide, multicenter, clinic-based cross-sectional study. Diabetes. 2013;62: Erlich H, Valdes AM, Noble J, et al. HLA DR-DQ haplotypes and genotypes and type 1 diabetes risk: analysis of the type 1 diabetes genetics consortium families. Diabetes. 2008;57: Khalil I, d Auriol L, Gobet M, et al. A combination of HLA-DQ Asp57-negative and HLA DQ Arg52 confers susceptibility to insulin-dependent diabetes mellitus. J Clin Invest. 1990;85: Aly TA, Ide A, Jahromi MM, et al. Extreme genetic risk for type 1A diabetes. Proc Natl Acad Sci USA. 2006;103: Desai M, Zeggini E, Horton VA, et al. An association analysis of the HLA gene region in latent autoimmune diabetes in adults. Diabetologia. 2007;50: Howson JM, Rosinger S, Smyth DJ, Boehm BO, Todd JA. Genetic analysis of adult-onset autoimmune diabetes. Diabetes. 2011;60: Andersen MK, Lundgren V, Turunen JA, et al. Latent autoimmune diabetes in adults differs genetically from classical type 1 diabetes diagnosed after the age of 35 years. Diabetes Care. 2010;33: Vatay A, Rajczy K, Pozsonyi E, et al. Differences in the genetic background of latent autoimmune diabetes in adults (LADA) and type 1 diabetes mellitus. Immunol Lett. 2002;84: Xiang Y, Huang G, Zhou W, Che Z, Zhou P, Zhou Z. Prevalence of metabolic syndrome (MetS) in Chinese subjects gradually increased with impaired glucose homeostasis: a multicenter, clinical based, cross-sectional study. BMC Public Health. 2012;12: Wang JP, Zhou ZG, Lin J, et al. Islet autoantibodies are associated with HLA-DQ genotypes in Han Chinese patients with type 1 diabetes and their relatives. Tissue Antigens. 2007;70: Sayer D, Whidborne R, Brestovac B, Trimboli F, Witt C, Christiansen F. HLA-DRB1 DNA sequencing based typing: an approach suitable for high throughput typing including unrelated bone marrow registry donors. Tissue Antigens. 2001;57: Wirtz C, Sayer D. Data analysis of HLA sequencing using Assign- SBT v3.6 from Conexio. Methods Mol Biol. 2012;882: Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet. 2001;68: Stephens M, Donnelly P. A comparison of Bayesian methods for haplotype reconstruction from population genotype data. Am J Hum Genet. 2003;73: Khalil I, Deschamps I, Lepage V, al-daccak R, Degos L, Hors J. Dose effect of cis- and trans-encoded HLA-DQ heterodimers in IDDM susceptibility. Diabetes. 1992;41: Sanjeevi CB, Lybrand TP, DeWeese C, et al. Polymorphic amino acid variations in HLA-DQ are associated with systematic physical property changes and occurrence of IDDM. Members of the Swedish Childhood Diabetes Study. Diabetes. 1995;44: Noble JA, Valdes AM, Cook M, Klitz W, Thomson G, Erlich HA. The role of HLA class II genes in insulin-dependent diabetes mellitus: molecular analysis of 180 Caucasian, multiplex families. Am J Hum Genet. 1996;59: Pociot F, Akolkar B, Concannon P, et al. Genetics of type 1 diabetes: what s next? Diabetes. 2010;59: Hu CY, Allen M, Chuang LM, Lin BJ, Gyllensten U. Association of insulin-dependent diabetes mellitus in Taiwan with HLA class II DQB1 and DRB1 alleles. Hum Immunol. 1993;38: She JX. Susceptibility to type I diabetes: HLA-DQ and DR revisited. Immunol Today. 1996;17: Pettersen E, Skorpen F, Kvaloy K, Midthjell K, Grill V. Genetic heterogeneity in latent autoimmune diabetes is linked to various degrees of autoimmune activity: results from the Nord-Trondelag Health Study. Diabetes. 2010;59: Hosszufalusi N, Vatay A, Rajczy K, et al. Similar genetic features and different islet cell autoantibody pattern of latent autoimmune diabetes in adults (LADA) compared with adult-onset type 1 diabetes with rapid progression. Diabetes Care. 2003;26: Dorman JS, LaPorte RE, Stone RA, Trucco M. Worldwide differences in the incidence of type I diabetes are associated with amino acid variation at position 57 of the HLA-DQ chain. Proc Natl Acad Sci USA. 1990;87: Gutierrez-Lopez MD, Bertera S, Chantres MT, et al. Susceptibility to type 1 (insulin-dependent) diabetes mellitus in Spanish patients correlates quantitatively with expression of HLA-DQ Arg 52 and HLA-DQ non-asp 57 alleles. Diabetologia. 1992;35: