why is type 1 diabetes uncommon in asia?

Why Is Type 1 Diabetes Uncommonin Asia?

YONGSOO PARK

Division of Endocrinology and Metabolism, Department of Internal Medicine,Hanyang University Hospital, Hanyang University College of Medicine,471-020 Seoul, Korea

Department of Bioengineering, Hanyang University College of Engineering,471-020 Seoul, Korea

ABSTRACT: T1D (type 1 diabetes) incidence rates are extremely low inAsian populations. The prevalences of islet-specific autoantibodies arereported to be low compared with Caucasians. Although the clinicaland immunologic characteristics of T1D in Asians appear to be differ-ent from those of Caucasians, if we apply correct patient definition andstandardized methods, the typical T1D patients are very similar, in theimmunologic as well as genetic perspectives. Although the association ofindividual allele seems to be different between populations, if we comparethe identical DR-DQ haplotypes, the association and transmission to di-abetic offspring were similar for Asians and Caucasians. The high-riskHLA genotypes/haplotypes were found to be independent determinantsof diabetes in the first-degree relatives of individuals with T1D, particu-larly in the presence of autoantibodies. A different genetic susceptibilityincluding a low frequency of high-risk HLA alleles could explain the lowerprevalence of islet-specific autoantibodies and the low incidence of T1D,or different genetic and environmental interactions might be involvedin the etiology of T1D. It is certain that DR-DQ linkage disequilibrium(LD) is an important factor explaining the difference in T1D incidencein different countries. LD between highly susceptible DRB1 alleles andprotective DQB1 alleles, and vice versa, is the major contributing factorto the low incidence of T1D in Asians. We also suggested that differentgenetic/environmental interactions might operate in the etiology of T1Dbetween Caucasians and Asians. It would be of great help for primaryprevention to investigate to what degree genetic determinants influencethe well-known regional differences in incidences, since we can identifyenvironmental risk factors that may either initiate the autoimmune pro-cess or promote already ongoingb cell damage in different countries. Forthis, population-based epidemiological studies are necessary to identifyrisk determinants that may be useful for primary prevention strategies.

Address for correspondence: Yongsoo Park, M.D., Department of Internal Medicine, Hanyang Uni-versity Hospital, 249-1 Kyomun-dong, Kuri, Kyunggi-do, 471-020, Korea. Voice: 82-31-560-2239;fax: 82-31-553-7369.

e-mail: [email protected]

Ann. N.Y. Acad. Sci. 1079: 31–40 (2006). C© 2006 New York Academy of Sciences.doi: 10.1196/annals.1375.005

31

32 ANNALS NEW YORK ACADEMY OF SCIENCES

KEYWORDS: type 1 diabetes; incidence; genetic determinants; environ-mental etiologies

Is Type 1 Diabetes (T1D) Uncommon in Asia?

Immune-mediated T1D is an etiologic subtype of diabetes caused by autoim-mune destruction of the insulin-secreting b cells of the islets of Langerhans.Unlike T2D, a more common form of diabetes with onset mostly, but not ex-clusively in adults, T1D is among the most prevalent life-threatening disordersaffecting children around the world. The incidence of T1D in different ethnicgroups is extremely variable, suggesting the involvement of genetic as well asenvironmental elements. The incidence rate of T1D occurring below the ageof 15 years has been registered in many parts of the world, owing to the activi-ties of the WHO-DIAMOND collaborative studies.1 In Caucasian populations,T1D incidence rates are high with rates in excess of 20 cases/year/100,000 in-dividuals, while Asian countries have extremely low T1D incidence rates withthe rate less than 1 case/year/100,000 individuals. A mail survey to investigatethe T1D incidence has been conducted in all hospitals with more than 20 bedsin Seoul, Korea since the year 1986.2,3 All T1D patients had been registeredthereafter. The annual incidence rates did not show significant temporal vari-ation ranging from 0.6 to 2.2 cases per 100,000. Additional T1D incidencedata in Japan with a high degree of ascertainment have demonstrated that theincidence is similar with no temporal variation.4 From these, we can indicatethat T1D is much less frequent in Asia than in countries with a predominantlyCaucasian population.

The average ascertainment level of TID in Asian countries is low, and in-correct patient definition might lead to a false conclusion on the demographiccharacteristics of T1D in Asia. There is a difference in the relative proportion ofT1D by age of onset in Asia and Europe.5 The percentage of T2D in young dia-betic subjects in Asians is higher than in Caucasoid population. In accordancewith the view that development of T1D involves heterogeneous mechanisms,different clinical courses ofb cell destruction have been reported in Asian T1Dpatients. The most prevalent clinical manifestation is abrupt onset with severeclinical symptoms, very high blood glucose levels and ketonemia at diagnosis,which shares similar clinical features with that seen in Caucasian populations.Another subtype of T1D in the Asian population is a slow-onset form. Theeventual clinical features of these patients include ketosis proneness, unstableblood glucose levels, and extremely diminished 24-h urinary C-peptide ex-cretion rate. The majority of Asian patients with T2D have been reported tobe a non-obese type with low insulin secretion. It is also true that the clinicaldistinction between T1D and T2D has been difficult to determine in some ofthe Asian population. In our Seoul T1D genetic consortium, we would like tomonitor every single development of T1D prospectively.6 Having sera as wellas DNA, we could define the patient more correctly and we may be able to getan access to the etiology of T1D. We have the hypothesis that the high-risk hap-

PARK: TYPE 1 DIABETES INCIDENCE AND GENETIC SUSCEPTIBILITY FACTORS 33

lotypes/genotypes are independent determinants of diabetes in the first-degreerelatives, particularly in the presence of islet-specific antibodies. Concordantsiblings or T1D patients and multiple autoantibody (+) siblings appeared toshare the same 2 HLA DR-DQ haplotypes. We are following up every singlecase prospectively whether he or she has autoantibodies.

Is T1D in Asia also Caused by Autoimmunity?

T1D is now perceived as an autoimmune disease characterized by selectivedestruction of pancreatic b cells. In the Asian populations, it has been knownthat the clinical and immunologic characteristics of T1D are quite differentfrom those of Caucasians. Besides the low incidence rate of T1D in Asia, theprevalences of islet cell cytoplasmic antibody (ICA) in Asian T1D patients hadbeen reported to be extremely low compared with those of the Caucasians.7

However, not only is the detection of the islet-specific autoantibodies in aconventional way rather complicated and time-consuming, but incorrect patientdefinition might lead to false conclusion on the immunologic characteristics ofT1D in Asia.8 Previous studies have shown a low prevalence of ICA (14–16%)in recent-onset Japanese patients with T1D, suggesting that the autoimmunecomponent may not be so important in the pathogenesis of T1D in Japanese.7

However, a series of recent well-standardized study on the prevalence of ICAin Asians with T1D demonstrated that more than 80% of recent-onset patientshad ICA.8,9

During the past decade investigators have identified, cloned, and expresseda series of islet autoantigens, including GAD65 and IA-2 and applied as analternative to ICA. We also applied a new radioligand-binding assay in T1Dpatients, whose age at onset was less than 15 years and the prevalences ofIA-2 autoantibodies, anti-GAD65, and ICA in recent-onset patients (duration< 1 year) reached those of Caucasian levels.9 In contrast, in a subset of these pa-tients with long standing diabetes (duration > 3 years), the prevalence of IA-2autoantibodies, anti-GAD65, and ICA were decreased to 22%, 49%, and25%, respectively. The overall prevalence of islet-specific autoantibodies inthis childhood-onset T1D in Korea was comparable to Caucasians, espe-cially among those in recent-onset cases (duration less than 1 year). In ourKorean patients, IA-2 autoantibodies as well as anti-GAD65 were signifi-cantly associated with the presence of ICA, especially in those with onset age< 15 years. Although there should be some other T1D related with insulin de-ficiency, the major part of T1D with onset age less than 15 is mainly caused byautoimmunity.

Is T1D in Asia also Determined by the Genetic Susceptibility?

T1D is much less frequent in Asia than in countries with a predominantlyCaucasian population.3,4 To investigate whether the genetic determinants


influence the development of T1D in these low incidence countries, Ikegamiet al. studied siblings of T1D probands with age at onset under 20 years.10 Theratio of the risk for siblings of T1D patients and the population prevalence (ls),often used to assess the degree of familial clustering of a disease is more than200, a much higher value than that in Caucasian populations. Familial cluster-ing of a disease does not necessarily indicate a genetic component, since it maybe caused by sharing of the same environment among family members. If agenetic factor is responsible for the high ls value for T1D with low populationprevalence in Asians, then susceptibility alleles whose frequencies are verylow in the general population may be segregating in T1D families.

Is HLA also Important in Asia?

HLA Association in Asians

The HLA class II alleles on chromosome 6p21 are the most highly related toT1D susceptibility.11–13 Affected siblings-pairs (ASPs) with T1D share HLAalleles more often than expected. Multiple candidate genes in the HLA to-gether with strong linkage disequilibrium (LD) in the HLA make it difficultto pinpoint the locus responsible for T1D susceptibility. To overcome this, wehave used simultaneous analysis of multiple candidate gene polymorphisms inthe same patients. Even in Asians, both the DR and DQ alleles are the ones tohave the highest association with T1D.6,8,11 In Korean T1D cases, HLA DR3and DR9 were increased. DR4 as a group was not significantly increased in di-abetic patients compared to controls. Among the DR4 subtypes, DRB1∗0401and DRB1∗0405 had increased frequencies in patients. Two DR4 subtypes(0403 and 0406) had lower frequencies in patients. As expected, DR15 confersstrong protection. DR12 was also strongly protective. When the HLA DQB1alleles were identified in the T1D patients, the only DQB1∗0201 allele had sig-nificantly higher frequencies in patients, while three DQB1 alleles (0301, 0601,0602) had significantly lower frequencies in patients compared to controls. Fivehaplotypes (DRB1∗03-DQB1∗0201, DRB1∗0401-DQB1∗0302, DRB1∗0405-DQB1∗0302, DRB1∗0407-DQB1∗0302, and DRB1∗0901-DQB1∗0303) hadsignificantly increased frequencies in diabetic patients. Four other haplotypes(DRB1∗15-DQB1∗0601, DRB1∗15-DQB1∗0602, DRB1∗08-DQB1∗0601,and DRB1∗10-DQB1∗05) had significantly lower frequencies in patients.11

HLA Genotypic and Haplotypic Associations

It has been proposed that the contribution of the HLA DQ moleculeto overall disease susceptibility may be genotype dependent. We analyzedthe association of DQ8 (DQA1∗0301-DQB1∗0302) and DQ4 (DQA1∗0301-DQB1∗0401) haplotypes in Asian patients with T1D.14 Although the


prevalence of the DR4-DQ8 haplotype did not differ in patients versus controls,the HLA DR3/4-DQ8 genotype had an increased risk indicating a synergisticeffect. When DR3/4 diabetic patients are compared to all other DR4 posi-tive patients, a significant association of DQ8 with the DR3/4 genotype wasfound. In contrast, a significant association of DQ4 with DR4/X (X: otherthan 1, 3, 4) was found. High-risk DR4 subtypes (DRB1∗0401, ∗0402, ∗0405)were predominant in DR4/X, whereas protective DR4 subtypes (DRB1∗0403,∗0406) were observed mainly in the DR3/4 genotype. The association of DR4haplotypes with diabetes varies depending on the haplotype borne on the ho-mologous chromosome. This might contribute not only to the synergistic effectof DR3/4, but also to the susceptibility influence of DQ4 haplotypes confinedto DR4/X.

Common Transmission of HLA Haplotypes Across Ethnicity

It has been suggested that HLA alleles of Asian patients associated with T1Ddiffer from those of Caucasians. It is apparent that population frequencies ofDR and DQ alleles and haplotypes vary dramatically between ethnic groups.Because of this, some highly “diabetogenic” DR or DQ alleles, which arerelatively uncommon in a population may be mistakenly considered neutral inpopulation association studies. Only haplotype analyses can reveal the effectof DR and DQ alleles simultaneously. Although the nature of association for aparticular allele may be very different in various ethnic groups, the effect of agiven allele (or a haplotype) on T1D susceptibility is probably consistent in allpopulations. In our transracial study, for all parental haplotypes, which wereidentical at DRB1 and DQB1, the association with T1D and the transmissionof the haplotypes from nondiabetic parents to diabetic offspring was similar forKorean and Caucasian families.15 Thus, the influence of class II susceptibilityand resistance alleles appears to transcend ethnic and geographic diversity ofT1D incidence.

Other Intra-HLA Susceptibility Genes

It is known that more than one genetic locus even within the HLA is impor-tant for disease risk. The highest risk genotype for T1D consists of individualsheterozygous for DR3- and DR4-associated haplotypes. This HLA DR3/4genotype can only explain some 20% or 12% of the total genetic contribu-tion in Caucasians, assuming a 30% or 50% concordance rate in monozygotictwins, respectively.13,16 In Asians, the contribution of this genotype to overallsusceptibility of T1D appeared to be less than that in Caucasians.6,8 Moreover,the population frequency of this genotype in Caucasians is still 10–20 timeshigher than the prevalence of T1D associated with this genotype. DRB1 sub-typing might influence the risk of T1D in this high-risk DQ population, though


it is assumed to account for only 10% of additional familial aggregation. Thisimplies that additional protective risk genes (HLA or non-HLA) and/or envi-ronmental factors influence susceptibility to the disease, especially in Asians.We found independent susceptibility of D6S273 and D6S2223 microsatellitemarkers in DQ8/DQ2 heterozygous individuals from the Caucasian DiabetesAutoimmunity Study in the Young (DAISY) repository.17 In our Koreans,MIC-A and tumor necrosis factor-a (TNF-a) microsatellites were also asso-ciated with T1D independently.6,18

Are Susceptible Genes Similar Everywhere?

T1D is one of the first disorders with a complex genetic basis that researchershave begun to unravel. More than 20 years ago, the HLA region was found tocontain a major locus that influences predisposition to T1D, and a decade agoa locus with a smaller effect was identified in the insulin gene region. Withthe advent of numerous microsatellite markers suitable for genome screening,additional 20 loci that influence susceptibility to T1D have been reported.8,19

Some of the new loci appear to predispose people to T1D independently ofHLA and may be important factors in families with T1D who lack strongHLA susceptibility. Other loci may interact to cause susceptibility, and specificcombinations may be diabetogenic. Some highly “diabetogenic” genes, whichare relatively uncommon in a population may be mistakenly considered neutralin population association studies. Because of these interactions, the effectof some susceptibility genes, whose influence is evident in one population,becomes obscure in other population. Notwithstanding, although isolating theactual predisposing genes in T1D is more difficult than isolating those involvedin single-locus genetic disorders, the fact that the genes can be identified withthe use of a reasonable number of multiethnic families is very encouraging forfuture research on other genetically complex disorders.

One of the confirmed T1D susceptibility loci is IDDM12 located on chromo-some 2q33. However, there are still inconsistencies between different studies.Transmission disequilibrium test (TDT) revealed highly significant deviationof transmission for alleles at the (AT)n microsatellite marker and the A/G poly-morphism within the CTLA4 gene in the data sets with Mediterranean origins(Italian, Spanish, French, and Mexican Americans).20 In contrast to a negativeresult in Japanese and Chinese, a positive result was also observed in a smallKorean data set.21 Recent large-scale studies performed in Caucasians indi-cated CTLA4 gene to be responsible for T1D as well as autoimmune thyroiddisease (ATD) susceptibility. We studied the association of CTLA4 polymor-phism with T1D, ATD, and T1D patients with ATD in a large genetically distinctKorean cohort. The +6230 G>A polymorphism of CTLA4 was significantlyassociated with Graves’ disease patients and T1D patients with ATD, but notwith T1D without ATD (unpublished observation). Given the high prevalenceof ATD in patients with T1D, we may suggest that ATD may be etiologically


and genetically distinct from T1D and T1D patients should be separately clas-sified according to the ATD phenotype in genetic studies as well as in clinicaltrials.

We reported highly significant evidence for association between T1D andmultiple single nucleotide polymorphisms (SNPs) from 180 kb of genomicDNA within the IDDM5 interval in a large multiethnic family.22 We also per-formed fine mapping using a high density of SNPs in our Korean population.23

We found several SNPs associated with T1D. Importantly, M55V polymor-phism is one of the highest associations. Although there have been incon-sistent reports, mostly from the Caucasians, we confirmed the susceptibilityinfluence of M55V polymorphism in our Korean case–control studies, notonly in T1D patients23 but also in Graves’ disease patients (unpublished ob-servation). This susceptibility influence was also confirmed in the Japanesepopulation.24 A single amino acid substitution (M55V) of SUMO4 encodinga small ubiquitin-like modifier type 4 protein was found to be strongly as-sociated with T1D. SUMO4 was shown to be able to conjugate to IkBa andnegatively regulate NF-kB transcriptional activity. The M55V substitution re-sults in a considerable increase of NF-kB transcriptional activity and abouttwofold higher expression of NF-kB-dependent genes. These findings suggesta novel pathway implicated in T1D pathogenesis.

Concannon et al. reported the results of a genome screen for linkage withT1D and analyzed the data by multipoint linkage methods.19 An initial panelof 212 ASPs were genotyped for 438 markers spanning all autosomes, andadditional 467 ASPs were used for follow-up genotyping. Other than the well-established linkage with the HLA region at 6p21.3, they found only 1 region,located on 1q. Being ascribed to a weak effect of the disease genes, geneticheterogeniety or random variation, these sorts of differences will give us bigobstacle for the confirmation and fine mapping of susceptibility intervals andidentification of etiologic mutations. If oligogenicity rather than polygenicityapplies to human T1D, studies of large multiplex families from geneticallyand culturally homogeneous populations are likely to improve the prospectsof identifying susceptibility genes by genetic linkage studies followed by po-sitional cloning.

From these, it may be evident that susceptible genes are working everywhere,although their contribution might vary across geographical regions or ethnicity.However, different genetic make-up, LD pattern, or varying interaction withother genes may influence differentially which diseases are popular in onepopulation and which genes are important in susceptibility. Comorbidities arealso important in determining susceptibility.

Why Is T1D so Uncommon in Asia?

The frequencies of T1D-associated HLA antigens are correlated with theworldwide diabetes incidences, which differ by ethnicity as well as geographic


regions. Dorman et al. plotted the association of the frequency of DQB1non-ASP with T1D incidences.25 At that time Japan and Korea were the ex-ceptions. Because they ignored the contribution of DRB1 alleles, that plotcounts as one of oversimplified view. The Sardinians and the Scandinavianpopulation have the highest T1D incidence in the world (17.6–28.6/100,000per year). The high frequencies of DRB1∗0301, DRB1∗0401, the 0301/0302,and 0301/0201 DQ dimers correlate with the high incidence of T1D in Scan-dinavians, while the high frequency of DRB1∗0301 is consistent with the highdisease incidence in Sardinians. In contrast, the Asian populations, such asJapanese and Koreans, have a very low incidence. The absence of the DR3haplotype also correlates with the low incidence of T1D in Japanese. The con-tribution of DR3/4 to T1D susceptibility appears less important in Korea thanin Caucasoid population. Instead, other rare genotypes are important in accor-dance with the prevalent haplotypes in the general population. Among the DRgenotypes, DR3/9, DR9/9, and DR3/X (X: other than 3, 4) were susceptiblegenotypes. In Asians, more moderate risk haplotypes rather than highly sus-ceptible DR3 and DR4 haplotypes are prevalent and there are less chances ofthe DQ0201/0302 dimers to be formed. Therefore, it becomes evident that T1Dis uncommon in Asia considering the frequency of DRB1-DQB1 genotypes inthe general population. Moreover, LD between highly susceptible DRB1 alle-les and protective DQB1 alleles, and vice versa, is also the major contributingfactor to the low incidence of T1D in Japanese, Korean, and Chinese. Weanalyzed the association of the different DR4 subtypes with susceptible DQ8(DQA1∗0301-DQB1∗0302) and neutral DQ4 (DQA1∗0301-DQB1∗0401) hap-lotypes in Asian general population. In Asians, DQ8 is usually combined withprotective DR4 subtypes (DRB1∗0403, ∗0406), while DQ4 is in LD with sus-ceptible DRB1∗0405. High-risk DR4 subtypes (DRB1∗0401, ∗0402, ∗0405)were predominant in DR4/X, whereas protective DR4 subtypes (DRB1∗0403,∗0406) were observed mainly in the DR3/4 genotype. The low incidence ratein the Asian population may be explained by the counterbalancing effect ofthe DRB1 and DQB1 alleles in the general population.

However, the genetic contribution does not explain the whole portion of T1Dincidence variability across population. In genetically rather similar popula-tions, such as in Estonia and Finland, up to sixfold variation in incidence rateshave been found. In addition, in the genetically very homogenous Swedishpopulation, significant and consistent incidence variability is present pointingto the importance of nongenetic risk factors. This suggestion is also supportedby the finding that the incidence is increasing with time in many countries thathave population-based registries. Korea has the lowest documented incidenceof T1D in the world, and prevalences of islet-specific autoantibodies are re-ported to be low compared with Caucasians. A different genetic susceptibilityincluding a low frequency of high-risk HLA alleles could explain the lowerprevalence of islet-specific autoantibodies to GAD and islet cells cytoplasm(ICA) and the low incidence of T1D, or different genetic and environmental


interactions that might be involved in the etiology of T1D. Although Asians maybe genetically protected against T1D compared with the Caucasians, it is un-known to what degree genetic determinants influence the well-known regionaldifferences in incidence. We suggested that different genetic–environmentalinteractions do operate in the etiology of T1D between Caucasians and Asians.

ACKNOWLEDGMENTS

This work was supported by the grants from Korea Science and Engineer-ing Foundation (R01-2001-00177 and R01-2005-10075) and a grant from theKorea Health 21 R and D Project, Ministry of Health and Welfare, Republicof Korea (A05-0463-B50704-05N1-00030B).

REFERENCES

1. WHO DIAMOND PROJECT GROUP. 1990. WHO multinational project for childhooddiabetes. Diabetes Care 13: 1062–1068.

2. KO, K. et al. 1994. The incidence of IDDM in Seoul from 1985 to 1988. DiabetesCare 17: 1473–1475.

3. PARK, Y. et al. 1996. Molecular epidemiology of insulin-dependent diabetes mel-litus (IDDM) in Korea. Diabetes Res. Clin. Pract. 34: S55–S59.

4. TAJIMA, N. et al. 1985. A comparison of the epidemiology of youth-onset insulin-dependent diabetes mellitus between Japan and the United States. Diabetes Care11: 17–23.

5. KUZUYA T. et al. 2002. Report of the committee on the classification and diagnosticcriteria of diabetes mellitus. Diabetes Res. Clin. Pract. 55: 65–85.

6. PARK, Y. 2004. Prediction of the risk of type 1 diabetes from polymorphisms incandidate genes. Diabetes Res. Clin. Pract. 66(Suppl):S19–S25.

7. NOTSU, K. et al. 1983. A population survey of pancreatic islet cell antibodies andantithyroid antibodies. Tohoku J. Exp. Med. 141: 261–263.

8. PARK, Y. & G.S. EISENBARTH. 2001. Genetic susceptibility factors of type 1 diabetesin Asians and their functional evaluation. Diabetes Metab. Res. Rev. 17: 2–11.

9. PARK, Y. et al. Evaluation of the efficacy of multiple autoantibody screening inKorean patients with IDDM. Acta Diabetologica 37: 213–217.

10. IKEGAMI, H. & T. OGIHARA. 1996. Genetics of insulin-dependent diabetes mellitus.Endocr. J. 43: 605–613.

11. PARK, Y. et al. 1998. Combination of HLA DR and DQ molecules determine thesusceptibility to insulin-dependent diabetes mellitus in Koreans. Hum. Immunol.59: 794–801.

12. TODD, J. et al. 1987. HLA-DQB gene contributes to susceptibility and resistanceto insulin-dependent diabetes mellitus. Nature 329: 599–604.

13. RISCH, N. 1987. Assessing the role of HLA-liked and unlinked determinants ofdisease. Am. J. Hum. Genet. 40: 1–14.

14. PARK, Y. et al. 2001. Transracial evidence for the influence of the homologousHLA DR-DQ haplotype on transmission of HLA DR4 haplotypes to diabeticchildren. Tissue Antigens 57: 185–191.


15. PARK, Y. et al. 2000. Common susceptibility and transmission pattern of HLADRB1-DQB1 haplotypes to Korean and Caucasian patients with type 1 diabetes.J. Clin. Endocrinol. Metab. 85: 4538–4542.

16. NOBLE, J. et al. 1996. The role of HLA class II genes in insulin-dependent diabetesmellitus: molecular analysis of 180 Caucasians, multiplex families. Am. J. Hum.Genet. 59: 1134–1148.

17. PARK, Y. et al. 1999. D6S2223 alleles associated with risk of IDDM in DR3-DQ2/DR4-DQ8 heterozygotes. Diabetes 48: A183.

18. PARK, Y. et al. 2001. MICA polymorphism is associated with type 1 diabetes inKorean Population. Diabetes Care 24: 33–38.

19. CONCANNON, P. et al. 1998. A second-generation screen of the human genome forsusceptibility to insulin-dependent diabetes mellitus. Nat. Genet. 19: 292–296.

20. MARRON, M. et al. 1997. Insulin-dependent diabetes mellitus (IDDM) is associatedwith CTLA4 polymorphisms in multiple ethnic groups. Hum. Mol. Genet. 8:1275–1282.

21. MARRON, M. et al. 2000. Genetic and physical fine-mapping of a type 1 diabetessusceptibility gene (IDDM12) to a 100-kb phagemid artificial chromosome clonecontaining D2S72-CTLA4-D2S105 on chromosome 2q33. Diabetes 49: 492–499.

22. GUO, D. et al. 2004. A new IkBa modifier, SUMO-L, is associated with type 1diabetes. Nat. Genet. 36: 837–841.

23. PARK, Y. et al. 2005. Assessing the validity of the association between the SUMO4M55V variant and risk of type 1 diabetes. Nat. Genet. 37: 112.

24. NOSO, S. et al. 2005. Genetic heterogeneity in association of the SUMO4 M55Vvariant with susceptibility to type 1 diabetes. Diabetes 54: 3582–3586.

25. DORMAN, J. et al. 1990. Worldwide differences in the incidence of type I diabetesare associated with amino acid variation at position 57 of the HLA-DQ betachain. Proc. Natl. Acad. Sci. USA 87: 7370–7374.

why is type 1 diabetes uncommon in asia?

Documents