Diabetes Mellitus, Insulin-Dependent

Description

The type of diabetes mellitus called IDDM is a disorder of glucose homeostasis that is characterized by susceptibility to ketoacidosis in the absence of insulin therapy. It is a genetically heterogeneous autoimmune disease affecting about 0.3% of Caucasian populations (Todd, 1990). Genetic studies of IDDM have focused on the identification of loci associated with increased susceptibility to this multifactorial phenotype.

The classical phenotype of diabetes mellitus is polydipsia, polyphagia, and polyuria which result from hyperglycemia-induced osmotic diuresis and secondary thirst. These derangements result in long-term complications that affect the eyes, kidneys, nerves, and blood vessels.

The term diabetes mellitus is not precisely defined and the lack of a consensus on diagnostic criteria has made its genetic analysis difficult. Diabetes mellitus is classified clinically into 2 major forms of the primary illness, insulin-dependent diabetes mellitus (IDDM) and noninsulin-dependent diabetes mellitus (NIDDM; 125853), and secondary forms related to gestation or medical disorders.

Appearance of the IDDM phenotype is thought to require a predisposing genetic background and interaction with other environmental factors. Rotter and Rimoin (1978) hypothesized that there are at least 2 forms of IDDM: a B8 (DR3)-associated form characterized by pancreatic autoimmunity, and a B15-associated form characterized by antibody response to exogenous insulin. Interestingly, the DR3 and DR4 alleles seem to have a synergistic effect on the predisposition to IDDM based on the greatly increased risk observed in persons having both the B8 and B15 antigens (Svejgaard and Ryder, 1977). Rotter and Rimoin (1979) hypothesized a combined form. Tolins and Raij (1988) cited clinical and experimental evidence to support the idea that those IDDM patients in whom diabetic nephropathy (see 603933) eventually develops may have a genetic predisposition to essential hypertension.

Gambelunghe et al. (2001) noted heterogeneity of the clinical and immunologic features of IDDM in relation to age at clinical onset. Childhood IDDM 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. 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: 138275) autoantibodies and/or islet cell antibodies.

Nepom et al. (1987) studied the mechanism of the exaggerated susceptibility to IDDM in DR3/DR4 heterozygotes, and concluded that its basis is the formation of hybrid molecules of the closely linked DQ-alpha (HLA-DQA1; 146880) and -beta (HLA-DQB1; 604305) chains. The DR-alpha molecules are not polymorphic, and mixed DR alpha-beta dimers would not result in novel HLA molecules. On the other hand, both the alpha and beta chains of DQ are polymorphic, and a DQ alpha-beta dimer composed of transcomplementing chains would be unique to a heterozygous individual and not expressed in either parent. In the mouse, such transcomplementation has been demonstrated structurally, and epitopes newly formed in the resulting hybrid molecules allow for an altered functional immune response different from that of either parent.

The human MHC class II molecule encoded by DQA1*0102/DQB1*0602 (termed DQ0602) confers strong susceptibility to narcolepsy (161400) but dominant protection against type I diabetes. To elucidate the molecular features underlying these contrasting genetic properties, Siebold et al. (2004) determined the crystal structure of the DQ0602 molecule at 1.8-angstrom resolution. Structural comparisons to homologous DQ molecules with differential disease associations highlighted a previously unrecognized interplay between the volume of the P6 pocket and the specificity of the P9 pocket, which implies that presentation of the expanded peptide repertoire is critical for dominant protection against type I diabetes. In narcolepsy, the volume of the P4 pocket appears central to the susceptibility, suggesting that the presentation of a specific peptide population plays a major role.

Hyperglycemia, the basic metabolic abnormality in IDDM, is caused by abnormally increased gluconeogenesis and insufficient glucose disposal. Ketosis results from the accumulation of free fatty acids and their oxidation.

McCorry et al. (2006) found an association between IDDM and idiopathic generalized epilepsy (EIG; 600669) in a population-based survey in the U.K. Among 518 EIG patients aged 15 to 30 years, 7 also had IDDM. In contrast, there were 465 IDDM patients among an age-matched cohort of 150,000 individuals. The findings suggested that the prevalence of IDDM is increased in patients with EIG (odds ratio of 4.4).

Type 1 diabetic patients have diminished responses following T-cell activation. By immunoblot analysis, Nervi et al. (2000) found reduced levels of phosphorylated CD3Z (186780) in IDDM1 patients after T-cell stimulation. Immunoblot, immunoprecipitation, and densitometric analyses revealed significantly reduced LCK expression in unstimulated peripheral blood cells of IDDM1 patients compared to controls. The reduced LCK expression correlated with a lower proliferative response. Very low LCK expression may also correlate with the HLA-DQB1*0201/0302 (see 604305) genotype. Confocal microscopy demonstrated normal plasma membrane expression of LCK in patients and controls. Downstream signal transducing molecules were not affected in these patients.

Kent et al. (2005) examined T cells from pancreatic draining lymph nodes, the site of islet cell-specific self-antigen presentation. They cloned single T cells in a nonbiased manner from pancreatic draining lymph nodes of patients with type I diabetes and from nondiabetic controls. A high degree of T-cell clonal expansion was observed in pancreatic lymph nodes from long-term diabetic patients but not from controls. The oligoclonally expanded T cells from diabetic patients with DR4, a susceptibility allele for type I diabetes, recognized the insulin A 1-15 epitope restricted by DR4. Kent et al. (2005) concluded that their results identified insulin-reactive, clonally expanded T cells from the site of autoinflammatory drainage in long-term type I diabetics, indicating that insulin may indeed be the target antigen causing autoimmune diabetes.

Porter and Barrett (2005) reviewed monogenic syndromes of abnormal glucose homeostasis, focusing on 3 mechanisms: insulin resistance, insulin secretion defects, and beta-cell apoptosis.

Stechova et al. (2012) reported a family with naturally conceived monozygotic female quadruplets, in which type 1 diabetes was diagnosed in 2 of the quadruplets simultaneously and a third quadruplet was diagnosed as pre-diabetic. All 4 quadruplets were positive for anti-islet cell autoantibodies to GAD65 (138275) and to IA-2 (601773), indicating an ongoing anti-islet autoimmunity in the nondiabetic quadruplets. Serologic examination confirmed that all the quadruplets and their father had recently undergone an enteroviral infection of the EV68-81 serotype. Immunocompetent cells from all family members were characterized by gene expression arrays, immune-cell enumerations, and cytokine-production assays. The microarray data provided evidence that the viral infection and IL27 (608273) and IL9 (146931) cytokine signaling contributed to the onset of T1D in 2 of the quadruplets. Stechova et al. (2012) stated that the propensity of stimulated immunocompetent cells from nondiabetic members of the family to secrete high levels of IFN-alpha (IFNA1; 147660) further corroborated their conclusion. They observed that the number of T-regulatory cells as well as plasmacytoid and/or myeloid dendritic cells was diminished in all family members. Stechova et al. (2012) concluded that this family supported the so-called 'fertile-field' hypothesis proposing that genetic predisposition to anti-islet autoimmunity, if 'fertilized' and precipitated by a viral infection, results in full-blown type 1 diabetes.

IDDM exhibits 30 to 50% concordance in monozygotic twins, suggesting that the disorder is dependent on environmental factors as well as genes. The average risk to sibs is 6% (Todd, 1990). Recessive, dominant, and multifactorial hypotheses have been advanced, as well as 'susceptibility' hypotheses (Rotter, 1981). Genetic and environmental influences in IDDM were reviewed by Craighead (1978). Usually in genetic disease the most severe form of a disorder shows the clearest genetic basis. It is therefore surprising to find that the genetics of IDDM is less clear than that of NIDDM. Concordance in NIDDM was 100% for identical twins in which the index case had onset of diabetes after age 45 years, and nearly half had a diabetic parent, while discordance was found in half the pairs with earlier onset, few of whom had a family history of diabetes (Tattersall and Pyke, 1972).

Nilsson (1964) commented on the difficulties of distinguishing dominant and recessive inheritance when gene frequency is high. He considered autosomal recessive inheritance to be most likely, with a gene frequency of about 0.30 and a lifetime penetrance of about 70% for males and 90% for females. A gene frequency of about 0.05 and a penetrance of 25 to 30% would be required to account for the findings on a dominant hypothesis. Hodge et al. (1980) proposed a 3-allele model based on a susceptibility locus (S) tightly linked to the HLA complex. Thomson (1980) espoused a 2-locus model. See 125850 for a clear example of an autosomal dominant type of diabetes mellitus: maturity-onset diabetes of the young (MODY).

Cudworth and Woodrow (1975) found that the relative risk of IDDM was 2.12 for HLA-A 8 and 2.60 for W15. Rubinstein et al. (1977) found that diabetic sibs shared their HLA genes with a significantly increased frequency, leading them to postulate a recessive gene linked to HLA (and specifically to HLA-D as indicated by 3 informative cases with recombination within the HLA). They estimated the penetrance at 50% because half the HLA-identical sibs of index cases were diabetic. This conclusion fits with published observations of 6-10% risk to sibs of patients when both parents are normal. As an appendix to their paper, they presented a table of risk to relatives on the basis of the above hypotheses. Barbosa et al. (1978) also concluded that IDDM is a recessive with 50% penetrance and with linkage to HLA (theta = 0.13, lod = 3.98) on the basis of the study of 21 families with 2 or more affected sibs and normal parents.

Vadheim et al. (1986) pointed out that several studies suggested a higher incidence of IDDM among the offspring of affected males than among those of affected females. To test the hypothesis that differential transmission by the father of genes predisposed to diabetes may explain this phenomenon, Vadheim et al. (1986) examined parent-to-offspring transmission of HLA haplotypes and DR alleles in 107 nuclear families in which a child had IDDM. They found that fathers with a DR4 allele were significantly more likely to transmit this allele to their diabetic or nondiabetic children than were mothers with a DR4 allele. No difference between parents was observed for HLA-DR3; however, DR3 was transmitted significantly more than 50% of the time from either parent. Field et al. (1986) reconfirmed the fact that sharing of 2 HLA haplotypes by sibs with diabetes mellitus was increased in comparison to mendelian expectations. Whereas sharing of GM-region genes was not different from mendelian expectations in the total sampled, affected pairs who shared 2 HLA haplotypes did show significantly increased sharing of GM-region genes.

MacDonald et al. (1986) studied families with IDDM in parent and child. The proportion of diabetic parents who transmitted DR4 to diabetic offspring (78%) was significantly higher (P less than 0.001) than the gene frequency of DR4 in the overall diabetic population (43%). The proportion of nondiabetic parents who transmitted DR4 to diabetic offspring (22%) was not significantly different from the gene frequency in the nondiabetic population but significantly lower (P less than 0.05) than the gene frequency in the overall IDDM population. This was taken to indicate a strong dominant effect of DR4. The proportion of nondiabetic parents who transmitted DR3 was similar to the gene frequency of DR3 in the overall diabetic population, but it was significantly higher than the gene frequency of DR 3 in the nondiabetic population (15%; P less than 0.005). The percentage of diabetic offspring who were DR3/DR4 (35%) was identical to that in the overall IDDM population (35%). MacDonald et al. (1986) interpreted this to mean that DR3 plays an enhancing role, with DR4 playing the main role.

Thomson et al. (1988) analyzed the results from 11 studies involving 1,792 Caucasian probands with IDDM. Antigen genotype frequencies in patients, transmission from affected parents to affected children, and the relative frequencies of HLA-DR3 and -DR4 homozygous patients all indicated that DR3 predisposes in a 'recessive'-like and DR4 in a 'dominant'-like or 'intermediate' fashion, after allowing for the synergistic effect of the 2 HLA types. DR2 showed a protective effect, DR1 and DRw8 showed predisposing effects, and DR5 showed a slight protective effect. They found evidence that only subsets of DR3 and DR4 are predisposing. The presence or absence of asp at position 57 of the DQ-beta gene was shown to be insufficient of itself in explaining the inheritance of IDDM. They suggested that the distinguishing features of the DR3-associated and DR4-associated predisposition remain to be identified at the molecular level.

Using an overall sib risk of 6%, Thomson et al. (1988) estimated that the risks for those sharing 2, 1, or 0 haplotypes are 12.9%, 4.5%, and 1.8%, respectively. The highest sib risk was 19.2% for sibs sharing 2 haplotypes with a DR3/DR4 proband. Field (1988) put this study in perspective with a discussion of other factors, including nongenetic factors. Sheehy et al. (1989) likewise concluded that susceptibility to diabetes is best defined by a combination of HLA-DR and HLA-DQ alleles.

In a study of 266 unrelated white patients with IDDM, Baisch et al. (1990) extended the assessment of the role of HLA-DQ alleles in susceptibility to the disease. They used allele-specific oligonucleotide probes and PCR to study HLA-DQ beta-chain alleles. Two major findings emerged. First, HLA-DQw1.2 was protective; it was found in only 2.3% of IDDM patients and in 36.4% of controls. This was 'dominant protection,' i.e., it did not matter what other allele was present. Second, HLA-DQw8 increased the risk of IDDM and the effect was one of 'dominant susceptibility' except that persons who were HLA-DQw1.2/DQw8 had a relative risk of 0.37, demonstrating that the protective effect of HLA-DQw1.2 predominated over the effect of HLA-DQw8. Segall and Bach (1990) reviewed the significance of these findings. See also review by Todd (1990).

The Eurodiab Ace Study Group and the Eurodiab Ace Substudy 2 Study Group (1998) studied the characteristics of familial type I diabetes mellitus, i.e., cases in which more than one affected first-degree relative was diagnosed before the age of 15 years. They used data from an international network of population-based registries and from a case-control study conducted in 8 of the network's centers. They found a positive association between the population incidence rate of type I diabetes and the prevalence of type I diabetes in fathers of affected children. A similar association was observed with the prevalence in sibs, but the association with prevalence in mothers was weaker and not significant. Pooling results from all centers showed that a greater proportion of fathers (3.4%) of affected children had type I diabetes than mothers (1.8%) giving a risk ratio of 1.8. Affected girls were more likely to have a father with type I diabetes than affected boys, but there was no evidence of a similar finding for mothers or sibs. Familial type I diabetes patients had a younger age at onset than nonfamilial patients.

Krischer et al. (2003) determined the extent to which different screening strategies could identify a population of nondiabetic relatives of a proband with type 1 diabetes who had 2 or more immunologic markers from the group consisting of islet cell antibodies (ICA), microinsulin autoantibodies (MIAA), GAD65 (138275) autoantibodies (GAA), and ICA512 (601773) autoantibodies (ICA512AA). Screening for any 3 antibodies guaranteed that all multiple antibody-positive subjects were detected. Screening for 2 antibodies at once and testing for the remaining antibodies among those who were positive for 1 resulted in a sensitivity of 99% for GAA and ICA, 97% for GAA and MIAA or GAA and ICA512AA, 93% for ICA512AA and ICA, 92% for MIAA and ICA, and 73% for ICA512AA and MIAA. From a laboratory perspective, screenings for GAA, ICA512AA, and MIAA are semiautomated tests with high throughput that, if used as initial screen, would identify at first testing 67% of the 2.3% of multiple antibody-positive relatives (100% if antibody-positive subjects are subsequently tested for ICA) as well as 4.7% of relatives with a single biochemical autoantibody, some of whom may convert to multiple autoantibody positivity on follow-up. Testing for ICA among relatives with 1 biochemical antibody would identify the remaining 33% of multiple antibody-positive relatives. They concluded that further follow-up and analysis of actual progression to diabetes will be essential to define actual diabetes risk in this large cohort.

General

Clerget-Darpoux et al. (1981) concluded that the data in 30 multiplex families best fitted a model with a susceptibility gene which was not linked to but interacted with the HLA system. Under 3 different genetic models for IDDM, Hodge et al. (1981) found evidence for linkage with 2 different sets of marker loci: HLA, properdin factor B, and glyoxalase-1 on chromosome 6, and Kidd blood group (then thought to be on chromosome 2, but later shown to be on chromosome 18). Thus, 2 distinct disease-susceptibility loci may be involved in IDDM, a situation also postulated for Graves disease (275000).

Bell et al. (1984) described an association between IDDM and a polymorphic region in the 5-prime flanking region of the insulin gene (INS; 176730). This polymorphism (Bell et al., 1981) arises from a variable number of tandemly repeated (VNTR) 14-bp oligonucleotides. When divided into 3 size classes, a significant association was seen between the short-length (class I) alleles and IDDM. Several studies were unable to demonstrate linkage of these VNTR alleles to IDDM in families, but this may in part be attributable to the fact that the disease-associated allele is present at high frequency in the general population. Several disease-associated polymorphisms were identified and the boundaries of association were mapped to a region of 19 kb on 11p15.5. Ferns et al. (1986) studied 14 families in which 13 had 2 cases of IDDM and found no linkage to polymorphic loci 5-prime to the insulin gene or to those 3-prime to the HRAS gene. Association with HLA was again found; persons who were HLA identical to the diabetic proband were more likely to be diabetic than those who were nonidentical. From studies of allele sharing in affected sib pairs, Cox et al. (1988) found evidence of HLA-linked susceptibility to IDDM but no evidence of a contribution of similar magnitude by the insulin-gene region. This failure of family studies to demonstrate linkage is difficult to reconcile with the association demonstrated between alleles at the VNTR locus in the 5-prime region of the insulin gene on 11p (Bell et al., 1984; Bell et al., 1985). Donald et al. (1989) used DR and DQ RFLPs for linkage analysis and demonstrated very close linkage of an IDDM-susceptibility locus. No evidence was found of any effect of the insulin gene.

Raum et al. (1979) found a rare genetic type of properdin factor B (F1) in 22.6% of patients with IDDM but in only 1.9% of the general population. If, as the authors suggested, this is an indication of linkage disequilibrium, not association, some populations should not show the relationship.

Based on a study in mice (Prochazka et al., 1987) it may be that corresponding recessive genes are located on chromosomes 6 and 11 in man; the THY1 (188230) and the APOA1 (107680) genes are on human 11q. By use of an affected sib pair method, Hyer et al. (1991) excluded the possibility of an IDDM susceptibility gene on 11q.

Lucassen et al. (1993) presented a detailed sequence comparison of the predominant haplotypes found in the region of 19 kb on 11p15.5 in a population of French-Canadian IDDM patients and controls. Identification of polymorphisms, both associated and unassociated with IDDM, permitted a further definition of the region of association to 4.1 kb. Ten polymorphisms within this region were found to be in strong linkage disequilibrium with each other and extended across the insulin gene locus and the VNTR situated immediately 5-prime to the insulin gene. These represent a set of candidate disease polymorphisms, one or more of which may account for the susceptibility to IDDM.

Using 96 affected sib pairs and a fluorescence-based linkage map of 290 marker loci (average spacing 11 cM), Davies et al. (1994) searched the human genome for genes that predispose to type I (insulin-dependent) diabetes mellitus. A total of 18 different chromosomal regions showed some positive evidence of linkage to the disease, strongly suggesting that IDDM is inherited in a polygenic fashion. Although the authors determined that no genes are likely to have as large effects as IDDM1 (in the major histocompatibility complex on 6p21), significant linkage was confirmed in the insulin gene region on 11p15 (IDDM2; 125852) and established to 11q (IDDM4; 600319), 6q (600320), and possibly to chromosome 18. Possible candidate genes within regions of linkage include GAD1 (605363) and GAD2 (138275), which encode the enzyme glutamic acid decarboxylase; SOD2 (147460), which encodes superoxide dismutase; and the Kidd blood group locus. Linkage of IDDM susceptibility to the region of the FGF gene on chromosome 11q13 was also reported by Hashimoto et al. (1994).

Genetic analysis of a mouse model of major histocompatibility complex-associated autoimmune type I (insulin-dependent) diabetes mellitus showed that the disease is caused by a combination of a major effect at the MHC and at least 10 other susceptibility loci elsewhere in the genome (Risch et al., 1993).

In a genomewide scan of 93 affected sib pair families from the UK, Davies et al. (1994) found a similar genetic basis for human type I diabetes, with a major component at the MHC locus (IDDM1) explaining 34% of the familial clustering of the disease. Mein et al. (1998) analyzed a further 263 multiplex families from the same population to provide a total UK dataset of 356 affected sib pair families. Only 4 regions of the genome outside IDDM1/MHC, which was still the only major locus detected, were not excluded, and 2 of these showed evidence of linkage: 10p13-p11 (maximum lod score = 4.7) and 16q22-q24 (maximum lod score = 3.4). They stated that these and other novel regions, including 14q12-q21 and 19p13-q13, could potentially harbor disease loci.

Concannon et al. (1998) reported the results of a genome screen for linkage with IDDM and analyzed the data by multipoint linkage methods. An initial panel of 212 affected sib pairs were genotyped for 438 markers spanning all autosomes, and an additional 467 affected sib pairs 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 and not previously reported, where the lod score exceeded 3.0. Lods between 1.0 and 1.8 were found in 6 other regions, 3 of which had been reported in other studies.

Cox et al. (2001) reported a genome scan using a new collection of 225 multiplex families with type I diabetes and combining the data with those from previous genome scans (Davies et al., 1994; Concannon et al., 1998; Mein et al., 1998). The combined sample of 831 affected sib pairs, all with both parents genotyped, provided 90% power to detect linkage. Three chromosome regions were identified that showed significant evidence of linkage with lod scores greater than 4: 6p21 (IDDM1); 11p15 (IDDM2); and 16q22-q24; 4 other regions showed suggestive evidence of linkage with lod scores of 2.2 or greater: 10p11 (IDDM10, 601942); 2q31 (IDDM7, 600321; IDDM12, 601388; IDDM13, 601318); 6q21 (IDDM15, 601666); and 1q42. Exploratory analyses, taking into account the presence of specific high-risk HLA genotypes or affected sibs' ages at disease onset, provided evidence of linkage at several additional sites, including the putative IDDM8 (600883) locus on 6q27. The results indicated that much of the difficulty in mapping type I diabetes susceptibility genes results from inadequate sample sizes, and pointed to the value of international collaborations to assemble and analyze much larger datasets for linkage in complex diseases.

Paterson and Petronis (2000) used data from a genomewide linkage study of 356 affected sib pairs with type I diabetes to perform linkage analyses using parental origin of shared alleles in subgroups based on sex of affected sibs and age of diagnosis. They found that evidence for linkage to IDDM4 occurred predominantly from opposite sex sib pairs and that for linkage to a locus on chromosome 4q occurred in sibs where one was diagnosed before age 10 years and one after age 10. Paterson and Petronis (2000) concluded that these methods might help reduce locus heterogeneity in type I diabetes.

Using DNA from 253 Danish IDDM families, Bergholdt et al. (2005) analyzed the chromosomal region 21q21.3-qter, which had been previously linked to IDDM by the European Consortium for IDDM Genome Studies (2001). Multipoint nonparametric linkage analysis showed a peak score of 3.61 at marker D21S1920 (p = 0.0002), and a '1-lod drop' interval of 6.3 Mb was identified between markers D21S261 and D21S270. No association was found with 74 coding SNPs from 32 candidate genes within the '1-lod drop' interval.

Using 2,360 SNP markers in the 4.4-Mb human major histocompatibility complex (MHC) locus and the adjacent 493 kb centromeric to the MHC, Roach et al. (2006) mapped the genetic influences for type 1 diabetes in 2 Swedish samples. They confirmed previous studies showing association with T1D in the MHC, most significantly near HLA-DR/DQ. In the region centromeric to the MHC, they identified a peak of association within the inositol 1,4,5-triphosphate receptor 3 gene (ITPR3; 147267). The most significant single SNP in this region was at the center of the ITPR3 peak of association. The estimated population-attributable risk of 21.6% suggested that variation within ITPR3 reflects an important contribution to T1D in Sweden. Two-locus regression analysis supported an influence of ITPR3 variation on T1D that is distinct from that of any MHC class II gene.

The Wellcome Trust Case Control Consortium (2007) described a joint genomewide association study using the Affymetrix GeneChip 500K Mapping Array Set, undertaken in the British population, which examined approximately 2,000 individuals and a shared set of approximately 3,000 controls for each of 7 major diseases. Case-control comparisons identified 7 independent association signals in type 1 diabetes at p values of less than 5.0 x 10(-7).

In a study of 4,000 individuals with type 1 diabetes, 5,000 controls, and 2,997 family trios independent of the Wellcome Trust Case Control Consortium (2007) study, Todd et al. (2007) confirmed the previously reported associations of rs2542151 in the PTPN2 gene (176887) on chromosome 18p11, rs17696736 in the C12ORF30 gene on chromosome 12q24, rs2292239 in the ERBB3 gene (190151) on chromosome 12q13, and rs12708716 in the KIAA0350 gene (CLEC16A; 611303) on chromosome 16p13 (p less than or equal to 10(-9); combined with WTCCC p less than or equal to 1.15 x 10(-14)), leaving 8 regions with small effects or false-positive associations. The association with rs17696736 led to the identification of a nonsynonymous SNP (rs3184504) in the SH2B3 gene (605093) that was sufficient to model the association of the entire region (p = 1.73 x 10(-21); see IDDM20, 612520).

To identify genetic factors that increase the risk of type 1 diabetes, Hakonarson et al. (2007) performed a genomewide association study in a large pediatric cohort of European descent. In addition to confirming previously identified loci, they found that type 1 diabetes was significantly associated with variation within a 233-kb linkage disequilibrium block on chromosome 16p13 that contains the KIAA0350 gene, which is predicted to encode a sugar-binding, C-type lectin. Three common noncoding variants of this gene (rs2903692, rs725613, and rs17673553) in strong linkage disequilibrium reached genomewide significance for association with type 1 diabetes. A subsequent transmission disequilibrium test replication study in an independent cohort confirmed the association. The combined P values for these SNPs ranged from 2.74 x 10(-5) to 6.7 x 10(-7). Hakonarson et al. (2007) noted that the Wellcome Trust Case Control Consortium (2007) had identified the KIAA0350 gene as a type 1 diabetes locus in a genomewide association study.

Smyth et al. (2008) evaluated the association between type 1 diabetes and 8 loci related to the risk of celiac disease in 8,064 patients with type 1 diabetes, 2,828 families providing 3,064 parent-child trios, and 9,339 controls. The authors found significant association between type 1 diabetes and rs1738074 in the TAGAP gene on chromosome 6q25 (see IDDM21, 612521) and confirmed association with rs3184504 in the SH2B3 gene (605093) on chromosome 12q24 (see IDDM20, 612520).

Cooper et al. (2008) performed a metaanalysis of 3 genomewide association studies, combining British type 1 diabetes (T1D) case-control data (Wellcome Trust Case Control Consortium, 2007) with T1D cases from the Genetics of Kidneys in Diabetes study (Mueller et al., 2006) for a total of 3,561 cases and 4,646 controls. Cooper et al. (2008) found support for a previously detected locus on chromosome 4q27 at rs17388568 (p = 1.87 x 10(-8); see IDDM23, 612622). After genotyping an additional 6,225 cases, 6,946 controls, and 2,828 families, they also found evidence for 4 previously unknown and distinct risk loci: at rs11755527 in intron 3 of the BACH2 gene (605394) on chromosome 6q15 (p = 4.7 x 10(-12)); at rs947474, near the PRKCQ gene (600448) on chromosome 10p15 (p = 3.7 x 10(-9)); at rs3825932 in intron 1 of the CTSH gene (116820) on chromosome 15q24 (p = 3.2 x 10(-15)); and at rs229541, located between the C1QTNF6 and SSTR3 (182453) genes on chromosome 22q13 (p = 2.0 x 10(-8)).

Barrett et al. (2009) reported the findings of a genomewide association study of type 1 diabetes, combined in a metaanalysis with 2 previously published studies (Wellcome Trust Case Control Consortium, 2007; Cooper et al., 2008). The total sample set included 7,514 cases and 9,045 reference samples. Forty-one distinct genomic locations provided evidence for association with type 1 diabetes in the metaanalysis (P less than 10(-6)). Using an analysis that combined comparisons over the 3 studies, they confirmed several previously reported associations, including rs2476601 at chromosome 1p13.2 (P = 8.5 x 10(-85)), rs7111341 at 11p15.5 (P = 4.4 x 10(-48)), rs2292239 at 12q13.2 (P = 2.2 x 10(-25)), and rs3184504 at 12q24.12 (P = 2.8 x 10(-27)). Barrett et al. (2009) further tested 27 novel regions in an independent set of 4,267 cases and 4,463 controls, and 2,319 affected sib pair families. Of these, 18 regions were replicated (P less than 0.01; overall P less than 5 x 10(-8)) and 4 additional regions provided nominal evidence of replication. A region on 1q32.1 represented by SNP rs3024505 (combined P = 1.9 x 10(-9)) contains the immunoregulatory cytokine genes IL10 (124092), IL19 (605687), and IL20 (605619). The strongest evidence of association among these 27 novel regions was achieved at rs10509540 on chromosome 10q23.31; see IDDM24, 613006.

Wallace et al. (2010) used imputation to assess association with T1D across 2.6 million SNPs in a total of 7,514 cases and 9,405 controls from 3 existing GWA studies (Wellcome Trust Case Control Consortium, 2007; Cooper et al., 2008; Barrett et al., 2009). They obtained evidence of an association at rs941576, a marker in the imprinted region of chromosome 14q32.2, for paternally inherited risk of T1D (p = 1.62 x 10(-10); ratio of allelic affects for paternal versus maternal transmissions = 0.75). Wallace et al. (2010) suggested that rs941576, which is located within intron 6 of the maternally expressed noncoding RNA gene MEG3 (605636), or another nearby variant alters the regulation of the neighboring functional candidate gene DLK1 (176290).

Inflammatory bowel disease (see 266600), including Crohn disease (CD) and ulcerative colitis (UC), and T1D are autoimmune diseases that may share common susceptibility pathways. Wang et al. (2010) examined known susceptibility loci for these diseases in a cohort of 1,689 CD cases, 777 UC cases, 989 T1D cases, and 6,197 shared control subjects of European ancestry. Multiple previously unreported or unconfirmed disease-loci associations were identified, including CD loci (ICOSLG, 605717; TNFSF15, 604052) and T1D loci (TNFAIP3; 191163) that conferred UC risk; UC loci (HERC2, 605837; IL26, 605679) that conferred T1D risk; and UC loci (IL10, 124092; CCNY, 612786) that conferred CD risk. T1D risk alleles residing at the PTPN22 (600716), IL27 (608273), IL18RAP (604509), and IL10 loci protected against CD. The strongest risk alleles for T1D within the major histocompatibility complex (MHC) conferred strong protection against CD and UC. The authors suggested that many loci involved in autoimmunity may be under a balancing selection due to antagonistic pleiotropic effects, and variants with opposite effects on different diseases may facilitate the maintenance of common susceptibility alleles in human populations.

HLA Associations

IDDM, although called the juvenile-onset type of diabetes, has its onset after the age of 20 years in 50% of cases. Caillat-Zucman et al. (1992) investigated whether the association of IDDM with certain HLA alleles, well documented in pediatric patients, also holds for adults. Interestingly, they found quite different HLA class II gene profiles, with a significantly higher percentage of non-DR3/non-DR4 genotypes and a lower percentage of DR3/4 genotypes in older patients. Although the non-DR3/non-DR4 patients presented clinically as IDDM, they showed a lower frequency of islet cell antibodies (ICA) at diagnosis and a significantly milder insulin deficiency. These data (1) suggest these subjects probably represent a particular subset of IDDM patients in whom frequency increases with age; (2) confirm the genetic heterogeneity of IDDM; and (3) prompt caution in extrapolating the genetic concepts derived from childhood IDDM to adult patients.

Nerup et al. (1974) found that IDDM (but not NIDDM) is associated with 2 particular HLA-A types (142800)--HLA-A8 and W15. Woodrow and Cudworth (1975) interpreted the association of HLA-A8 and W15 with IDDM as resulting from linkage disequilibrium between genes for these antigens and a gene determining susceptibility of diabetes.

To test for linkage between HLA and a locus for susceptibility to this disease, Clerget-Darpoux et al. (1980) studied 28 informative families with at least 1 child suffering from juvenile-onset IDDM. The 28 families were pooled with 21 from the literature and autosomal recessive inheritance was assumed. Maximum lod scores (6.00 to 7.36) were obtained for recombination fractions from 4% to 16%, according to the level of assumed penetrance (from 90% down to 10%). These high estimates of the recombination fraction are not consistent with the hypothesis that the association between IDDM and specific HLA haplotypes is a consequence of simple linkage disequilibrium between HLA and a susceptibility locus.

Spielman et al. (1980) did HLA-typing on all members of 33 families in which 2 or more sibs had IDDM. They interpreted the results as supporting the hypothesis that, closely linked to the HLA region, there is a locus (symbolized S by them) for susceptibility to insulin-dependent diabetes. (S(d) was their symbol for the susceptibility allele and S(a) for all other alleles.) They estimated penetrance for the homozygote for S(d) to be 71% and for the heterozygote 6.5%. The recombination fraction between S and HLA was estimated to be under 3%.

Rubinstein et al. (1981) analyzed 3 sets of published data on HLA-typed families with IDDM in which no significant heterogeneity was detected. Autosomal recessive inheritance and incomplete penetrance were assumed. A maximum lod score of 7.40 at theta = 0.05 was found. The segregation of HLA and GLO in 5 affected sib pairs (4 of the 5 pairs were HLA-identical and GLO-different), in which one of the sibs carried an HLA-GLO recombinant, placed the IDDM locus closer to HLA than to GLO.

Dunsworth et al. (1982) performed complex segregation and linkage analysis in 182 families with at least 1 IDDM proband. All families were typed for HLA-B antigens and 118 for HLA-DR. The recessive model best fitted the data, with the maximum likelihood estimate of recombination between HLA-DR and the diabetes susceptibility factor being 0.019. Substantial heterogeneity was suggested; the smallest recombination was for families whose probands had 2 high-risk D alleles. Using RFLPs of the HLA-DR-alpha gene, Stetler et al. (1985) could show a higher association than is found with serologic markers.

Rich et al. (1987) studied linkage of IDDM with HLA and factor B (138470) in combination with segregation analysis. They found evidence of strong linkage disequilibrium with the B-BF-D haplotype, with IDDM probably tightly linked to HLA-DR. The recombination fraction between the postulated major locus for IDDM and HLA was 0 in all models. They concluded that the best fitting genetic model of diabetic susceptibility is that of a single major locus with 'near recessivity' on a scale of standardized genetic liability, with a gene frequency of the IDDM susceptibility allele of approximately 14%.

Julier et al. (1991) studied polymorphisms of INS and neighboring loci in random diabetics, IDDM multiplex families, and controls. They found that HLA-DR4-positive diabetics showed an increased risk associated with common variants at polymorphic sites in a 19-kb segment spanned by the 5-prime INS VNTR and the third intron of the gene for insulin-like growth factor II (147470). In multiplex families the IDDM-associated alleles for polymorphisms in this region were transmitted preferentially to HLA-DR4-positive diabetic offspring from heterozygous parents. The effect was strongest in paternal meioses, suggesting a possible role for maternal imprinting. Julier et al. (1991) suggested that the results strongly support the existence of a gene or genes affecting HLA-DR4 IDDM susceptibility in a 19-kb region of INS-IGF2. Their approach may be useful in mapping susceptibility loci in other common diseases.

The fact that the association between IDDM and certain HLA-DQ alleles is even stronger than that with certain DR alleles and that there is little association with HLA-DP provides a boundary of disease association to the 430 kb between DQ and DP. In further studies of disease association with TAP (transporter associated with antigen processing) genes (170260), which map approximately midway between DP and DQ, Jackson and Capra (1993) found a higher association of a TAP allele with IDDM than with any single HLA-DP allele but the risk was lower than with HLA-DQB1*0302. These data provided new limits for IDDM susceptibility to the 190-kb interval between TAP1 and HLA-DQB1.

In a 2-stage approach to fine mapping, Herr et al. (2000) evaluated linkage in 385 affected sib-pair families using 13 evenly spaced polymorphic microsatellite markers spanning 14 Mb. Evidence of disease association was found for D6S2444, located within the 95% confidence interval of 1.7 cM obtained by linkage. Analysis of an additional 12 flanking markers revealed a highly specific region of 570 kb associated with disease that included the HLA class II genes. The peak of association was as close as 85 kb centromeric of HLA-DQB1. Recombination within the major histocompatibility complex was rare and nearly absent in the class III region. The authors concluded that the majority of disease association in the region can be explained by linkage disequilibrium with the class II susceptibility genes.

Greenbaum et al. (2000) noted that the presence of HLA haplotype DQA1*0102-DQB1*0602 is associated with protection from type I diabetes. The Diabetes Prevention Trial-type I has identified 100 islet cell antibody (ICA)-positive relatives with this protective haplotype, far exceeding the number of such subjects reported in other studies worldwide. Comparisons between ICA+ relatives with and without DQB1*0602 demonstrated no differences in gender or age; however, among racial groups, African American ICA+ relatives were more likely to carry this haplotype than others. The ICA+ DQB1*0602 individuals were less likely to have additional risk factors for diabetes (insulin autoantibody (IAA) positive or low first phase insulin release (FPIR)) than ICA+ relatives without DQB1*0602. However, 29% of the ICA+ DQB1*0602 relatives did have IAA or low FPIR. Hispanic ICA+ individuals with DQB1*0602 were more likely to be IAA positive or to have low FPIR than other racial groups. The authors conclude that the presence of ICA found in relatives suggests that whatever the mechanism that protects DQB1*0602 individuals from diabetes, it is likely to occur after the diabetes disease process has begun. In addition, they suggest that there may be different effects of DQB1*0602 between ethnic groups.

Redondo et al. (2000) used the transmission disequilibrium test to analyze haplotypes for association and linkage to diabetes within families from the Human Biological Data Interchange type I diabetes repository (1,371 subjects) and from the Norwegian Type 1 Diabetes Simplex Families study (2,441 subjects). DQA1*0102-DQB1*0602 was transmitted to 2 of 313 (0.6%) affected offspring (P less than 0.001, vs the expected 50% transmission). Protection was associated with the DQ alleles rather than DRB1*1501 in linkage disequilibrium with DQA1*0102-DQB1*0602: rare DRB1*1501 haplotypes without DQA1*0102-DQB1*0602 were transmitted to 5 of 11 affected offspring, whereas DQA1*0102-DQB1*0602 was transmitted to 2 of 313 affected offspring (P less than 0.0001). The authors concluded that both DR and DQ molecules (the DRB1*1401 and DQA1*0102-DQB1*0602 alleles) can provide protection from type IA diabetes.

Li et al. (2001) assessed the prevalence of families with both type I and type II diabetes in Finland and studied, in patients with type II diabetes, the association between a family history of type I diabetes, GAD antibodies (GADab), and type I diabetes-associated HLA-DQB1 genotypes. Further, in mixed type I/type II diabetes families, they investigated whether sharing an HLA haplotype with a family member with type I diabetes influenced the manifestation of type II diabetes. Among 695 families with more than 1 patient with type II diabetes, 100 (14%) also had members with type I diabetes. Type II diabetic patients from the mixed families more often had GADab (18% vs 8%) and DQB1*0302/X genotype (25% vs 12%) than patients from families with only type II diabetes; however, they had a lower frequency of DQB1*02/0302 genotype compared with adult-onset type I patients (4% vs 27%). In the mixed families, the insulin response to oral glucose load was impaired in patients who had HLA class II risk haplotypes, either DR3(17)-DQA1*0501-DQB1*02 or DR4*0401/4-DQA1*0301-DQB1*0302, compared with patients without such haplotypes. This finding was independent of the presence of GADab. The authors concluded that type I and type II diabetes cluster in the same families. A shared genetic background with a patient with type I diabetes predisposes type II diabetic patients both to autoantibody positivity and, irrespective of antibody positivity, to impaired insulin secretion. Their findings also supported a possible genetic interaction between type I and type II diabetes mediated by the HLA locus.

Linkage data implicating other disease susceptibility loci for type I diabetes are conflicting. This is likely due to (1) the limited power for detection of contributions of additional susceptibility loci, given the limited number of informative families available for study, (2) factors such as genetic heterogeneity between populations, and (3) potential gene-gene and gene-environment interactions. To circumvent some of these problems, the European Consortium for IDDM Genome Studies (2001) conducted a genomewide linkage analysis for type I diabetes mellitus-susceptibility loci in 408 multiplex families from Scandinavia, a population expected to be homogeneous for genetic and environmental factors. In addition to verifying the HLA and INS susceptibility loci, the study confirmed the locus of IDDM15 (601666) on chromosome 6q21. Suggestive evidence of additional susceptibility loci was found on 2p, 5q, and 16p. For some loci, the support for linkage increased substantially when families were stratified on the basis of HLA or INS genotypes, with statistically significant heterogeneity between the stratified subgroups. These data support both the existence of non-HLA genes of significance