Type 1 Diabetes Genetics Consortium

Abstract

Type 1 diabetes (T1D) results from the autoimmune destruction of the insulin-producing β cells. Genetic factors account for approximately 50% of the risk for T1D but, by the late 1990s, the genetic basis was limited. The Type 1 Diabetes Genetics Consortium (T1DGC) was formed in 2002 to accelerate discovery of genes contributing to T1D risk through a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) to assemble existing data and samples from affected sib-pair families and to establish new collections. In recognition of the 75th anniversary of the NIDDK, this manuscript highlights the contributions made by the T1DGC to understanding the genetic basis of T1D using both family (for linkage) and case-control (for genome-wide association) designs. The T1DGC conducted large-scale genetic research and used fine mapping to define risk regions. The T1DGC data, results, and samples have been made available to the scientific community, leading to the discovery of more than 100 loci associated with T1D risk, many with small effects and relevant to autoimmune pathways. The T1DGC not only expanded the list of genes contributing to disease risk but also identified noncoding genetic variation in disease-relevant cell types that contribute to the etiology of T1D. The success of the T1DGC and the NIDDK investment in the global consortium is highlighted in its continuing effect on mapping genetic variants to their function and identifying pathways that provide new targets for the prediction, prevention, and treatment of T1D.

The formation of the Type 1 Diabetes Genetics Consortium (T1DGC) dates to a meeting in Skagen, Denmark, in April 2000. The “Diabetes Genetics Y2K Workshop” was hosted by Drs Jørn Nerup and Flemming Pociot and provided a venue for a small group of scientists to discuss the needs of the field to detect genes that determine risk for type 1 diabetes (T1D). Participants in the meeting included Drs Pat Concannon, Nancy Cox, Cécile Julier, Jurg Ott, Stephen Rich, and John Todd. Topics discussed included the (then current) status of the field and approaches to identifying complex disease susceptibility genes. The meeting generated a consensus that collaboration was needed as well as the assembly of a unified collection of existing samples and data and recruitment of new samples from global resources.

Following the Skagen workshop, a message was sent by Dr Stephen Rich to Dr Robert Goldstein (chief scientific officer, Juvenile Diabetes Research Foundation [JDRF]) on May 24, 2000, suggesting that the JDRF and the NIDDK could facilitate the development of a collaboration to identify susceptibility genes for T1D. A meeting on “Genetics of Type 1 Diabetes” was held in Bethesda, Maryland, USA, on November 20, 2000, with presentation by experts in T1D, discussion by additional investigators in the field, and NIDDK/JDRF staff who organized the meeting. The consensus from the meeting was that a consortium was necessary to establish resources, support management and coordination, and stimulate collaboration among scientists. Consequently, the T1DGC Steering Committee was to be formed with guidance from the NIDDK and JDRF staff.

Meeting participants agreed to establish a resource to expand the ascertainment of families with members affected with T1D and to facilitate the collection of data and biospecimens from consenting participants with the ultimate goal of enhancing the identification of genes associated with T1D. The NIDDK extended invitations to global experts in T1D genetics to establish the T1DGC Steering Committee, with Dr Stephen Rich to serve as chair of the Steering Committee. Members of the committee agreed to guide the project's development, participating in face-to-face meetings and conference calls to achieve the study's goals. Steering Committee members discussed content of a National Institutes of Health (NIH) grant application to secure funding for the T1DGC objectives. The Steering Committee members included NIDDK (Beena Akolkar) and JDRF (Marie Nierras) staff, who were not involved with the development of the T1DGC grant, and Patrick Concannon (Benaroya Research Institute), Henry Erlich (Roche Molecular Systems), Cécile Julier (Institut Pasteur de Paris and Inserm), Grant Morahan (Walter & Eliza Hall Research Institute, Australia), Jørn Nerup (Steno Diabetes Center, Denmark), Flemming Pociot (Steno Diabetes Center, Denmark), John Todd (University of Cambridge, UK), and Stephen Rich (chair of the Steering Committee, principal investigator of the grant application, Wake Forest University, USA).

An initial research grant from NIDDK was awarded to Dr Rich to combine the genome scan data from the North American, United Kingdom, and European studies (individually published) and provide the preliminary data for an NIH application. Support for T1DGC Steering Committee meetings to plan the NIH application was provided by the JDRF (for non-US meetings and workshops) and the NIDDK (for US meetings). Administrative supplements from the NIDDK to an existing NIDDK award to Dr Rich provided support for administrative planning, to establish the T1DGC website to streamline member communications, and to initiate a Coordinating Center at Wake Forest University. The T1DGC application contained requests for the establishment of global networks (Asia-Pacific [AP], European [EU], North American [NA], and UK) for ascertainment of study participants; laboratories for sample processing, genetic, and immune data generation; processes for data sharing; and resource utilization. The T1DGC application was submitted in November 2001 and initial funding was awarded by the NIDDK on September 15, 2002, through the Special Diabetes Program (https://www.niddk.nih.gov/about-niddk/research-areas/diabetes/type-1-diabetes-special-statutory-funding-program/about-special-diabetes-program). In recognition of the NIDDK's 75th anniversary, this manuscript recounts the research conducted by the T1DGC that has expanded the genetic landscape for T1D through discovery of more than 100 loci associated with T1D risk, identification of key pathogenic cell types, and recognition of the importance of noncoding variation in the etiology of T1D.

Type 1 Diabetes Genetics Consortium–Specific Goals and Organizational Structure

The status of the genetic basis of T1D prior to the T1DGC was characterized by a small number of replicated candidate genes, mainly in the human major histocompatibility complex (HLA-A, HLA-B, HLA-DR, HLA-DQ, and HLA-DP) and the insulin (INS) gene. Numerous other candidate genes had been reported in small, underpowered studies, but generally failed to replicate in independent studies (that were also limited by small sample size and limited statistical power). The primary goal of the T1DGC was to better understand the genetic basis of T1D by identifying potential causal variants in global populations. The T1DGC would perform this task by conducting research and providing resources to the community to improve the power of genetic studies (1).

To provide the appropriate infrastructure for the logistical aspects of the study, the T1DGC Coordinating Center was established at Wake Forest University Health Sciences in Winston Salem, North Carolina, USA. The Coordinating Center facilitated the day-to-day operations of the T1DGC in 3 general areas: (1) operations, (2) statistics, and (3) systems. The Coordinating Center was staffed by faculty with expertise in project management, statistical genetics, genetic epidemiology, epidemiology, biostatistics, clinical trials, ethical and legal issues, bioinformatics, and laboratory procedures. The Coordinating Center was responsible for fiscal aspects, maintained the T1DGC website, prepared reports and meeting materials, tracked recruitment progress, developed operations and procedure manuals, and created data analysis plans.

Recruitment of Participants

Based on previous genome-wide linkage scans in T1D and estimates of familial clustering, it was determined that at least 4000 affected sib-pair (ASP) families would be needed to provide 80% power to identify chromosome regions with locus-specific effects of the sibling risk ratio (λs) greater than 1.3 (for a causal allele frequency of 0.3 under a multiplicative model corresponding to a relative risk >2.3, which would exceed that of 2 known non-HLA loci, PTPN22 and INS). To reach this goal, the T1DGC identified existing resources of ASP families (∼1400) with appropriate consent for broad sharing; however, an additional 2500 ASP families were needed through new collection. To identify, recruit, and collect the samples and data from these new ASP families, the T1DGC developed 4 regional networks (2, 3): NA, EU, UK, and AP. Recruitment targets varied by network: AP and UK networks each targeted approximately 200 ASP families, while EU and NA networks each targeted approximately 1200 ASP families. In 2000, these regional networks were tasked with recruiting study participants. Each network established a recruitment center and staff, using multiple approaches to recruit volunteers. Within each network, field centers identified, ascertained, and collected samples and data from participating families.

The large number of samples and the international locations of the T1DGC networks required the establishment of laboratories at multiple sites. Each network established a DNA repository (for DNA isolation and Epstein-Barr virus [EBV]-transformed B-cell lines), an HLA laboratory to genotype each study participant, and an islet autoantibody laboratory. The T1DGC Coordinating Center certified these sites, aided staff, and provided protocols for harmonizing methods across the study. Descriptions of the consent process, template informed consent documents for participants, and the T1DGC agreement for members have been described elsewhere (4).

Samples and data from the T1DGC have been deposited in the NIDDK Central Repository (https://repository.niddk.nih.gov/home), and summary statistics from analyses are provided in the database of Genotypes and Phenotypes (dbGaP, https://dbgap.ncbi.nlm.nih.gov/aa/wga.cgi?page=login, accession Nos. phs000180, phs000910, phs000911, phs001222, phs001426, and phs002468) as well as the European Genome-Phenome Archive (EGA, https://ega-archive.org/) and the T1D Knowledge Portal (https://t1d.hugeamp.org/).

Findings, Publications, and Data Sharing

The T1DGC provided data and resources to bona fide investigators to enable research on the genetic basis of T1D. As one of the initial NIH-supported consortia, the T1DGC focused on obtaining informed consent from participants to permit sharing of data as well as biological materials with limited restrictions. Informed consents across the T1DGC networks used a template that was modified based on each site's cultural sensitivities and legal restrictions, resulting in restriction of sharing to studies of T1D, autoimmunity, and complications of T1D. Biological resources were grouped into two forms, those that were considered “renewable” (eg, data collected from standardized forms, assays, and DNA from EBV-transformed B-cell lines) and “nonrenewable” (eg, serum and plasma samples, and peripheral blood). For renewable resources, investigators could request and receive the materials without review; however, for nonrenewable resources, a request would be reviewed by an external committee and access determined.

An objective of the T1DGC was to make collected and generated resources and data as widely available as possible. All publications from the T1DGC were generated by a writing committee and would include the “T1DGC” as a banner coauthor, with all members cited in an appendix. Publications that used T1DGC resources by members and nonmembers would be required to acknowledge the funding of the T1DGC and, depending on the publication, have “T1DGC” as either a coauthor or as an acknowledged contributor of materials or data. Publications by the T1DGC include genetic analysis reports, outcomes from T1DGC-supported workshops, and numerous studies carried out by individual researchers.

Linkage Analysis

The T1DGC conducted a series of genome-wide linkage scans. In the first report (5), 4 sets of European ancestry families (3 previously published and 1 newly collected) provided genome-wide linkage scan data for the combined analyses, for a total of 1435 families (6358 individuals; 3072 with T1D). At that time, microsatellite markers were assessed within families and were merged across studies for a total of 1190 markers used in the combined analysis. The strongest evidence for linkage to T1D was on chromosome 6p21 in the major histocompatibility complex (MHC) (encoding the HLA genes) with 9 non–HLA-linked regions having nominal evidence (P < .01), including loci in the regions of 2q31-q33, 10p14-q11, and 16q22-q24.

With increased sample size (2496 families) and improvements in genotyping technologies, the T1DGC transitioned from microsatellite markers to approximately 6000 single-nucleotide variations (SNVs, formerly single-nucleotide polymorphisms [SNPs]). By genotyping ASP families on SNV arrays, T1DGC confirmed known associations at INS (6, 7), IFIH1 (2q24) (8), and KIAA0350 (CLEC16A, 16p13) (9) and discovered a new association at the UBASH3A locus (21q22.3) (10). Subsequently, the T1DGC collected DNA from additional families and increased the sample size for linkage to 4422 ASPs from 3892 families (11). This was the largest linkage scan ever conducted and provided further evidence for linkage in the HLA region and the INS locus, and with suggestive support for 5 other loci (12-14).

Genome-wide Association Scans

While family-based approaches to gene mapping were limited due to difficulties in the recruitment of families with parents and at least 2 children with T1D, wide support intervals of loci containing many genes, and low statistical power, the systematic characterization of SNVs across the human genome (15), along with the identification of haplotype blocks showing strong linkage disequilibrium (16, 17) and haplotype tagging (18), led to the development of array-based SNV genotyping that could be applied to DNA collections “at scale.” With the improvement in genotyping technologies, the field quickly shifted to genome-wide association studies using case-control designs. The shift was driven by the relative ease of collecting individual samples compared to family-based collections.

The Wellcome Trust Case-Control Consortium (WTCCC) provided the first application for a large-scale genome-wide association scan (GWAS) approach (19). The WTCCC performed SNV genotyping in 7 common diseases, including T1D (19, 20), each providing 2000 cases and compared to 3000 shared controls. The WTCCC was designed to detect common SNVs associated with disease that had an odds ratio (OR) greater than 1.5. In retrospect, this design was unlikely to provide a more complete description of the genetic architecture of T1D, where estimates suggested that the ORs for associated SNVs are typically between 1.1 and 1.3 (7-9, 19-22). Nevertheless, 4 new T1D loci were identified and replicated, including early efforts toward fine mapping of a region to identify the causal variants/genes as well as those shared with autoimmune thyroid disease (20).

The T1DGC conducted a well-powered GWAS meta-analysis (23) combining existing and newly ascertained cases with T1D (n = 7514) and unaffected reference (control) samples (n = 9045). In this new study, a total of 41 distinct genomic regions were identified at P less than or equal to 5 × 10⁻⁸. Of these, 27 novel loci were selected for replication in an independent set of 4267 cases, 4463 controls, and 2319 ASP families. Many of the novel associations reported by the T1DGC were within, or adjacent to, promising candidate genes for T1D (eg, IL10, IL19, IL20, GLIS3, CD69, and IL27) and were of much smaller ORs (less than 1.15). The T1DGC's commitment to data sharing empowered the scientific community, providing data to investigators even during the primary manuscript preparation (eg, 8, 24, 25), and continuing to the present (26-31). Thus, T1DGC GWAS data have significantly contributed to advancing research into the genetic basis of T1D, motivating further analyses and discoveries.

Fine Mapping

The T1DGC participated in the design of the ImmunoChip (32), a targeted genotyping array of approximately 186 000 SNVs clustered in 186 regions of the genome with previous evidence of association to 1 or more of 12 autoimmune diseases. The T1DGC conducted fine mapping with the ImmunoChip to identify additional risk loci for T1D that were shared with other autoimmune diseases, refine existing T1D risk loci, and determine regulatory sequences in tissues and cell types relevant for T1D (33). T1DGC's first ImmunoChip project focused on European ancestry cohorts; 138 229 SNVs were analyzed using data from 6670 cases (20) and 6523 controls from the British 1958 Birth Cohort (34), 2893 controls from the UK National Blood Service (19), and 2846 controls from the NIHR Cambridge Biomedical Research Centre Cambridge BioResource (35). Additionally, data from 2 601 T1DGC ASP families and 69 T1DGC trio families (36) were incorporated. The first stage of the ImmunoChip study revealed associations with T1D in 44 genomic regions, including 38 previously recognized loci and 4 newly identified loci (1q32.1, CAMSAP2; 2q13, ACOXL; 4q32.3, CPE/TLL1; and 5p13.2, IL7R), providing a robust set of T1D-associated candidate SNVs for further investigation.

With the SNVs providing statistically significant evidence of association with T1D, the associated SNVs were compared with those SNVs associated with 15 other immune-mediated diseases (33). The T1D SNVs are more genetically similar in effect direction and size to diseases characterized by organ-specific autoantibodies, such as juvenile idiopathic arthritis, Graves disease, and rheumatoid arthritis, rather than to disorders such as ulcerative colitis and Crohn disease. The strongest enrichment for associated SNVs was with juvenile idiopathic arthritis, while the strongest negative enrichment was with ulcerative colitis. Further, T1D-associated SNVs were enriched in enhancer chromatin states in immunologically relevant tissues (eg, thymus, CD4⁺ and CD8⁺ T cells, CD19⁺ B cells, and CD34⁺ stem cells) with little evidence of enrichment in promoter sequences (33). These results strongly suggest that the majority of functional variation critical to T1D risk resides in enhancer sequences. Systems genetics analysis of T1D-associated SNVs showed that most were not in strong linkage disequilibrium with any SNVs that encoded protein changes but instead were associated with gene expression changes in T or B lymphocytes, mediating these effects in cis as well as in trans (37). The ImmunoChip array has subsequently been applied to other NIDDK-supported projects, including TrialNet (38) and The Environmental Determinants of Diabetes in the Young (TEDDY) initiatives (39). The T1DGC ImmunoChip data have been used to develop T1D genetic risk scores (40) that have been employed in multiple studies for the prediction of those at genetic risk of developing T1D (41, 42).

Multiancestry Studies in Type 1 Diabetes

Throughout the field of human genetics, there is a recognized need for increased genetic diversity in study populations, not only for improved ability to discover new genes associated with complex human phenotypes but also for equitable use of resources that can improve the human condition across social/cultural domains (43). The historical view of T1D was that the disease primarily affected European ancestry populations; however, the most significant increase in recent incidence of T1D is in non-European populations (44). The first genomic study of T1D in those of African ancestry included 227 individuals with T1D and 471 controls genotyped on the ImmunoChip (45). With the small sample size, no statistically significant finding other than those in the HLA region were identified. Nevertheless, for dissection of the complex associations with T1D in the MHC, even these small studies (eg, from Japanese (46) and Afro-Caribbean (47) populations) are important.

Subsequently, the T1DGC assembled the largest available collection of case and control participants with African ancestry for analysis with the ImmunoChip (48), comprising 1021 cases and 2928 controls. While this data set is smaller than the GWAS conducted in European ancestry samples a decade earlier, it remains an important resource. Notably, SNVs in 6 of the T1D risk regions identified in European-ancestry cohorts (2q33.2, CTLA4; 6p21.3, HLA; 6q22.32, CENPW; 11p15.5, INS; 12q13.2, IKZF4-RPS26-ERBB3; and 17q12, IKZF3-ORMDL3-GSDMB) attained genome-wide significance in African-ancestry participants. For most of these regions, the most statistically significant SNV in the African-ancestry population was not the same as that identified in populations of European ancestry, and the most associated SNV in participants of European ancestry was often more difficult to determine (given greater linkage disequilibrium and larger haplotype block size) than that seen in those of African ancestry.

In African-ancestry populations, rs3842727 in the INS locus was the most significantly associated non-HLA SNV. In contrast, in European-ancestry populations, rs689 (tagging the INS promoter VNTR) had the largest effect in the locus (48). The 12q13.2 locus is a complex region with several plausible candidates for T1D susceptibility genes (IKZF4, RPS26, and ERBB3). In European-ancestry individuals, GWAS results support ERBB3 while fine mapping supported the IKZF4-RPS26 region. In the African-ancestry samples, rs705705 (near RPS26) has the strongest association with T1D and rs2292239 (ERBB3) was not associated with T1D. The 17q12 locus has numerous SNVs associated with T1D in European-ancestry populations and contains several candidate genes (IKZF3, ORMDL3, and GSDMB). In African-ancestry participants, the rs56380902 SNV (intronic in GSDMB) was significantly associated with T1D; in contrast, the most significant European-ancestry SNV (rs12453507) was not associated with T1D risk in African ancestry data. Inclusion of data from participants with African ancestry centered the T1D risk gene in this locus to GSDMB.

T1DGC fine-mapping analyses have included participants of diverse genetic ancestry, including a large fine-mapping and functional evaluation report (49) and a GWAS of T1D risk and age at onset (30). Despite progress made through inclusion of genetically diverse population samples, the work by the T1DGC and others emphasizes the need for significantly increased recruitment of participants with genetic ancestries underrepresented currently in T1D research (50). Not only are larger sample sizes of genetically diverse populations needed, the sampling of sites across continents needs to be extensive to capture both genetic and environmental heterogeneity. The inclusion of diverse ancestry populations in the future should accelerate gene discovery in T1D (now approaching 50 years; Fig. 1), and increase the number of loci (now >100, Fig. 2) to open new avenues for prediction and intervention.

Figure 1.

Timeline of gene discovery in type 1 diabetes (T1D). The earliest genetic association with T1D occurred in 1973 (HL-A) with relatively few discoveries over the next 30 years (INS, CTLA4, PTPN22, IL2RA, IFIH1) until advances in genotyping technology and assembly of large case-control series (Wellcome Trust Case Control Consortium) stimulated discovery and fine mapping of genomic regions associated with T1D.

Figure 2.

Genes associated with type 1 diabetes (T1D) risk and protection. Current status of candidate genes associated with T1D risk, ranked by the size of their effect on risk. There are now more than 100 candidate genes in the human genome associated with T1D using data generated by the Type 1 Diabetes Genetics Consortium (10, 23, 28, 33, 49).

Structural Variants

The T1DGC focused its attention on DNA variation that, except for highly polymorphic HLA loci and early microsatellites, typically involve a single base change (SNV). Structural variants include copy number variants and variable number of tandem repeats (VNTRs) such as the INS VNTR, the second most strongly associated T1D locus (51). The scan confirmed the known T1D association at the INS locus; the INS VNTR association with T1D was indistinguishable from the tagging SNV rs689. However, the study did not identify novel T1D structural variation loci.

Multiomic Analysis in Type 1 Diabetes

The range of molecular tools to characterize the functional significance of genetic variants associated with T1D has expanded rapidly. In T1D, studies have been conducted on gene expression (first by arrays and later by bulk and single-cell RNA sequencing), epigenomics (assessment of the types of modifications to DNA that regulate gene function without change in DNA sequence), and proteomics (the structure and function of the proteins resulting from genomic, transcriptomic, and epigenomic modification). These studies have used whole blood (a mixture of cell types) or cells associated with the diseases process (such as T-cell subsets, B cells, and islet β cells). The T1DGC processed whole blood, stored aliquots of peripheral blood mononuclear cells, and established EBV-transformed B cell lines that have enabled research to be conducted in these areas.

The general concept for application of multiomic approaches to determining the candidate causal variants (from GWAS and fine mapping) to cell type–specific function involves several steps. An example of multiomic approaches using T1DGC resources (49) followed a multiethnic meta-analysis framework, conducting fine mapping to refine the T1D associations across regions covered by the ImmunoChip. From this multiethnic analysis, 2 intronic variants, rs72928038 and rs6908626, in the BACH2 locus (6q15) emerged as likely causal variants. Both alternative alleles were associated with increased risk and chromatin-state annotations across cell types from the BLUEPRINT Consortium (52), and the Roadmap Epigenomics Project annotated both SNVs as likely functional, with rs72928038 overlapping a T cell–specific active enhancer and rs6908626 residing in the BACH2 promoter. Using Assay for Transposase-Accessible Chromatin with high-throughput sequencing (ATAC-seq) to identify chromatin-accessible quantitative trait loci (caQTLs), rs72928038 had decreased accessibility of its BACH2 enhancer while rs6908626 did not alter accessibility at the BACH2 promoter (49). In 14 participants heterozygous for rs72928038, only 4% (5/121) of ATAC-seq reads overlapping the enhancer site contained the T1D risk allele, suggesting the risk allele leads to restricted accessibility. In contrast, chromatin accessibility at rs6908626 (promoter) did not exhibit a similar allelic bias in heterozygotes. These data suggest that rs72928038, rather than rs6908626, is the site in BACH2 that is functionally relevant and contributes to T1D risk. Additional work (49) implicated rs72928038 in decreased expression of BACH2 in whole blood and other immune-cell types, overlapping binding sites for the STAT1 and the ETS family of transcription factors. Furthermore, rs72928038 had allele-specific ETS1 binding but no STAT1 binding, thereby prioritizing rs72928038 alternative allele (A) disrupting ETS1 binding that leads to decreased enhancer activity and BACH2 expression in naive CD4⁺ T cells (49). The T1DGC has provided data and sample resources that have been used to increase understanding of T1D pathogenesis (53-56).

Workshops Supported by the Type 1 Diabetes Genetics Consortium

As part of the outreach component of the T1DGC, a series of workshops was established, providing training in bioinformatics, with a focus on tools and utilization of public databases for advanced interrogation of genomic features (often specific to the MHC) and open to T1DGC members and their staff/trainees. The workshops provide value to individual investigators in multiple countries who contribute to the study, allowing staff/trainees access to tools that would allow them to use the data generated by the T1DGC.

The T1DGC launched the MHC Fine Mapping Workshop to more comprehensively examine the genetic basis of T1D in the extended MHC region (encompassing 4 Mb between 29 and 34 Mb on chromosome 6p21.3). Characterization of genes in this region used classic HLA genotyping and a framework map of 66 microsatellite markers and more than 3000 SNVs in 2321 ASP families. The reports were published in Diabetes, Obesity and Metabolism 11 (Supplement 2), 2009, reflecting multiple analytic approaches applied to a common data set (57). Stepwise conditional logistic regression with recursive partitioning identified significant contributions to risk of T1D by HLA-B and HLA-A independent of the large effects of HLA-DQB1, HLA-DRB1, and HLA-DPB1. There was evidence supporting an MHC effect by year of diagnosis, as well as MHC seasonality effects, but little support for maternal-offspring HLA compatibility, parent-of-origin, or noninherited maternal effects at the 8 classic HLA loci.

The T1DGC Rapid Response Workshop was established to evaluate candidate gene associations in a large collection of T1DGC ASP families (58). The project consisted of 2298 ASP families (11 159 individuals) with 382 SNVs genotyped in 21 candidate genes associated with T1D, chosen based on the published literature, including those considered “confirmed” or “candidates.” As all SNV data were generated on 2 platforms, the project demonstrated a major effect of small rates of genotyping errors on association results. Association with T1D was confirmed for INS, PTPN22 (as well as an association independent of the known causal R620W variant), IL2RA, IFIH1, and CTLA4, while not replicating the SUMO4 association identified in populations of Asian ancestry. Individual candidate gene analyses and multiple gene analyses were presented in Washington, DC, USA, and 17 manuscripts were published in Genes and Immunity Volume 10 (Supplement 1) December 2009.

The T1DGC Autoantibody Workshop was established to encourage the genetics community to use T1DGC phenotypic, genotypic, and autoantibody data on ASP families to discover genes accounting for variation in the presence of autoantibodies. The T1DGC partnered analysts with content experts in a working group setting, providing the working groups with autoantibody and genetic data on 9976 individuals from 2321 ASP families. Six reports were published (Diabetes Care, 38: Supplement 2, 2015), ranging from candidate gene analyses of selected autoantibodies to evaluation of regions of genetic variants associated with autoimmunity on the collection of autoantibodies (59). The HLA genes were the major contributors to the presence of islet and nonislet autoantibodies in the context of T1D. SNVs in IFIH1, PTPN22, SH2B3, BACH2, and CTLA4 were associated with the occurrence of multiple autoantibodies, while other SNVs were associated with single autoantibodies. Together, HLA-DRB1*01:01, HLA-DRB1*04:04, and the PTPN22 R620W variant were consistently associated with autoimmunity in the workshop data.

Future Opportunities

Members of the T1DGC and others have reflected on future research opportunities in predicting the genetic contribution to risk/protection in T1D, including initiation of islet autoimmunity, progression of β-cell destruction, and development of clinical disease (eg 60, 61). As T1D is likely heterogeneous, defining its genetic basis can contribute to understanding its heterogeneity. This characterization can aid in application of precision approaches, especially as it relates to the prediction of risk, diagnosis, and identification of those who may respond to pharmacological intervention (eg, teplizumab, approved by the US Food and Drug Administration in 2022), and prognosis of likely risk of diabetic complications.

The natural history of T1D starts with an underlying genetic risk that may vary with ancestry and that interacts with environmental exposures. There is a need to establish underlying phenotypes that reflect consequences of variation in complex genetic networks operating within a dynamic environmental framework. Further genetic and functional evaluations are necessary to establish and confirm involvement of such networks in T1D to fully elucidate the biological mechanisms of the networks and to identify the strongest risk phenotypes (60), some that will be regulated by genetic variation associated with T1D and others that may affect initiation of the autoimmune process or progression of islet autoimmunity.

The pathogenesis of T1D may represent the perturbation of multiple functional modules that are products of genetic variation but amplified by epigenetic modification in such a way that the targeted network component can produce differing effects on risk through developmental and/or physiological instability. Understanding these networks and consequences of perturbations of networks may provide insights for novel intervention targets, including delay or prevention of T1D diagnosis in individuals positive for multiple islet autoantibodies (62, 63).

There are gaps in understanding the genetic basis of T1D, particularly with respect to populations of diverse genetic ancestry, in comprehensive analysis of gene expression (in cells and relevant tissues), epigenetic modification, and development of biomarker panels through proteomic analyses (61). While the T1DGC has contributed to identifying the majority of genetic risk, there remains “missing heritability.” Additional focused research into, for example, rare variants of large effect and mechanisms that alter transcription, is important. The resources established by the T1DGC as well as other NIDDK-supported and other resources, including TEDDY (https://teddy.epi.usf.edu/teddy/), the Human Islet Research Network (HIRN) (https://hirnetwork.org), the Human Pancreas Analysis Program (HPAP) (https://hirnetwork.org), and the Network for Pancreatic Organ Donors with Diabetes (nPOD, https://npod.org) provide a basis for future discovery of events from early to late stage in the disease process. These resources provide the potential for identification of new targets for intervention or biomarkers for monitoring the effects and outcomes of potential therapeutic agents.

Summary

To provide adequate statistical power to detect genes of even smaller effect that contribute to T1D risk, the T1DGC was formed in 2002 with support by the NIDDK to assemble existing data and samples from ASP families as well as to establish new collections. In recognition of the NIDDK's 75th anniversary, this report highlights the T1DGC and how it has contributed to the expansion of the T1D genetic landscape to more than 100 risk regions.

It is now well established that a substantial portion of the genetic risk for T1D is encoded in the HLA locus, with a few additional loci (eg, INS and PTPN22) that make moderate contributions to risk. However, the vast majority of T1D risk loci make only minor individual contributions to risk. Because of their small individual effect sizes, these loci are difficult for individual investigators or laboratories to identify because of the need for very large sample sizes to discover and then confirm their contribution with adequate statistical evidence. The T1DGC, with its extensive resources, has made a substantial contribution to the identification of these loci with small effect sizes (see Fig. 2).

With both the collected biospecimens and the data from prior T1DGC studies available through NIDDK-supported repositories, individual laboratories can now discover and/or confirm additional T1D risk loci as well as fine-map those that are already established. The effect of the T1DGC is likely to continue as investigators tackle more complex issues, such as the integration of genomics with gene expression, epigenetics, and 3-dimensional mapping of interactions within the genome, that will be needed to determine the likely target effector genes involved in T1D pathogenesis.

Abbreviations

AP

Asian Pacific

ASP

affected sib-pair

ATAC-seq

Assay for Transposase-Accessible Chromatin with high-throughput sequencing

EBV

Epstein-Barr virus

EU

European

GWAS

genome-wide association scan

HLA

human major histocompatibility complex

INS

insulin gene

JDRF

Juvenile Diabetes Research Foundation

MHC

major histocompatibility complex

NA

North American

NIDDK

National Institute of Diabetes and Digestive and Kidney Diseases

NIH

National Institutes of Health

OR

odds ratio

SNV

single-nucleotide variation

T1DGC

Type 1 Diabetes Genetics Consortium

VNTRs

variable number of tandem repeats

WTCCC

Wellcome Trust Case-Control Consortium

Disclosures

F.P. has received consulting and lecture fees from Sanofi. J.A.T. is a consultant for GSK, Avammune, Vesalius, and Immunocore, and codirector of the Oxford-GSK Institute for Molecular and Computational Medicine. S.S.R. is a consultant for Westat and has received lecture fees from Sanofi. S.O.G., P.C., B.A., H.A.E., C.J., G.M., and C.R.N. report no disclosures.

Data Availability

Data sharing is not applicable to this article as no data sets were generated or analyzed during the present study. Data from the T1DGC are accessible from public repositories identified in the article.