Current status and the future for the genetics of type I diabetes

Abstract

The Type I Diabetes Genetics Consortium (T1DGC) is an international collaboration whose primary goal is to identify genes whose variants modify an individual’s risk of type I diabetes (T1D). An integral part of the T1DGC’s mission is the establishment of clinical and data resources that can be used by, and that are fully accessible to, the T1D research community (http://www.t1dgc.org). The T1DGC has organized the collection and analyses of study samples and conducted several major research projects focused on T1D gene discovery: a genome-wide linkage scan, an intensive evaluation of the human major histocompatibility complex, a detailed examination of published candidate genes, and a genome-wide association scan. These studies have provided important information to the scientific community regarding the function of specific genes or chromosomal regions on T1D risk. The results are continually being updated and displayed (http://www.t1dbase.org). The T1DGC welcomes all investigators interested in using these data for scientific endeavors on T1D. The T1DGC resources provide a framework for future research projects, including examination of structural variation, re-sequencing of candidate regions in a search for T1D-associated genes and causal variants, correlation of T1D risk genotypes with biomarkers obtained from T1DGC serum and plasma samples, and in-depth bioinformatics analyses.

Introduction

The Type I Diabetes Genetics Consortium (T1DGC) has performed both candidate gene/region studies and genome-wide explorations for factors that modify risk for type I diabetes (T1D). The candidate gene/region studies have recently taken the format of T1DGC-sponsored workshops. These workshops used the same subset of T1DGC samples (9976 subjects from 2321 affected sib-pair (ASP) families). One workshop focused on the major histocompatibility complex on chromo-some 6p21, containing the HLA genes and numerous other genes related to the immune response. The results of that workshop have been published¹ and the samples and results are available (http://www.t1dgc.org) and the results readily viewed online (http://www.t1dbase.org).² Access to the individual data is available through a number of points of entry (see http://www.t1dgc.org and http://www.t1dbase.org) (Figure 1). The second workshop used the same samples, but focused on published reports that supported specific candidate genes that may affect risk of T1D. An additional set of genes were included in the second workshop that were suggested by the Wellcome Trust Case Control Consortium T1D results,³ as candidates in other autoimmune diseases, and from scans of type II diabetes that identified genes that could affect risk through β-cell function. These genes were subjected to either extensive evaluation of the complete gene using tag single nucleotide polymorphisms (SNPs), or to replication studies of specific published polymorphisms. The results of that workshop are provided in this volume. As expected, the much larger sample size provided by the T1DGC workshop permitted more robust testing of candidate genes, with some genes being replicated and others not.

Figure 1.

Current status of human genes and regions that may predict risk for type I diabetes (T1DBase, 1 January 2009 view; http://www.t1dbase.org).

The T1DGC has also provided genome-wide coverage for detection of T1D susceptibility using both a family design (ASPs) and a case-control design. The family-based approach (linkage) has been considered as a useful method to detect T1D risk genes/regions that have relatively large effect but may be rare in the general population (but important in individual families). The results of the T1DGC suggest that there are few, if any, remaining genes of large effect yet to be discovered beyond the HLA region:^4,5 the region containing CTLA4 on chromosome 2q33, the region containing the insulin gene (INS) on 11p15.5, and a region on chromosome 6q (possibly distinct from HLA on 6p21) assigned earlier as IDDM15, and two regions on chromosome 19 with suggestive evidence of linkage. The third and final installment of ~1600 T1DGC ASP families (6000–7000 DNA samples) will be genotyped following the completion of recruitment, likely in the first quarter of 2009. These data will complete the genome-wide linkage scan conducted by the T1DGC, permitting additional studies to be performed, including the analyses of association of regions and parent-of-origin effects.

Recently, the T1DGC conducted a genome-wide association scan to detect variants that have small effects in the population but are also common in frequency. Earlier genome-wide association scan and candidate gene studies with smaller sample sizes have identified at least 18 regions with evidence of association to T1D with odds ratios between 1.15 and 2.^6-8 The T1DGC meta-analysis of over 7500 cases and over 9000 controls identified 42 distinct genomic locations with P-values <10⁻⁶. Fifteen of these regions were reported earlier, whereas the remaining 26 regions (and one on the X chromosome) were chosen for replication.⁹ Replication genotyping was performed in an additional series of 4267 cases, 4670 controls, and 2319 T1DGC families. Sixteen regions exhibited convincing replication and six of the remaining nine had nominal significance. The data from the T1DGC provide the most accurate description of the complex genetic basis of susceptibility to T1D. Nevertheless, there remain many avenues to explore.

Localization of T1D candidate genes and causal variants

Although a targeted examination of coding regions (exons) represents one approach of gene discovery, it assumes that the causal variant may have an obvious function. Alternatively, the focus on exons ignores potentially important variation in the majority of the human genome. In genome-wide approaches, the most linked or associated common variants (mostly SNPs) with the strongest effects are often intronic or intergenic. This situation often exists with SNPs that also do not cause an amino acid change or likely splicing alteration. Thus, these variants provide no clear indication as to the ‘causal’ gene or ‘causal variant.’ To interrogate fully the region that could harbor a T1D variant, a complete sequence of the region from many individuals is required to catalog the candidate SNPs/other variants for testing in large population samples. The T1DGC provides an outstanding resource and collaborative network for this ‘next generation’ sequencing technology to be applied to highly characterized samples for detection and association analyses of low frequency and common variants, complete with classical HLA genotyping, autoantibody characterization, Epstein-Barr virus (EBV)-transformed cell lines, extensive SNP data, and plasma and serum samples.

Structural variants

The T1DGC has focused its attention on DNA polymorphisms that, with the exception of the complex genotyping of the highly polymorphic HLA loci, typically involve a single base change (SNP). The assembled resources of the T1DGC have permitted both examination of linkage and association with T1D using these singlesite polymorphisms. However, another source of variation in the genome has not yet been assessed by the T1DGC—especially the potential function of structural variants, including copy number variants (CNVs). Recently, a high-resolution genome-wide analysis of a sample of multiplex families with autism was performed to detect CNVs. Screening in the multiplex families permitted the detection of autosomal CNVs that could influence susceptibility to autism with follow-up analyses of multiple replicate populations.¹⁰ Regions in the genome were identified, including 16 that had at least one de novo CNV over-represented in affected children that were not present in either parent, suggesting the possibility of new mutations that could lead to disease. In addition to the report of rare structural variants in autism, both micro-deletions and micro-duplications larger than 100 kb were identified from 150 individuals with schizo-phrenia and 268 ancestry-matched controls.¹¹ Novel deletions and duplications of genes were present in 5% of controls compared with 15% of cases (20% of young-onset cases). These findings suggest that multiple (perhaps) individually rare, structural variants alter genes in neurodevelopmental pathways that contribute to schizophrenia risk. It has been estimated that at least 10% of the genome could be affected by structural variants, including CNVs.¹² The T1DGC has assembled and assessed both ASP families and case–control collections that can be used to examine the function of common structural variation in T1D. The T1DGC collections represent a powerful approach to extract the maximum CNV data possible by transmission analyses within families and by increasing the initial genome-wide dataset for follow-up studies.

Summary

The T1DGC has conducted both genome-wide and candidate studies of large collections of families and individuals (cases and controls) to detect regions of the genome that harbor susceptibility loci. These efforts have resulted in the replication of established T1D risk loci, but have not confirmed other loci claimed earlier. At the same time, many new and intriguing candidate genes have emerged as worthy of further study. The current view of the next steps in the delineation of the genetic architecture of T1D involves re-sequencing and further genotyping of candidate regions, sequencing in ethnically diverse samples to detect both unrecognized common SNPs and rare SNPs, and genotyping in large numbers of well-phenotyped samples of T1D cases, controls, and families to further define risk variants. In addition to the exploration of structural variation on T1D risk, the identification of potential disease variants requires the follow-up studies of gene function, regulation, interactions with other genes (leading to complex systems approaches), and epigenetic mechanisms. Although the T1DGC has accomplished much to identify genes that may modify risk of T1D, much remains to be accomplished to understand the function of these genes and the biology that ultimately will help in the discovery of therapeutic pathways.