Abstract
Objective:
Our aim was to study the influence of type 1 diabetes (T1D) genetic risk factors on the transition through preclinical stages of T1D.
Methods:
In TrialNet participants who have been genotyped with the TEDDY-T1DExomeChip array (Illumina HumanCoreExome Beadarray with custom content), we evaluated the influence of the overall T1D genetic risk score (GRS2), its HLA and non-HLA components, HLA-DR3 and HLA-DR4 haplotypes and 90 single nucleotide polymorphisms previously associated with islet autoimmunity and/or T1D on three transitions between diabetes stages: from single confirmed autoantibody positive to stage 1 (N=4,314), from stage 1 to stage 2 (N=3,066), and from stage 2 to stage 3 (clinical) T1D (N=2,045).
Results:
The T1D GRS2 was associated with all three transitions with hazard ratios(HR) of 1.11(1.09-1.14) for single autoantibody positivity to stage 1, HR 1.05(1.03-1.08) for stage 1 to 2, and HR 1.13(1.09-1.17) for stage 2 to 3 T1D. The T1D GRS2 HLA and HLA class II components were associated with all three transitions. The HLA class I component and the HLA-DR4 haplotype were associated with the transition from single autoantibody positivity to stage 1 and from stage 2 to stage 3 T1D, while HLA-DR3 was only associated with the latter.
Conclusions:
Genetics influence transitions through each stage of preclinical T1D, with main contributions from HLA class II. These results increase our understanding of T1D development and support incorporating the T1D GRS2 to enhance the prediction of progression through the preclinical stages of T1D.
Keywords: early type 1 diabetes markers, genetic risk factors, genetic risk score, prediction of T1D, type 1 diabetes
Introduction
Type 1 diabetes (T1D) is a heterogeneous disease and further refining stages of T1D is needed to apply precision medicine to the prediction and prevention of this disease. While genes coding for variants of Human Leucocyte Antigens (HLA) contribute to about 50% of the overall genetic risk (1,2), more than 50 non-HLA susceptibility single nucleotide polymorphisms (SNPs) have been associated with diabetes risk (3,4). More recently, genetic risk scores (GRS) that combine information from multiple HLA and non-HLA T1D-associated genetic variants have been developed and associated with progression to T1D (5–7). In autoantibody-positive relatives of individuals participating in the Type 1 Diabetes TrialNet Study, the T1D GRS1 was shown to improve the prediction of T1D risk overall as well as the progression from single to multiple autoantibody positivity and from multiple autoantibody positivity to clinical T1D (6). In The Environmental Determinants of Diabetes in the Young (TEDDY), GRS have been able to identify infants without a family history of T1D with a greater than 10% risk for early-stage T1D (7), and a combined risk score incorporating autoantibody status, GRS2, and first-degree relative status could further enhance prediction of T1D among susceptible children over horizons up to 8 years of age (8).
Previous prospective birth cohort studies, such as the Diabetes Autoimmunity Study in the Young (DAISY) and the TEDDY studies, have tested the associations of non-HLA SNPs with the onset of islet autoimmunity (IA) and the progression from IA to T1D. In the DAISY study, PTPN22 rs2476601 and two UBASH3A SNPs (rs11203203 and rs9976767) were associated with development of IA, while SNPs in INS (rs689), UBASH3A (rs11203203), and IFIH1 (rs1990760) were significantly associated with progression from IA to T1D after adjusting for family history and HLA-DR3/4 genotype (9,10). In TEDDY participants carrying high-risk HLA genotypes, eight SNPs were significantly associated with the development of IA using time-to-event analysis (P < 0.05), with four remaining significant after adjustment for multiple testing, including rs2476601 (PTPN22), rs2292239 (ERBB3), rs3184504 (SH2B3), and rs1004446 (INS) (11).
These studies have shown that genetic factors involved in the initiation of IA differ from those identified once the disease has been established. However, the influence of genetic factors through the preclinical stages of T1D has not been studied. Type 1 diabetes stages were proposed in 2015 (12) with stage 1 defined by the presence of multiple islet autoantibodies, followed by abnormal glucose tolerance (stage 2) before stage 3 (clinical) diabetes. The T1D stages are widely used to describe the preclinical progression of T1D, and to determine eligibility for clinical trials (13–15) and treatments to prevent or delay the onset of clinical T1D (16). Therefore, further investigation is needed to evaluate the genetic risk factors that modulate the progression of subclinical disease through the different stages of T1D. The aim of the current study was to evaluate the influence of the T1D GRS2 (17), the HLA region and SNPs in non-HLA loci previously associated with IA and/or T1D on the stages of diabetes to enhance the prediction through the preclinical stages of T1D.
Materials and Methods
Study Population and Design
The TrialNet Pathway to Prevention study screens for the presence of islet autoantibodies and follows relatives of individuals with T1D at increased risk of diabetes as previously described (18). All subjects are tested for autoantibodies to glutamic acid decarboxylase (GADA), insulin (IAA), and insulinoma-associated antigen 2 (IA-2A); if any of these are detected, autoantibodies to zinc transporter 8 (ZnT8A) and islet cell antibodies (ICA) are measured as well. This study included TrialNet Pathway to Prevention monitoring participants and participants in any of the TrialNet prevention trials who have been genotyped with the “TEDDY-T1DExomeChip” (Illumina HumanCoreExome Beadarray with custom content of ~90K SNPs) with the exception of participants who received teplizumab in the TN10 trial (16), who were excluded from this analysis because of a significant effect of teplizumab on the natural history of T1D. In the monitoring phase, participants who are positive for two or more islet autoantibodies undergo metabolic assessments to determine glycemic status and T1D staging (12), and are eligible for prevention trials. TrialNet prevention trials included the following: Oral Insulin for Prevention of Diabetes in Relatives At Risk for Type 1 Diabetes Mellitus (TN07 trial), Anti-CD3 Monoclonal Antibody (Teplizumab) for Prevention of Diabetes in Relatives At-Risk for Type 1 Diabetes Mellitus (TN10), CTLA-4 Ig (Abatacept) for Prevention of Abnormal Glucose Tolerance and Diabetes in Relatives At-Risk for Type 1 Diabetes Mellitus (TN18 trial), Exploring Immune Effects of Oral Insulin in Relatives at Risk for Type 1 Diabetes Mellitus (TN20) and Hydroxychloroquine for Prevention of Abnormal Glucose Tolerance and Diabetes in Individuals At-Risk for Type 1 Diabetes Mellitus (TN22) (13–16).
In this study, we evaluated the influence of the overall T1D GRS2 and its components, including its HLA, non-HLA, HLA Class I and HLA Class II components, and the HLA-DR3 and HLA-DR4 haplotypes. In addition, we studied 90 SNPs at the individual level, including 67 SNPs from the T1D GRS2 (17) and 23 pre-specified SNPs that prospective studies, such as DAISY and TEDDY, have associated with the development or progression of islet autoimmunity and T1D. Table 1 lists the 23 pre-specified SNPs, their locations, and effect allele frequencies. We assessed the association of these genetic factors with transitions from single confirmed autoantibody positivity to stage 1 T1D (N=4,314), from stage 1 to stage 2 T1D (N=3,066) and from stage 2 to stage 3 (clinical) T1D (N=2,045). All study participants gave informed consent, and the study was approved by the ethics committee at each study site.
Table 1:
List of pre-specified single nucleotide polymorphisms (SNPs) not included in GRS2
| SNP | Gene | Chromosome | Position | Alleles | MAF |
|---|---|---|---|---|---|
| rs111776337 | MRPS21-PRPF3 | 1 | 150319828 | C>T | 0.048 |
| rs1990760 | IFIH1 | 2 | 162267541 | T>C | 0.389 |
| rs1534422 | MIR3681HG | 2 | 12500615 | A>G | 0.469 |
| rs55900661 | NRIR | 2 | 6808849 | G>A | 0.055 |
| rs77967786 | COL6A6 | 3 | 130594375 | G>A | 0.034 |
| rs11705721 | PXK/PDHB | 3 | 58414687 | T>C | 0.353 |
| rs10517086 | LINC02357/RBPJ | 4 | 26083889 | G>A | 0.303 |
| rs2327832 | TNFAIP3 | 6 | 137651931 | A>G | 0.204 |
| rs73043122 | RNASET2 | 6 | 166969779 | G>C | 0.018 |
| rs7804356 | SKAP2 | 7 | 26852046 | T>C | 0.224 |
| rs7020673 | GLIS3 | 9 | 4291747 | G>C | 0.457 |
| rs113306148 | PLEKHA1 | 10 | 122400322 | C>T | 0.012 |
| rs12251307 | IL2RA | 10 | 6081532 | C>T | 0.108 |
| rs689 | INS | 11 | 2160994 | T>A | 0.201 |
| rs1004446 | INS | 11 | 2148913 | G>A | 0.323 |
| rs3184504 | SH2B3 | 12 | 111446804 | C>T | 0.491 |
| rs2292239 | ERBB3 | 12 | 56088396 | G>T | 0.378 |
| rs3825932 | CTSH | 15 | 78943104 | C>T | 0.346 |
| rs4788084 | IL27 | 16 | 28528527 | C>T | 0.385 |
| rs763361 | CD226 | 18 | 69864406 | C>T | 0.495 |
| rs12151883 | TASP1 | 20 | 13394404 | T>A | 0.050 |
| rs3788013 | UBASH3A | 21 | 42421219 | C>A | 0.457 |
| rs11203203 | UBASH3A | 21 | 42416077 | G>A | 0.389 |
MAF: minor allele frequency based on the single confirmed autoantibody positive participants (N = 4,314)
Position: Based on the Genome Reference Consortium Human Build 38
Genotyping
The genotyping was performed using the TTEDDY-T1DExomeChip (an Illumina HumanCoreExome Beadarray with ~90,000 SNPs as additional custom content, selected from regions of the genome robustly associated with autoimmune diseases and pathways relevant to T1D initiation and progression), as previously described (19). The T1D GRS2 (17) includes 67 SNPs, 30 of which were directly genotyped. Hence, we imputed 32 SNPs with a median R2 of 0.997 (min = 0.858 and max = 0.999) using the TOPMed Imputation Server with the multiethnic TOPMed reference panel, which includes 97,256 reference samples and > 308 million genetic variants. We also imputed 5 SNPs in the HLA region (rs72848653 (R2 = 0.999), rs9266268 (R2 = 0.999), rs16899379 (R2 = 0.998), rs2524277 (R2 = 0.995), and rs9268500 (R2 = 0.925)) using the Michigan Imputation Server with the high-resolution HLA reference panel spanning five global populations (n = 21,546) based on whole-genome sequencing data. Code to generate the HLA interaction part of the T1D GRS2 is freely available online (https://github.com/sethsh7/PRSedm).
Statistical Analysis
Cox proportional hazard models were used to test the effect of each T1D-GRS2 score, score components, HLA haplotype or SNP on time from single autoantibody positivity to stage 1, from stage 1 to stage 2, and from stage 2 to stage 3 T1D. Each SNP in the model was defined according to the number of risk alleles present (0, 1, or 2) and was treated as a continuous variable in the model. HLA DRB1*0301, DQA1*0501, DQB1*0201 (HLA-DR3) and HLA DRB1*04, DQA1*0301 or 0303, DQB1*0302 (HLA-DR4) haplotypes were categorized as present/absent. Besides the overall T1D GRS2, the following components of the T1D GRS2 were analyzed separately: HLA Class I (21 SNPs), HLA Class II (14 SNPs), HLA (combining HLA Class I and II) and non-HLA (32 SNPs). Both unadjusted and adjusted analyses for age at baseline, sex, and the first 3 principal components estimated from the TEDDY-T1DExomeChip array to account for population stratification (ancestral heterogeneity) were performed.
The predictive accuracy of the models for time to progression for the various stages was evaluated for GRS2 using the Inverse Probability of Censoring Weighting (IPCW) estimation of the time-dependent receiver operating characteristic (ROC) curve and the corresponding area under the curve (AUC) with the censoring by the additive Aalen model of timeROC package in R. Time horizons were calculated for a range of 5 years. The optimal cutoff points for the T1D GRS2 were identified using the Youden index with the Youden function of the “cenROC” R package (i.e., cut-point that achieves the maximum of the sum of the sensitivity + specificity - 1) (20). The Kaplan-Meier survival curves were generated to examine the effects of the T1D GRS2 on the transition from single autoantibody positivity to stage 1, stage 1 to stage 2, and stage 2 to stage 3 T1D. Differences between the curves were assessed with log-rank tests, with P < 0.05 considered statistically significant. To correct for multiple hypothesis testing, the false discovery rate (FDR) adjusted p-values were calculated. . All the analyses were performed using the R version 4.4.0 (“timeROC”, “cenROC”, “survival”, and “survminer” packages).
Results
The characteristics of study participants in each of the three sets of analyses are shown in Table 2. At baseline, the mean age (± SD) for participants who were confirmed single autoantibody positive, at stages 1 or 2 were, respectively, 14.2 (± 12.1), 12.3 (± 10.3), and 13.5 (± 10.3) years of age. About 50% of the participants were male in the confirmed single autoantibody positive, while the proportion of male was slightly higher (54-55%) in those at stage 1 and stage 2. The majority of the participants were White (87-88%) and non-Hispanic (86-87%).
Table 2:
Characteristics of autoantibody positive participants by type 1 diabetes stage at baseline
| Single antibody positive | Stage 1 | Stage 2 | |
|---|---|---|---|
|
| |||
| N | 4,314 | 3,066 | 2,045 |
|
| |||
| Baseline age (yrs), mean (SD) | 14.2 (± 12.1) | 12.3 (± 10.3) | 13.5 (± 10.3) |
|
| |||
| Sex, N (%) | |||
| Male | 2,184 (50.6) | 1,682 (54.9) | 1,111 (54.3) |
| Female | 2,128 (49.3) | 1,383 (45.1) | 934 (45.7) |
| Missing | 2 (0.1) | 1 (0) | 0 (0) |
|
| |||
| Race, N (%) | |||
| White | 3,734 (86.6) | 2,686 (87.6) | 1,801 (88.1) |
| African American | 109 (2.5) | 84 (2.7) | 48 (2.3) |
| Other | 162 (3.8) | 101 (3.3) | 72 (3.4) |
| Missing | 309 (7.1) | 195 (6.4) | 124 (6.2) |
|
| |||
| Ethnicity, N (%) | |||
| Non-Hispanic | 3,708 (86) | 2,666 (87) | 1,773 (86.7) |
| Hispanic | 416 (9.6) | 253 (8.3) | 172 (8.4) |
| Missing | 190 (4.4) | 147 (4.7) | 100 (4.9) |
|
| |||
| GRS2, mean (SD) | 13.4 (± 2.1) | 13.8 (± 1.9) | 13.9 (± 1.8) |
|
| |||
| Follow-up (yrs), mean (SD) | 0.97 (± 1.68) | 1.6 (± 2.2) | 2.4 (± 2.6) |
We tested the effect of the overall T1D GRS2, its HLA, HLA Class I, HLA Class II, and non-HLA components, HLA-DR3 and HLA-DR4 haplotypes and 90 SNPs (previously associated with IA and/or T1D) on time to development from single autoantibody positivity to stage 1, from stage 1 to stage 2, and from stage 2 to stage 3 T1D. Table 3 shows genetic associations from single autoantibody positivity to stage 1 T1D, Table 4 shows genetic associations from stage 1 to stage 2 T1D, and Table 5 shows genetic associations from stage 2 to stage 3 T1D. Both unadjusted and adjusted analyses for age at baseline, sex, and the top 3 principal components of genetic ancestry were calculated for each three stages. The T1D GRS2 was significantly associated with all three transitions with hazard ratio (HR) 1.11 (1.09-1.14) for single autoantibody positivity to stage 1, HR 1.05 (1.03-1.08) for stage 1 to stage 2, and HR 1.13 (1.09-1.17) for stage 2 to stage 3 T1D. The HLA component of the T1D GRS2 was significantly associated with all three transitions with HR 1.11 (1.08-1.14) for single autoantibody positivity to stage 1, HR 1.05 (1.02-1.07) for stage 1 to stage 2, and HR 1.13 (1.09-1.18) for stage 2 to stage 3 T1D. The HLA class II component of the T1D GRS2 was also significantly associated with all three transitions with HR 1.10 (1.07-1.13) for single autoantibody positivity to stage 1, HR 1.04 (1.02-1.07) for stage 1 to stage 2, and HR 1.11 (1.07-1.15) for stage 2 to stage 3 T1D. The HLA class I component of the T1D GRS2 was only associated with transition from single autoantibody positivity to stage 1 (HR 1.10 (1.05-1.16)) and with transition from stage 2 to stage 3 T1D (HR 1.11 (1.04-1.18)), while the non-HLA component of the T1D GRS2 was only associated with transition from single autoantibody positivity to stage 1 (HR 1.09 (1.03-1.16)).
Table 3:
Genetic associations from single autoantibody positivity to stage 1 type 1 diabetes
| SNP | Gene | Allele | Hazard Ratio (95% CI) | FDR corrected p values | Adjusted Hazard Ratio (95% CI) | FDR corrected p values |
|---|---|---|---|---|---|---|
| GRS2 | 1.11 (1.09-1.14) | 1.88 x 10−17 | 1.09 (1.07-1.12) | 1.42 x 10−11 | ||
| HLA score | 1.11 (1.08-1.14) | 1.91 x 10−15 | 1.09 (1.06-1.11) | 1.29 x 10−9 | ||
| HLA class II | 1.10 (1.07-1.13) | 2.37 x 10−11 | 1.08 (1.05-1.11) | 9.26 x 10−7 | ||
| HLA-DR4 | 1.42 (1.29-1.57) | 7.95 x 10−11 | 1.36 (1.23-1.50) | 4.26 X 10−8 | ||
| rs9275490 | HLA-DQ81 | C>G | 1.28 (1.19-1.37) | 3.88 x 10−10 | 1.24 (1.16-1.34) | 6.82 x 10−8 |
| rs9271347 | HLA | G>A | 0.52 (0.40-0.66) | 2.98 x 10−6 | 0.57 (0.44-0.73) | 0.0001 |
| rs9273032 | HLA-DQ62 | G>A | 0.54 (0.42-0.69) | 8.92 x 10−6 | 0.59 (0.46-0.75) | 0.0003 |
| HLA class I | 1.10 (1.05-1.16) | 0.0004 | 1.09 (1.04-1.14) | 0.0036 | ||
| rs72848653 | HLA | C>T | 0.82 (0.72-0.92) | 0.0088 | 0.81 (0.72-0.91) | 0.0058 |
| rs9273369 | HLA-DQ25 | T>C | 0.89 (0.82-0.96) | 0.0226 | 0.87 (0.80-0.94) | 0.0036 |
| Non-HLA score | 1.09 (1.03-1.16) | 0.0256 | 1.09 (1.03-1.15) | 0.0396 | ||
| rs9469200 | HLA-DQ75 | T>C | 0.78 (0.66-0.93) | 0.0397 | 0.80 (0.67-0.95) | 0.0603 |
| rs6934289 | HLA | T>C | 0.83 (0.72-0.95) | 0.0475 | 0.83 (0.72-0.95) | 0.0512 |
| rs1281935 | HLA-DQ73 | G>T | 1.17 (1.04-1.32) | 0.0615 | 1.18 (1.04-1.33) | 0.0409 |
| rs10517086 | LINC02357/RBPJ | G>A | 0.91 (0.85-0.98) | 0.0769 | 0.90 (0.84-0.97) | 0.0423 |
FDR: false discovery rate
Only showing results where FDR corrected p value < 0.05
Table 4:
Genetic associations from stage 1 to stage 2 type 1 diabetes
| SNP | Gene | Allele | Hazard Ratio (95% CI) | FDR corrected p value | Adjusted Hazard Ratio (95% CI) | FDR corrected p value |
|---|---|---|---|---|---|---|
| GRS2 | 1.05 (1.03-1.08) | 0.0025 | 1.05 (1.03-1.08) | 0.0047 | ||
| HLA score | 1.05 (1.02-1.07) | 0.0438 | 1.04 (1.02-1.07) | 0.0718 | ||
| rs3788013 | UBASH3A | C>A | 0.91 (0.85-0.96) | 0.0438 | 0.91 (0.86-0.97) | 0.0860 |
| HLA class II | 1.04 (1.02-1.07) | 0.0438 | 1.04 (1.01-1.07) | 0.0867 | ||
| rs3087243 | CTLA4 | G>A | 0.90 (0.85-0.96) | 0.0438 | 0.91 (0.85-0.97) | 0.0860 |
| rs2289702 | CTSH | C>T | 0.84 (0.76-0.94) | 0.0438 | 0.84 (0.76-0.94) | 0.0860 |
FDR: false discovery rate
Only showing results where FDR corrected p value < 0.05
Table 5:
Genetic associations from stage 2 to stage 3 type 1 diabetes
| SNP | Gene | Allele | Hazard Ratio (95% CI) | FDR corrected p value | Adjusted Hazard Ratio (95% CI) | FDR corrected p value |
|---|---|---|---|---|---|---|
| GRS2 | 1.13 (1.09-1.17) | 2.03 x 10−9 | 1.12 (1.08-1.17) | 1.22 x 10−8 | ||
| HLA score | 1.13 (1.09-1.18) | 2.03 x 10−9 | 1.12 (1.08-1.17) | 5.80 x 10−8 | ||
| HLA class II | 1.11 (1.07-1.15) | 4.87 x 10−6 | 1.10 (1.06-1.15) | 2.15 x 10−5 | ||
| rs9269173 | HLA | T>A | 1.45 (1.25-1.69) | 3.15 x 10−5 | 1.39 (1.19-1.62) | 0.0007 |
| HLA-DR3 | 1.27 (1.12-1.43) | 0.0021 | 1.23 (1.09-1.38) | 0.0197 | ||
| rs9273369 | HLA-DQ25 | T>C | 1.17 (1.06-1.29) | 0.0251 | 1.14 (1.04-1.26) | 0.0717 |
| rs72848653 | HLA | C>T | 1.26 (1.09-1.45) | 0.0251 | 1.23 (1.07-1.43) | 0.0551 |
| HLA class I | 1.11 (1.04-1.18) | 0.0316 | 1.09 (1.02-1.16) | 0.1167 | ||
| HLA-DR4 | 1.21 (1.06-1.38) | 0.0467 | 1.20 (1.05-1.37) | 0.0670 | ||
| rs9469200 | HLA-DQ75 | T>C | 0.74 (0.60-0.93) | 0.0751 | 0.72 (0.58-0.90) | 0.0483 |
| rs72727394 | RASGRP1 | C>T | 1.14 (1.03-1.26) | 0.0764 | 1.16 (1.05-1.28) | 0.0483 |
FDR: false discovery rate
Only showing results where FDR corrected p value < 0.05
Tables 3, 4 and 5 show all SNPs, HLA haplotypes and genetic scores with FDR corrected p-values < 0.05. The HLA-DR4 haplotype was significantly associated with transitions from single autoantibody positivity to stage 1 (HR 1.42 (1.29-1.57)) and from stage 2 to stage 3 T1D (HR 1.21 (1.06-1.38)). The HLA-DR3 haplotype was only significantly associated with the transition from stage 2 to stage 3 T1D (HR 1.27 (1.12-1.43)). Among the individual HLA SNPs, rs9275490, rs9271347, rs9273032, rs72848653 and rs9273369 were significantly associated with the transition from single autoantibody positivity to stage 1 T1D (Table 3), and rs9269173 was significantly associated with the transition from stage 2 to stage 3 T1D. For the individual non-HLA SNPs, none of the borderline significant results from stage 1 to stage 2 T1D in unadjusted analyses for rs3788013 (UBASH3A), rs3087243 (CTLA4) and rs2289702 (CTSH) stayed significant in the adjusted analyses. Results for adjusted analyses were otherwise similar to the unadjusted results.
The predictive accuracy of the models for time to progression for each transition was evaluated for GRS2. At a time horizon of 2 years, the T1D GRS2 performed with AUC of 57.45%, 54.51% and 57.47%, respectively for the transitions from single autoantibody positive to stage 1,stage 1 to stage 2, and stage 2 to stage 3 T1D. The optimal cutoff points for T1D GRS2 were 12.80, 13.85, and 14.03, respectively, for the transitions from single positive to stage 1, stage 1 to stage 2, and stage 2 to stage 3, with the corresponding Youden indices being, respectively, 0.11, 0.07, and 0.11. Kaplan-Meier curves were generated to examine the effect of the T1D GRS2 on each of the transitions under study. The risk of progression to stage 1 in those initially identified with a single positive autoantibody was 47.4% vs 34.3% at 3 years, for those with GRS2 ≥12.80 vs <12.80 respectively (p < 0.0001, Figure 1A). The risk of progression from stage 1 to stage 2 T1D was 70.6% vs 62.1% at 3 years, for those with GRS2 ≥13.85 vs <13.85 respectively (p = 0.0074, Figure 1B). The risk of progression from stage 2 to stage 3 T1D in 3 years was 59.6% in those with GRS2 ≥14.03; in contrast, it was only 47.9% in participants with GRS2 <14.03 (p < 0.0001, Figure 1C).
Figure 1: Kaplan Meier survival curves of the GRS2 on the transition from single autoantibody positivity to stage 1 (A), stage 1 to stage 2 (B), and stage 2 to stage 3 (C).
Single Ab+ = single confirmed autoantibody positive
Discussion
To our knowledge, this is the first and largest prospective study to date to analyze the influence of the overall T1D GRS2, its HLA and non-HLA components, HLA-DR3 and HLA-DR4 haplotypes and SNPs previously associated with IA and/or T1D on the transition through the preclinical stages of T1D. We found that, in autoantibody-positive relatives of individuals with T1D, the T1D GRS2 was significantly associated with each of the three transitions, i.e., from single autoantibody positivity to stage 1, from stage 1 to stage 2, and from stage 2 to stage 3 T1D. Furthermore, the T1D GRS2 HLA and HLA class II components were associated with all three transitions. This data supports the additional contribution of genetics in refining the progression of T1D through the stages in relatives at increased risk.
A previous study evaluated the utility of the T1D GRS1 in 1,244 TrialNet Pathway to Prevention study participants and found that a higher T1D GRS1 was predictive of progression from single to multiple positive autoantibodies and from multiple autoantibodies to T1D (6). However, this study was limited by a relatively small number of participants, and an initial version of the T1D GRS. While the initial T1D GRS1 is based on 30 SNPs (21), the T1D GRS2 includes 67 SNPs (17) and better captures the complex interactions between the HLA DR-DQ haplotypes, assigning weights correspondingly in case of interactions. Importantly, previous studies did not consider the transition between stages of T1D, which are widely used today to classify the preclinical status of individuals and determine eligibility for clinical trials and treatments to prevent progression to clinical T1D (13–16).
We observed a significant although relatively modest effect of the T1D GRS2 on all stages of T1D, underscoring that genetic factors continue to play a role in later stages of T1D progression. A significant AUC of 0.57 for the T1D-GRS2 indicates that there is an association between the T1D GRS2 and disease progression. However, it also suggests that the T1D GRS2 alone is insufficient for predicting T1D and it needs to be considered alongside other risk factors. This effect seems to be mainly captured by SNPs in the HLA component of the T1D GRS2 and more specifically the HLA class II subcomponent. This is consistent with the well-known contribution of HLA class II genes to T1D risk (2,22,23). However, it is interesting to note that this effect is not only significant for the transition from single autoantibody positivity to stage 1 T1D but for the transition through all three preclinical stages of T1D, including when dysglycemia is present. In particular, the HLA-DR3 haplotype was only associated with transition from stage 2 to stage 3 T1D. In contrast, the HLA-DR4 haplotype had a significant effect on transition from single autoantibody positivity to stage 1 and from stage 2 to stage 3 T1D. The TEDDY study found that the HLA DR4 haplotype predicted seroconversion to IAA-first (i.e. insulin as the first appearing autoantibody) as well as progression to multiple autoantibodies but not to clinical diabetes (24).
Currently, there are numerous screening efforts for T1D underway both for relatives of individuals with T1D as well as for the general population. To date, all screening strategies include measurement of islet autoantibodies, with some of them initially using T1D GRS to identify infants at risk, such as PLEDGE (general population screening program through Sanford Health) and CASCADE (Newborn Screening for type 1 diabetes and celiac disease in Washington State). Our study supports that genetic factors also influence the progression through subsequent preclinical stages of T1D, and T1D GRS2 along with specific SNPs could be used to refine risk prediction. Incorporating genetic risk predictions can help tailor intervention studies for individuals with T1D; indeed some trials include high-risk HLA haplotypes as inclusion criteria, such as the methyldopa trial in HLA-DR4 positive individuals (25) and GAD GAD-alum trial in new onset HLA-DR3 positive individuals (DIAGNODE-3, NCT05018585). It is possible that T1D GRS or specific SNPs could enhance the prediction of response to disease modifying therapies in T1D, but this remains to be tested.
The limitations of this study include that participants are not followed since birth and, therefore, the time of seroconversion is often unknown and early age transitions are under-represented in TrialNet. However, this characteristic of TrialNet enhances generalizability to the general population, where screening is offered to individuals with unknown prior history of islet autoimmunity. It has to be noted that TrialNet Pathway to Prevention enrolls relatives of individuals with T1D, and thus, validation studies in the general population are needed. Analyses were limited to participants with at least one confirmed positive autoantibody. Furthermore, the majority of TrialNet participants are non-Hispanic white, and further studies are needed to see whether these findings apply to other races and ethnicities. As the SNPs from the T1D GRS2 and the other pre-specified SNPs are based upon initial findings in previous mainly non-Hispanic white studies, it is possible that other HLA haplotypes and non-HLA risk variants contribute to these transitions.
In conclusion, this is the first large prospective study analyzing the influence of the T1D GRS2, its HLA and non-HLA components, HLA Classes I and II, HLA-DR3 and HLA-DR4 haplotypes and SNPs known to be associated with IA and/or T1D risk on progression through the preclinical stages of T1D. With the increased availability of genetic data and the decreased costs of genotyping, it may be warranted to include a T1D GRS to further refine T1D risk prediction in the early stages of T1D. Our study supports that the use of genetics facilitates a personalized medicine approach for individuals with early-stage T1D.
Acknowledgments
This study was supported by the National Institutes of Health (NIH) R01 DK121843. The sponsor of the study was the Type 1 Diabetes TrialNet Pathway to Prevention Study Group. The Type 1 Diabetes TrialNet Study Group is a clinical trials network currently funded by the National Institutes of Health (NIH) through the National Institute of Diabetes and Digestive and Kidney Diseases, the National Institute of Allergy and Infectious Diseases, and The Eunice Kennedy Shriver National Institute of Child Health and Human Development, through the cooperative agreements U01 DK061010, U01 DK061034, U01 DK061042, U01 DK061058, U01 DK085461, U01 DK085465, U01 DK085466, U01 DK085476, U01 DK085499, U01 DK085509, U01 DK103180, U01 DK103153, U01 DK103266, U01 DK103282, U01 DK106984, U01 DK106994, U01 DK107013, U01 DK107014, U01 DK106993, UC4 DK117009, and Breakthrough T1D (formerly JDRF).
Footnotes
Disclosure summary: The authors have no relevant conflicts of interest to disclose.
Data Availability
Original data generated and analyzed during this study are included in this published article or in the data repositories listed in References.
References
- 1.Parkes M, Cortes A, van Heel DA, Brown MA. Genetic insights into common pathways and complex relationships among immune-mediated diseases. Nat Rev Genet. 2013;14(9):661–673. [DOI] [PubMed] [Google Scholar]
- 2.Redondo MJ, Steck AK, Pugliese A. Genetics of type 1 diabetes. Pediatr Diabetes. 2018;19(3):346–353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Burren OS, Adlem EC, Achuthan P, Christensen M, Coulson RM, Todd JA. T1DBase: update 2011, organization and presentation of large-scale data sets for type 1 diabetes research. Nucleic Acids Res. 2011;39(Database issue):D997–1001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Barrett JC, Clayton DG, Concannon P, Akolkar B, Cooper JD, Erlich HA, Julier C, Morahan G, Nerup J, Nierras C, Plagnol V, Pociot F, Schuilenburg H, Smyth DJ, Stevens H, Todd JA, Walker NM, Rich SS. Genome-wide association study and meta-analysis find that over 40 loci affect risk of type 1 diabetes. Nat Genet. 2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Winkler C, Krumsiek J, Buettner F, Angermuller C, Giannopoulou EZ, Theis FJ, Ziegler AG, Bonifacio E. Feature ranking of type 1 diabetes susceptibility genes improves prediction of type 1 diabetes. Diabetologia. 2014;57(12):2521–2529. [DOI] [PubMed] [Google Scholar]
- 6.Redondo MJ, Geyer S, Steck AK, Sharp S, Wentworth JM, Weedon MN, Antinozzi P, Sosenko J, Atkinson M, Pugliese A, Oram RA, Type 1 Diabetes TrialNet Study G. A Type 1 Diabetes Genetic Risk Score Predicts Progression of Islet Autoimmunity and Development of Type 1 Diabetes in Individuals at Risk. Diabetes Care. 2018;41(9):1887–1894. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bonifacio E, Beyerlein A, Hippich M, Winkler C, Vehik K, Weedon MN, Laimighofer M, Hattersley AT, Krumsiek J, Frohnert BI, Steck AK, Hagopian WA, Krischer JP, Lernmark A, Rewers MJ, She JX, Toppari J, Akolkar B, Oram RA, Rich SS, Ziegler AG, Group TS. Genetic scores to stratify risk of developing multiple islet autoantibodies and type 1 diabetes: A prospective study in children. PLoS Med. 2018;15(4):e1002548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ferrat LA, Vehik K, Sharp SA, Lernmark A, Rewers MJ, She JX, Ziegler AG, Toppari J, Akolkar B, Krischer JP, Weedon MN, Oram RA, Hagopian WA, Group TS. A combined risk score enhances prediction of type 1 diabetes among susceptible children. Nat Med. 2020;26(8):1247–1255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Steck AK, Wong R, Wagner B, Johnson K, Liu E, Romanos J, Wijmenga C, Norris JM, Eisenbarth GS, Rewers MJ. Effects of non-HLA gene polymorphisms on development of islet autoimmunity and type 1 diabetes in a population with high-risk HLA-DR,DQ genotypes. Diabetes. 2012;61(3):753–758. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Steck AK, Dong F, Wong R, Fouts A, Liu E, Romanos J, Wijmenga C, Norris JM, Rewers MJ. Improving prediction of type 1 diabetes by testing non-HLA genetic variants in addition to HLA markers. Pediatr Diabetes. 2014;15(5):355–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Torn C, Hadley D, Lee HS, Hagopian W, Lernmark A, Simell O, Rewers M, Ziegler A, Schatz D, Akolkar B, Onengut-Gumuscu S, Chen WM, Toppari J, Mykkanen J, Ilonen J, Rich SS, She JX, Steck AK, Krischer J, Group TS. Role of Type 1 Diabetes-Associated SNPs on Risk of Autoantibody Positivity in the TEDDY Study. Diabetes. 2015;64(5):1818–1829. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Insel RA, Dunne JL, Atkinson MA, Chiang JL, Dabelea D, Gottlieb PA, Greenbaum CJ, Herold KC, Krischer JP, Lernmark A, Ratner RE, Rewers MJ, Schatz DA, Skyler JS, Sosenko JM, Ziegler AG. Staging Presymptomatic Type 1 Diabetes: A Scientific Statement of JDRF, the Endocrine Society, and the American Diabetes Association. Diabetes Care. 2015;38(10):1964–1974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Writing Committee for the Type 1 Diabetes TrialNet Oral Insulin Study G, Krischer JP, Schatz DA, Bundy B, Skyler JS, Greenbaum CJ. Effect of Oral Insulin on Prevention of Diabetes in Relatives of Patients With Type 1 Diabetes: A Randomized Clinical Trial. JAMA. 2017;318(19):1891–1902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Russell WE, Bundy BN, Anderson MS, Cooney LA, Gitelman SE, Goland RS, Gottlieb PA, Greenbaum CJ, Haller MJ, Krischer JP, Libman IM, Linsley PS, Long SA, Lord SM, Moore DJ, Moore WV, Moran AM, Muir AB, Raskin P, Skyler JS, Wentworth JM, Wherrett DK, Wilson DM, Ziegler AG, Herold KC, Type 1 Diabetes TrialNet Study G. Abatacept for Delay of Type 1 Diabetes Progression in Stage 1 Relatives at Risk: A Randomized, Double-Masked, Controlled Trial. Diabetes Care. 2023;46(5):1005–1013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Libman I, Bingley PJ, Becker D, Buckner JH, DiMeglio LA, Gitelman SE, Greenbaum C, Haller MJ, Ismail HM, Krischer J, Moore WV, Moran A, Muir AB, Raman V, Steck AK, Toledo FGS, Wentworth J, Wherrett D, White P, You L, Herold KC, Type 1 Diabetes TrialNet Study G. Hydroxychloroquine in Stage 1 Type 1 Diabetes. Diabetes Care. 2023;46(11):2035–2043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Herold KC, Bundy BN, Long SA, Bluestone JA, DiMeglio LA, Dufort MJ, Gitelman SE, Gottlieb PA, Krischer JP, Linsley PS, Marks JB, Moore W, Moran A, Rodriguez H, Russell WE, Schatz D, Skyler JS, Tsalikian E, Wherrett DK, Ziegler AG, Greenbaum CJ, Type 1 Diabetes TrialNet Study G. An Anti-CD3 Antibody, Teplizumab, in Relatives at Risk for Type 1 Diabetes. N Engl J Med. 2019;381(7):603–613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sharp SA, Rich SS, Wood AR, Jones SE, Beaumont RN, Harrison JW, Schneider DA, Locke JM, Tyrrell J, Weedon MN, Hagopian WA, Oram RA. Development and Standardization of an Improved Type 1 Diabetes Genetic Risk Score for Use in Newborn Screening and Incident Diagnosis. Diabetes Care. 2019;42(2):200–207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Mahon JL, Sosenko JM, Rafkin-Mervis L, Krause-Steinrauf H, Lachin JM, Thompson C, Bingley PJ, Bonifacio E, Palmer JP, Eisenbarth GS, Wolfsdorf J, Skyler JS. The TrialNet Natural History Study of the Development of Type 1 Diabetes: objectives, design, and initial results. Pediatr Diabetes. 2009;10(2):97–104. [DOI] [PubMed] [Google Scholar]
- 19.Triolo TM, Parikh HM, Tosur M, Ferrat L, You L, Gottlieb PA, Oram RA, Onengut-Gumuscu S, Krischer JP, Rich SS, Steck AK, Redondo MJ. Genetic Associations with C-peptide Levels before Type 1 Diabetes Diagnosis in At-Risk Relatives. J Clin Endocrinol Metab. 2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ruopp MD, Perkins NJ, Whitcomb BW, Schisterman EF. Youden Index and optimal cut-point estimated from observations affected by a lower limit of detection. Biom J. 2008;50(3):419–430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Oram RA, Patel K, Hill A, Shields B, McDonald TJ, Jones A, Hattersley AT, Weedon MN. A Type 1 Diabetes Genetic Risk Score Can Aid Discrimination Between Type 1 and Type 2 Diabetes in Young Adults. Diabetes Care. 2016;39(3):337–344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Noble JA, Erlich HA. Genetics of type 1 diabetes. Cold Spring Harb Perspect Med. 2012;2(1):a007732. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Thomson G, Valdes AM, Noble JA, Kockum I, Grote MN, Najman J, Erlich HA, Cucca F, Pugliese A, Steenkiste A, Dorman JS, Caillat-Zucman S, Hermann R, Ilonen J, Lambert AP, Bingley PJ, Gillespie KM, Lernmark A, Sanjeevi CB, Ronningen KS, Undlien DE, Thorsby E, Petrone A, Buzzetti R, Koeleman BP, Roep BO, Saruhan-Direskeneli G, Uyar FA, Gunoz H, Gorodezky C, Alaez C, Boehm BO, Mlynarski W, Ikegami H, Berrino M, Fasano ME, Dametto E, Israel S, Brautbar C, Santiago-Cortes A, Frazer de Llado T, She JX, Bugawan TL, Rotter JI, Raffel L, Zeidler A, Leyva-Cobian F, Hawkins BR, Chan SH, Castano L, Pociot F, Nerup J. Relative predispositional effects of HLA class II DRB1-DQB1 haplotypes and genotypes on type 1 diabetes: a meta-analysis. Tissue Antigens. 2007;70(2):110–127. [DOI] [PubMed] [Google Scholar]
- 24.Krischer JP, Liu X, Lernmark A, Hagopian WA, Rewers MJ, She JX, Toppari J, Ziegler AG, Akolkar B, Group TS. Predictors of the Initiation of Islet Autoimmunity and Progression to Multiple Autoantibodies and Clinical Diabetes: The TEDDY Study. Diabetes Care. 2022;45(10):2271–2281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ostrov DA, Gottlieb PA, Michels AW. Rationally designed small molecules to prevent type 1 diabetes. Curr Opin Endocrinol Diabetes Obes. 2019;26(2):90–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Original data generated and analyzed during this study are included in this published article or in the data repositories listed in References.