Abstract
The risk of Type 1 Diabetes (T1D) comprises both genetic and environmental components. We investigated whether genetic susceptibility to T1D could be mediated by changes in DNA methylation, an epigenetic mechanism that potentially plays a role in autoimmune diabetes. Using data from a non-diabetic population comprising blood samples taken at birth (n=844), childhood (n=911) and adolescence (n=907), we evaluated the associations between 65 top GWAS single nucleotide polymorphisms (SNPs) and genome-wide DNA methylation levels. We identified 159 proximal SNP-cytosine phosphate guanine (CpG) pairs (cis), and 7 distal SNP-CpG associations (trans) at birth, childhood, and adolescence. There was also a systematic enrichment for methylation related SNPs to be associated with T1D across the genome, after controlling for genomic characteristics of the SNPs, implying that methylation could either be on the causal pathway to T1D or a non-causal biomarker. Combining the principles of Mendelian Randomization and genetic colocalization analysis, we provided evidence that at 5 loci, ITGB3BP, AFF3, PTPN2, CTSH and CTLA4, DNA methylation is potentially on the causal pathway to T1D.
- Abbreviations
- ARIES
- Accessible Resource for Integrated Epigenomic Studies
- ALSPAC
- Avon Longitudinal Study of Parents and Children
- BOX
- Bart’s Oxford family study of Type 1 Diabetes
- CpG
- Cytosine-phosphate-guanine dinucleotides
- GWAS
- Genome-wide association study
- HLA
- Human leukocyte antigen
- JLIM
- Joint likelihood mapping
- LD
- Linkage disequilibrium
- SNP
- Single nucleotide polymorphism
- T1D
- Type 1 diabetes
- MAF
- Minor allele frequency
- MR
- Mendelian Randomization
- mQTL
- methylation quantitative trait loci
- RA
- Rheumatoid arthritis
- 2SMR
- Two Sample Mendelian Randomization