The changing frequency of Type 1 diabetes-associated genes over time

: Insights into how our changing environment influences risk

Abstract

Type 1 diabetes (T1D) is a chronic autoimmune disease. Both genetic and environmental risk factors contribute to its development. The incidence of the condition is increasing at a rate of 3% per year in most westernised populations. Counterintuitively, while the incidence has been increasing, the proportions of individuals positive for high-risk Human leucocyte antigen (HLA) class II genes have been decreasing. This might suggest an environmental impact on changing incidence. It is, however, unknown whether this change has continued into the 21st century. In addition, a panel of non-HLA genes also contribute to the risk of T1D, but their frequencies over time have not been monitored. The aim of this study is to explore 1) whether HLA class II frequency continues to change over time, 2) the changes in non-HLA SNP frequencies over time and 3) whether a report of non-HLA SNPs associated with the development of islet autoimmunity, but not diabetes, can be replicated in this study and if the dynamics over time have changed for these SNPs
Methods
Two populations were available for this study, individuals with diabetes participating in the long-running Bart’s Oxford study (BOX) recruiting since 1985 and the “Golden Years” cohort (GY) diagnosed between 1922 and 1948. HLA class II genotyping was carried out using well-established PCR-SSP. Non-HLA SNPs associated with T1D were analysed using Taqman genotyping. Autoantibodies to glutamic acid decarboxylase (GAD), insulin, insulinoma-associated protein-2 (IA-2), IA-2β and Zinc transporter 8 (ZnT8) were measured using radio-binding assays. Descriptive statistics and chi-square for trend were used to analyse changes to genotype frequency over time while chi-square analysis and logistic regression were used to test for associations between the SNPs and islet autoantibodies.
Results
HLA class II data from 1737 cases diagnosed with diabetes under the age of 21 years between 1985 and 2015 were analysed, and results do not support a change in HLA class II frequency over this time span. Genotyping of 19 non-HLA T1D associated SNPs in DNA samples from 1992 BOX and 285 GY participants did not provide evidence for changes in frequency over time. Serum collected a median of one day from diagnosis (range -61 to 90 days) was available from 550 people with diabetes (median age at diagnosis 10.9, range 0.7 to 21 years) participating in the BOX study. ZnT8A, IA-2A and IA-2βA were more common (p<0.001) in newly-diagnosed patients homozygous for FCRL3 rs3761959 CC, compared with other genotypes. Furthermore, all 20 individuals homozygous for the disease-associated allele of RELA had GAD autoantibodies, compared with 79% of those without this allele (p=0.055). Logistic regression showed that FCRL3 was associated with IA-2A, ZnT8A, and IA-2βA (p<0.002 for all), independently of age at diagnosis, sex, and HLA class II genotype. In addition to FCRL3 and RELA, the LPP GG genotype had the highest association with GADA positivity (p=0.018) and positive association with IA-2βA (p=0.036), independent of age at diagnosis, sex, and HLA class II genotypes.

Conclusion
This study has shown for the first time that the well-documented change in HLA class II frequency over time is specific for HLA class II and not shared by non-HLA loci associated with T1D. There is no evidence that HLA class II frequency is continuing to decrease in the twenty-first century, but a larger study incorporating data from other long-term longitudinal studies would be beneficial. Associations of three non-HLA SNPs with islet autoimmunity was confirmed in this study.

Date of Award	2 Dec 2021
Original language	English
Awarding Institution	University of Bristol
Supervisor	Kathleen M Gillespie (Supervisor) & Alistair Williams (Supervisor)