Genetics point way to type 1 and monogenic diagnosis and prediction
medwireNews: Studies presented in a session on the genetics of diabetes, at the EASD annual meeting in Lisbon, Portugal, demonstrated the heterogeneity of type 1 diabetes risk and the predictive and diagnostic power of a genetic risk score (GRS).
In the first presentation of the session, Nicholas Thomas, from the University of Exeter in the UK, reported that the type 1 diabetes risk associated with the human leukocyte antigen (HLA) class 2 risk alleles varies according to age.
He said that although being heterozygous for the DR3 and DR4 risk alleles is known to carry the greatest overall risk, this knowledge comes from studies of young patients, when in fact 40% of type 1 diabetes patients are diagnosed beyond the age of 25 years. And his team’s analysis found that homozygosity for the DR4 risk alleles assumes greater importance in this older age bracket. Although DR3/DR4 accounted for the largest proportion of cases overall, only 41% of study participants with this combination were older than 25 years when diagnosed with diabetes, compared with 64% of those with DR4/DR4.
Indeed, after this age, DR4/DR4 assumed comparable importance to DR3/DR4, with the alleles accounting for a respective 42% and 37% of cases diagnosed in this age group. The protective DR15 allele remained equally so at all ages, however.
The study data came from 379,511 White Europeans in the UK Biobank, and Thomas noted that the high risk alleles are present in only 6.3% of this population, with DR4/DR4 present in only 1.0%. However, he said that the study demonstrates the heterogeneity of type 1 diabetes risk and, on a practical level, may help to reduce uncertainty around late diagnoses.
Two presentations concerned the predictive power of GRSs. The first was the type 1 diabetes GRS, which is based on 30 genetic variants, including both HLA and non-HLA risk variants, and can accurately discriminate between type 1 and type 2 diabetes.
Maria Redondo (Baylor College of Medicine, Houston, Texas, USA) reported research showing that this GRS can also predict the development of type 1 diabetes in at-risk people from Type 1 Diabetes TrialNet. Of the 291 who were positive for one autoantibody at inclusion, 156 progressed to being positive for multiple autoantibodies, with 55 of these progressing to type 1 diabetes. Of the 953 initially positive for multiple autoantibodies, 421 progressed to diabetes.
The team found that the GRS significantly predicted time to diabetes progression. During 5 years of follow-up, 74% of people with a score below 0.25 remained free of diabetes, compared with 56% of those with a higher score. The score was able to refine risk even among people with low diabetes risk according to the Diabetes Prevention Trial-Type 1 Risk Score (DPTRS ≤7) after stratifying by the number of autoantibodies; among those with multiple autoantibodies, for example, 90% of those with a GRS below 0.25 remained free of diabetes, compared with 72% of those with a higher score.
It also predicted progression from single to multiple autoantibodies and from multiple autoantibodies to diabetes, again being able to refine risk in patients with a low DPTRS score.
And a presentation from Matthew Johnson, again from the University of Exeter, showed that a separate GRS, based on 10 variants in known autoimmune monogenic diabetes genes, can differentiate between this form of diabetes and that caused by polygenic clustering. Accurate diagnosis is important for patients with monogenic forms, he noted, because specific immunotherapies may be available, and it also allows for prenatal testing.
The team obtained their results in a cohort of 79 patients with type 1 diabetes and at least one other autoimmune disease (most frequently autoimmune enteropathy and hypothyroidism), diagnosed before the age of 5 years.
Sequencing confirmed monogenic diabetes in 47% of the cohort; these patients had a significantly lower GRS than other patients and were also found likely to have previously unknown monogenic diabetes variants (in recognized diabetes genes). The GRS had a discriminative accuracy of 82% for monogenic versus polygenic diabetes, compared with 76% for age at onset; combining the two increased it to 94%.
In the same session, Natalie Zubkova (Endocrinology Research Center, Moscow, Russia) reported a high prevalence of previously undiagnosed maturity-onset diabetes of the young (MODY) in a cohort of patients with gestational diabetes mellitus (GDM). Indeed, it was present in 46.0% of the 50 women who had overt GDM (fasting glucose ≥7 mmol/L) and 25.5% of 153 with GDM (≥5.1 mmol/L). Commenting from the audience, Andrew Hattersley (University of Exeter) suggested this high rate could be due to the low BMI of the patients (average 24 kg/m2), relative to other published cohorts, which would maximize the proportion of women identified with a strong genetic component to their GDM.
In keeping with other studies from Russia, the majority of MODY mutations were in the GCK gene, present in 34.0% of patients with overt GDM and 13.7% of those with milder GDM.
medwireNews is an independent medical news service provided by Springer Healthcare. © 2017 Springer Healthcare part of the Springer Nature group