Genetic Research Informs on Timing of Type 1 Diabetes Development

Miriam E Tucker

October 03, 2017

LISBON, PORTUGAL — New research may point to better ways of diagnosing type 1 diabetes in children and adults and predicting who will develop it in the future.

Findings from two studies on the genetics of type 1 diabetes were presented recently at the European Association for the Study of Diabetes (EASD) 2017 Annual Meeting.

In one study, information from the UK Biobank revealed differences in human leukocyte antigen (HLA) allele patterns between people who developed type 1 diabetes in adulthood — a group that hasn't been genetically profiled before — and those with childhood-onset type 1 diabetes.

Those results were presented by Dr Nicholas JM Thomas of the Institute of Biomedical and Clinical Science, University of Exeter Medical School, United Kingdom.

The latter follows Dr Thomas's startling UK Biobank data report at the 2016 EASD meeting that half of all type 1 — formerly known as "juvenile" — diabetes develops after 30 years of age.

"In time, the hope is that we can develop an algorithm to help primary-care physicians to be able to stratify at diagnosis to get a clear idea of the likely type of diabetes," Dr Thomas told Medscape Medical News.

In the other study, a previously derived genetic risk score — developed by the scientists at University of Exeter that incorporates HLA and non-HLA T1D-associated single nucleotide polymorphisms (SNPs) — was found to predict the risk of progression of islet autoimmunity and the development of type 1 diabetes among participants of the TrialNet Pathways to Prevention study.

These individuals were positive for one or more type 1 diabetes–associated autoantibodies but did not have the condition at baseline. Those data were presented by Maria J Redondo, MD, PhD, of Baylor College of Medicine, Houston, Texas.

"Genetic information about type 1 diabetes is very complex. We want to include genetic information in predictive models to use for research and also for clinical applicability," she told Medscape Medical News.

Asked to comment, session comoderator Katharine R Owen, MD, of the Diabetes Research Laboratories, University of Oxford, United Kingdom, said: "I think there's a bit more validation to be done, but yes, in theory absolutely we should be using HLA in some form to help in diagnosing type 1 diabetes because it does seem to add something.

"I'd like to see more data on how these risk scores can help us in our gray-area cases.…There's quite a big group between the ages of 25 and 45 where it's really hard to tell sometimes if it's type 1 or type 2," she added.

Adults who develop type 1 diabetes are often misdiagnosed with type 2 and not immediately treated with insulin. But because type 1 diabetes often has a slower onset in adults — sometimes called "latent autoimmune diabetes of adulthood (LADA)" — affected individuals may have some preserved beta-cell function. The need for insulin eventually becomes obvious, but ideally doctors should consider the possibility of type 1 in such adults at diagnosis.

HLA Allele Pattern Differs in Older- vs Younger-Onset T1D

Dr Thomas's research examined 379,511 white Europeans aged 40 to 70 at recruitment into the UK Biobank for the prevalence of the HLA alleles DR3 and DR4, long known to be highly correlated with type 1 diabetes based on data from studies of people who developed the condition in childhood. He and his colleagues also looked at the prevalence of the strongly protective alleles DR15 and DQ6.

The DR3 and DR4 alleles (in any combination) were found in 6.4% of the overall UK Biobank population but contributed to 61% of all cases of type 1 diabetes.

Of those with DR3/DR4-associated type 1 diabetes, 59% were diagnosed younger than 25 years of age, while 37% developed the condition aged 26 to 50 years.

In contrast, for DR4/DR4, just 24% had younger onset, while 42% were diagnosed after 25 years of age. Those with DR3/DR3 were equally likely to develop the condition at either time point (21% for both age groups).

Mean age at diagnosis was 21 years for those with DR3/DR4, 26 for DR3/DR3, and 31 for DR4/DR4.

At the same time, DR15 remained protective against the development of type 1 diabetes throughout life.

Dr Thomas and coauthors conclude: "While all three major genotypes greatly increase risk of T1D throughout life, population analysis has shown for the first time that DR4/DR4 specifically predisposes to T1D over 30 years of age, and carriers of this genotype have the highest risk for development of late-onset T1D.

 "This is clear evidence that type 1 diabetes after 30 years is not just a delayed version of type 1 diabetes before 30. Further work is needed to understand these differences."

Risk Score Predicts Type 1 Diabetes Progression

In the TrialNet study, 5381 individuals with a median age of 11 were genotyped using the Illumina Immunochip SNP array. They included 291 who were positive for one autoantibody, of whom 156 subsequently progressed to two or more antibodies and 55 who eventually developed type 1 diabetes.

Of another 953 with two or more autoantibodies at baseline, 421 developed type 1 diabetes. Median follow-up time was 5.5 years. 

The type 1 diabetes genetic risk score (GRS) was calculated and based on 30 T1D-associated SNPs, including both HLA and non-HLA T1D-associated SNPs.

The GRS predicted the time to development of type 1 diabetes (hazard ratio, 1.49; = .0016), independently of antibody status, gender, HLA DR3/DR4 status, and the Diabetes Prevention Trial-Type 1 Risk Score (DPTRS), which includes glycemia, age, body-mass index, and C-peptide levels (Diabetes Care. 2014;37:979–984).

Participants with GRS cutoffs of less than 0.25 had a 74% likelihood of not developing T1D within 5 years, vs a 56% chance among those with GRS scores of 0.25 or above (hazard ratio, 0.57; = .0012).

The GRS was particularly useful among participants with DPTRS scores of 7 or below, since those with scores above 7 are already highly likely to develop type 1 diabetes (70% at 5 years, 93% at 10 years).

But among those with DPTRS scores of 7 or below, the GRS plus antibody status improved the prediction capability.

For example, the probability of not developing type 1 diabetes within 10 years was 97.4% for those with DPTRS of <7, just a single positive antibody, and GRS below 0.25, compared with just a 46% chance of being T1D-free at 10 years for those with DPTRS <7 but also having multiple antibodies and GRS of >0.25.

The GRS was similarly predictive of progression from single to multiple antibody status and from multiple antibody status to type 1 diabetes.

What Are the Clinical Implications?

Asked during a press briefing — during which the UK Biobank study results were discussed — what the clinical implications of this research might be, Dr Thomas replied, "It may be that when you've got adults developing type 1 diabetes it might change the odds ratios in the [genetic] risk score (which was originally developed by the Exeter team)."

But, he added, "The most important thing is it tells us there's heterogeneity in type 1 diabetes in the immune process.…That's important when it comes to studying T1D and also in the future when it comes to using immunotherapies.…It might suggest different treatments to treat subtly different diseases."

In the meantime, Dr Owen said, "In terms of being clearer about diagnosis, we probably just need to be better clinicians.

"What we see as secondary-care physicians is very often that not enough notice has been taken of the actual clinical features and presentation and progression."

Dr Redondo, Dr Thomas, and Dr Owen have no relevant financial relationships.  

European Association for the Study of Diabetes (EASD) 2017 Annual Meeting; September 12, 2017; Lisbon,  Portugal. Abstract 49, Abstract 51

For more diabetes and endocrinology news, follow us on Twitter and on Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.