The recognition of a monogenic form of diabetes dates back to the 1970s, when Robert Tattersall first described maturity-onset diabetes of the young (MODY).¹ As knowledge of the condition has expanded, the initial criteria outlined (autosomal dominant inheritance, age of diagnosis <25 years and no insulin requirement) have been found to recognise some, but not all cases of MODY.²

Today, the term monogenic diabetes incorporates not just MODY but other single gene defects causing diabetes, including neonatal diabetes (diabetes presenting at age <6 months) and mitochondrial diabetes (maternally inherited mutations in mitochondrial DNA).

The prevalence of monogenic diabetes varies according to the population studied and how they were selected for testing. MODY accounts for between roughly 1 and 4% of diabetes in those diagnosed below 30 years,^2,3 whilst neonatal diabetes is less common and affects approximately 1 in 100,000 births.⁴

BENEFITS OF DIAGNOSIS

The impetus to identify monogenic diabetes comes from variable treatment options that depend on the affected gene. In many cases, these options are superior to the standard care provided to people with type 1 or type 2 diabetes, so there is individual clinical benefit in identifying cases. Mutations in hepatocyte nuclear factor-1α (HNF1A) and -4α (HNF4A), β-cell transcription factors, are best managed with low doses of sulphonylureas.^5,6 Glucokinase (GCK) MODY, involving a key enzyme in glycolysis, requires no treatment at all^7,8 and mutations in hepatocyte nuclear factor-1β (HNF1B) are usually insulin-requiring.⁹ In neonatal diabetes, the most commonly affected genes are KCNJ11 and ABCC8, encoding the two subunits of the ATP-sensitive potassium channel of the β-cell.⁴ High dose sulphonylurea therapy can be used in >90% of cases, without the need for insulin therapy and associated complications in affected neonates.¹⁰

In addition to being more precise about treatment,^5,11,12 making a diagnosis of monogenic diabetes has other benefits, including cascade testing in family members to identify other cases, and early identification of other features that may be associated with the genetic defect, for example renal abnormalities in people with HNF1B MODY.⁹ Additionally, delays in diagnosis can impact treatment success. People with HNF1A mutations achieved better glycaemic control the earlier a genetic diagnosis was made,¹³ and in children with permanent neonatal diabetes, failure to transfer to sulphonylurea treatment completely was associated with a longer duration at the time of transfer from insulin.¹⁰

WHAT ARE THE CURRENT CHALLENGES?

1. Considering the possibility of diagnosis

A significant proportion of MODY cases remain misdiagnosed as type 1 or type 2 diabetes, whilst new cases are not always recognised, even at diagnosis.¹⁴ The marked geographical variation in frequency of genetic test requests signals variation in clinical practice, but this heterogeneity is likely to improve following efforts from the Exeter Molecular Genetics team to improve education, and a network of genetic diabetes nurses to facilitate training and testing.¹⁵

2. Deciding to test

A stratification process is required to identify the cases most likely to have monogenic diabetes. For neonatal diabetes, this is relatively straightforward, as the phenotype of developing diabetes in the first 6 months of life is unambiguous, nearly always monogenic, and testing is available free of charge.⁴ For cases of MODY, clinical stratification (selecting those diagnosed below 30 years) with biomarkers (selecting only those negative to pancreatic autoantibodies, with some evidence of endogenous C-peptide production) seems to be the most effective approach.²

3. Diagnosis in ethnic groups

Whilst initial studies to identify monogenic diabetes have focused on white European populations, there is now increasing recognition of monogenic diabetes in people of other ethnicities.¹⁶ However, a 2016 study revealed that the detection rate for MODY in South Asian people referred for genetic testing was much lower than in people of white ethnicity.¹⁷ It is likely that more people will need to be tested from ethnic groups with a high prevalence of young-onset type 2 diabetes, as the separation of MODY from type 2 diabetes is challenging on the basis of clinical features and existing biomarkers.¹⁸

4. Understanding test results

Anyone undertaking genetic testing in their practice will have come across novel mutations or variants of unknown significance (VUS) in genetic testing reports.¹⁹ Understanding what these results mean, whether they should be conveyed to the patient and how they might impact treatment is challenging. Co-segregation studies in family members can be a powerful tool, along with detailed clinical phenotyping, to disentangle pathogenicity.²⁰ Sometimes clarity around diagnosis cannot be provided, and it is good clinical practice to keep a record of patients with these variants which can be revisited regularly with the molecular genetics service, as more knowledge becomes available.

NEW DEVELOPMENTS IN THE FIELD

To date, the field of diabetes genetics has focused on detecting monogenic disorders, but now attention is turning to utilisation of information about the polygenic risk of diabetes. Single nucleotide polymorphisms (SNP) can modify the risk of type 1 and type 2 diabetes, both polygenic conditions. Summing up the risk of individual SNP genotypes can provide a composite polygenic risk score. In diabetes, these risk scores have been used to support diabetes classification, given the challenge of differentiating type 2 diabetes from adult-onset type 1 diabetes and also MODY.^21,22 The impact of using this clinical information and its applicability in all ethnicities have yet to be determined, and those in the field await further studies.²³

Shivani Misra, Consultant in Diabetes and Metabolic Medicine, Imperial College Healthcare NHS Trust, London

REFERENCES

1. Tattersall RB 1974 Quarterly Journal of Medicine 43 339–357.
2. Shields BM et al. 2017 Diabetes Care 40 1017–1025.
3. Johansson BB et al. 2017 Diabetologia 60 625–635.
4. De Franco E et al. 2015 Lancet 386 957–963.
5. Shepherd M et al. 2003 Diabetes Care 26 3191–3192.
6. Pearson ER et al. 2003 Lancet 362 1275–1281.
7. Steele AM et al. 2014 JAMA 311 279–286.
8. Spyer G et al. 2009 Diabetic Medicine 26 14–18.
9. Clissold RL et al. 2014 Nature Reviews Nephrology 11 102–112.
10. Babiker T et al. 2016 Diabetologia 59 1162–1166.
11. Pearson ER et al. 2006 New England Journal of Medicine 355 467–477.
12. Gloyn AL et al. 2004 New England Journal of Medicine 350 1838–1849.
13. Shepherd MH et al. 2018 Diabetologia 61 2520−2527.
14. Shields BM et al. 2010 Diabetologia 53 2504–2508.
15. Shepherd M et al. 2014 Clinical Medicine 14 117–121.
16. Pihoker C et al. 2013 Journal of Clinical Endocrinology & Metabolism 98 4055–4062.
17. Misra S et al. 2016 Diabetologia 59 2262–2265.
18. Misra S & Owen KR 2018 Current Diabetes Reports 18 141.
19. Kawasaki E et al. 2000 Journal of Clinical Endocrinology & Metabolism 85 331–335.
20. Althari S & Gloyn AL 2015 Review of Diabetic Studies 12 330–348.
21. Patel KA et al. 2016 Diabetes 65 2094−2099.
22. Oram RA et al. 2016 Diabetes Care 39 337−344.
23. Martin AR et al. 2018 bioRχiv doi:10.1101/441261.