Expanding knowledge of the genetic and phenotypic heterogeneity of congenital hyperinsulinism

Abstract

Congenital hyperinsulinism (HI) is a monogenic disorder characterised by inappropriate secretion of insulin from the pancreatic beta-cells despite low blood glucose, leading to hypoglycaemia. The aim of this thesis was to expand knowledge of the genetic and phenotypic heterogeneity of HI, using the world’s largest cohort of internationally referred individuals living with HI (n=4,181).

The first section of this thesis introduces HI, with a focus on the clinical presentation and mechanisms underpinning monogenic forms of disease. The next section describes the methodologies relevant throughout the thesis and not covered by the specific methods section of each chapter.

In chapter 1, we identified 173 cases diagnosed with HI in childhood (1-16 years old) and sequenced the known genetic causes to determine how common monogenic forms of disease were in those diagnosed outside of infancy. We found that although the pickup rate for genetic diagnoses in childhood was lower than that of those diagnosed in infancy, it was still substantial, with 24% of those clinically diagnosed in childhood receiving a genetic diagnosis. We investigated the reason for these later diagnoses and conclude that they likely form the tail-end in the normal range in age at diagnosis for monogenic HI.

Chapter 2 is a case report detailing a clinically mild form of HADH-HI caused by a likely pathogenic variant (c.709+39C>G). The described case initially presented at 22 months old with hypoglycaemia, the oldest reported age at diagnosis of hypoglycaemia in HADH-HI. Their disease progressed with multiple episodes of hypoglycaemia following protein and exercise, not requiring medical management. Insulin had not been measured at the time of hypoglycaemia meaning a biochemical diagnosis of hyperinsulinaemic hypoglycaemia could not be made. We performed an extensive literature search and compared the present case to reported cases with HADH-HI and unpublished data on cases with the same variant in the Exeter cohort. This demonstrated that this case is a rare example of HADH-HI characterised by a late age at onset and mild episodic hypoglycaemia.

In Chapter 3 we assessed the performance of REVEL, an in silico predictive tool used in variant classification, on variants acting via a loss of function compared to gain of function mechanism. We used a highly curated set of disease-causing loss- (n=66) and gain of function (n=65) genetic variants in ABCC8, GCK, and KCNJ11 that result in the opposite phenotypes of HI or monogenic diabetes. We show that REVEL is better at predicting the pathogenicity of loss of function variants than gain of function variants (55% vs 35% of variants met the threshold to be used as strong evidence in variant classification, respectively, P=0.0352). This means that variants acting via a gain of function mechanism are less likely to reach the higher evidence weightings used in variant classification, having implications for clinicians and scientists performing genetic testing in HI and monogenic diabetes but also for other rare diseases.

In chapter 4, we assessed the prevalence and phenotypic variability of a novel non-coding genetic cause of HI in which variants disrupting a cis-regulatory element of HK1 cause the gene to be inappropriately expressed in pancreatic beta-cells, where it should be silenced. To do this, we screened 1,763 individuals with genetically unsolved HI from three European genomics laboratories. Disease causing variants were identified in 94 probands (5%) and 28 clinically affected family members. We report reduced penetrance and a wide range in birth weight, age at diagnosis and treatment. The phenotypic variability is associated with the location of the variant within the regulatory element. This, to our knowledge, is therefore the first genotype-phenotype association in the non-coding genome for a Mendelian disease.

In chapter 5, we screened individuals with genetically unsolved HI for variants in the SLC16A1 promoter to better assess the contribution of these variants to the aetiology of HI, as no variants have been reported since the initial discovery in 2007. The three initially reported promoter variants were hypothesised to cause the inappropriate expression of SLC16A1 in pancreatic beta-cells, where it is selectively repressed in healthy individuals. We identified a novel 94bp deletion in the promoter of SLC16A1 causing HI in 15 individuals from eight families. We identified a wide range in age at diagnosis (1.8-65 years) and disease associated with fasting and exercise for individuals with this variant. Immunohistochemical analysis of pancreatic tissue obtained from three of the cases with the 94bp deletion confirmed the inappropriate expression of SLC16A1 in the pancreatic beta-cells.

The final section of this thesis provides an overview of each data chapter, the implications and importance of the findings in the wider context of HI research, and suggests future avenues for research in each specific area.

This thesis has provided new insights into the clinical presentation and genetic causes of HI. I have identified novel presentations of the known genetic causes, including both milder disease and later disease onset in some individuals. I have identified specific genotype-phenotype associations in HI and identified novel non-coding variants that cause the disease. These new insights have improved our understanding of the genetics of HI, helping to inform which cases should be genetically tested, given insights into the interpretation of genetic variants causing HI, and provide exciting avenues for future research.