Congenital hyperinsulinism is characterised by the inappropriate release of insulin during hypoglycaemia. This potentially life-threatening disorder can occur in isolation, or present as a feature of syndromic disease. Establishing the underlying aetiology of the hyperinsulinism is critical for guiding medical management of this condition especially in children with diazoxide-unresponsive hyperinsulinism where the underlying genetics determines whether focal or diffuse pancreatic disease is present. Disease-causing single nucleotide variants affecting over 30 genes are known to cause persistent hyperinsulinism with mutations in the KATP channel genes (ABCC8 and KCNJ11) most commonly identified in children with severe persistent disease. Defects in methylation, changes in chromosome number, and large deletions and duplications disrupting multiple genes are also well described in congenital hyperinsulinism, further highlighting the genetic heterogeneity of this condition. Next-generation sequencing has revolutionised the approach to genetic testing for congenital hyperinsulinism with targeted gene panels, exome, and genome sequencing being highly sensitive methods for the analysis of multiple disease genes in a single reaction. It should though be recognised that limitations remain with next-generation sequencing with no single application able to detect all reported forms of genetic variation. This is an important consideration for hyperinsulinism genetic testing as comprehensive screening may require multiple investigations.

Introductory paragraph Gene expression is tightly regulated with many genes exhibiting cell-specific silencing when their protein product would disrupt normal cellular function 1 . This silencing is largely controlled by non-coding elements and their disruption might cause human disease 2 . We performed gene-agnostic screening of the non-coding regions to discover new molecular causes of congenital hyperinsulinism. This identified 14 non-coding de novo variants affecting a 42bp conserved region encompassed by a regulatory element in intron 2 of Hexokinase 1 ( HK1 ). HK1 is widely expressed across all tissues except for liver and pancreatic beta-cells and is thus termed a “disallowed gene” in these specific tissues. We demonstrated that the variants result in a loss of repression of HK1 in pancreatic beta-cells, thereby causing insulin secretion and congenital hyperinsulinism. Using epigenomic data accessed from public repositories, we demonstrated that these variants reside within a regulatory region that we determine to be critical for cell-specific silencing. Importantly, this has revealed a disease mechanism for non-coding variants that cause inappropriate expression of a disallowed gene.

Background Hyperinsulinism results from inappropriate insulin secretion during hypoglycaemia. Down syndrome is causally linked to a number of endocrine disorders including Type 1 diabetes and neonatal diabetes. We noted a high number of individuals with Down syndrome referred for hyperinsulinism genetic testing, and therefore aimed to investigate whether the prevalence of Down syndrome was increased in our hyperinsulinism cohort compared to the population. Methods We identified individuals with Down syndrome referred for hyperinsulinism genetic testing to the Exeter Genomics Laboratory between 2008 and 2020. We sequenced the known hyperinsulinism genes in all individuals and investigated their clinical features. Results We identified 11 individuals with Down syndrome in a cohort of 2011 patients referred for genetic testing for hyperinsulinism. This represents an increased prevalence compared to the population (2.5/2011 expected vs. 11/2011 observed, p = 6.8 × 10−5). A pathogenic ABCC8 mutation was identified in one of the 11 individuals. Of the remaining 10 individuals, five had non‐genetic risk factors for hyperinsulinism resulting from the Down syndrome phenotype: intrauterine growth restriction, prematurity, gastric/oesophageal surgery, and asparaginase treatment for leukaemia. For five individuals no risk factors for hypoglycaemia were reported although two of these individuals had transient hyperinsulinism and one was lost to follow‐up. Conclusions Down syndrome is more common in patients with hyperinsulinism than in the population. This is likely due to an increased burden of non‐genetic risk factors resulting from the Down syndrome phenotype. Down syndrome should not preclude genetic testing as coincidental monogenic hyperinsulinism and Down syndrome is possible.

Gene expression is tightly regulated with many genes exhibiting cell-specific silencing when their protein product would disrupt normal cellular function. This silencing is largely controlled by non-coding elements and their disruption might cause human disease. We performed gene-agnostic screening of the non-coding regions to discover new molecular causes of congenital hyperinsulinism. This identified 14 non-coding de novo mutations affecting a 42bp conserved region encompassed by a regulatory element in intron 2 of Hexokinase 1 (HK1), a pancreatic beta-cell disallowed gene. We demonstrated that these mutations resulted in expression of HK1 in the pancreatic beta-cells causing inappropriate insulin secretion and congenital hyperinsulinism. These mutations identify a regulatory region critical for cell-specific silencing. Importantly, this has revealed a new disease mechanism for non-coding mutations that cause inappropriate expression of a disallowed gene.

Objective Mutations in the KATP channel genes, ABCC8 and KCNJ11, are the most common cause of congenital hyperinsulinism. The diagnosis of KATP-hyperinsulinism is important for the clinical management of the condition. We aimed to determine the clinical features that help to identify KATP-hyperinsulinism at diagnosis. Design We studied 761 individuals with KATP-hyperinsulinism and 862 probands with hyperinsulinism of unknown aetiology diagnosed before 6 months of age. All were referred as part of routine clinical care. Methods We compared the clinical features of KATP-hyperinsulinism and unknown hyperinsulinism cases. We performed logistic regression and ROC analysis to identify the features that predict KATP-hyperinsulinism. Results Higher birth weight, diazoxide unresponsiveness and diagnosis in the first week of life were independently associated with KATP-hyperinsulinism (adjusted Odds Ratio 4.5 (95% CI, 3.4-5.9), 0.09 (0.06-0.13) and 3.3 (2.0- 5.0) respectively). Birth weight and diazoxide unresponsiveness were additive and highly discriminatory for identifying KATP-hyperinsulinism (ROC area under the curve for birth weight 0.80, diazoxide responsiveness 0.77, and together 0.88, 95% CI 0.85-0.90). 86% born large for gestation and 78% born appropriate for gestation who did not respond to diazoxide treatment had KATP-hyperinsulinism. In contrast, of those individuals born small for gestation, none who were diazoxide responsive and only 4% of those who were diazoxide unresponsive had KATP-hyperinsulinism. Conclusions Individuals with hyperinsulinism born appropriate or large for gestation and unresponsive to diazoxide treatment are most likely to have an ABCC8 or KCNJ11 mutation. These patients should be prioritised for genetic testing for KATP channel genes.