Comprehensive screening shows that mutations in the known syndromic genes are rare in infants presenting with hyperinsulinaemic hypoglycaemia

Summary Objective Hyperinsulinaemic hypoglycaemia (HH) can occur in isolation or more rarely feature as part of a syndrome. Screening for mutations in the “syndromic” HH genes is guided by phenotype with genetic testing used to confirm the clinical diagnosis. As HH can be the presenting feature of a syndrome, it is possible that mutations will be missed as these genes are not routinely screened in all newly diagnosed individuals. We investigated the frequency of pathogenic variants in syndromic genes in infants with HH who had not been clinically diagnosed with a syndromic disorder at referral for genetic testing. Design We used genome sequencing data to assess the prevalence of mutations in syndromic HH genes in an international cohort of patients with HH of unknown genetic cause. Patients We undertook genome sequencing in 82 infants with HH without a clinical diagnosis of a known syndrome at referral for genetic testing. Measurements Within this cohort, we searched for the genetic aetiologies causing 20 different syndromes where HH had been reported as a feature. Results We identified a pathogenic KMT2D variant in a patient with HH diagnosed at birth, confirming a genetic diagnosis of Kabuki syndrome. Clinical data received following the identification of the mutation highlighted additional features consistent with the genetic diagnosis. Pathogenic variants were not identified in the remainder of the cohort. Conclusions Pathogenic variants in the syndromic HH genes are rare; thus, routine testing of these genes by molecular genetics laboratories is unlikely to be justified in patients without syndromic phenotypes.


| INTRODUC TI ON
Congenital hyperinsulinaemic hypoglycaemia (HH) is a disorder where episodes of hypoglycaemia are caused by unregulated insulin secretion despite low blood glucose levels. Prompt treatment is crucial in order to avoid serious lifelong complications such as seizures and permanent brain injury. 1 Genetic testing is important for the clinical management of this condition as identifying the underlying genetic aetiology will inform on the pancreatic histology which, for patients who are unresponsive to medical treatment, will help to determine whether a lesionectomy or a near-total pancreatectomy is required. 2 HH is most commonly the result of a monogenic aetiology with mutations in at least eight genes reported to cause isolated disease. 3 Routine screening of these genes using a combination of rapid Sanger sequencing and targeted next-generation sequencing identifies a mutation in approximately 40%-50% of cases with persistent HH. 3 This suggests that further genetic aetiologies remain to be discovered.
HH has also been reported as a feature in at least 20 different rare genetic syndromes. 1 The most common is Beckwith-Wiedemann syndrome where HH occurs in approximately 50% of cases. 4 More rarely, HH has been described in individuals with other overgrowth disorders including Sotos syndrome 5 and growth delay syndromes such as Kabuki. 6 HH has also been reported in some cases with congenital disorders of glycosylation and chromosome abnormalities such as Turner syndrome and Patau syndrome. [7][8][9][10] For patients with syndromic HH, an early genetic diagnosis is important as this will guide medical management and provide information on prognosis and recurrence risk.
Screening of the syndromic genes is not routinely performed as part of the genetic testing strategy for individuals with HH. The targeted analysis of a gene is usually only performed when there is a clinical suspicion of a specific syndrome in an individual and therefore acts to confirm the clinical diagnosis. 11 Consequently, the prevalence of mutations in these genes in HH is not known. As many patients are referred for genetic testing at diagnosis of HH, it is possible that some individuals with a mutation in a syndromic gene will not have developed additional extra-pancreatic features at the time of referral for genetic testing. These patients may therefore not receive the most appropriate genetic testing.
In order to assess the frequency of mutations in the known syndromic genes in individuals with HH of unknown genetic cause, we performed a comprehensive analysis of genome sequencing data from 82 affected infants. Screening patients for mutations in the "syndromic" HH genes irrespective of clinical features has the advantage that a genetic diagnosis can precede development of clinical features and guide clinical management, rather than being confirmatory.

| Patient details
82 infants with HH diagnosed within the first 12 months of life were referred for genetic testing to the Molecular Genetics Laboratory at the Royal Devon and Exeter NHS Foundation Trust. All patients in the cohort had received a biochemical diagnosis of HH in their local centre. All patients had a blood glucose of <2.8 mmol/L (median 1.9 mmol/L) with a concomitant insulin of >15 pmol/L (median 96 pmol/L) and/or a c-peptide of >300 pmol/L (see Table 1). In all cases, the HH had persisted for greater than 6 months or had required pancreatic resection following a poor response to treatment.
In 15 patients, extra-pancreatic features were present at the time of referral for genetic testing. None of these patients had received a clinical diagnosis consistent with a known syndromic form of HH at the time of study. Mutations in the ABCC8, KCNJ11, HADH, HNF4A, HNF1A, GLUD1, GCK, and SLC16A1 genes had been excluded in all patients using targeted next-generation sequencing. 11 The coding regions of the HH candidate genes, UCP2 and HK1, were also screened but no likely pathogenic variants were identified in the cohort. [12][13][14] The study was approved by the North Wales Research Ethics Committee. Consent was obtained from each patient after full explanation of the purpose and nature of all procedures used.

| Gene panel and variant calling
We utilized the Phenomizer browser to identify disease entries annotated for hyperinsulinaemic hypoglycaemia (HPO id: 0000825). 15 A survey of the literature was also undertaken to identify further genes in which mutations have been reported to cause HH as part of a syndrome. The sequencing data were analysed using an approach based on the GATK best practice guidelines. 16 This involved aligning the reads to the hg19/GRCh37 human reference genome with BWA mem, applying Picard for duplicates removal, and GATK IndelRealigner for local realignment and running the base quality score realignment. GATK TA B L E 1 Clinical characteristics of the 82 patients included in this study. All patients had received a clinical diagnosis of HH from their referring clinician, with biochemical testing undertaken in their local laboratories. When applicable, median values are given with the range.  TA B L E 2 Syndromes in which hyperinsulinaemic hypoglycaemia (HH) has been reported as a feature. The 18 genes in which mutations have been reported to cause syndromic HH plus the three genomic regions known to be affected by copy number variants (CNVs) or uniparental isodisomy (UPD) are provided.

| RE SULTS
We identified 20 genetic syndromes in which HH has been reported as a feature (Table 2). In 82 patients, we screened the coding regions and intron/exon boundaries of the 18 genes associated with these syndromes. We also searched for copy number variations (CNVs) and evidence of uniparental isodisomy at genomic regions associated with Beckwith-Wiedemann, Patau and Turner syndromes.
We identified 70 nonsynonymous (nonsense, frameshift, splice site, missense) variants in the 18 genes. As the incidence of HH in outbred populations is estimated to be between 1 in 27 000 and 1 in 50 000, 18 we excluded all variants in dominant genes that were present in gnomAD controls 19 at a frequency greater than 1 in 27 000. In addition, we excluded variants that did not fit the known inheritance

| D ISCUSS I ON
We screened for genetic aetiologies where HH has been reported to feature as part of a complex syndrome in 82 patients with HH of unknown cause. We identified one patient with a pathogenic variant in KMT2D that confirmed a diagnosis of Kabuki syndrome. 6 At referral, the patient was reported to the genetics laboratory as suggesting that HH may be a more common presenting feature than previously recognized. Their screening of an additional 100 isolated HH patients identified one further patient (who was retrospectively recognized to have the classic syndromic phenotype), which is a similar frequency to our study.
Mutations in the "syndromic" HH genes were not identified in the majority of our cohort (81/82 patients). Our patients were selected due to the presence of persistent HH rather than a syndromic phenotype. Our low pickup rate suggests that mutations in these genes are rare in patients presenting with isolated HH.
It is likely that for the majority of cases with syndromic disease, additional features are evident from birth allowing for a clinical diagnosis followed by confirmatory genetic testing. This would consequently reduce the prevalence of cases within our cohort as the majority of patients are referred for routine screening of the 8 known isolated HH genes. Heterozygous variants in genes that cause a dominant disease were excluded if they were inherited from an unaffected parent; however, it is possible that these could be pathogenic but displaying incomplete penetrance in the parents. We can also not rule out the possibility that some patients have mutations in other genes known to cause multisystem disease which was not screened in this study as HH has not been recognized as a common feature. Furthermore, it is possible that some patients with Beckwith-Wiedemann syndrome have a methylation defect in the absence of a structural abnormality that was not detected by our analysis.
The finding that the patient with Kabuki syndrome had HH as the presenting feature highlights the potential benefit of screening for the syndromic genes in individuals newly diagnosed with HH.
Identifying the underlying genetic aetiology is important for these patients as it will inform on prognosis which will allow for better clinical management. A genetic diagnosis also provides important information on recurrence risk.
We found that mutations causing these 20 syndromes are rare.
The recent adoption of targeted next-generation sequencing by mo- Thus, routine testing of this panel of genes may not be justified for patients without syndromic features as we have shown that mutations in these genes are rare in such a cohort. It is anticipated that variant interpretation will become easier through data sharing initiatives such as ClinVar, 22 and as the number of publically available control data sets increase, more variants can be excluded by frequency. 19 This may make testing of these genes in all patients with HH more feasible in future.
Screening large cohorts for mutations in these genes will provide further information on the prevalence of mutations in HH, which will help to guide future genetic screening strategies for this condition and may also identify individuals with "nonclassical" features, which would expand the phenotype associated with these syndromes.
In conclusion, while the clinical impact of finding a mutation in a syndromic gene is high as it can inform clinical management, mutations in the syndromic HH genes are rare in individuals referred for routine testing for HH. Given the time-consuming nature of a comprehensive screen, routine testing of these genes by molecular genetics laboratories is unlikely to be justified in patients without syndromic phenotypes.

CO N FLI C T O F I NTE R E S T
The authors have no conflict of interests to declare.