Congenital Hyperinsulinism (HI) is the most common cause of persistent hypoglycemia in infants and children. In most cases, it results from the disruption of genes that control insulin secretion from the pancreatic beta-cell (1, 2).

Over 30 genetic forms of HI have been described which cause an isolated or syndromic form of the condition (3, 4). Many different types of genetic variation which disrupt the known HI genes/genetic loci are reported. These include single nucleotide variants (SNVs), indels, methylation defects, aneuploidies and large deletions or duplications (referred to as copy number variants; CNVs) (5).

Partial or whole gene deletions and/or duplications have been reported in at least 5 HI disease genes (6–10). Deletions and duplications which encompass multiple genes have also been described in individuals with HI, who typically have a syndromic condition (11–18). These variants are referred to as contiguous gene CNVs.

The molecular mechanism of HI is apparent when the contiguous gene CNV disrupts a known HI gene. For example, in Usher syndrome type 1C, the 122-kb homozygous deletion on chromosome 11p15.1 encompasses ABCC8 (13, 19), and in the chr20p deletion syndrome, HI results from heterozygous loss of the FOXA2 gene or its regulatory elements (14). For other contiguous gene CNVs, the critical region for the HI may not have been defined (12, 20) or the deletions may be coincidental or linked to phenotypic risk factors (21).

The size of the disease-causing contiguous gene CNV will vary between individuals, with their phenotype determined by the combination of genes disrupted by the copy number change. Variable penetrance and variable expressivity have also been reported (14, 22, 23). In some children, the presence of syndromic features may raise clinical suspicion of a contiguous gene CNV, prompting genome-wide copy number analyses by karyotyping, microarray or whole genome sequencing (WGS). In contrast, children who present with HI may undergo HI genetic testing as a first-line investigation.

Many genomic laboratories perform targeted analysis of the coding regions of the known HI disease genes during routine screening. While most will employ the necessary bioinformatic pipelines to detect large duplications or deletions affecting a gene on the HI panel, they often do not screen for contiguous gene CNVs where the critical region has not been determined, as these are harder to target and analyze using conventional methods. Consequently, the overall contribution of these variants to the etiology of HI has not been established although some studies have reported the prevalence of individual contiguous gene CNVs in HI cohorts (12, 14, 20).

In this study, we assessed the overall contribution of disease-causing contiguous gene CNVs to the etiology of HI. To do this we performed copy number analysis in a large cohort of patients referred to our laboratory for routine genetic testing for HI.

Contiguous gene CNVs affecting 5 of the 6 genomic loci were identified in 53 probands: 28 were from the genetically solved cohort and 25 were from the genetically unsolved cohort ( Figure 1 ).

We detected a contiguous gene CNV affecting the X chromosome in 24 females (24/53; 45%). Microarray analysis or karyotyping had already established a diagnosis of Turner syndrome in 14 of these children prior to referral for HI genetic testing. At least 11 individuals had a deletion that encompassed KDM6A (the exact coordinates of the variants were not available for 13 individuals). We identified a terminal chr9p deletion in 15 individuals which ranged in size from ~7.2Mb to 19.4Mb. Microarray analysis had detected the deletion in 6 of these individuals prior to referral for HI genetic testing. Five patients had a chr20p11.2 deletion ranging in size from ~3Mb to 8Mb. In one child, we identified a 4.8Mb duplication on the paternal allele at the chr11p.15.5 differentially methylated region, confirming a diagnosis of Beckwith-Wiedemann syndrome (15). SNP array analysis of all 8 individuals in the solved cohort with a deletion of ABCC8 exons 1-22, demonstrated the deletion extended over exons 1-19 of the adjacent USH1C gene, confirming a diagnosis of Usher syndrome type 1c. We did not detect any CNVs encompassing the NSD1 gene at chr5q35, which would have been consistent with Sotos syndrome (15). We previously reported clinical and genetic data for 9 of the 15 individuals with a chr9p deletion and the 5 individuals with a chr20p11.2 deletion (12, 14).

Identifying 25 contiguous gene CNVs in the genetically unsolved HI cohort increased the pick-up rate for disease-causing variants from 50.9% (95% CI: 49.3-52.5%, n=1,916/3,763) to 51.6% (95% CI: 50.0-53.2%, n=1,941/3,763). These variants had an overall prevalence of 2.7% (95% CI: 2.1-3.6%, n=53/1,941) in the solved cases, which is similar to the prevalence of variants in other routinely tested HI genes ( Figure 2 ).

The median age at diagnosis of HI in the 53 individuals with a contiguous gene CNV was 1 day (range, 0 days – 6 years) and their median birthweight was +0.41 SDS (range, -2.3 to +3.1 SDS). We obtained treatment data for 43 individuals at referral for genetic testing. Among them, 35 children were treated with diazoxide, 3 with octreotide and 1 with a combination of diazoxide and octreotide. Four children with Turner syndrome were not receiving treatment at referral because their HI was in remission, and two children, both with Usher syndrome type 1C, had undergone near total pancreatectomy.

Clinicians reported additional features in 32 of the 53 individuals (60%) at the time of referral for genetic testing ( Figure 1 ). In all but two cases, at least some of the additional features reported were consistent with the genetic diagnosis, though not necessarily pathognomonic for the condition ( Figure 3 ).

The median age at referral of individuals within the solved cohort (i.e. those where the contiguous gene CNV that had already been detected prior to referral for HI testing) was 409 days (range, 11–4262 days). This was higher but not significantly different to those who underwent HI genetic testing as a first-line test (median 191 days, range 12–4065 days) (p > 0.05, Mann Whitney U).

We comprehensively screened for contiguous gene CNVs in a large cohort of patients referred for routine HI genetic testing and identified variants in 53 individuals across five distinct genetic loci. Collectively, these variants accounted for 2.7% of genetic diagnoses, a prevalence comparable to that of variants in HNF4A, HK1, and GCK (3.2%, 2.9% and 2.2% respectively; Figure 2 ); genes which are routinely included on gene panels for HI.

Twenty-one probands (40%) did not have extra-pancreatic features listed on the genetic testing request form, suggesting that HI was the presenting feature of their condition (30, 31). This finding highlights the need to screen for contiguous gene CNVs in all individuals with HI, regardless of their phenotype. Detecting these variants early is important because it informs prognosis, allowing for proper monitoring and the prompt diagnosis and medical management of additional clinical features if/when they develop.

The most common contiguous gene CNVs we identified were deletions on the X chromosome. These were identified in 24 females. In at least 11 patients, the deletion extended over the HI and Kabuki syndrome gene (KDM6A). Haploinsufficiency of this gene has been proposed as a mechanism for the increased prevalence of HI observed in girls with Turner syndrome (20). In 14 children, clinicians noted additional features prior to referral for HI genetic testing. These had prompted karyotyping or microarray analysis as a first line genetic test which confirmed Turner syndrome. Six of the 15 probands with a chr9p deletion had also undergone microarray analysis prior to referral for HI genetic testing which detected the deletion.

There are several reasons why patients who had already received a genetic diagnosis might have been sent for additional HI genetic testing. While the first papers describing the association between HI and Turner syndrome and HI and the 9p deletion syndrome were published in 1979 and 1996 respectively (32, 33), it has only been in the last few years that the role of these contiguous gene CNVs in the etiology of HI has been firmly established (12, 20). In keeping with this, 13 of the 14 individuals with a prior diagnosis of Turner syndrome were referred for HI genetic testing before the 2018 publication by Gibson et al., which described 12 girls with HI and X chromosome abnormalities (20). Similarly, all 6 patients with previously confirmed chr9p deletions were referred prior to the 2019 publication by Banerjee et al. (12 ),. In addition, it was possible that some children may have had a contiguous gene CNV and co-incidental HI (digenic disease). Because of this, clinicians may have considered comprehensive HI genetic testing necessary, particularly for those children with persistent disease that responded poorly to treatment as finding a pathogenic variant in a K-ATP channel gene for example would inform medical management (34).

We identified contiguous gene CNVs in 1.4% (95% CI: 1.1%-1.8%, n=53/3,763) of our whole HI cohort. This figure likely reflects a minimum prevalence due to several referral biases. For example, our cohort has an underrepresentation of children with Beckwith-Wiedemann syndrome as this condition is often diagnosed clinically prior to confirmatory methylation testing. We also did not search for mosaic CNVs, as these would have been called with less confidence potentially resulting in false positive calls. Therefore, it is possible that we missed some clinically relevant variants especially those causing Turner syndrome, where mosaicism has been observed (20). Finally, we limited our search to include CNVs with strong evidence for their causative role in HI, other HI disease-causing CNVs will therefore not have been detected.

In this study, we screened for contiguous gene CNVs using WGS data or off-target reads generated from routine tNGS analysis. Until recently, many laboratories have employed more cost-effective targeted gene panels or exome sequencing for screening the HI genes due to the higher costs of WGS (5). While both methods can be adapted to target these contiguous gene CNVs, their large size makes them difficult to screen for using these approaches. As the cost of WGS continues to decrease, laboratories will likely adopt it as the gold standard first-line genetic test for HI, making the detection of these contiguous gene CNVs easier. Until then, we recommend considering a CNV screen when a large deletion is detected in the ABCC8, KDM6A (females only), or FOXA2 genes, as these might indicate Usher, Turner and 20p11.2 deletion syndromes respectively.

While we have shown that our in-house software, SavvyCNV can call off-target CNVs from small tNGS panels with high precision and recall, we recognize that this methodology does not meet diagnostic standards for CNV calling (26). We used this approach in this study as it allowed for CNV analysis using existing data at no additional cost. As such, when we identified a novel contiguous gene CNV, we informed the clinicians and advised them to request orthogonal testing, which confirmed our findings in all cases.

In conclusion, we have shown that contiguous gene CNVs are an important cause of HI that should be considered by laboratories and clinicians managing children with this condition.