A novel CHD7 variant disrupting acceptor splice site in a patient with mild features of CHARGE syndrome: a case report

Rapid advancement of high-throughput next generation sequencing (NGS) technologies has provided the necessary conditions for the identification of disease-causing variants in the human genome [1]. Even though the technique for detecting variants is now becoming routine, the challenge of performing a detailed analysis and determining the function of the altered gene is still open [2].

According to data from the 1000 Genomes Project, a typical human genome contains 149–182 sites with protein truncating variants, more than 10,000 sites with peptide sequence altering variants, and about 459,000–565,000 variant sites affecting known regulatory regions [3]. Previous studies have shown that many genetic diseases have been caused by coding and non-coding variants disrupting splice sites [4, 5]. Variants altering pre-mRNA splicing account for at least 10% of disease-causing gene alterations [6].

The CHD7 (Chromodomain Helicase DNA Binding Protein 7; MIM #608892; NM_017780.4) gene, which is located on chromosome 8q12, encodes protein that contains several helicase family domains [7]. Heterozygous variants within this gene are the major cause of CHARGE syndrome (MIM #214800; ORPHA: 138). This syndrome is a rare autosomal dominant disorder estimated to occur in about 1:10,000 births worldwide [8]. CHARGE syndrome is characterised by a number of different congenital anomalies, including Coloboma, Heart defects, Atresia of the choanae, Retardation of growth and development, Genital hypoplasia, Ear anomalies, and deafness [9, 10]. CHARGE syndrome shares several major findings and phenotypic overlap with Mandibulofacial dysostosis, Guion-Almeida type (MFDGA; MIM #610536; ORPHA: 79113) [11, 12] and Kabuki syndrome 1 (KABUK1; MIM #147920; ORPHA: 2322) [13], which are caused by heterozygous variants in EFTUD2 (MIM #603892) and KMT2D (MIM #602113), respectively. There is some evidence that the phenotype can be caused by a variant in the SEMA3E (Semaphorin-3E; MIM #608166) gene [14], but almost 60–70% of patients have a pathogenic variant in the CHD7 gene [15, 16]. In the CHD7 database [17, 18], 801 pathogenic variants in CHD7 are currently recorded. According to the Human Gene Mutation Database (HGMD) [19, 20], the most frequent variants in the CHD7 gene are nonsense, small deletions, and insertions. Moreover, a number of different pathogenic variants in CHD7 splice sites have also been recorded.

Variants that affect pre-mRNA splicing and eventually translated proteins have been shown to be one of the most frequent causes of hereditary disorders [4]. The characterisation of splice site variants of unknown consequence and clinical significance is critical for appropriate and comprehensive genetic testing [34]. The most common consequence of splicing variants is skipping of one or more exons, but in certain cases the activation of cryptic 5′ (donor) or 3′ (acceptor) splice site and partial or full retention of specific introns may occur [35]. However, the effects of these variants on splicing are often not pursued as a cause of disease. Since there are different consequences of these variants, they are especially difficult to predict without mRNA or cDNA analysis [5].

In this study, we presented the characterisation of a previously unreported de novo heterozygous acceptor splice site c.5535-1G > A variant in the CHD7 gene. Earlier studies of variants in the 3′ splice site have shown that transitions, as well as in this case guanine (G) substitution for adenine (A), occur much more frequently than transversions, and the most frequently mutated base of an aberrant 3′ splice site is G [36]. The patient carrying this acceptor splice site variant fulfils major diagnostic characteristics of CHARGE syndrome outlined by Blanke et al. [37] and updated by Verloes [38]: the proband has three major characteristics (unilateral choanal atresia, facial palsy, temporal bone abnormalities) and three minor characteristics (delayed puberty, delayed mylestones, cardiovascular malformation). The haploinsufficiency of the CHD7 gene is assumed to be the pathogenic mechanism of CHARGE syndrome [10]. This mechanism is supported by studies in which a heterozygous knock-out mouse model was engineered. In these studies, heterozygous Chd7 mice presented features similar to those associated with CHARGE syndrome, providing strong support to the importance of the hemizygosity of CHD7 in this syndrome [39‐41]. Heterozygous variants of the CHD7 gene play important roles in a number of vitally important processes such as embryonic development, cell cycle arrest, apoptosis, transcriptional regulation, DNA repair, replication, and recombination. Furthermore, the CHD7 protein is essential for regulation of distal genomic elements that control the expression of critical neural crest transcriptional factors. For this reason, it is hypothesised that the loss of 50% of the CHD7 protein is insufficient to maintain normal function of this gene, thus resulting in clinical manifestation [10, 16, 42].

Our further investigation of the acceptor splice site c.5535-1G > A variant detected by WES aimed to elucidate the pathogenicity of the novel variant in the CHD7 gene. In silico analysis predicted this variant to be likely pathogenic. Moreover, the sequence alignment of the CHD7 protein across seven evolutionarily distant species revealed that the region containing the variant at the acceptor splice site is highly conserved, thus suggesting that this region is fundamentally important in vertebrates and alterations could be pathogenic (Fig. 2b). Our following research was based on cDNA analysis of the CHD7 gene transcript. Unexpectedly, the data obtained on the patient’s cDNA showed that the c.5535-1G > A variant does not lead to adjacent exon skipping or intron retention; instead it activates a cryptic 3′ splice site only one nucleotide downstream from the pathogenic variant site. Eng et al. [43] presented a similar study of the acceptor splice site c.4437-1G > A variant in the ATM gene, which caused the same consequences to the cDNA sequence. In both cases, the first nucleotide of a particular exon is a G. The cryptic 3′ splice site has therefore been created using the mutated A of the intron and the G of the exon. These findings could suggest that if the original 3′ splice site is mutated the subunits of U2 spliceosomal snRNAs recognises the next AG site, which is located downstream of the original acceptor splice site. However, Moon et al. [44] reported two abnormal transcripts of the HRPT2 gene, also known as the CDC73 gene, resulting from the same c.238-1G > A variant of intron 2. The first transcript eliminated the whole exon 3, while the another lacked the first 23 nucleotides of exon 3 due to the use of a cryptic acceptor splice site in exon 3. For this reason, the exact mechanism for how one cryptic splice site is selected over another within the exon remains unclear. According to literature that we reviewed and the results of our experiment, the variant at − 1 position in the acceptor splice site and the first G nucleotide in the adjacent exon sequence is quite frequent [45, 46]. However, such variants usually disturb the reading frame by exon skipping [35] instead of the slippage of the splicing mechanism only 1 nt downstream as in our case. Moreover, the possible effect of altered splice sites is usually not based on mRNA or cDNA analysis and is predicted by computational algorithms [45, 47, 48] though the exact effect of the specific variant should be verified in functional studies [49].

Further in silico analysis showed that the activation of the cryptic 3′ splice site, including the first G nucleotide of exon 27, consequently disrupted the reading frame of the patient’s cDNA sequence, resulting in a truncated 1,869 amino acid protein NP_060250.2:p.(Gly1846Glufs*23). The original CHD7 protein (UniProtKB: Q9P2D1), which is 2,997 amino acids in length, has two N-terminal chromodomains and an SNF2 family domain that are involved in a variety of cellular processes but mainly affect the regulation of transcription. It also contains a C-terminal conserved helicase domain as well as two C-terminal BRK domains. The function of the BRK domain is still unknown (Fig. 2), but it is hypothesised that this domain may interact with chromatin components that are important and unique to higher eukaryotes [50, 51]. In our study, the truncated CHD7 protein is predicted to result in the loss of two BRK domains.

Based on these findings, our hypothesis is that the heterozygous acceptor splice site c.5535-1G > A variant in the CHD7 gene results in partial haploinsufficiency of the CHD7 gene in a patient with suspected CHARGE syndrome, because the variant causes premature termination of the translation of the mutated allele. The partial haploinsuficiency of the CHD7 gene could also clarify the mild features of CHARGE syndrome that are characteristic of our patient. The results reported here explain the pathogenicity of this novel variant and predict the possible effect on the CHD7 protein, but further proteomic experiments are needed to clarify and validate the molecular effects caused by the c.5535-1G > A splice variant on the CHD7 protein.