Introduction

Sensorineural hearing loss (SNHL) is the most common sensory disorder in humans [1]. Genetic deficits, structural anomalies, some viral infections, and teratogens account for the etiology of Congenital hearing loss (HL) [2,3,4]. Approximately 1 to 2 out of every 1000 live births are affected by congenital HL. About 67% O HLs etiology attributed to genetic defects [5]. HL is classified into two types: syndromic, 30% of the cases in which, HL occurs as a feature of a syndrome, in more than 400 syndromes, and non-syndromic which contains 70 of the remaining cases [6].

To date, mutations in more than 150 genes have been identified to cause non-syndromic HL [7]. Inheritance modes of SNHL include: autosomal dominant (DFNA, about 17%), autosomal recessive (DFNB, about 80%), X-linked (DFNX, about 1–2%), and mitochondrial (about 1%). Therefore, the autosomal recessive mode of inheritance is the most common cause of the HL among studied cases [8]. Autosomal recessive SNHL (ARSNHL) can occur as a result of homozygous and compound or double heterozygous mutations in some genes [9]. Mutations in GJB2 and GJB6 gene have been identified as the prevalent causes of SNHL in many studied populations [10,11,12]. To date, there is no formal report of GJB6 mutations from Iran, while GJB2 mutations are frequently reported from different parts of Iran among different ethnics and races [13, 14].

At the molecular level, SNHL is the result of defects in genes with diverse functions such as those encoding: transcription factors, POU3F4 and ESRRB [15], gap junction proteins GJB2 and GJB6), motor proteins, MYO6A, MYO7A [16], cell–cell adhesion molecules, CHD23, anion transporters, integral membrane proteins, TMIE, TMC1, extracellular matrix proteins TECTA, OTOA [17, 18], signaling molecules, PJVK[19], methyl transferase LRTOMT [20], serine protease, TMPRSS3, and cytoskeleton proteins such as TRIOBP.

CDH23 is an atypical member of cadherin superfamily; its encoding gene (CHD23 (NM_022124) located on 10q22.1 and includes 70 exons. CDH23 is a calcium dependent cell–cell adherent glycoprotein with a long extracellular tail. It plays essential roles in stereocilia organization and inner ear function. Mutations in CDH23 can affect mechano-transduction and tip links of sensory hair cells in mice [21,22,23].

OTOGL, which is located on human chromosome long arm 12 (12q.21.31( encodes a protein containing 2344 amino acids known as Otogelin-like protein (OTOGL) [24, 25]. OTOGL is a component of tectorial membrane that is expressed in the inner ear of vertebrates in high levels in embryonic stages. OTOGL and Stereocilin interactions are essential in the development of outer hair cells of the cochlea, vestibular system, and inner ear cells, mutations in which would lead to congenital mild-to-moderate hearing loss. Knockdown of OTOGL in zebra fish has been reported to result in moderate SNHL [26,27,28,29]. In human, mutations in different parts of the OTOGL have been reported as the cause of mild-to-moderate non-progressive autosomal recessive HL DFNB84B [30] and retinal disease [31].

The next-generation sequencing (NGS) is a high-throughput technology, that helps the researchers discover the genetic etiology of human diseases and genome variants, especially in the case of complex and high heterogeneous disorders such as HL [32, 33] and intellectual disability [34]. Due to high locus heterogeneity of HLs, whole exome sequencing (WES) as the most common approach of NGS was carried out to define the genetic etiology of a profound congenital form of HL in one single extended Iranian family.

Material and methods

Subjects, DNA extraction, GJB2 analysis

This study was approved by the Ethics Committee of the Isfahan University of Medical Sciences (IR.MUI.RESEARCH.REC.1397.109).

Two patients suffering from two forms of ARNSHL including congenital profound (IV3) and moderate (III3) HL and normal members of a large inbreed Iranian family with multi consanguineous marriages from Isfahan, center of Iran, were recruited to examine for HL and retinopathy and evaluation of thyroid function. Genetic counseling was done and family information was obtained, pedigree was drawn by our genetic counselor (Fig. 1).

Fig. 1
figure 1

Family pedigree: cosegregation for the CDH23 novel variant is shown by red and for the novel OTOGL variant is shown by green (Existence or lack of both variants are shown by + or − symbols in each person, respectively.)

Full size image

After taking informed consent, 5 ml peripheral blood sample was taken from all the consent members of the family and conserved in EDTA -continuing tubes. DNA was extracted from peripheral blood samples by a DNA extraction kit (BioGenet Korea). Purity and concentration of DNA samples were determined using 1% agarose gel and Nanospec cube biophotometer (Nanodrop 2000 Thermoscientific, USA), respectively. The GJB2 gene was amplified using our previous studied primers [35]. Sanger sequencing was done to find the probable mutations in the GJB2 gene by comparing the sequences results with the genomic reference sequence, NG_008358.1.

Whole exome sequencing, bioinformatics, and in silico ANALYSIS

DNA ample was sent to Macrogen (South Korea) (www.macrogen.com) for WES using Novaseq 6000 platform (Illumina San Diego, CA, USA) with 151 bp paired end reads.

In this procedure, DNA was fragmented to create DNA library suite Illumina platform, then fragments were captured to cover all exons, exon/intron junctions, splice sites and flanking intron sequences of all genes (Agilent Sure Select V6 post). All of the fragments were amplified and sequencing was done (mean depth of coverage 100 X for more than 92% 0f the sequences). For bioinformatics analysis, FASTQ file was produced by converting the raw data. In silico, analysis was performed by GATK (Genome Analysis Toolkit) (https://gatk.broadinstitute.org/) and variant calling by BWA (Burrows-Wheeler Aligner) (http://bio-bwa.sourceforge.net/).

In this study, genome alignment and variant detection was based on genome assembly hg19 (hg19, NCBI Build 37), and Picard algorithm was used to mark duplicate reads. Variant filtering was done based on homozygous missense start codon change, splice site, nonsense, stop loss, and indel variants with MAF < 1% in population- based databases such as: db SNP version 147, 1000 genomes project phase 3 database (https://www.internationalgenome.org/), NHLBI GO exome sequencing project (ESP) (https://evs.gs.washington.edu/), exome aggregation consortium (ExAC) (http://exac.broadinstitute.org), and Iranome (http://www.iranome.ir/) and our local exome database (Named GTaC) containing about 1500 exome sequenced samples.

Then, evaluation of the reported variants was done using the following online in silico tools; PROVEAN (http://provean.jcvi.org/), PANTHER (http://www.pantherdb.org/), mutation Taster (http://www.mutationtaster.org/), SIFT (https://sift.bii.a-star.edu.sg/), CADD (https://cadd.gs.washington.edu/) were used to predict their deleterious effect on protein function. In addition, the identified variants were assessed by Human Splicing Finder (HSF) version 3.1 (http://www.ubm.be/hsf/) to predict the effects of them on splice phenomenon. Variant interpretation was accomplished based on the American college of medical genetics (ACMG) (www.acmg.net) guidelines and variant description was based on the Human genome variation society (HGVS) (www.hgvs.org) recommendations.

Variants confirmation

To confirm the WES results and co-segregation analysis, primers were designed by the online web site (http://primer3.ut.ee/), encompassing at least 60 flanking nucleotides of the identified regions of the CDH23 gene (Table 1). Standard polymerase chain reaction (PCR) was applied to amplify the target regions detected by WES. Sanger sequencing was done for the products of PCR.

Table 1 Primer sequences for CDH23 and OTOGL identified variants
Full size table

Results

Clinical findings

There were two affected individuals in the family: a 6-years-old boy (IV3), and a 36-years-old female (III3), that she was his aunt (Fig. 1). The proband was the only offspring of a first cousin consanguineous marriage, which suffered from congenital profound HL. HL in the second patient showed a moderate phenotype, which was different from HL phenotype in the proband. Evaluation of the clinical history showed no other complications in both patients. Retinal examination and the thyroid hormone analysis were performed to relate the HL to Usher or Pendered syndromes, respectively, but the results were normal. Non-syndromic hearing loss was suggested for genetic testing.

WES, Co-segregation analysis

Due to the negative results for GJB2 mutation, WES was applied to define the genetic etiology of HL in the family. Two novel variants were detected in the proband (a homozygous missense variant, c.2961 T > G (NM_022124) in exon 26 of the CDH23 gene, and a heterozygous Splice site variant c.1863-1G > T in OTOGL, (NM_173591)). To confirm the WES results Sanger sequencing was carried out for affected and non-affected individuals of family. The c.2961 T > G variant was detected in the father and mother of the proband as heterozygous state. Proband and his mother were heterozygous for the identified variant in OTOGL too, while the father was homozygous for the wild type variant (Figs. 2 and 3). The second affected member (the proband's aunt) refused to participate in our more examinations.

Fig. 2
figure 2

Heterozygous stat of c.2961T > G in the parents in the CDH23 gene (above), homozygous state of the variant in proband (below)

Full size image
Fig. 3
figure 3

Heterozygous stat of c.1863-1G > T in the proband and his mother in the OTOGL gene (above), homozygous state of the wild type variant in father (below)

Full size image

Bioinformatics analysis and variant interpretations

Deleterious effect of the two identified variants was approved by Bioinformatics tools (Table 2). The effect of variant on splicing was evaluated by HSF (http://umd.be/Redirect.html), and alteration of wild type acceptor site was predicted by the software (Fig. 4). Furthermore, both detected variants have not been reported in the literatures (1000 genomes project phase 3, NHLBI GO ESP, dbSNP version 147, HGMD, ExAC, Clinvar databases and Iranome). Based on the American College of Medical Genetics and Genomics (ACMG) guidelines, c.2961T > G in CDH23 was categorized as likely pathogenic.

Table 2 In silico analysis of the identified variants in CDH23 and OTGL
Full size table
Fig. 4
figure 4

HSF prediction for c.1863-1G > T variant in OTOGL

Full size image

Discussions

So far, mutations in more than 150 genes have been reported to cause autosomal recessive non-syndromic sensorineural hearing loss (ARNSNHL). After intellectual disability, non-syndromic hearing loss (NSNHL) is the second cause of inherited disabilities in Iran [36, 37]. Mutations in CDH23 lead to the Usher syndrome type1 (UHS1) or ARNSNHL, type DFNB12 depend on the mutation site and type [38, 39]. Mutations in OTOGL can lead to moderate symmetric, non-progressive form of NSNHL, type DFNB84A [40].

CDH23 protein contains 27 extracellular domains, there are some calcium binding sites (LDRE, DXNDN, and DXD) in these domains, which play essential roles in cell–cell adhesion in the hair cells of the cochlea [41]. Therefore, mutations in the extracellular domains of CDH23 and/or structural changes in these sites can affect calcium binding affinity and protein function. The effect of CDH23: c.2961 T > G variant on amino acid sequence is to change aspartate to glutamate at the position 987 of CDH23 (p. Asp987Glu); this change is occurred in one of the calcium binding sites of the CDH23 protein. It is implied that in the heterozygous state inheritance of the change cannot affect the protein function; so, it cannot lead to hearing loss in the father and the mother of the proband. The homozygous inheritance of the mutation interrupts the function of CDH23 in the proband, and it is the cause of HL due to the loss of function mechanism.

Although, future functional studies are necessary to reveal the real effect of the CDH23:c.2961 T > G variant on protein structure and/or function, mutations in calcium binding sites of the CDH23 lead to ARNSNHL, by disruption of protein function or affecting the calcium-binding affinity in human and mice [42, 43]. Based on the providing reported variant; c.2959G > A (p.Asp987Asn) in the same position and c.2956A > T (p.Lys986Ter) in the same domain of the protein have been led to non-syndromic hearing loss but not the Usher syndrome [38].

The identified variant c.1863-1G > T in the OTOGL gene is a frameshift variant, which interrupts the splicing phenomenon through changing the wild type acceptor site at the junction of intron 18/exon 19. This may lead to a truncated protein, by exon skipping mechanism or longer non-functionally protein, as a result of intron remaining. This variant has been investigated by ET Lim et al., in association with autism spectrum disorders [44].

There are some reports of mutations in OTOGL associated with involvement of visual system [45], but in the current study visual impairment was not observed in the family. In this study, the heterozygous variant in the OTOGL gene was not responsible for the phenotype of the proband. However, it might be the disease etiology in the proband’s aunt but the blood sample was not available for further analysis. Mutations in CDH23 and OTOGL genes are responsible for two different forms of ARNSHL. Also, in some cases of ARNSHL di-genic inheritance of defected alleles (double heterozygosis state) have been led to deafness [35]. There is no evidence for di-genic inheritance of OTOGL/CDH23 genes to cause hearing loss. In this study, the proband’s healthy mother completely rules out the possibility of di-genic inheritance of the cited two variants leading to deafness.

Conclusions

In this study, we reported two novel variants, a homozygous likely pathogenic in the CDH23 gene and a heterozygous splice site in the OTOGL gene.

This study confirmed no ARNSHL phenotype due to double heterozygous for the two mentioned variants. Also, the function of the CDH23 protein is not interrupted in the optical system as a result of homozygous CDH23: c.2961 T > G variant, thereby ruling out the Usher syndrome. Furthermore, double heterozygosis state of the CDH23: c.2961 T > G/OTOGL; c.1863-1G > T variants cannot interrupt the function of optical system. The findings highlighted the importance of genetic counseling and NGS technology in diagnosis of heterogeneous hereditary disorders such as HL.