Abstract

Background: Aberrations in cadherin-related 23 (CDH23) account for a significant proportion of familial autosomal recessive non-syndromic hearing loss (DFNB12), a common subtype of hereditary hearing loss worldwide.

Aims: To elucidate the molecular basis and pathogenic mechanism of DFNB12 in an affected girl from a nine-member pedigree.

Study Design: Family-based genetic study with pedigree analysis.

Methods: Clinical whole-exome sequencing combined with pedigree analysis was used to identify disease-causing mutations. The potential functional consequences of these mutations were investigated using structural bioinformatic approaches, including homology modeling, molecular dynamics simulations, and other relevant tools.

Results: The proband carried compound heterozygous variants: a known pathogenic maternal variant (c.6049G > A) and a paternal haplotype comprising two linked variants (c.3262G > A and c.6911G > A), each individually classified as benign. Pedigree segregation analysis demonstrated that the paternal haplotype acts as a single pathogenic allele.

Conclusion: Two individually benign variants can combine to form a novel pathogenic haplotype (c.3262A–c.6911A). This mechanism may be under-recognized in routine variant interpretation pipelines. Our findings underscore the importance of evaluating the combined effects of linked benign variants to ensure accurate genetic counseling.

INTRODUCTION

Non-syndromic hearing loss (NSHL) is one of the most frequently observed neurosensory impairments in newborns worldwide.^1,2 Although NSHL can result from a wide range of causes, hereditary factors account for 50–60% of reported cases. Approximately 80% of NSHL cases are classified as autosomal recessive NSHL (DFNB12), which is associated with severe to profound hearing impairment.^3,4 To date, 105 loci linked to DFNB12 have been identified, involving 67 disease-causing genes (http://hereditaryhearingloss.org).^5-7 The most prevalent genes associated with hereditary hearing loss include GJB2, SLC26A4, MYO15A, OTOF, and cadherin-related 23 (CDH23).⁸

Among these, CDH23 plays an important role in DFNB12. The gene is located on the long arm of chromosome 10 (10q22.1, Gene ID 64072) and encodes CDH23, a glycoprotein involved in Ca²⁺-dependent cell–cell adhesion.^9,10 CDH23 is highly expressed in the stereocilia of inner ear sensory neurons and in retinal photoreceptor cells.^11,12 It is often located at the junction of the inner and outer segments of cilia and participates in the calyx-like process of pyramidal photoreceptors.¹³ CDH23 plays a crucial role in maintaining cell connections, mediating cell communication, and ensuring proper morphogenesis of hair bundles in inner ear neurosensory cells.¹² Disruption of hair cell stereocilia in animal models, mimicking CDH23 variants, results in noise-induced hearing loss and age-related deafness.¹⁴

Given its importance in auditory perception, deleterious mutations in CDH23 account for most hearing loss in DFNB12 patients. However, some pathogenic variants remain undiscovered.

MATERIAL AND METHODS

Patients and subjects

The study included members of a single pedigree, comprising one affected patient and eight unaffected healthy relatives. The proband was a 4-year-old girl who was first diagnosed with bilateral congenital profound hearing loss in 2019 during her initial hospital visit. The study was approved by the Ethics Committee of the Chongqing Medical University (approval number: 20200631, date: 20.05.2020). Written informed consent was obtained from all participants and their legal guardians. Clinical examinations, including auditory steady-state response (ASSR) audiogram, distortion product otoacoustic emission (DPOAE), and auditory brainstem response (ABR), were performed following standard clinical procedures in a double-blind manner.

Clinical whole-exome sequencing (cWES), Sanger sequencing, and pedigree analysis

Genomic DNA from the proband was sequenced using the BioelectronSeq4000 platform at Capitalbio Medlab (Beijing, China) according to a standard clinical whole-exome sequencing (cWES) workflow (Supplementary Figure 1). Potential pathogenic variants identified through high-throughput sequencing were validated by PCR-Sanger sequencing in the proband and her family members. The sequence-specific primers used were as follows:

• For CDH23 c.3262G > A: F, 5’-GCAGCTGCTAACACCTGTCT-3’; R, 5’-TTCAGGGATGTCCTCAGGGA-3’
• For CDH23 c.6049G > A: F, 5’-CAAGACATCTACGCCGTGGT-3’; R, 5’-GTCCAGCACTGAGAAGGGAG-3’
• For CDH23 c.6911G > A: F, 5’-CAAGCTGACTGTCAACGTCCT-3’; R, 5’-TGTAGTGACCAGCCCATTA-3’

Bioinformatics analyses of CDH23 variants

Bioinformatics analyses are described in detail in the Supplementary Materials and Methods. The physicochemical properties of the protein were characterized using multiple parameters. A linear model was then proposed to evaluate the clinical outcomes of combined variants based on the quantification of multidimensional structural characteristics, including secondary structure, flexibility, root-mean-square deviation (RMSD), isoelectric point (pI)/molecular weight, stability, surface charge distribution, hydrogen bonding, and hydrophobicity. A scoring formula quantifying the potential deleterious effects of combined mutations on molecular function was defined as follows:

The variable ai,j represents the effect of mutation on structural parameter i, which can be assigned a numeric value. The variable ki,j denotes the weight of ai,j, which is independent of other mutations and structural parameters in an additive model but dependent on them in a cooperative model. The value m indicates the number of amino acid residue substitutions, and n denotes the number of structural parameters included in the model. The clinical significance of combined mutations can be inferred by comparing their scores with those of known variants, which serve as internal references.

RESULTS

Clinical description of hearing characteristics and genetic analyses

Nine family members—four males and five females—spanning three generations were recruited for this study. The proband was first identified as unresponsive to sound at 10 months of age. ABR, ASSR, and DPOAE tests confirmed severe hearing loss in the proband, leading to a diagnosis of congenital bilateral profound deafness (Figures 1a–c). No family history of congenital hearing loss was reported in the pedigree (Figure 1d). Temporal bone computed tomography and magnetic resonance imaging were performed to exclude potential inner ear malformations.

Subsequently, standard cWES was conducted on lymphocyte DNA extracted from the proband’s peripheral blood. Compound heterozygous variants were identified in CDH23, including a maternal c.6049G > A variant and a paternal haplotype comprising c.3262G > A in linkage with c.6911G > A. All three loci were genotyped via Sanger sequencing in all nine family members (Supplementary Figure 2). The potential deleterious effects of these CDH23 variants were assessed using conventional prediction tools typically included in the cWES protocol (Supplementary Tables 1 and 2).

According to the predictions, the c.6049G > A variant was classified as pathogenic or likely pathogenic, consistent with previous clinical reports. In contrast, the c.3262G > A and c.6911G > A variants were of uncertain significance or likely benign, making it difficult to draw definitive conclusions. Notably, the clinical significance of these two variants has not been previously reported.

Analysis of genotype–phenotype correlations within the family revealed an intrafamilial segregation pattern consistent with autosomal recessive inheritance (Figure 1e). The affected individual’s clinical phenotype aligned with CDH23-related prelingual deafness. Based on cWES and pedigree analysis, the compound heterozygous variants c.6049G > A (p.Glu2017Ser) and the paternal haplotype c.3262G > A (p.Val1088Met) in linkage with c.6911G > A (p.Arg2304Gln) in CD23 were confirmed as pathogenic.

Compound heterozygous mutations in conserved regions of CDH23

The three confirmed pathogenic variants in the CDH23 gene, initially identified by cWES (Figure 2a), include c.3262G > A (p.Val1088Met), c.6049G > A (p.Glu2017Ser), and c.6911G > A (p.Arg2304Gln). All three variants are located within conserved coding regions (Figure 2b). Sanger sequencing confirmed that the proband harbors compound heterozygous alleles of c.3262G > A, c.6049G > A, and c.6911G > A. The c.6049G > A variant was inherited from the mother, whereas the paternal haplotype carrying c.3262G > A in linkage with c.6911G > A was inherited from the father (Figures 1e and 2c).

The affected amino acid residues are highly conserved across species due to their location within essential domains. Sequence alignment of Homo sapiens CDH23 with seven other species using constraint-based multiple protein alignment tool revealed strong evolutionary conservation at these positions (Figure 2d).

Bioinformatics analysis of CDH23 mutations

To explore the potential molecular basis, a series of structural bioinformatics methods were applied. First, homology modeling of CDH23 variants was performed using SWISS-MODEL, followed by structural optimization through molecular dynamics simulation. An array of structural analyses was then conducted to assess the structural implications of the mutants.

The secondary structures of CDH23 extracellular (EC) domains were analyzed using SOPMA, which indicated that the alpha-helical content decreased in the EC19 and EC21 variants (Supplementary Table 3). According to PredyFlexy, the molecular flexibility of the EC19 variant was notably increased, whereas EC10 and EC21 showed minimal changes (Supplementary Figures 3a-c). Domain stability analysis revealed a significant decrease for the EC10 (p.Val1088Met) and EC19 (p.Glu2017Ser) variants ( Supplementary Table 4). RMSD analysis showed a value of 0.254 Å for EC19, which was higher than 0.076 Å for EC10 and 0.059 Å for EC21 (Supplementary Figure 3d). These analyses provided insights into the possible structural and functional impacts of the variants.

The 3D structures of the native and mutant EC domains were examined using Swiss PDB Viewer. Compared with native structures, the mutant domains showed fluctuations that could influence molecular dynamics and function (Supplementary Figure 4). The pl of EC21 was lower than those of EC10 and EC19 (Supplementary Figure 4b). Substitution of native amino acids with mutant residues also altered charge distributions in EC10 and EC19 (Supplementary Figure 4c).

Changes in hydrogen bonding were observed in EC19 and EC21 mutants. In EC19, a hydrogen bond formed between Ser55 and Ile23 that was absent in the native domain, while in EC21, a bond appeared between Gln4 and Glu91 (Supplementary Figure 4d). Hydrophobicity analysis using ProtScale indicated alterations in EC10 (Supplementray Figure 5).

Overall, the cumulative structural scores were consistent with the pedigree analysis (Table 1; Supplementary Figures 3–5; Supplementary Tables 3 and 4).

DISCUSSION

Congenital deafness is one of the most prevalent sensory impairments in newborns, with an estimated incidence of 1–3 per 1,000 across populations, irrespective of sex or racial background.^15,16 The recurrence risk increases substantially when both parents carry the same pathogenic variants associated with hearing loss.¹⁷ Among cases of non-syndromic hearing impairment (DFNB12), CDH23 variants are the most frequently observed genetic lesions.^18,19

The confirmed novel compound heterozygous variants of CDH23, c.6049G > A (p.Glu2017Ser) and c.3262G > A (p.Val1088Met) linked with c.6911G > A (p.Arg2304Gln), jointly contribute to autosomal recessive congenital hearing loss. While c.6049G > A (p.Glu2017Ser) alone is disease-causing, our study demonstrates that c.3262G > A (p.Val1088Met) in combination with c.6911G > A (p.Arg2304Gln) represents a novel pathogenic haplotype. These findings expand the known pathogenic mutation spectrum of CDH23 and have clinical significance for genetic counseling in hereditary deafness.

In recent years, an increasing proportion of phenotypic variability may be attributed to combinations of variants at multiple loci rather than a single locus. For example, a linear cis combination of single nucleotide polymorphisms (SNPs) explained significantly more variance than the best individual SNP.²⁰ The additive effects of linked variants (haplotypes) on biomolecular functions warrant further investigation.

CDH23 is a large protein consisting of 3,354 amino acid residues and plays a critical role in stereocilia organization and hair bundle formation. Although the biological effects of accumulated mutations on protein function remain unclear, our study is the first to reveal that substitutions of amino acid residues located far apart in the primary structure can result in CDH23 dysfunction. Individual substitutions did not significantly impair the normal function of CDH23, suggesting a minimal impact on protein activity. However, the cooperative interplay between the two substitutions appears to amplify the effect on protein function. In theory, such additive or cooperative effects could substantially disrupt polypeptide folding, conformational stability, and physicochemical properties, leading to functional deficiencies. Although pedigree and structural bioinformatics analyses provide strong evidence for pathogenicity, the lack of experimental validation remains a limitation, warranting functional corroboration in future studies.

CONCLUSION

Our pedigree-based study demonstrates that the combination of two benign variants can generate a novel pathogenic variant, a type of deleterious effect that has received little attention. Unlike cis-variants in GJB2, which often remain benign, this CDH23 haplotype represents a rare case in which individually benign variants synergistically induce pathogenicity. These findings highlight the importance of evaluating complex variant interactions to ensure accurate genetic diagnosis and counseling. Structural bioinformatics can play a critical role in interpreting such complex coding-region mutations and providing mechanistic insight into their clinical significance. While our conclusions are grounded in robust genetic evidence from the pedigree, we propose that integrated structural bioinformatics modeling may offer a valuable strategy for assessing the clinical relevance of complex variants in cases where pedigree data are unavailable. This approach is also applicable for interpreting de novo mutations, which are frequently encountered in clinical genetics.

Ethics Committee Approval: The study was approved by the Ethics Committee of the Chongqing Medical University (approval number: 20200631, date: 20.05.2020). The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

Informed Consent: Written informed consent was obtained from all participants and their legal guardians.

Data Sharing Statement: The datasets analyzed during the current study are available from the corresponding author upon reasonable request.

Authorship Contributions: Concept- J.Z.; Design- J.Z.; Supervision- J.Z.; Funding- J.Z.; Materials- Z.X.T., Y.T.Z.; Data Collection or Processing- Z.X.T., Z.X.W., Y.T.Z., Y.Y.H.; Analysis and/or Interpretation- Z.X.T., Z.X.W., X.S.; Literature Review- J.Z., X.S, Y.Y.H.; Writing- Z.X.T., J.Z.; Critical Review- J.Z.

Conflict of Interest: The authors declared that they have no conflict of interest.

Financial Disclosure: This research was funded by Chongqing science and technology bureau (CSTB2022NSCQ-MSX0956).

Supplementary Table 1:
https://balkanmedicaljournal.org/img/files/supp-table-1.xls

Supplementary Tables 2-4:
 https://balkanmedicaljournal.org/img/files/5-supplement-table-2-4-y.pdf

Supplementary Figures:
 https://balkanmedicaljournal.org/img/files/supplement-figure-1-4.pdf

Supplementary Materials and Methods:
https://balkanmedicaljournal.org/img/files/5-supplementary-Materials-and-Methods-y.pdf

REFERENCES

1. Ford CL, Riggs WJ, Quigley T, et al. The natural history, clinical outcomes, and genotype-phenotype relationship of otoferlin-related hearing loss: a systematic, quantitative literature review. Hum Genet. 2023;142:1429-1449.
2. Ma D, Shen S, Gao H, et al. A novel nonsense mutation in MYO15A is associated with non-syndromic hearing loss: a case report. BMC Med Genet. 2018;19:133.
3. Li MM, Tayoun AA, DiStefano M, et al.; ACMG Professional Practice and Guidelines Committee. Electronic address: [email protected]. Clinical evaluation and etiologic diagnosis of hearing loss: a clinical practice resource of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2022;24:1392-1406.
4. Kim BJ, Kim AR, Lee C, et al. Discovery of CDH23 as a significant contributor to progressive postlingual sensorineural hearing loss in Koreans. PLoS One. 2016;11:e0165680.
5. Del Castillo I, Morín M, Domínguez-Ruiz M, Moreno-Pelayo MA. Genetic etiology of non-syndromic hearing loss in Europe. Hum Genet. 2022;141:683-696.
6. Moser T, Chen H, Kusch K, Behr R, Vona B. Gene therapy for deafness: are we there now? EMBO Mol Med. 2024;16:675-677.
7. Hu H, Zhou P, Wu J, et al. Genetic testing involving 100 common mutations for antenatal diagnosis of hereditary hearing loss in Chongqing, China. Medicine (Baltimore). 2021;100:e25647.
8. Koohiyan M, Hashemzadeh-Chaleshtori M, Salehi M, et al. A novel cadherin 23 variant for hereditary hearing loss reveals additional support for a DFNB12 nonsyndromic phenotype of CDH23. Audiol Neurootol. 2020;25:258-262. 
9. Woo HM, Park HJ, Park MH, et al. Identification of CDH23 mutations in Korean families with hearing loss by whole-exome sequencing. BMC Med Genet. 2014;15:46.
10. Sotomayor M, Gaudet R, Corey DP. Sorting out a promiscuous superfamily: towards cadherin connectomics. Trends Cell Biol. 2014;24:524-536.
11. Richardson GP, de Monvel JB, Petit C. How the genetics of deafness illuminates auditory physiology. Annu Rev Physiol. 2011;73:311-34.
12. Schietroma C, Parain K, Estivalet A, et al. Usher syndrome type 1-associated cadherins shape the photoreceptor outer segment. J Cell Biol. 2017;216:1849-1864.
13. Cosgrove D, Zallocchi M. Usher protein functions in hair cells and photoreceptors. Int J Biochem Cell Biol. 2014;46:80-89.
14. Yasuda SP, Seki Y, Suzuki S, et al. c.753A>G genome editing of a Cdh23^ahlallele delays age-related hearing loss and degeneration of cochlear hair cells in C57BL/6J mice. Hear Res. 2020;389:107926. 
15. Arnos KS, Welch KO, Tekin M, et al. A comparative analysis of the genetic epidemiology of deafness in the United States in two sets of pedigrees collected more than a century apart. Am J Hum Genet. 2008;83:200-207.
16. Ortore RP, Leone MP, Palumbo O, et al. Compound heterozygosity for OTOAtruncating variant and genomic rearrangement cause autosomal recessive sensorineural hearing loss in an Italian family. Audiol Res. 2021;11:443-451. 
17. Ma D, Zhang J, Luo C, et al. Genetic counseling for patients with nonsyndromic hearing impairment directed by gene analysis. Mol Med Rep. 2016;13:1967-1974.
18. Usami SI, Isaka Y, Miyagawa M, Nishio SY. Variants in CDH23 cause a broad spectrum of hearing loss: from non-syndromic to syndromic hearing loss as well as from congenital to age-related hearing loss. Hum Genet. 2022;141:903-914.
19. Bork JM, Peters LM, Riazuddin S, et al. Usher syndrome 1D and nonsyndromic autosomal recessive deafness DFNB12 are caused by allelic mutations of the novel cadherin-like gene CDH23. Am J Hum Genet. 2001;68:26-37.
20. Ehret GB, Lamparter D, Hoggart CJ; Genetic investigation of anthropometric traits consortium; Whittaker JC, Beckmann JS, Kutalik Z. A multi-SNP locus-association method reveals a substantial fraction of the missing heritability. Am J Hum Genet. 2012;91:863-871.