Molecular epidemiology of Chinese Han deaf patients with bi-allelic and mono-allelic GJB2 mutations

Background Recessive mutations in GJB2 is the most common cause of genetic hearing loss worldwide. The aim of this study is to determine the spectrum and frequency of GJB2 variants in Chinese Han deaf patients and to investigate the underlying causative genes in patients with mono-allelic GJB2 mutations. Methods We analyzed the mutation screening results of GJB2 in 1852 Chinese Han probands with apparently autosomal-recessive hearing loss in our laboratory. Targeted next-generation sequencing of 139 known deafness-related genes were performed in 44 probands with mono-allelic GJB2 mutations. Results Bi-allelic GJB2 mutations was identified in 25.65% of patients, in which the c.235delC (p.L79Cfs*3) mutation is the most frequent cause for both severe-to-profound (84.93%) and mild-to-moderate hearing loss (54.05%), while the c.109G > A (p.V37I) mutation is another frequent cause for mild-to-moderate hearing loss (40.54%). In 3.89% of patients only one mutant allele can be identified in GJB2. Targeted next generation sequencing in 44 such probands revealed digenic heterozygous mutations in GJB2/GJB6 and GJB2/GJB3 as the likely pathogenic mechanism in three probands. In 13 probands, on the other hand, pathogenic mutations in other deafness-associated genes (STRC, EYA1, MITF, PCDH15, USH2A, MYO15A, CDH23, OTOF, SLC26A4, SMPX, and TIMM8A) can be identified as the independent genetic cause, suggesting that the mono-allelic GJB2 mutations in those probands is likely co-incidental. Conclusions Our results demonstrated that GJB2 should be a primary target for mutation screening in Chinese Han deaf patients, and those with mono-allelic GJB2 mutations should be further screened by next generation sequencing.


Introduction
Hearing loss is a heterogeneous disorder that affects language acquisition and social skill development in children. It is estimated that 50%~60% cases of hearing loss have a genetic etiology [1]. To date, there have been over 100 genes identified to cause non-syndromic hearing loss and over 700 genetic syndromes described with features of hearing loss. Despite this, mutations in a single gene GJB2 (OMIM 121011) account for a large proportion of nonsyndromic hearing loss in most populations worldwide [2].
The GJB2 gene codes for a gap junction protein connexin-26 (Cx26), which is essential for the physiological function of supporting cells in the cochlea [3]. About 200 GJB2 pathogenic mutations have been reported so far [4]. A number of missense mutations may lead to autosomal dominant non-syndromic hearing loss DFNA3 and autosomal dominant syndromic hearing loss associated with hyperproliferative epidermal disorders [5,6]. On the other hand, a majority of GJB2 mutations are inherited in a recessive form and lead to non-syndromic hearing loss DFNB1. The mutation spectrum of GJB2 and the frequencies of these mutations vary greatly across different ethnic groups [2,7], and the Chinese population has a quite distinct spectrum of GJB2 mutations from other populations [8].
With China having approximately one fifth of the world's population, evaluating the molecular epidemiology of GJB2 mutations in Chinese deaf patients has important implications in guiding genetic testing for deafness. In the present study, we analyzed the GJB2 mutation screening results and audiometric data of 1852 Chinese Han deaf probands to determine its GJB2 mutation spectrum and genotypephenotype correlation.
In addition, previous mutation screening of GJB2 in deaf patients revealed that a substantial number of them carried only one mutant allele [2,9,10]. The allele frequency of GJB2 mutations in heterozygous patients was significantly higher than expected in the general population. Possibly other mutations, either within the DFNB1 locus or in other unlinked genes, could contribute to the hearing loss in patients with mono-allelic GJB2 mutations. To this end, this study also used targeted next-generation sequencing (NGS) to detect single nucleotide variants, small insertions and deletions (indels) and copy number variations (CNVs) of 139 known deafness-related genes in 44 patients with mono-allelic GJB2 mutations. The results would provide important information for genetic testing and counselling, especially for those with mono-allelic GJB2 mutations.

Patients
We reviewed the records of patients with sensorineural hearing loss who received genetic testing for deafness in our laboratory in Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine. Included in this study were patients with bilateral, non-syndromic, sensorineural hearing loss. A total of 1852 unrelated deaf probands, 979 males and 873 females, were analyzed for GJB2 testing and audiological examination results. The familial cases were compatible with an autosomal-recessive inheritance and the rest cases were sporadic. The ages of the subjects ranged from 2 months to 68 years, with the median age of 12 years old. All subjects were of Chinese Han ethnicity. The severity of hearing loss was classified based on the better hearing ear as mild (21~40 dB), moderate (41~70 dB), severe (71~95 dB), and profound (> 95 dB).

Ethics statement
A written informed consent was obtained from each subject or their guardians to participate in this study. This study was approved by the Ethics Committee of Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine.

Mutation analysis of the GJB2 gene
Genomic DNA was extracted from blood samples using the Genomic DNA Extraction Kit (Tiangen Biotech, Beijing, China). The GJB2 (NM_004004.5) coding exon (exon 2) and flanking regions as well as the non-coding exon 1 and its flanking splice sites were amplified by polymerase chain reaction (PCR), and the PCR product was then Sanger sequenced in both directions. Sequence data were analyzed using Sequencher 5.4.5, and primer sequences are provided in Additional file 1: Table S4.

Targeted next-generation sequencing
For library preparation, 2μg genomic DNA was randomly fragmented to 150-200 bp fragments by ultrasound shearing. End-repair, adenylation, adapter ligation and PCR amplification were completed according to the standard Illumina protocol. The amplified DNA was captured with a Deafness-related Gene Panel (WuXi NextCODE, Shanghai, China) designed to capture all exons and splicing sites of 139 deafness genes. Sequencing of the enrichment libraries were then performed on the Illumina HiSeq high-throughput platform.
The raw reads were mapped to the human reference genome (UCSC hg19), and the Sentieon software suite was used to call Single Nucleotide Variants (SNVs) and small insertions or deletions (InDels). Copy number variation detection was carried out with CNVkit [11] and ExomeDepth [12] tools that detect copy number variations based on read depth. The SNVs and InDels were annotated with an inhouse developed annotation pipeline developed by WuXi NextCODE using Variant Effect Predictor (VEP) software.

Variant filtering and interpretation
With the exception of three known common mutations in Chinese Hans, c.235delC and c.109G > A in GJB2 and c.919-2A > G in SLC26A4, an in-house Chinese Han allele frequency database was used to exclude variants with minor mutation frequency (MAF) higher than 0.005 in the general population. ClinVar, OMIM and HGMD database were used to annotate known pathogenic variants. In addition, multiple computational tools (SIFT, Polyphen2, PROVEAN, Muta-tionTaster and PANTHER) were used to predict the functionality of nonsynonymous variants. Segregation analysis was performed when DNA samples from the family members were available. The reported variants and CNVs were validated by Sanger sequencing (primer sequences for PCR amplification are provided in Additional file 1: Table S4).
When analyzing the hearing loss levels in these subjects, we found that 92.21% (438/475) of patients with bi-allelic GJB2 mutations exhibited severe-to-profound hearing loss ( Table 2). The c.235delC mutation alone, was identified in 372 (84.93%) such probands in at least one allele. On the contrary, the c.235delC and c.109G > A mutations are the major causes for the remaining 37 probands with mild-to-moderate hearing loss, accounting for 54.05% (20/37) and 40.54% (15/ 37) of probands in at least one allele, respectively.

Additional or alternative causes in patients with monoallelic GJB2 mutations
In our cohort, we also found 72 (3.89%) deaf probands carrying only a single recessive pathogenic mutation of GJB2 (Additional file 1: Table S1, the heterozygous c.109G > A variant is not included in this group due to its incomplete penetrance and high frequency in the general population of Chinese Hans [13]). To elucidate the molecular etiology of the hearing loss in probands with mono-allelic GJB2 mutations, we further sequenced 139 known deafnessrelated genes (Additional file 1: Table S2) by targeted NGS in 44 such probands with good quality and quantity of DNA samples.
Despite that genomic deletions containing GJB6 and upstream regions of GJB2 was frequently detected in several ethnic groups [14,15], such genomic deletion was not detected in our CNV analysis based on read depth of the NGS. Instead, in three probands (D592, C290 and D1028) with mono-allelic c.235delC mutation in GJB2, we identified an additional heterozygous mutation c.538C > T (p. R180*) in GJB3, c.547G > A (p.E183K) in GJB3 and c.228delG (p. L79Cfs*3) in GJB6, respectively (Table 3). These GJB2/GJB3 and GJB2/ GJB6 mutations may combine to cause hearing loss in a digenic inheritance pattern as previously reported [15,16].

Discussion
In this study, we presented an overview on the mutation spectrum of GJB2 in a large cohort (n = 1852) of patients with hearing loss in Chinese Hans. Biallelic mutations in GJB2 are responsible for up to 25.65% of patients, representing the most frequent cause for genetic hearing loss in our cohort. The most prevalent GJB2 mutations identified in this study were c.235delC and c.109G > A, accounting for 65.16 and 11.79% of the mutant alleles. Most (92.21%) of the patients with bi-allelic GJB2 mutations had severe-to-profound hearing loss, in which the c.235delC is the predominant causes (84.93%). Interestingly, our results showed that c.235delC also contributes to mild-to-moderate hearing loss in a significant percentage (54.05%) of such patients, with c.109G > A being another major contributor (40.54%, Table 2). In comparison with previous studies of other Chinese ethnicities such as the Uyghur population [28], the mutation spectrum of GJB2 is considerably different in Chinese Hans, as c.35delG, a common GJB2 mutation in both Uyghurs and Caucasians, was detected in only one proband in our cohort.
It has long been puzzling that mutation screening of GJB2 in a large proportion (6-15%) of patients with autosomal recessive hearing loss would identify only one pathogenic mutant allele [9,29,30]. In our cohort, we also identified 72 (3.89%) subjects carrying only a single recessive pathogenic mutation in GJB2, and that excludes those carrying the incompletely penetrant c.109G > A variant, which has a carrier frequency of 12.2% in Chinese Han normal hearing controls [31]. In our cohort, the carrier rate of mono-allelic mutations in GJB2 (3.89% overall, 2.97% for c.235delC) is higher than that previously reported in the Chinese Han general population (2.45% overall, 1.78% for c.235delC) [32], suggesting that at least in some cases a second unidentified pathogenic mutation may act either in cis or in trans to the GJB2 mutation to lead to the hearing loss. This hypothesis has subsequently been proved by our targeted NGS in 44 probands with mono-allelic GJB2 mutations. In three probands, digenic inheritance of GJB2/GJB3 and GJB2/GJB6 mutations was identified as the likely pathogenic cause for their hearing loss (Table 3). On the other hand, two dominant and a series of recessive mutations in 11 deafness-associated genes were also identified as the independent pathogenic causes in 13 additional probands, suggesting that those probands are coincidental carriers of the GJB2 mutations.
Overall our targeted NGS resolved the pathogenic cause in 16 (36.36%) probands with mono-allelic GJB2 mutations, validating the importance of highthroughput sequencing in such patients. For the remaining unresolved cases, possible pathogenic causes may include: 1) a second mutant allele in GJB2 may exist deeply in the introns or the noncoding regulatory regions uncovered by the targeted NGS; 2) mutation in a yet unknown deafnessassociated gene may lead to the hearing loss in coordination with or independent to the GJB2 mutation; and 3) in some sporadic cases environmental factors may contribute to the hearing loss.  Segregating with HL, determined on basis of segregation analysis in affected and/or unaffected family members