Whole-exome sequencing of de novo genetic variants in a Chinese family with a sporadic case of congenital nonsyndromic hearing loss

Patient details

The family included in this study is of Han Chinese heritage and resides in Chengdu City of Southwest China. The proband was a 22-month-old girl with NSHL who had previously been born in our hospital by spontaneous delivery at full term. Both of her parents were healthy during pregnancy. The baby failed the newborn hearing examination but no prenatal or postnatal risk factors for hearing loss were identified. Similarly, no family history of hearing abnormalities was reported. When the parents brought the 22-month-old child back to the hospital in October 2018, physical, biochemical, and otoscopic examinations were carried out. A CT scan of the temporal bones and MR analysis of the child’s head were also done to search for any organic brain lesions, and pure tone audiometry was performed in the girl and her parents.

Written informed consent was obtained from both parents for them and their daughter to participate in the study. The work was approved by the Research Ethics Committee of Sichuan Provincial People’s Hospital, School of Medicine, University of Electronic Science and Technology of China.

Whole-exome and mitochondrial DNA sequencing

Blood genomic DNA and mitochondrial DNA were extracted from all family members according to standard procedures (Abcam, Cambridge, UK) and stored in -20°C. The DNA concentration and quality were examined using a NanoDrop 2000 (Thermo, USA).

The whole-exome sequencing, the entire mitochondrial DNA and genetic variations analysis are described in our previous work.¹² The fragmented genomic DNA was enriched using a NimbleGen probe capture array SeqCap EZ Exome Kit v3.0 (Roche NimbleGen, Inc. Madison, WI). The kit using the SeqCap advanced design algorithm coupled with 2.1 million long oligonucleotide probes to achieve superior target enrichment performance, and detect genetic variants with ~98% sensitivity and 99% specificity. The enriched DNA fragments passed the qPCR test, and the size distribution and concentration of these DNA fragments were examined using the Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA). The samples were sequenced on an Illumina NovaSeq 6000 (Illumina, San Diego, CA), and two parallel reactions were performed. Raw image files were processed by the BclToFastq (Illumina) for base calling to generate the raw data. The low-quality variations were filtered out using the quality score = 20 (Q20). The sequencing reads were aligned to the NCBI human reference genome (hg19) using Burrows-Wheeler Aligner (version 0.6.2). SAMtools and Pindel were used to analyze single nucleotide polymorphisms (SNPs) and insertion/deletion of the sequence. The coding variants and CNVs were filtered out in the dbSNP135, Exome Variant Server, 1000 Genomes, and in-house database with more than 100,000 Chinese exomes (Joy Oriental Co. Beijing, China). The variants and CNVs were also searched in the Human Gene Mutation Database (HGMD), ClinVar, and the Online Mendelian Inheritance in Man database (OMIM).

The entire mitochondrial DNA was enriched by long-range PCR followed by massively parallel sequencing.

The related primers are listed in Table 1.

Sanger sequencing

Sanger sequencing was used to verify the variations of the candidate genes in the family members.¹² The primers for amplifying the targeted region of candidate genes are also shown in Table 1.

Variant functional assay

The wild-type GJB2 cDNA and GJB2 cDNA with the c.262G>C variant were amplified with the primers shown in Table 1. The HA-tagged wild-type and mutant coding sequences were inserted into the pcDNA3.1(+) vector (Invitrogen, Carlsbad, CA) using the Mut Express® II Fast Mutagenesis kit V2 (Vazyme, Nanjing, China). Human H1299 cells (ATCC, Manassas, VA) were transfected to express the vectors using the jetPRIME Transfection Kit (Polyplus, Illkirch, France) according to the manufacturer’s instructions. After 48 hours of transfection, the cells were collected for immunoblotting and immunohistochemical analysis.

Protein analysis

The online Clustal Omega and Conseq software programs were used to align the amino acid sequences in a variety of species. The Polyphen-2, SIFT and MutationTaster programs were used to predict the variants as “damaging” or “possibly damaging”. The clinical interpretation of genetic variants by the American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines was followed to classify the variants into “benign”, “likely benign”, “uncertain significance”, “likely pathogenic”, and “pathogenic”.¹³

Immunoblotting and western blot analysis

For immunoblotting analysis, the cells were lysed with RIPA buffer containing protease inhibitor cocktail (Roche Diagnostics GmgH, Mannheim, Germany). The lysate was centrifuged, collected, and boiled in SDS loading buffer. Then, the proteins were separated on 10% SDS-polyacrylamide gels. After the proteins were transferred onto polyvinylidene difluoride membranes (Millipore, USA), the membranes were blocked and incubated with rabbit polyconal anti-HA antibody (Dilution: 1:1000, Cat No.: 51064-2-AP, Proteintech, Chicago, USA) and the secondary antibodies (Dilution: 1:10000, Cat No.: BA1055, Boster Wuhan, China), and the protein bands were visualized using an HRP chemiluminescent substrate kit (Millipore) and a ChemiDoc XRS+ System (Bio-Rad Company, Berkeley, CA).

For immunohistochemical analysis, the cells were fixed with 4% paraformaldehyde in 0.1 M PBS for 20 min and then rinsed three times in PBS. Then, the coverslips were immersed in cold methanol for 15 min at -20°C. The primary antisera and dilutions were as follows: rabbit anti-HA antibody at 1:100 (Proteintech) for WT/MUT GJB2. After incubation with primary antiserum at 4°C overnight, the cells were rinsed in PBS three times before adding Alexa Fluor 488- and/or Alexa Fluor 594-conjugated secondary antibodies (Dilution: 1:500, Cat. No.: A-11008, Invitrogen). ER was stained with ER-Tracker Red at 1:2000 dilutions (Beyotime, Shanghai, China) for 10 min at room temperature. Preimmune rabbit serum was used as the primary antibody for the negative controls. The images were visualized using a Zeiss Axio Imager Z2 microscope (Carl Zeiss, Jena, Germany).

Clinical characteristics of the patient’s family

The pedigree of the family is shown in Figure 1A. The kinship connection between the proband and parents is confirmed by the exome sequence data.²¹ The proband had normal physical, biochemical and otoscopic evaluations. No abnormality was found in her cranium by MR examination or in her cochlear, vestibular, and semicircular canals by CT scan. Pure tone audiometry indicated that her left and right hearing thresholds were 78 dB and 87 dB, respectively, with severe hearing loss in both ears (Figure 1B). Since there was no family history of HL and the child’s parents had normal hearing, the affected infant is considered to be a sporadic case of NSHL.

Figure 1. Genetic characteristics of the family, the pedigree of the family (A) and audiograms for the proband (B).

The horizontal axis of the audiogram shows the tone frequency (Hz) and the vertical axis displays hearing level (dBHL). Severe hearing loss was classified as a pure-tone average between 70-95 dBHL. х, left ear, о, right ear.

Analyses of variants detected in GJB2 gene in the patient’s exome

The mitochondrial sequencing showed no NSHL-causing variants or large deletions. The exome sequences revealed no known NSHL-causing variants in the family except that the proband had a de novo heterogeneous variant c.262G>C in the GJB2 gene (MAF unknown, Table 2 and Figure 2), whereas her parents were wild type. The c.262G>C variant led to a missense variant of p.A88P, which was graded to be “damaging” with a SIFT score of 0.00 and a Polyphen-2 score of 1.00. To examine the effect of the c.262G>C variant on the protein expression level and cellular localization, we transiently transfected the GJB2 c.262G>C mutant into H1299 cells and found that the mutant was expressed weakly and displayed a punctate distribution in the cytoplasm and cytomembrane. In contrast, wild-type GJB2 was expressed robustly and was distributed mainly in the cytomembrane (Figures 3A & B). This result confirmed that the GJB2 p.A88P mutant may fail to locate into the cell membrane and subsequently reduce the formation of gap junctions in quantity.

Figure 2. DNA and protein sequence analysis of GJB2.

(A) The DNA sequence electropherograms (I1 father, I2 mother, II1 daughter) revealing wild-type sequence of the parents and de novo 262G to C transversion from their daughter (black arrow). (B) The schematic diagram of Cx26, where M1-M4 are transmembrane domains, E1-E2 are two extracellular loops, CL is intracellular loop, and NH2 and COOH is N- and C-cytoplasmatic termini respectively. The non-pathogenic c.79G>A (p.V27I), c.341A>G (p.E114G) is in both the M1 and intracellular loop. The c.262G>C (p. A88E) is in the M2 of Cx26. (C) The alignment of the Cx26 amino acid sequences among the different species. The alanine at codon 88 is highly conserved.

Figure 3. Expression of the p.A88P Cx26 mutants in cells.

The immunofluorescence staining showed the wild-type GJB2 was expressed robustly and distributed mainly in the cytomembrane, while the p.A88P mutant was expressed weakly and displayed a punctate distribution in the cytoplasm and cytomembrane (A, arrow); the ER tracker red demonstrated the cytoplasm (A). The immunoblotting confirmed the wild-type GJB2 largely localized in the cytomembrane, while the p.A88P mutants co-localized in the cytoplasm and cytomembrane (B).

Because heterozygous c.262G>C missense variants were previously found in both patients and healthy carriers in the clinic,^2,4 we rechecked the exome sequences of the family to search for any other possible genetic causes of NSHL. We failed to find any other NSHL-causing variants, but noticed that the mother was a heterozygote of GJB2 c.79G>A (p.V27I), c.341A>G (p.E114G), the father was wild type, and the affected infant was a heterozygote of c.79G>A (p.V27I) and c.341A>G (p.E114G). As mentioned before, although no significant loss of function has been detected when VG and IG gap junctions coexist with the VE and IE types, the VG and IG types have displayed a moderate deficit in biochemical coupling and reduced channel activity in vitro.¹¹ Hence, we deduced that the de novo p.A88P mutants in the infant dislocated from the cell membrane, exacerbating GJB2 loss-of-function in the context of the p.[V27I; E114G], whereas the wild-type p.A88 in her mother could compensate for the loss, thus the infant’s compound heterozygosity at p.A88P and p.[V27I; E114G] was affected while her mother is an asymptomatic carrier of p.[V27I; E114G]. This result indicates that the multiple genetic variants in GJB2 could influence protein function additively.

Analyses of de novo variants detected in the patient’s exome

As we noticed that the c.262G pathogenic variant was de novo, we examined the de novo variants in the patient’s exome. It showed that there were approximately 47,000 variants, of which 143 variants were de novo (0.34%, 143/47,000). Among these 47,000 variants, 74 (0.016%, 74/47,000) were predicted to be pathogenic or likely pathogenic. Remarkably, 23 de novo variants were predicted to be pathogenic or likely pathogenic, including 21 heterozygous and two homozygous variants. The de novo adverse variants accounted for approximately one-third (23/74) of all pathogenic or likely pathogenic variants (Table 2). Compared with the frequency of de novo variants in total being only 0.34%, the frequency of de novo adverse variants in all de novo variants reached above 16% (23/143) which is surprisingly high. The 23 de novo adverse variants were distributed in 19 different genetic areas, including 12 known genetic regions, two unclassified gene zones (LOC100509263 and LOC81691) and five other chromosome domains without defined roles (Table 2). Except for the de novo GJB2 c.262G>C variant, the other adverse variants have not yet been reported to be NSHL-causing.

Notably, eight copy number variations (CNVs) - nearly half of the infant’s 17 total adverse CNVs - were de novo. The de novo pathogenic CNVs accounted for 42% (8/19) of the total de novo adverse mutated genes (Table 2); however, the minor allele frequencies of these CNVs were below 0.05 in the Human Gene Mutation Database and our in-house database (Joy Oriental Co., Table 2). Of these CNVs, six lacked the relevant information about their function, except the exon deletion in Rho Guanine Nucleotide Exchange Factor 5 (ARHGEF5, exon 2-12, 13,006 bp) and Opsin 1, Medium Wave Sensitive 2 (OPN1MW2, exon 1-6, 13,365 bp). The ARHGEF5 and OPN1MW2 gene products are crucial proteins that transduce external environmental cues into cellular signals across the cell membrane. Indeed, most CNVs do not encode important genes related to development and are thought to be subjected to adaptation to different environments.¹⁴ Here, these recurrent de novo pathogenic CNVs in the patient remind us about the environmental influence on genetic components.

In total, four missense, four frameshift, three noncoding, two splice-site, one in-frame deletion and one stop gain variant which were predicted to be pathogenic/likely pathogenic, were de novo (Table 2). Interestingly two heterozygous pathogenic variants in mitogen-activated protein kinase 8 interacting kinases 1 and 2 (MNK1 and MNK2) were de novo: one MNK1 noncoding variant c.679 C>T in chromosome 11 (exon 12, MAF = 0.000046) and the other MNK2 missense variant c.1816 G>A in chromosome 22 (exon 6, MAF unknown). Both MNK1 and MNK2 are serine/threonine kinases from the Ca²⁺/calmodulin-dependent kinase family and take part in initiating mRNA translation in response to MAPK signaling, accordingly playing important roles concerning environmental stress and cytokines.¹⁵ Also, four de novo pathogenic variants, including one homozygous missense variant c.11801C>T (MAF = 0.016), accumulated in the cell surface-associated Mucin 4 gene (MUC4). Mucins are integral membrane glycoproteins on the cell surface. As the major constituents of mucus, mucins protect epithelial cells from outward stimuli. Additionally, two de novo heterozygous pathogenic variants c.1162-4(IVS9) insG (MAF unknown) and c.1162-5(IVS9) A>T (MAF = 0.000008) were detected in the Rad1-like checkpoint DNA exonuclease gene (RAD21). The RAD21 gene encodes the major cohesion subunit, known as the component of a heterotrimeric cell cycle checkpoint complex, regulating the segregation of sister chromatids in cell cycle progression and connecting inducible gene expression in response to diverse stimuli.¹⁶ It is assumed that the proband’s genes with de novo pathogenic variants, including the disease-causing GJB2 c.262G>C (p.A88P), were the key participants immediately linking the external stimuli and cellular signals. Therefore, we think that the causes of all these de novo adverse variants in the affected infant might.be directly linked to the fetal/maternal environmental factors.

Underlying data

Open Science Framework: Whole-exome sequencing of de novo genetic variants in a Chinese family with a sporadic case of congenital nonsyndromic hearing loss. https://doi.org/10.17605/OSF.IO/DS7TW.²²

This project contains the following underlying data:

NCBI Gene: Exome sequencing of a Chinese family with a sporadic congenital NSHL. Accession number PRJNA688744.

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).