Five affected individuals in two unrelated families were recruited in the present study. Clinical tests were performed as routine for patients with hearing loss and additional conditions. Personal and familial medical history, including hearing loss, tinnitus, vestibular symptoms, use of aminoglycosides, and other clinical abnormalities, were collected. Written informed consents were obtained from participants or parents. This study and associated research protocols were approved by the Ethic Committee of participating Hospitals.

Genomic DNAs were extracted from peripheral blood cells. Whole-exome was captured by SureSelect Human All Exon Kit (Agilent), followed by high-throughput sequencing by HiSeq2000 sequencer (Illumina Inc.). After strict quality control, the clean reads were aligned to the human reference genome (hg19) for SNP calling and deleterious variants were predicted by multiple commonly used programs, such as MutationTaster, Polyphen-2, and SIFT, as described in previous study (Yu et al., 2016) . Detected variants with minor allele frequency (MAF) >0.001 in gnomAD or in-house Chinese Exome Database were excluded. ACMP/AMP classification of each of the variants were provided based on a standard criterion (Richards et al., 2015). Finally, potential pathogenic variants were verified by further Sanger sequencing with specific primers.

To analyze a potential splicing variant, total RNAs were isolated from peripheral blood lymphocyte cells using TRIzol and reverse-transcribed using M-MLV reverse transcriptase (Invitrogen, Beijing section). RT-PCR was performed to analyze the c.1590T>G variant-containing region of TJP2 using primers (forward: 5′-AGG​AAA​GGC​CAA​GTT​CCA​GA-3′ and reverse: 5′-GCA​TCC​TCC​CGC​ACT​AAT​CC-3′). The RT-PCR products were examined by 1.5% agarose gel electrophoresis and Sanger sequencing.

In family 1, the 6-year-old boy (Figures 1A,B) was first referred to the clinic when he was 20-month-old due to congenital hearing-loss. His hearing and speech performance became normal after receiving unilateral cochlear implantation at the age of 20 months. Physical examination showed bilateral nail anomalies in his fingers and toes. The fingernails of the first and fifth digits were absent, aplasia nails were exhibited on his index fingers, and the nails surfaces of the remaining fingers were rough and bumpy. The proband’s thumbs were fingerlike but not triphalangeal. His fifth fingers were abnormally short. X-ray analysis revealed bilateral absence of the distal phalanges of the index finger (Figures 1C,D). In his right foot, there were absent toenails on the first and second digits with severe hypoplasia toenails on the third to fifth digits. In the left foot, there were absent toenails on the digits 1–3, rudimentary and rough nails of the digits 4–5. His feet examination and roentgenograms showed bilateral hypoplasia of the distal phalanx of the fifth digits (Figures 1E,F). Audiological evaluations by behavioral audiometry, auditory brainstem response (ABR), multifrequency steady-state evoked potential (ASSR), and distorted product otoacoustic emission (DPOAE) revealed bilateral profound sensorineural hearing loss (SNHL) (Supplementary Figure S1). Temporal bone CT scan and brain MRI were normal. The 20-year-old mother (Figure 1G) with normal cognitive function also showed similar phenotypes, such as bilateral profound congenital sensorineural hearing loss and onychodystrophy (Figures 1H,I). Her toenail development was severely affected, showing no toenails at all (Figures 1J,K). There was no hearing intervention for the mother, and she communicates mainly through sign language.

In family 2 (Figure 2A), the 7-year-old girl (Figure 2B) did not pass the neonatal hearing screening. Audiological examinations by behavioral audiometry, ABR, ASSR, and DPOAE confirmed that she had bilateral profound sensorineural hearing loss. At 3 years of age, she received unilateral cochlear implantation. Her Categorical Auditory Performance (CAP) and Speech Intelligibility Rating (SIR) scores were 4 and 3 points, respectively. She was born vaginally at term (39W⁺³) after an uneventful pregnancy. There was no history of ototoxic drug use during pregnancy and no asphyxia and no history of injury at birth. Her growth measurements in terms of body weight, height, and head circumference were within normal ranges. She was not good at active expression and showed poor memory. Physical examination revealed her bilateral nail anomalies in fingers and toes (Figures 2C–F). Her fingernails in the first and fifth digits were absent, and the remaining fingers were apparently hypoplastic. Her thumbs were fingerlike but not triphalangeal; her left fifth digit was abnormally short. Her toenails were totally absent. X-ray showed normal phalanxes development. Her feet were flat (pes planus). Dental examinations showed her tooth hypoplasia. Physical development and neurological examinations by EEG, Griffith intelligence test, and behavioral evaluation were normal. Temporal bone CT scan and brain MRI did not show any abnormalities.

The 37-year-old father (Figure 2G) showed similar phenotypes as his daughter, including profound congenital SNHL and onychodystrophy (Figures 2H, I). In addition, his foot X-ray showed hypoplasia of the distal phalanx in his second right toe (Figures 2J,K). His cognitive function and other development processes were normal. The 28-year-old mother was diagnosed with profound congenital bilateral sensorineural hearing loss, infrequent epilepsy, and mild intellectual disability. Brain MRI did not show abnormalities. Both her parents communicate with sign language. Her grandparents were phenotypically normal.

Trio-WES analysis identified a known pathogenic variant (c.1516C>T; p.R506*) in the ATP6V1B2 gene (GenBank: NM_001693) from four affected individuals with hearing loss and onychodystrophy, i.e., the affected mother and son in family 1 and the affected father and daughter in family 2. This variant was previously reported as a de novo variant in multiple cases with deafness and onychodystrophy (Yuan et al., 2014) as well as multiple cases with DOORS syndrome (deafness, onychodystrophy, osteodystrophy, mental retardation, and seizures) (Beauregard-Lacroix et al., 2021). In the present two families, this LoF allele could be passed to the next generation consistent with the form of autosomal dominant disorder.

Based on ACMG/AMP classification (Richards et al., 2015), this variant can be classified to PVS1 (based on previous functional study) (Yuan et al., 2014), PP3 (predicted to be deleterious), PP5 (pathogenic in ClinVar and DVD), and PP1 (co-segregation in family members). This variant is therefore classified as pathogenic.

Further analysis of the trio-WES of the family 1 revealed that the affected mother and son also carried a variant (c.1590T>G, p.D530E; g.108844T>G; chr9:71845067T>G; Figure 3A) in the TJP2 gene (GenBank: NM_004817), a known candidate gene for intrahepatic cholestasis and sensorineural hearing loss (Kim et al., 2014) (Supplementary Table S1). The c.1590T>G variant was predicted to be disease-causing (MAF = 0 in gnomAD) by MutationTaster and several commonly used algorithms. In addition, we identified a splicing site alteration in exon 9 in the TJP2 gene (wildtype: CTGGTGGCAATGATG; mutant: CTGGTGGCAATGAGG), thereby increasing a donor site at g.108836. To confirm the predicted splicing site, RT-PCR analysis was conducted to examine the c.1590T>G-containing region using the proband’s peripheral blood lymphocyte cells and normal control sample. The 406-bp RT-PCR product as a normal control was clearly detected by electrophoresis and further confirmed by Sanger sequencing (Figures 3E,F, arrowhead in lane C), while PCR product in patient was shown as multiple bands (lane P) and no single peak in Sanger sequencing, suggesting a mixed PCR product due to a splicing event. Quantification analysis using ImageJ showed at least 2-fold higher expression of 406-bp RT-PCR product (arrowhead) in control group (lane C) than that in patient (lane P), suggesting a potential effect of mutation on the accurate splicing of pre-mRNA into mRNA (Figure 3E). This variant in ACMG/AMP can be classified to PM2 (not shown in gnomAD or ChinaMap) and PP3 (predicted to be deleterious by multiple algorithms analyses). This variant is therefore classified as a VUS (variants of uncertain significance). Clinic spectrum comparison based on current medical data between these two patients and reported cases did not show significant differences. However, clinic follow-up is required to examine if these patients develop intrahepatic cholestasis or additional conditions.

WES analysis of the affected spouse in family 2 (individual II-2, Figure 2A) identified a missense variant (c.733A>G, p.M245V in the KIF11 gene) (GenBank acc. no., NM_004523). This variant is predicted to be disease-causing and is not recorded in gnomAD or ChES database (>10,000 Han population). At least 100 different mutations of the gene have been reported in patients with autosomal dominant microcephaly with or without chorioretinopathy, lymphedema, or mental retardation (MIM: 152950 and HGMD). Among them, two missense mutations (p.S235C and p.H244Y) in KIF11 were previously identified in two unrelated patients with severe hearing loss in addition to chorioretinopathy and microcephaly (Jones et al., 2014; Mirzaa et al., 2014). Unexpectedly, we found that these two hearing-loss-associated variants and the current variant in KIF11 (Figure 4A) are all mapped to a small region within a single β-sheet (amino acids 235–245, Figure 4B) in the catalytic domain of Kinesin motor (Smart Motif: mart.embl-heidelberg.de; PDB database: 3K5E), suggesting a specific portion of the protein implicated in hearing function. According to ACMG/AMP classification, this variant can be classified to PM2 (not detected in gnomAD and ChinaMap) and PP3 (predicted to be deleterious by multiple commonly used algorithms). This variant is therefore classified as a VUS (variants of uncertain significance).