A novel missense variant in CEACAM16 gene causes autosomal dominant nonsyndromic hearing loss

Hearing loss is the most common sensorineural disorder in humans. It is estimated that more than half of cases of hearing loss are attributable to hereditary causes. Autosomal dominant nonsyndromic hearing loss (ADNSHL) is a paradigm of genetic heterogeneity with more than 45 causative genes identified to date (http://hereditaryhearingloss.org/). These genes play an essential role in the morphology and development of hair cells, synaptic transmission of the auditory nerve, and other roles in the inner ear. For example, the transmembrane channel-like gene1 (TMC1) is specifically required for hair cell mechanoelectrical transduction and is a functionally redundant stereocilia component (Kawashima et al., 2011; Pan et al., 2013) and eyes absent 4 (EYA4), a member of the vertebrate EYA gene family of transcriptional activators, is important for the development and maturation of the organ of Corti (Wayne et al., 2001). However, many genes causing deafness remain unidentified (Ding et al., 2020; Fu et al., 2021; Qian et al., 2020; Zhang et al., 2021; Zhu et al., 2018). It is therefore important to identify these genes and their functions in the inner ear.

Carcinoembryonic antigen-related cell adhesion molecule 16 (CEACAM16, MIM 614591) is a member of the CEACAM family, which is known to play a role in tissue architecture and homeostasis, cell growth and differentiation, angiogenesis, and tumor suppression (Kuespert et al., 2006). CEACAM16 maps to the DFNA4B locus on chromosome 19q13.32 and encodes a protein of 425 amino acids. Although the specific functions of the protein encoded by CEACAM16 are still not clear, available evidence indicates that CEACAM16 is crucial for hearing maintenance as a structural component of the tectorial membrane (Kammerer et al., 2012). CAMCAM16 mutations can lead to late-onset bilateral progressive sensorineural hearing loss that begin in the first or second decade (Wang et al.,2015). To date, three CEACAM16 missense variants associated with ADNSHL have been reported (Hofrichter et al., 2015; Wang et al., 2015; Zheng et al., 2011). Additionally, Booth et al. (2018) and Dias et al. (2019) described loss of function variants in CEACAM16 that can cause autosomal recessive nonsyndromic hearing loss (ARNSHL).

In this study, we identified a novel missense variant in the CEACAM16 gene, c.763A>G; (p.Arg255Gly) in a Chinese family using targeted next-generation sequencing approach. The phenotype was consistent with ADNSHL. Further functional experiments were carried out to evaluate the pathogenesis resulting from this mutation.

The tectorial membrane is essential for normal hearing due to its exclusive physical properties and its role in modulating the flow of fluid in the subtectorial space (Nowotny & Gummer, 2011), amplifying the basilar membrane motion (Booth et al., 2019; Zwislocki & Kletsky, 1979), and shaping the cochlear tuning (Teudt & Richter, 2014). The tectorial membrane is composed of collagen and noncollagenous proteins including alpha-tectorin (TECTA), beta-tectorin (TECTB), otogelin (OTOG), otogelin-like (OTOGL), and CEACAM16. Research in both humans and animal models have shown that pathogenic variations in genes encoding for these proteins can alter the properties of the tectorial membrane and eventually lead to hearing loss (Cheatham et al., 2014; Ghaffari et al., 2013; Legan et al., 2014; Schraders et al., 2012; Yariz et al., 2012).

CEACAM16 protein is one of the proteins of the tectorial membrane which is thought to be essential for its function. As mentioned earlier, CEACAM16 is a secreted glycoprotein encoded by CEACAM16 gene and belongs to the CEACAM immunoglobulin superfamily (Zheng et al., 2011). Unlike other members of the family, CEACAM16 is well conserved in mammals and is selectively expressed in the inner ear (Kammerer et al., 2012). The structure of the CEACAM16 protein (Figure 2b) contains two Ig constant-like domains (domain A and domain B) and two Ig variable-like domains (domain N1 and domain N2) with the N- and the C-terminus separately, and lacks both a C-terminal transmembrane domain and a glycosylphosphatidylinositol anchor (Zebhauser et al., 2005). Its unique structure allows crosslink with TECTA and TECTB which involves the formation of the striated-sheet matrix, a laminated component that associates with the frequency-dependent mechanical properties of the tectorial membrane (Goodyear & Richardson, 2018; Jones et al., 2015). One study observed that the striated-sheet matrix was absent from the tectorial in the Ceacam16^βgal/βgal mice (Cheatham et al., 2014). Data from mice lacking Ceacam16 found high and low frequency hearing loss in young mice (Kammerer et al., 2012), and in humans, a similar phenotype was reported in families harboring variants in the CEACAM16 gene (Booth et al., 2018; Hofrichter et al., 2015; Wang et al., 2015; Zheng et al., 2011).

In our research, we identified a missense variant in CEACAM16 (p.Arg255Gly) which caused ADNSHL in a Chinese family. All affected individuals showed bilateral progressive sensorineural hearing loss, a very similar phenotype to the one reported in an American family (1070) and another Chinese family (SY-026) (Wang et al., 2015; Zheng et al., 2011). However, the earliest onset age of hearing loss in the family reported in our study was 7 years old (IV-3), substantially earlier than that reported in the other families (the youngest was 10 years old) (Booth et al., 2018; Hofrichter et al., 2015; Wang et al., 2015; Zheng et al., 2011). These findings suggests that the onset of hearing loss caused by the CEACAM16 variants can occur within the first decade of life in humans. These findings appear consistent with those seen in the Ceacam16 null mouse, which also displayed hearing loss at a young age (∼4 weeks old) (Kammerer et al., 2012). Based on our bioinformatic analysis, the p.Arg255Gly variant cosegregated with the phenotype of DFNA4B in the family and was absent from multiple population databases (Table 1) and the 200 normal hearing controls cohort. The arginine residue at 255 in CEACAM16 was not found to be highly conserved across species (lack of conservation in dogs) (Figure 2A). The p.Arg255Gly variant is located within the Ig constant-like domain B (Figure 2B), which is crucial for stabilizing the Ig-like conformation and increasing the affinity of the homophilic binding (Athanasia & Andreas, 2014; Bonsor et al., 2015). Therefore, we speculate that this novel missense variant might damage the structure and function of CEACAM16.

Transfection of the mutant protein (p.Arg255Gly) in HEK293T cells, visualized using immunofluorescence, suggest similar intracellular and extracellular protein expression between mutant and wild type proteins. These results are in agreement with our findings using Western blot, and suggest that the p.Arg255Gly variant does not interfere with the normal subcellular localization of CEACAM16. Our findings are also consistent with those reported by Wang et al. (2015) and Zheng et al. (2011). Furthermore, we performed quantitative analysis of the mutant proteins extracted from the cell lysate and culture medium by Western blot and ELISA, respectively. Interestingly, in both cases the amount of mutant protein was higher than that of WT, indicating that the new variant in CEACAM16 increases the secretion of the mutant CEACAM16 protein, and suggesting it is a gain-of-function mutation. Subsequently, in order to investigate whether the variant p.Arg255Gly affected protein expression at the transcription level, we used PCR to detect the relative expression l CEACAM16 mRNA level in HEK293T cells. In contrast to the Western blot results, PCR results indicated that mRNA expression levels of mutant CEACAM16 were significantly lower than those of the WT group. There may be several reasons for this inconsistency. First, the difference in plasmid transfection efficiency of transiently transfected HEK293T cells may have masked the true mRNA levels of the mutant gene. Second, according to the research of Perl et al. (2017), protein levels are more conserved than mRNA levels in all datasets, and changes in transcription are associated with translational changes that exert opposite effects on the protein level. Third, some other factors may influence the correlation between mRNA and protein expression, such as mRNA degradation rate, variation of mRNA secondary structure, among others (Guo et al., 2008). To date, the specific function of the protein encoded by CEACAM16, as well as the pathogenetic mechanisms underlying hearing impairment of patients harboring variants in this gene, remain unclear. In previous studies, a missense variant (p.Thr140Pro) in CEACAM16 was identified in an ADNSHL pedigree and was predicted to disrupt a glycosylation site, interfering with protein stability (Zheng et al., 2011). In another family with hearing loss, the p.Gly169Arg variant at Ig-C like domain A was assumed to change the spatial structure of CEACAM16 and decrease stability and the polymerizing ability of the protein (Wang et al., 2015). However, these studies lack in vivo evidence to help validate their assumptions. To the best of our knowledge, CEACAM16 is a secreted glycoprotein with a signal peptide at the N terminus; guided by the signal peptide, the newly synthesized glycoprotein is translocated to the endoplasmic reticulum and then secreted (Blobel & Dobberstein, 1975). According to Kammerer et al. (2012), CEACAM16 is located in Deiters and interdental cells of the organ of Corti in the inner ear of mice, and released via the Deiters cell's projections, after which is incorporated into the matrix of the tectorial membrane. In the tectorial membrane, CEACAM16 interacts with TECTA and TECTB via its two Ig-V like domains where it forms the dark and light zone of the striated-sheet matrix (Goodyear & Richardson, 2018). The p.Arg255Gly mutation we report occurs within the Ig-C like domain B of CEACAM16. This Ig-C like domain contains conserved cysteine residues that stabilize the Ig-like conformation by forming a disulphide bridge (Bork et al., 1994; Williams & Barclay, 1988), and plays a key role in CEACAM16 dimer formation (Kammerer et al., 2012). By comparing the properties of arginine and glycine, we know that the mutant residue is smaller and more hydrophobic than the wild-type residue, and is neutral in contrast to the wild-type residue, which is negatively charged. This mutation introduces an amino acid with distinct properties, which can disturb the Ig-C like domain B and decrease protein stability (Cheng et al., 2006; Venselaar et al., 2010). Therefore, we speculate that the variant p.Arg255Gly might cause structural changes in the CEACAM16 protein, altering its intracellular synthesis and transport pathways, and leading to increased secretion. At the same time, due to the decreased stability of the extracellular protein, the mutation might interfere with CEACAM16 dimerization and affect the interaction between CEACAM16 and other proteins. This alteration might also compromise the physical properties of the tectorial membrane and eventually lead to hearing loss. This study also has some limitations. We used CEACAM16 antibody commercial kits and were unable to detect oligomers using Western blot. Further vivo experiments are necessary to investigate the pathophysiological significance of CEACAM16.

In summary, we implicate a novel variant in CEACAM16 as the genetic cause of progressive hearing loss in a large Chinese pedigree. We suggest that this variant leads to increased secretion of mutant CEACAM16 protein in vitro, uncovering a putative novel mechanism underlying DFNA4B-related hearing loss.

The authors declare no conflict of interest.

Dejun Zhang, Shasha Huang, and Pu Dai carried out the studies, participated in collecting data, and drafted the manuscript. Jie Wu, Yongyi Yuan, and Xiaohong Li performed the statistical analysis and participated in its design. Xue Gao, Mingyu Han, and Song Gao helped to draft the manuscript and performed the data analysis. All authors read and approved the final manuscript.