Hair cells are one of the sensory cells in the auditory epithelium of the mammalian inner ear which are necessary for transforming sound vibration and mechanical forces including gravity into neural signals that can be interpreted by the brain (Goutman et al., 2015). In mammals, there are two different hair cells classified as outer hair cells and inner hair cells. Outer hair cells that have three rows of cells can mechanically amplify sound-induced vibrations to increase the responsiveness of sensory epithelium, while inner hair cells have only one row of cells and are functional in transmitting the signals to the spiral ganglion neurons (Ma et al., 2018). Therefore, the number and function of hair cells are crucial for hearing (Liu et al., 2019; Qi et al., 2019; He et al., 2020). Hearing loss could be caused by different reasons including genetic factors, aging, infectious diseases, ototoxic drugs, and noise exposure (Zhu et al., 2018; Gao et al., 2019), among which the genetic factors contribute the most by leading to hereditary and progressive hearing loss. To date, the mechanisms of hair cell damage mainly include mechanical shearing forces and oxidative damage to hair cells, which eventually induce apoptotic cell death in hair cells (Liu et al., 2016; Ding et al., 2020; Zhou et al., 2020). The lower vertebrates have the hair cell regeneration ability though life times; while in mammals, although neonatal mice cochlea have limited hair cell regeneration ability, this regeneration ability is rapidly reduced with age (Wang et al., 2015; Zhang et al., 2020).

A typical characteristic of biological development is to separate the compositionally distinct extracellular fluids by sealing off the paracellular spaces (Madara, 1998). The essential structure of this seal formed with epithelial sheets is the continuous band-like networks of neighboring cells that are known as tight junctions (Nakano et al., 2009). Tight junctions that circumscribe the apical end of the cell can make the membranes of the adjacent cells closely connected and thereby split the cells into the basolateral and apical region, and then form a barrier to restrict paracellular permeability of ions and solutes (Dejana, 2004). In recent years, several integral membrane proteins have been identified to be involved in the function of the tight junctions and the specific expression of these proteins in different tissues may account for various “tightness” of tight junctions (Van Itallie and Anderson, 2004; Li et al., 2018). Nowadays, at least three types of transmembrane proteins, occludin, claudin family, and the junction adhesion molecule family, are reported to compose the tight junctions by interacting with the cytoskeleton and other membrane-associated proteins (Furuse et al., 1993; Martin-Padura et al., 1998; Tsukita and Furuse, 2000).

Claudins have been reported to be the primary components of the tight junctions in many tissues to form the backbone of tight junction strands (Tipsmark et al., 2008). The claudins are conserved in different species for the membrane topology with two extracellular loops, one N-terminal cytoplasmic domain and one C-terminal cytoplasmic domain (Li et al., 2011). The extracellular loops are essential for forming the paracellular barrier structure by polymerizing the intercellular claudins into strands, while the C-terminal cytoplasmic domains of claudins can coordinate with the scaffolding proteins to link the tight junctions to the actin cytoskeleton (Wen et al., 2004; Krause et al., 2008). The epithelia of mammalian nephric tubules utilize abundant tight junctions to transport ions selectively between the intercellular regions (Gong and Hou, 2017). Many different claudins are reported expressing in different segments of nephric tubules and are necessary for kidney function as selective paracellular pores or barriers to regulate the diffusion of different ions and water via tight junctions (Van Itallie et al., 2006; Angelow et al., 2007; Milatz et al., 2010; Li et al., 2011). For example, it has been reported that in the porcine kidney epithelial cell line, Claudin-7 can form a paracellular barrier to Cl-, but serve as a paracellular channel for Na⁺ (Alexandre et al., 2005). In addition, there are some claudins which are not merely expressed in tight junctions, but also in some other cellular parts play important roles in embryonic development (Gregory et al., 2001; Kollmar et al., 2001).

The zebrafish has become an excellent model organism in biological research to investigate otic vesicle and neural development which are closely related to the human diseases. The greatest advantage of using zebrafish for research is the convenience for recording and tracking of the developmental process by imaging (Downes and Granato, 2004; Bandmann and Burton, 2010). Moreover, the otic vesicle of zebrafish separates the distinct paracellular fluids compositionally and therefore becomes a great model for investigating the function of claudins and tight junctions on hearing and balance. Recently, loss of both claudin7b and claudin j was found to lead to the abnormal otolith formation and hair cell function and further cause the inner ear dysfunction in the zebrafish (Hardison et al., 2005; Li et al., 2018). However, the functions of other claudins in the inner ear remain unclear, and thus it promoted us to establish a loss of function model to study the role of claudins in hearing function.

In this study, we found that claudin h (claudin 3 in mammals) were expressed in the otic vesicle and neuromasts of zebrafish embryos by in situ hybridization. Loss of function treatments by either morpholino injection or CRISPR-cas9 could both cause hearing loss and vestibular dysfunction. We further found that these dysfunctions might be caused by abnormal otolith formation, hair cell loss, and otic vesicle morphological defects. Moreover, claudin h deficiency could induce hair cell apoptosis, which explained the decrease in number of the hair cells. In summary, our study proved that claudin h was essential for the formation of hair cell and otoliths and the normal hearing function of the inner ear in zebrafish.

It is well known that tight junction strands composed of claudins are crucial for the function of permeability barrier of the emunctory like the kidney. However, more and more studies found that many claudins, such as claudin a, claudin b, claudin j, claudin 7b, claudin 9, and claudin14, are expressed in the inner ear and necessary for its development, implying that tight junctions in the inner ear might also contain multiple claudins to tightly seal the different regions to regulate the function of the ear including hearing and balance perception during the development (Ben-Yosef et al., 2003; Hardison et al., 2005; Nakano et al., 2009; Li et al., 2018). In this study, the claudin h was also found to express in the otic vesicle and the lateral line neuromast, indicating claudin h might be also involved in the formation and function of otic vesicle and lateral line.

The development of otic vesicle and formation of otoliths were clearly impaired in the zebrafish lacking claudin h in this study. Similarly, it has been reported that though the deficiency of claudin j did not directly affect the generation of the calcium carbonate and protein complex, it remarkably prevented these crystallites from efficiently aggregating into the otoliths (Hardison et al., 2005). Moreover, claudin 7b was also reported to regulate the formation of otoliths in zebrafish (Li et al., 2018). As we know claudins are necessary for the tight junctions to separate the extracellular fluids by sealing off the paracellular spaces, and many studies have shown that altering or removing the claudins in the tight junctions can alter the ionic permeability barrier (Furuse et al., 2001; Colegio et al., 2003). During the development of zebrafish, the otic vesicle gradually develops into a well-balanced ionic and fluid-filled vesicle with compositional division of the distinct paracellular fluids with tight junctions and thereby to support the normal development of hair cells and otoliths (Li et al., 2018). Therefore, the otolith phenotype might be due to the loss of claudins which disrupted the ionic composition in the otic vesicle and further prevented the fusion of crystallites, which was indispensable for normal otolith formation (Hardison et al., 2005).

Otoliths can transmit acoustic vibrations and acceleration forces to the hair cells. The utricular otolith is indispensable for vestibular function, while saccular otoliths are only necessary for hearing (Popper and Fay, 1993; Riley and Moorman, 1999). It has been reported that in zebrafish, claudin j expressed in the otic placode and loss of function of claudin j caused severe reduction in otoliths size, insensitivity to tapping stimulus, and inability to control balance, which indicated that the hearing and vestibular function were significantly affected by the deficiency of claudin j (Hardison et al., 2005). Nowadays, eyes movement detected by VOR have been used to assess the vestibular function. In this study, significant vestibular dysfunction was also detected in the claudin h morphants by VOR, which was consistent with the loss of the utricle otolith in claudin h morphants. In addition, the startle response of zebrafish that has been used as a reliable evaluation of hearing (Yang et al., 2017) was also performed in this study, and revealed an auditory defect in the claudin h morphants (Figures 3F–H). Overall, these results clarified that the loss of claudin h could lead to the defects of otic vesicle and otolith, and further result in vestibular dysfunction and auditory handicap.

Sensory hair cells in the membranous labyrinth of otic vesicle and lateral line neuromasts in zebrafish serve as the receptors for the acoustic signals and pressure acceleration in the surrounding area which then transform these signals or mechanical pressure into chemical or electrical signals to activate auditory or vestibular circuitry (Baxendale and Whitfield, 2016). A large number of studies have shown that hair cell damage is a leading cause for hearing loss (He et al., 2016; Yu et al., 2017; Zhang et al., 2019). Therefore, investigating whether the development and formation of the hair cells were affected by the claudin h deficiency could further illustrate the potential cellular mechanisms of the vestibular and auditory dysfunction found in this study. It has been reported that claudin14 was expressed in the outer hair cells, cochlear and supporting cells in mice, and the deletion of this gene would lead to significant degeneration of both cochlear outer hair cells and inner hair cells which further cause the congenital hearing deficiency during the first 3 weeks of life (Ben-Yosef et al., 2003). Similarly, we also found the hair cell numbers were remarkably decreased in both otic vesicle and lateral line of claudin h morphants and mutants, though most of the hair cell clusters still existed. Moreover, the growth of the hair cell kinocilia was also significantly suppressed by the defects of claudin h in this study. These results indicated that the abnormal morphogenesis and formation of hair cells might be a crucial reason for the inner ear dysfunction. Nowadays, many claudin genes, such as claudin a, claudin b, claudin 7b, claudin f, and claudin j, have been reported to be expressed in otic vesicle or lateral line neuromasts during zebrafish larvae development¹ and mutation of them might cause different developmental defects. For example, claudin 7b was indispensable for the otic epithelial structure, the otolith formation and sound stimulation sensitiveness by sustaining initial integrity of otic epithelia during embryogenesis; nevertheless, it had no impact on the number and morphology of hair cells (Li et al., 2018). Moreover, although similar otolith phenotype and hearing defects were also found in the claudin j mutants, the epithelial structure of the otocyst does not seem to be protruding disrupted (Hardison et al., 2005). Taken together, these results indicated that multiple claudins are involved and properly orchestrated to regulate the development and function of the otic vesicle during zebrafish larvae development.

In this study, apoptosis analysis results showed that more cleaved caspase-3-positive cells that were colocalized with the hair cell emerged in the claudin h morphants, indicating the lack of hair cells was caused by the hair cell apoptosis. Apoptosis is one of the common pathways for cell death which was characterized by DNA degradation (Morrill and He, 2017). Nowadays, apoptosis is also considered to be one of the mechanisms for the sensory hair cell death induced by acoustic trauma and the caspase-3 activation is a key step during this process (Qian et al., 2020). As the most important components of the tight junctions, claudins play a central role in regulation of paracellular permeability and could create charge-selective channels in the paracellular space with varied combinations or expression levels of claudins (Ben-Yosef et al., 2003). It has been reported that claudin 9 is necessary to form a paracellular ion permeable barrier for Na⁺ and K⁺. Knocking out of the claudin 9 in mice would eliminate this ion barrier function and thereby increase the K⁺ concentration in the basolateral fluid of the hair cells (Nakano et al., 2009). Similarly, claudin 14 mutation was also reported to induce the hair cell degeneration by altering ionic permeability of the paracellular barrier for K⁺ to further cause human deafness (Ben-Yosef et al., 2003). Besides, longtime exposure in high concentration of K⁺ was significantly toxic to hair cells and could suppress hair cell repolarization which eventually led to the hair cells death (Hibino and Kurachi, 2006). Altogether, we hypothesize that the deficiency of claudin h in the otic vesicle might have changed the ionic composition through the disruption of paracellular ionic permeability which then induced the hair cell apoptosis and further caused the hearing loss and vestibular dysfunction. However, more details need to be confirmed in future studies. In summary, our characterization of the claudin h morphants and mutants not only revealed the biological significance of claudin h for the development and function of the hearing organs, but also provided new insights into the pathogenesis of human deafness due to loss of claudin proteins.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

The animal study was reviewed and approved by Administration Committee of Experimental Animals, Jiangsu Province, China.

DL conceived the project. JG, PQ, YH, CG, CW, CC, and HW performed most of the experiments. DL, JG, and HW analyzed the data and prepared the manuscript. All authors commented and approved the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.