The Dual Molecular Identity of Vestibular Kinocilia: Bridging Structural and Functional Traits of Primary and Motile Cilia

Reviewer #1 (Public review):

Summary

Xu et al. use transcriptomic comparisons of mouse cochlear and vestibular hair to show that the vestibular hair cells alone are enriched in gene expression for proteins necessary for cilia motility and to further argue that such motility is a normal function of the kinocilia.

Background:

Cilia are prominent in sensory receptors, including vertebrate photoreceptors, olfactory neurons, and mechanosensitive hair cells of the inner ear and lateral line. Cilia can be motile or nonmotile depending on their axonemal structure: motile cilia require dynein and the inner 2 singlet microtubules of the 9+2 array. Primary cilia, present early in development, are considered to have sensory functions and to be nonmotile (Mill et al., Nature Rev Gen 2023).

In hair cells, the kinocilium anchors and polarizes the mechanosensitive hair bundle of specialized microvilli. The kinocilium matures from the primary cilium of a newborn hair cell; behind it, the bundle of mechanosensory microvilli rises in a descending staircase of rows. During maturation of the mammalian cochlea, all hair cells lose the kinocilium, though not the associated basal body. The consensus for many years has been that most vertebrate kinocilia, and especially mammalian kinocilia, are nonmotile, based largely on the lack of spontaneous motility in excised mammalian vestibular organs, but also on the impression that the rare examples of spontaneous beating motility even in non-mammalian hair cells are associated with deterioration of the preparation (Rüsch & Thurm 1990).

Strengths

In comparing RNA expression across the 4 major types of mouse hair cells - 2 cochlear and 2 vestibular - Xu et al. noted that some ciliary genes related to motility are expressed by vestibular but not cochlear hair cells. They curated the ciliary genes into types known to be associated with different aspects of beating motility, and also investigated the expression of genes typical of primary cilia, which are considered to have sensory and cell signaling functions and to be nonmotile. They add immunostaining to back up some of the RNA data, and also evaluate relative expression by neonatal mouse cochlear and vestibular hair cells from a published dataset. The focus on kinociliary genes is an appropriate use of the comparative expression data for cochlear and vestibular hair cells, and the paper overall is readable and interesting. The transcriptome data are rounded off by comparing the authors' results in adult hair cells with published neonatal mouse cochlear and vestibular transcriptomes.

Weaknesses:

(1) Data:

a) The main weakness in the data is the lack of functional and anatomical data from mouse hair bundles. While the authors compensate in part for this difficulty with bullfrog crista bundles, those data are also fragmentary - one TEM and 2 exemplar videos. Much of the novelty of the EM depends on the different appearance of stretches of a single kinocilium - can we be sure of the absence of the central microtubule singlets at the ends?

b) While it was a good idea to compare ciliary motility expression in published P2 datasets for mouse cochlear and vestibular hair cells for comparison with the authors' adult hair cell data, the presentation is too superficial to assess (Figure 6C-E; text from line 336) - it is hard to see the basis for concluding that motility genes are specifically lower in P2 cochlear hair cells than vestibular hair cells. Visually, it is striking that CHCs have much darker bands for about 10 motility-related genes.

(2) Interpretation:

The authors take the view that kinociliary motility is likely to be normally present but is rare in their observations because the conditions are not right. But while others have described some (rare) kinociliary motility in fish organs (Rusch & Thurm 1990), they interpreted its occurrence as a sign of pathology. Indeed, in this paper, it is not clear, or even discussed, how kinociliary motility would help with mechanosensitivity in mature hair bundles. Rather, the presence of an autonomous rhythm would actively interfere with generating temporally faithful representations of the head motions that drive vestibular hair cells.

Could kinociliary beating play other roles, possibly during development - for example, by interacting with forming accessory structures (but see Whitfield 2020) or by activating mechanosensitivity cell-autonomously, before mature stimulation mechanisms are in place? Then a latent capacity to beat in mature vestibular hair cells might be activated by stressful conditions, as speculated regarding persistent Piezo channels that are normally silent in mature cochlear hair cells but may reappear when TMC channel gating is broken (Beurg and Fettiplace 2017). While these are highly speculative thoughts, there is a need in the paper for more nuanced consideration of whether the observed motility is normal and what good it would do.

Reviewer #2 (Public review):

Summary:

In this study, the authors compared the transcriptomes of the various types of hair cells contained in the sensory epithelia of the cochlea and vestibular organs of the mouse inner ear. The analysis of their transcriptomic data led to novel insights into the potential function of the kinocilium.

Strengths:

The novel findings for the kinocilium gene expression, along with the demonstration that some kinocilia demonstrate rhythmic beating as would be seen for known motile cilia, are fascinating. It is possible that perhaps the kinocilium, known to play a very important role in the orientation of the stereocilia, may have a gene expression pattern that is more like a primary cilium early in development and later in mature hair cells, more like a motile cilium. Since the kinocilium is retained in vestibular hair cells, it makes sense that it is playing a different role in these mature cells than its role in the cochlea.

Another major strength of this study, which cannot be overstated, is that for the transcriptome analysis, they are using mature mice. To date, there is a lot of data from many labs for embryonic and neonatal hair cells, but very little transcriptomic data on the mature hair cells. They do a nice job in presenting the differences in marker gene expression between the 4 hair cell types. This information is very useful to those labs studying regeneration or generation of hair cells from ES cell cultures. One of the biggest questions these labs confront is what type of hair cells develop in these systems. The more markers available, the better. These data will also allow researchers in the field to compare developing hair cells with mature hair cells to see what genes are only required during development and not in later functioning hair cells.

Author response:

Reviewer #1 (Public review):

Weaknesses:

(1) Data:

a) The main weakness in the data is the lack of functional and anatomical data from mouse hair bundles. While the authors compensate in part for this difficulty with bullfrog crista bundles, those data are also fragmentary - one TEM and 2 exemplar videos. Much of the novelty of the EM depends on the different appearance of stretches of a single kinocilium - can we be sure of the absence of the central microtubule singlets at the ends?

Our single-cell RNA-seq findings show that genes related to motile cilia are specifically expressed in vestibular hair cells. This has not been demonstrated before. We have also provided supporting evidence using electrophysiology and imaging from bullfrogs and mice. Although no ultrastructural images of mouse vestibular kinocilia were provided in our study, transmission electron micrograph of mouse vestibular kinocilia has been published (O’Donnell and Zheng, 2022). The mouse vestibular kinocilia have a “9+2” microtubule configuration with nine doublet microtubules surrounding two central singlet microtubules. This finding contrasts with a previous study, which demonstrated that the vestibular kinocilia from guinea pigs lack central singlet microtubules and inner dynein arms, whereas outer dynein arms and radial spokes are present (Kikuchi et al., 1989). The central pair of microtubules is absent at the end of the bullfrog saccular kinocilium (Fig. 7A). We would like to point out that the dual identity of primary and motile cilia is not just based on the TEM images. The kinocilium has long been considered a specialized cilium, and its role as a primary cilium during development has been demonstrated before (Moon et al., 2020; Shi et al., 2022).

In most motile cilia, the central pair complex (CPC) does not originate directly from the basal body; instead, it begins a short distance above the transition zone, a feature that already illustrates variation in CPC assembly across systems (Lechtreck et al., 2013). The CPC can also show variation in its spatial extent: for example, in mammalian sperm axonemes, it can terminate before reaching the distal end of the axoneme (Fawcett and Ito, 1965). In addition, CPC orientation differs across organisms: in metazoans and Trypanosoma, the CPC is fixed relative to the outer doublets, whereas in Chlamydomonas and ciliates it twists within the axoneme (Lechtreck et al., 2013). Such variation has been described in multiple motile cilia and flagella and is therefore not unique to vestibular kinocilia. What appears more unusual in our data is the organization at the distal tip, where a distinct distal head is present, similar to cilia tip morphologies recently described in human islet cells (Polino et al., 2023). Although this feature is intriguing, we interpret it primarily as a structural signature rather than as evidence for a specialized motile adaptation, and we will moderate our interpretation accordingly in the revision.

b) While it was a good idea to compare ciliary motility expression in published P2 datasets for mouse cochlear and vestibular hair cells for comparison with the authors' adult hair cell data, the presentation is too superficial to assess (Figure 6C-E; text from line 336) - it is hard to see the basis for concluding that motility genes are specifically lower in P2 cochlear hair cells than vestibular hair cells. Visually, it is striking that CHCs have much darker bands for about 10 motility-related genes.

We aimed to show that kinocilia in neonatal cochlear and vestibular hair cells are largely similar, except that neonatal cochlear hair cells lack key genes and proteins required for the motile apparatus. While these genes (e.g., Dynll1, Dynll2, Dynlrb1, Cetn2, and Mdh1) appear more highly expressed in P2 cochlear hair cells, they are not uniquely associated with the axoneme. For example, Dynll1/2 and Dynlrb1 are components of the cytoplasmic dynein-1 complex (Pfister et al., 2006), Cetn2 has multiple basic cellular functions beyond cilia (e.g., centrosome organization, DNA repair), and Mdh1 encodes a cytosolic malate dehydrogenase involved in central metabolic pathways such as the citric acid cycle and malate–aspartate shuttle. This contrasts with axonemal dyneins, which are uniquely required for cilia motility. To avoid ambiguity, we will mark such cytoplasmic or multifunctional genes with stars in both Figure 5G and Figure 6D together with legend in the revised manuscript.

Although those genes (i.e., Dynll1, Dynll2, Dynlrb1, Cetn2, and Mdh1) are highly expressed in neonatal cochlear hair cells, key genes for motile machinery are not detected. For example, Dnah6, Dnah5, and Wdr66 are not expressed in the P2 cochlear hair cells. Dnah6 and Dnah5 encode axonemal dynein and are part of inner and outer dynein arms while Wdr66 is a component of radial spokes. Importantly, we did not detect the expression of CCDC39 and CCDC40 in kinocilia of P2 cochlear hair cells. Axonemal CCDC39 and CCDC40 are the molecular rulers that organize the axonemal structure in the 96-nm repeating interactome and are required for the assembly of IDAs and N-DRC for ciliary motility (Becker-Heck et al., 2011; Merveille et al., 2011; Oda et al., 2014). We will modify Figure 6D to highlight the key difference between P2 cochlear and vestibular hair cells in the revised manuscript. We will also revise the text so that the key differences will clearly be described.

(2) Interpretation:

The authors take the view that kinociliary motility is likely to be normally present but is rare in their observations because the conditions are not right. But while others have described some (rare) kinociliary motility in fish organs (Rusch & Thurm 1990), they interpreted its occurrence as a sign of pathology. Indeed, in this paper, it is not clear, or even discussed, how kinociliary motility would help with mechanosensitivity in mature hair bundles. Rather, the presence of an autonomous rhythm would actively interfere with generating temporally faithful representations of the head motions that drive vestibular hair cells.

Spontaneous flagella-like rhythmic beating of kinocilia in vestibular HCs in frogs and eels (Flock et al., 1977; Rüsch and Thurm, 1990) and in zebrafish early otic vesicle (Stooke-Vaughan et al., 2012; Wu et al., 2011) has been reported previously. Based on Rüsch and Thurm (1990), spontaneous kinocilia motility occurred under non-physiological conditions and was interpreted as a sign of cellular deterioration rather than a normal feature. We speculate that deterioration under non-physiological conditions may lead to the disruption of lateral links between the kinocilium and the stereociliary bundle, effectively unloading the kinocilium and allowing it to move more freely. Additionally, fluctuations in intracellular ATP levels may contribute, as ciliary motility is highly ATP-dependent; when ATP is depleted, beating ceases. Similar phenomena have been documented in respiratory epithelia, where ciliary activity can temporarily pause. Nevertheless, the fact that kinocilia can exhibit spontaneous motility under these conditions indicates that they possess the motile machinery necessary for such beating. Irrespective of the condition, cilia without the molecular machinery required for motility will not be able to move.

We agree with the reviewer that, based on the present data, it is difficult to know the functional role of kinocilia and whether the presence of such autonomous rhythm would interfere with temporal fidelity. Spontaneous bundle motion, driven by the active process associated with mechanotransduction, was observed in bullfrog saccular hair cells (Benser et al., 1996; Martin et al., 2003). We will revise the discussion to clarify this important point of the reviewer. Specifically, we will emphasize that our observations of ciliary beating in the ex vivo conditions may not reflect its properties in the mature in vivo context, but rather a byproduct of motile machinery clearly present in the kinocilia. We speculate that this machinery in mature hair cells could operate in a more subtle mode—modulating the rigor state of dynein arms or related axonemal structures to influence kinociliary mechanics and, in turn, bundle stiffness in response to stimuli or signaling cues. Such a mechanism could either enhance sensitivity or introduce filtering properties, thereby contributing to the fine control of mechanosensory function without compromising temporal fidelity. Future studies using loss-of-function approach will be needed to reveal the unexplored role(s) of kinocilia for vestibular hair cells in vertebrates.

Could kinociliary beating play other roles, possibly during development - for example, by interacting with forming accessory structures (but see Whitfield 2020) or by activating mechanosensitivity cell-autonomously, before mature stimulation mechanisms are in place? Then a latent capacity to beat in mature vestibular hair cells might be activated by stressful conditions, as speculated regarding persistent Piezo channels that are normally silent in mature cochlear hair cells but may reappear when TMC channel gating is broken (Beurg and Fettiplace 2017). While these are highly speculative thoughts, there is a need in the paper for more nuanced consideration of whether the observed motility is normal and what good it would do.

We thank the reviewer for these excellent suggestions. We agree that kinociliary motility could plausibly serve roles during development, for example by guiding hair bundle formation or by contributing to early mechanosensitivity and spontaneous activity before mature stimulation mechanisms are established. It is also possible that the motility machinery represents a latent capacity in mature vestibular hair cells that could be reactivated under stress or pathological conditions. We will revise the Discussion to address these possibilities and to provide a more nuanced consideration of whether the observed motility is normal and what potential functions it might serve.

Reviewer #2 (Public review):

Summary:

In this study, the authors compared the transcriptomes of the various types of hair cells contained in the sensory epithelia of the cochlea and vestibular organs of the mouse inner ear. The analysis of their transcriptomic data led to novel insights into the potential function of the kinocilium.

Strengths:

The novel findings for the kinocilium gene expression, along with the demonstration that some kinocilia demonstrate rhythmic beating as would be seen for known motile cilia, are fascinating. It is possible that perhaps the kinocilium, known to play a very important role in the orientation of the stereocilia, may have a gene expression pattern that is more like a primary cilium early in development and later in mature hair cells, more like a motile cilium. Since the kinocilium is retained in vestibular hair cells, it makes sense that it is playing a different role in these mature cells than its role in the cochlea.

Another major strength of this study, which cannot be overstated, is that for the transcriptome analysis, they are using mature mice. To date, there is a lot of data from many labs for embryonic and neonatal hair cells, but very little transcriptomic data on the mature hair cells. They do a nice job in presenting the differences in marker gene expression between the 4 hair cell types. This information is very useful to those labs studying regeneration or generation of hair cells from ES cell cultures. One of the biggest questions these labs confront is what type of hair cells develop in these systems. The more markers available, the better. These data will also allow researchers in the field to compare developing hair cells with mature hair cells to see what genes are only required during development and not in later functioning hair cells.

We would like to thank reviewer 2 for his/her comments and hope that the datasets provided in this manuscript will be a useful resource for researchers in the auditory and vestibular neuroscience community.

Joint Recommendations:

We will make changes in the revision based on the joint recommendations of the two reviewers.

References

Becker-Heck, A., Zohn, I.E., Okabe, N., Pollock, A., Lenhart, K.B., Sullivan-Brown, J., McSheene, J., Loges, N.T., Olbrich, H., Haeffner, K., Fliegauf, M., Horvath, J., Reinhardt, R., Nielsen, K.G., Marthin, J.K., Baktai, G., Anderson, K.V., Geisler, R., Niswander, L., Omran, H., Burdine, R.D., 2011. The coiled-coil domain containing protein CCDC40 is essential for motile cilia function and left-right axis formation. Nat Genet 43, 79–84. https://doi.org/10.1038/ng.727

Benser, M.E., Marquis, R.E., Hudspeth, A.J., 1996. Rapid, Active Hair Bundle Movements in Hair Cells from the Bullfrog’s Sacculus. J. Neurosci. 16, 5629–5643. https://doi.org/10.1523/JNEUROSCI.16-18-05629.1996

Fawcett, D.W., Ito, S., 1965. The fine structure of bat spermatozoa. American Journal of Anatomy 116, 567–609. https://doi.org/10.1002/aja.1001160306

Flock, Å., Flock, B., Murray, E., 1977. Studies on the Sensory Hairs of Receptor Cells in the Inner Ear. Acta Oto-Laryngologica 83, 85–91. https://doi.org/10.3109/00016487709128817

Kikuchi, T., Takasaka, T., Tonosaki, A., Watanabe, H., 1989. Fine structure of guinea pig vestibular kinocilium. Acta Otolaryngol 108, 26–30.https://doi.org/10.3109/00016488909107388

Lechtreck, K.-F., Gould, T.J., Witman, G.B., 2013. Flagellar central pair assembly in Chlamydomonas reinhardtii. Cilia 2, 15. https://doi.org/10.1186/2046-2530-2-15

Martin, P., Bozovic, D., Choe, Y., Hudspeth, A.J., 2003. Spontaneous Oscillation by Hair Bundles of the Bullfrog’s Sacculus. J. Neurosci. 23, 4533–4548. https://doi.org/10.1523/JNEUROSCI.23-11-04533.2003

Merveille, A.-C., Davis, E.E., Becker-Heck, A., Legendre, M., Amirav, I., Bataille, G., Belmont, J., Beydon, N., Billen, F., Clément, A., Clercx, C., Coste, A., Crosbie, R., de Blic, J., Deleuze, S., Duquesnoy, P., Escalier, D., Escudier, E., Fliegauf, M., Horvath, J., Hill, K., Jorissen, M., Just, J., Kispert, A., Lathrop, M., Loges, N.T., Marthin, J.K., Momozawa, Y., Montantin, G., Nielsen, K.G., Olbrich, H., Papon, J.-F., Rayet, I., Roger, G., Schmidts, M., Tenreiro, H., Towbin, J.A., Zelenika, D., Zentgraf, H., Georges, M., Lequarré, A.-S., Katsanis, N., Omran, H., Amselem, S., 2011. CCDC39 is required for assembly of inner dynein arms and the dynein regulatory complex and for normal ciliary motility in humans and dogs. Nat Genet 43, 72–78. https://doi.org/10.1038/ng.726

Moon, K.-H., Ma, J.-H., Min, H., Koo, H., Kim, H., Ko, H.W., Bok, J., 2020. Dysregulation of sonic hedgehog signaling causes hearing loss in ciliopathy mouse models. eLife 9, e56551. https://doi.org/10.7554/eLife.56551

Oda, T., Yanagisawa, H., Kamiya, R., Kikkawa, M., 2014. A molecular ruler determines the repeat length in eukaryotic cilia and flagella. Science 346, 857–860. https://doi.org/10.1126/science.1260214

O’Donnell, J., Zheng, J., 2022. Vestibular Hair Cells Require CAMSAP3, a Microtubule Minus-End Regulator, for Formation of Normal Kinocilia. Front Cell Neurosci 16, 876805. https://doi.org/10.3389/fncel.2022.876805

Pfister, K.K., Shah, P.R., Hummerich, H., Russ, A., Cotton, J., Annuar, A.A., King, S.M., Fisher, E.M.C., 2006. Genetic Analysis of the Cytoplasmic Dynein Subunit Families. PLOS Genetics 2, e1. https://doi.org/10.1371/journal.pgen.0020001

Polino, A.J., Sviben, S., Melena, I., Piston, D.W., Hughes, J.W., 2023. Scanning electron microscopy of human islet cilia. Proceedings of the National Academy of Sciences 120, e2302624120. https://doi.org/10.1073/pnas.2302624120

Rüsch, A., Thurm, U., 1990. Spontaneous and electrically induced movements of ampullary kinocilia and stereovilli. Hearing Research 48, 247–263. https://doi.org/10.1016/0378-5955(90)90065-W

Shi, H., Wang, H., Zhang, C., Lu, Y., Yao, J., Chen, Z., Xing, G., Wei, Q., Cao, X., 2022. Mutations in OSBPL2 cause hearing loss associated with primary cilia defects via sonic hedgehog signaling [WWW Document]. https://doi.org/10.1172/jci.insight.149626

Stooke-Vaughan, G.A., Huang, P., Hammond, K.L., Schier, A.F., Whitfield, T.T., 2012. The role of hair cells, cilia and ciliary motility in otolith formation in the zebrafish otic vesicle. Development 139, 1777–1787. https://doi.org/10.1242/dev.079947

Wu, D., Freund, J.B., Fraser, S.E., Vermot, J., 2011. Mechanistic Basis of Otolith Formation during Teleost Inner Ear Development. Developmental Cell 20, 271–278. https://doi.org/10.1016/j.devcel.2010.12.00