Animal care and use were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the experiments received approval from the Emory University Institutional Animal Care and Use Committee. Rab11a conditional knockout mice and IFT88 conditional knockout mice were described previously (Kibar et al., 2001; Haycraft et al., 2007; Yu et al., 2014). The Looptail mouse strain with a missense mutation in Vangl2 was obtained from The Jackson Laboratory (Jax stock #000220; Bar Harbor, ME, United States). Rab11a and IFT88 conditional alleles were inactivated via Cre recombinase in the Foxg1^Cre+ mouse line as described previously (Hébert and McConnell, 2000; Pirvola et al., 2002). In this article, Rab11a^CKO/CKO indicates that Rab11a has been knocked out via Foxg1^Cre+. Similarly, Ift88^CKO/CKO indicates that Ift88 has been knocked out via Foxg1^Cre+. All Rab11a^+/+ mice in this research are littermates of mutants such as Rab11a^CKO/CKO mice.

Neonatal (P0) or postnatal day 2 (P2) C57Bl/6 mice were anesthetized on ice and disinfected with 75% alcohol and AnEr iodine. The heads of mice were removed and placed in a glass petri dish with sterile 1 × phosphate-buffered saline (PBS). The skull and brain were removed to expose the otic vesicles before fixing with 4% paraformaldehyde in PBS for 1 h at room temperature. The otic vesicles were washed with 1 × PBS three times and then stored at 4°C.

The freshly dissected otic vesicles were placed into sterile 1 × PBS. Microdissecting forceps were used to enter from the cochlea and the inner wall of the vestibule. The bone outside was removed to expose the utricle and saccule, and then the upper cap membrane was peeled off and the otoliths were cleaned carefully to obtain only the vestibular sensory epithelia.

The vestibular epithelia were incubated in blocking solution consisting of 10% donkey serum and in PBS-T (0.1% Triton X-100 in 1 × PBS) at room temperature for 1 h, followed by incubation with primary antibody in PBS-T supplemented with 5% donkey serum overnight at 4°C. After washing three times in PBS-T for 1 h each time, tissues were incubated in a solution containing secondary antibody with or without phalloidin in PBS-T supplemented with 5% donkey serum for 2 h at room temperature. Tissues were again washed three times in PBS-T, mounted in Fluoromount-G (#0100-01; SouthernBiotech, Birmingham, AL, United States) covered with 1.5-cm coverslips, and sealed.

The following primary antibodies were used: Rab11a (#2413, 1:200; Cell Signaling Technology, Danvers, MA, United States), LGN (1:200, a gift from Fumio Matsuzaki, RIKEN, Kobe, Japan) (Konno et al., 2008), β-spectrin (#612562, 1:200; BD Transduction Laboratories, Franklin Lakes, NJ, United States), acetylated tubulin (#23950, 1:400; Santa Cruz Biotechnology, Dallas, TX, United States), Prickle2 (1:500; a gift from Doris K Wu, NIDCD Laboratory of Molecular Biology, Bethesda, MD, United States), g-tubulin (#T6657, 1:200; Sigma Aldrich, St. Louis, MO, United States), and Vangl2 (#AF4815, 1:200; R&D Systems, Minneapolis, MN, United States).

The following secondary antibodies were used: donkey anti-rabbit-647 (#2492288, 1:1,000; The Jackson Laboratory), donkey anti-rabbit-555 (#2307443, 1:1,000; The Jackson Laboratory), donkey anti-mouse-647 (#2340862, 1:1,000; The Jackson Laboratory), donkey anti-mouse-555 (#2340813, 1:1,000; The Jackson Laboratory), donkey anti-sheep-488 (#2340750, 1:1,000; The Jackson Laboratory), FITC-conjugated phalloidin (#A12379, 1:1,000, Alexa Fluor; Thermo Fisher Scientific, Waltham, MA, United States), and Rho-conjugated phalloidin (#R415, 1:200, Alexa Fluor; Thermo Fisher Scientific).

The specimens were first pre-screened with an Olympus IX71 inverted fluorescence microscope. Confocal images were obtained with either a Leica TCS SP8 or Zeiss LSM510 with excitation wavelengths of 488, 543, and 633 nm.

Epithelia were washed three times in PBS for 10 min each time, and then fixed in 2.5% glutaraldehyde in 0.1 M PBS at room temperature for 30 min. Samples were then rinsed with 0.1 M cacodylate buffer followed by post-fixation with 1% osmium tetroxide in 0.1 M cacodylate for 1 h. Next, the samples were sequentially fixed in 30 and 50% ethanol for 15 min each, and 70% ethanol overnight at 4°C. The next day, the samples were fixed in 80, 90, and 100% ethanol for 15 min each. The specimens were placed into labeled microporous capsules and loaded into the sample boat of a chilled Polaron E3000 critical point drying unit (Quorum Tech, Laughton, United Kingdom). Samples were sputtered with gold for 3.5 min and scanned at 10 kV using a field emission scanning electron microscope (DS-130F; Topcon, Tokyo, Japan).

Pictures were processed using Adobe Photoshop CC2017 (San Jose, CA, United States). Statistical analyses of relevant angle data were performed using SPSS ver. 22 (IBM Corp., Armonk, NY, United States). Differences in the means between groups were analyzed using the t-test. In all analyses, P < 0.05 was taken to indicate statistical significance. GraphPad Prism6 (San Diego, CA, United States) was used to draw histograms. The rose function in MatLab software (MathWorks, Natick, MA, United States) was used to draw rose diagrams to show the polar distribution of utricle hair cells.

The kinocilium plays an important role in the development of hair cell cilia bundles. The extension and maintenance of the length of cilia relies on IFT to transport the required materials along the axons (Eatock and Songer, 2011). Specific knockout of IFT-related genes in the inner ear, including IFT88, Kif3a, and IFT20, leads to dysplasia or loss of kinocilia, and the stereocilia bundles become flattened (Jones et al., 2008; Sipe and Lu, 2011; May-Simera et al., 2015). Besides, deficiency of Ick/Cilk1, which encodes a ciliary kinase regulating IFT, results in kinocilia elongation and PCP defects including misshaping of stereocilia in the cochlea (Okamoto et al., 2017).

Knockout of Rab11 in the retinoblast cells in vitro resulted in damage to the cilia without affecting the vestibular utricle hair cells, but the number, length, and shape of the stereocilia were altered.

Cilia are rich in actin, and the protein synthesis process does not occur in ciliated axons, but kinocilia are aggregated and maintained through IFT (Bisgrove and Yost, 2006; Davis et al., 2006). In the process of cilia transport, the cell transfers various substances to the anchor point through the IFT complex, of which IFT88 is a subunit. Studies of the inner ear of IFT88/Polaris mutants have shown that after conditionally knocking out the IFT88 gene in cochlear hair cells, the position of the basal body changes. It is no longer polarized to the side of the cell but rather is present in the center of the hair cell. In some cells, kinocilia did not develop, the stereocilia were arranged in a circle, and the internal polarity of the cells was disrupted, while no morphological changes were observed in the vestibular hair cells (Jones et al., 2008).

The kinocilia of the utricle were missing or reduced in length in Foxg1^Cre+; Rab11a^fl/fl; IFT88^fl/+ mice. In addition, the phenotype of Rab11a and IFT mutant utricles included a higher proportion of abnormal cilia than the utricles of single Rab11a mutants. We concluded that the roles of Rab11a and IFT88 in the development of kinocilia are redundant, and their in stereocilia development partially overlap.

Disruption of the arrangement and polarity distribution of inner ear cilia bundles can cause hearing and balance disorders in mice and humans (Lempert et al., 1998; Littlewood Evans and Müller, 2000; Alagramam et al., 2001; Adato et al., 2005; Corey, 2009). In the inner ear sensory organs, the hair cells are arranged in a coordinated and regular manner, showing a unique PCP (Rida and Chen, 2009). The polar arrangement of hair cells plays an important role in the complete functioning of sensory organs (Jacobson et al., 2008). For the five vestibular organs, the relative arrangement of hair cells is indispensable for maintaining balance in three-dimensional space.

The PCP is reflected at three different anatomical levels in the vestibular organs. First, there is PCP at the intracellular level, the so-called intrinsic polarity of the cell. In mice, at about embryonic day 12.5, the hair cells begin to develop characteristics different from the precursor cells. A fibril appears in the center of the hair cell, and it is surrounded by micropili that develop into stereocilia. The original cilia gradually elongate into kinocilia, which is mediated by the core PCP protein. For example, core PCP protein migrates to the side of the hair cell, marking the establishment of the internal polarity of the cell (Deans, 2013). Second, there is intercellular polarity, marked by the localization of Dvl2/3, Fzd3, Pk, and Vangl2 that coordinate the arrangement of stereocilia between adjacent cells (Denman-Johnson and Forge, 1999). Our results showed that in the Rab11a^CKO/CKO utricle, the direction of the hair cells was disordered compared to the controls (Figure 4). In addition, immunofluorescence staining demonstrated changes in the expression of Vangl2 and Prickle-2 in the Rab11a^CKO/CKO utricle. Thus, Rab11a appears to affect the intercellular polarity by affecting Vangl2 and Prickle-2.

Finally, there is tissue polarity. In both the utricle and saccule, the LPR divides the sensory epithelial cells into two groups with opposite stereociliary bundle polarities that are able to detect movements in opposite directions (Deans, 2013). One hypothesis is that hair cells located on opposite sides of the LPR have unique transcriptional profiles and can respond to the PCP-based tissue polarity information in opposite manners. Emx2 is a candidate regulator for transcriptional patterning because the LPR did not form in Emx2 mutants (Holley et al., 2010). The mechanisms regulating tissue polarity and patterning of the LPR have not been elucidated. However, we observed abnormal LPR formation in the Rab11a^CKO/CKO utricle in the present study but did not investigate the underlying mechanisms.