Daple deficiency causes hearing loss in adult mice by inducing defects in cochlear stereocilia and apical microtubules

The V-shaped arrangement of hair bundles on cochlear hair cells is critical for auditory sensing. However, regulation of hair bundle arrangements has not been fully understood. Recently, defects in hair bundle arrangement were reported in postnatal Dishevelled-associating protein (ccdc88c, alias Daple)-deficient mice. In the present study, we found that adult Daple−/− mice exhibited hearing disturbances over a broad frequency range through auditory brainstem response testing. Consistently, distorted patterns of hair bundles were detected in almost all regions, more typically in the basal region of the cochlear duct. In adult Daple−/− mice, apical microtubules were irregularly aggregated, and the number of microtubules attached to plasma membranes was decreased. Similar phenotypes were manifested upon nocodazole treatment in a wild type cochlea culture without affecting the microtubule structure of the kinocilium. These results indicate critical role of Daple in hair bundle arrangement through the orchestration of apical microtubule distribution, and thereby in hearing, especially at high frequencies.

www.nature.com/scientificreports/ in HCs in the OC 9 , suggesting the involvement of microtubules in the formation of apical structures in the HCs of the OC in the context of PCP. However, the role of apical microtubules in apical morphogenesis in HCs of the OC remains to be elucidated. In this study, we analyzed Daple-deficient mice, from neonatal to adult stages, to determine the role of Daple in HC apical morphogenesis, especially via microtubules. We show the presence of hearing disturbances at all frequencies examined using the auditory brainstem response (ABR) test, especially at high frequencies. Reflecting the ABR results, malformation of hair bundles was found to be more severe in the basal area, indicating that Daple plays a consistent role in the cochlea for hearing. Our findings also unravel the role of apical microtubules in HC apical differentiation, which is consistent with the results obtained upon nocodazole administration. Finally, Daple seems to be essential, especially during the morphogenesis of hair bundles, because malformation of hair bundles was consistent from birth to adulthood in Daple-deficient mice.

Results
Auditory brainstem response (ABR) testing revealed hearing defects in Daple −/− adult mice, especially at higher frequencies. Although Daple −/− mice, in the embryonic and neonatal stages, have previously been reported to have defects in the arrangement of hair bundles in cochlear HCs 8 , hearing-potential and morphological changes with age have not been analyzed. Here, we performed ABR tests on 8-12-week-old mice and found lower sensitivity to sounds in Daple −/− mice than in Daple +/+ mice at all frequencies, ranging from 4 to 32 kHz. A highly significant hearing disturbance was detected around 24 kHz (42.1 ± 9.1 in +/+ vs. 74.3 ± 8.4 dB in −/− , p < 0.003, Fig. 1A). The gross shape and size of the cochleae in Daple −/− mice were comparable to those of Daple +/+ mice (Fig. 1B), which is consistent with a previous report regarding neonatal cochleae 8 . In addition, we found that the gross shape and size of the cochleae in Daple −/− mice were also similar to those of the cochleae in Daple +/+ mice at 4 weeks of age. These results suggested that sound wave transmission along the snail-shaped OC from the tympanic membrane occurs in the same way in the cochleae of both Daple +/+ and Daple −/− mice 10 . As localization and expression of Daple was similar between neonates and adults, along the apex to the base, in the cochlea (Fig. 1C,D; Supplementary Fig. S1), the hearing disturbances observed at all sound frequencies in Daple −/− mice are reasonable, although disruption at higher frequencies suggested the presence of certain functional and/or structural failures involving HCs in the more basal regions of the OC.
Hair bundle arrangement was affected in adult Daple −/− mice, especially in the basal region. Based on the abovementioned observations, we next examined the possible associations between the ABR results and morphological defects in hair bundles in adult Daple −/− mice, aged 8-12 weeks using scanning electron microscopy (SEM). In adult mouse outer hair cells (OHCs), deformities were observed in the hair bundles ( Fig. 2A), along with various morphological alterations similar to those observed in prior studies on neonatal Daple −/− mice 8 ; these were classified as follows: normal (Fig. 2B a,b), flat (Fig. 2B c,d), split (Fig. 2B e,f), and other dysmorphic bundles (Fig. 2B g,h). Although defects in the arrangement of hair bundles were observed over the entire length of the cochlea, the number of cells with defects in hair bundles was higher in the more basal cochlear regions in Daple −/− mice (% ratios of each dysmorphic bundle type from the apex to the base region, respectively: 17.4, 25.6, 28.8 in flat; 7.8, 24.5, 39.1 in split; 6.5, 7.5, and 17.1, respectively) (Fig. 2C). In contrast, in Daple +/+ mice, almost all the HCs exhibited normal V-shaped bundles ( Fig. 2C Daple +/+ ; apex 91 cells, middle 95 cells, base 94 cells). The increase in the ratio of abnormal hair bundle arrangement cells from 32% in the apex to 58% in the middle to 85% in the base suggested an association between the ABR results and morphological defects in hair bundles in 8-12-week-old adult Daple-deficient mice.
Deformed apical structures, comparable to those in 8-12-week-old adult mice, were present in postnatal day (P)3 Daple −/− mice. To compare morphological changes in hair bundles with age, we next focused on the cochleae of neonatal mice. Various deformities in apical hair bundles were also detected in P3 neonatal Daple −/− mice (Fig. 3A, Supplementary Fig. S2). As for the kinocilia, they were not always in the center of hair bundles (Fig. 3B). Some kinocilia were located at the center of hair bundles, while the bundles were split (Fig. 3B b). To evaluate apical morphological deformities, we classified kinocilia into three groups based on their localization relative to the hair bundle: normal (centered), off-centered, and poorly determined. The term "poorly examined" was used when the position could not be determined as centered or off-centered (Fig. 3B a). The results revealed that > 50% of the kinocilia were localized away from the center of the hair bundles, exhibiting a disrupted correlation between kinocilia and hair bundles (Fig. 3C). In terms of the arrangements of hair bundles, approximately 80% of HCs were abnormal in the basal region of the cochleae, which is the most developmentally mature cochlear region.

Apical microtubules emanate irregularly from disorderly aggregated structures in murine Daple −/− cochlear hair cells. Prior studies on ependymal cells have demonstrated that Daple functions
as a regulator of microtubule polymerization in mouse ventricles 7 . In cochlear HCs, apical microtubules are also important regulators of the apical morphogenesis of HCs of the OC. In this study, we focused on the apical microtubule networks of HCs of the OC to clarify their roles in apical morphogenesis in Daple −/− mice.
In P0 Daple +/+ HCs, apical microtubules were laterally localized, consistent with the lateral positioning of the kinocilium along the mediolateral axis in HCs. The microtubules spread from the pericentriolar region, the base of the kinocilium, toward the cell cortex, to attach to the lateral side of the plasma membrane ( Fig. 4A; Daple +/+ ). In contrast, the opposing ends of microtubules formed ring-like structures around the centrosome. In Daple −/− HCs, densely aggregated microtubule structures were observed at random positions within the cells, and many microtubules were not attached to the lateral membrane (Fig. 4A). As shown in Fig. 4B  www.nature.com/scientificreports/ To examine the 3D structure of the apical microtubule network in detail, we compared the z-series of images for microtubules/EB-1 (Fig. 4C). In Daple +/+ mice, the microtubule network, the center of which has a ring-like region, radially and dominantly emanated to the lateral side, whereas smaller amounts of microtubules were diffusely directed to the medial side. The microtubules diffused like an umbrella in Daple +/+ HC. Upon comparing the distribution of EB-1, a microtubule plus-ended binding protein, to that of microtubule, EB-1 was found to be diffusely distributed through the cytoplasm relatively merged with the distribution of microtubule. In contrast, in murine Daple −/− HCs, microtubule were disorderly aggregated, without ring-like regions in the centers of aggregation, and a smaller amount of the microtubule network, diffused within the cytoplasm, was observed compared to that in Daple +/+ HCs. When z-sliced images were observed, microtubule aggregates were clearly present at very high densities in Daple −/− mice. The results are illustrated in Fig. 5A and suggest that Daple plays a role in forming correct microtubule networks in mouse HCs.
Disordered microtubules were observed to run through the HC apical planes in transmission electron microscopy (TEM) images of Daple −/− mice. To obtain a clearer distribution of the microtubule network, we next performed thin-section TEM (Fig. 5B). In the thin-section TEM images of Daple +/+ mice, microtubules surrounding the centrosome were clearly observed, but those around the centrosome seemed relatively sparse, suggesting that this area may be the ring-like region observed in immunofluorescence images.  www.nature.com/scientificreports/ Microtubules were found to emanate from the pericentriolar region and attached to the lateral cell membrane in HCs. A smaller number of microtubules seemed to run in the medial direction. In contrast, in Daple −/− mice, as evident from the immunofluorescence images (Fig. 4), a higher density of microtubules compared with that in cochleae of Daple +/+ mice was present around the centrosomes without ring-like structures, forming a sparse region. Furthermore, some microtubules were unnaturally elongated in the cytoplasm. These results showed that the deficiency in Daple induced disturbed microtubule arrays in mice.

Cochlear hair bundle abnormalities were induced by nocodazole, a microtubule polymerization inhibitor.
The abovementioned results suggest that disordered microtubules contribute to the deformity in hair bundles in Daple −/− mice. To prove the validity of this hypothesis, we performed nocodazole treatment-based experiments on organ cultures of OC cells (Fig. 6A). The base region of the cochlea from embryonic day (E)17.5 mice was dissected, and the epithelial layer with three arrays of OHCs and one array of inner hair cells (IHC) were mechanically isolated using a stereo microscope for subsequent organ culture. DMSO-treated samples were prepared for use as a control ( Fig. 6B; DMSO). Without nocodazole, almost all HCs developed normally. Upon exposure to nocodazole (400 nM) for 2 days at 37 °C in a 5% CO 2 incubator, stereocilia developed mis-shaped hair bundles, with various kinds of changes, including the presence of flat, dysmorphic, or off-centered kinocilia in HCs, similar to those in Daple −/− mice (Fig. 6B,C). Approximately half of the cochlear culture HCs treated with 400 nM nocodazole did not develop correct hair bundles and, instead, had dysmorphic bundle patterns (Fig. 6D). No changes in the expression of PCP core protein were observed under these conditions, suggesting that these results were not related to tissue PCP but to cellular signals ( Supplementary Fig. S3).

Discussion
Here, first, hearing defects across a broad range of frequencies, especially at frequencies > 24 kHz, were found in 8-12-week-old adult Daple −/− mice. As sounds with a frequency > 24 kHz can cause approximately 30% of the basal region of the membrane to vibrate 10 , this result was consistent with the observation that more severe defects in hair bundle arrangement were present in more basal areas of the OC, as detected by SEM imaging. By comparing adult and neonatal mice (approximately P3), we also found that defects did not clearly progress in the apical structure of hair cells after the neonatal stage. This suggested that Daple is required for the acquisition of hearing ability in adult stages. There was a recent report concerning the special role of Gαi 3 activation through the GBA domain 11 , and mislocalization of apical Gαi 3 in HCs in Daple-deficient mice 8 . Gαi 3 mutant mice exhibited more severe mis-shaped hair bundle arrangements in the basal area of the cochlea duct 12 . This led us to speculate Daple deficiency and Gαi 3 function , may be correlated, specifically in the basal region of the OC.
In ventricular ependymal cells, Daple is reported to function in microtubule dynamics 7 . In previous reports, microtubules were suggested to be important for the apical arrangement of HCs 9 . The conditional knockout of several microtubule-related proteins, such as Lis1, a dynein activating microtubule-binding protein, disturbs the organization of microtubules by impairing developmental stage-specific connections between the microtubules and plasma membranes through the LGN (Gpsm2)-Gαi-dynein complex 9 . This might lead to the formation of an apical microtubule-rich bare zone and the stabilization of Gαi3-Daple-Dvl complexes on the lateral side of the plasma membrane in HCs. Our results regarding the dysregulation of EB1 and focal localization in Daple −/− mice also support this notion. This also suggests a close relationship between microtubules and apical structures, including stereocilia/kinocilia morphogenesis. On the other hand, Dvl deficiency induces various degrees of hair bundle malformations because of defects in the combination of Dvl 1-3, as reported previously 13,14 . Disturbance in Dvl subtype-specific interactions with Daple may have induced severe changes in the basal area of Daple −/− cochlea.
Several reports have demonstrated that other PCP pathways and actin cytoskeletal dysfunction are involved in various forms of cochlear malformation that are different from those seen in Daple-deficient mice. PCP proteins, such as Vangl, have a normal V-shape, but disorientated, hair bundle arrangements that are different from those in Daple −/− mice 15 . As for actin, it would be informative to investigate Rho-family protein-deficient mice, because of the upstream regulatory role of Rho-family proteins in the organization of actin filaments. Rac1-deficient mice have defects in the arrangement of hair bundles similar to those in Daple-deficient mice, but these are different from those in Daple −/− mice in that the cochlear duct is significantly shortened and the fragmentation of hair bundles is severely progressive around birth 16 . Moreover, the cochlear abnormalities observed in Daple −/− mice were different from those in mice deficient in actomyosin-related proteins, such as myosin2 17 , RhoA 18 , and Cdc42 19,20 .
In conclusion, we validated the hypothesis that Daple regulates the organization of microtubules in HCs of the OC, as well as in ependymal cells. E17.5 mouse cochlear organ culture revealed that the defects in the arrangements of stereocilia bundles after treatment with nocodazole were similar to those in Daple-deficient mice. This is the first study regarding cochlear HC differentiation employing microtubule polymerization inhibitors, except for the examination of apical surface rigidity 21 . Further studies examining cochlear cytoskeletal maturation processes in shorter intervals around birth may reveal the sequence of underlying molecular mechanisms and their related signals.

Methods
All methods were conducted in accordance with ARRIVE guidelines.

Auditory brainstem response (ABR) test. The details of the ABR test and the method have been
reported previously 22 . We injected ketamine (100 mg/kg) and xylazine (10 mg/kg) into the peritoneal cavity of mice and put mice into a sound isolation chamber. Subcutaneous needle electrodes were inserted in the pinna and vertex, with a ground electrode near the tail. Responses to tone pip stimuli were recorded at 4, 8, 12, 24, and 32 kHz, intensities ranging from 0 to 100 dB-SPL instep of 5 dB, in 8-10-week-old mice using a Power Lab 2/25 (AD Instruments, Australia) and a TDT Auditory Workstation (Tucker-Davis Technologies, Alachua, Florida, USA). The duration of tone bursts was 1 ms. We amplified and averaged 500 responses. All ABRs were measured without knowing the profiles or genotypes of the mice. Scanning electron microscopy (SEM). Inner ears obtained from Daple + / + or Daple-/-mice were fixed with 2% PFA and 2.5% glutaraldehyde in 0.1 M HEPES (pH 7.4) for 1 h at RT. They were then washed with 0.1 M HEPES and fixed in 1% OsO4 for 1 h on ice, incubated in 1% tannic acid overnight, and fixed in 1% OsO4 for 1 h on ice again. The organ of Corti was micro-dissected, dehydrated, dried at the critical point, sputter-coated, and observed using SEM (S-4800 microscope; Hitachi).

Assessment of
Transmission electron microscopy (TEM). Inner ears obtained from Daple + / + or Daple-/-mice were fixed with 2% PFA and 2.5% glutaraldehyde and treated with 2% tannic acid in 0.1 M HEPES (pH 7.4) for 1 h at RT. They were then washed with 0.1 M HEPES and fixed in 1% OsO4 for 2 h on ice. The organ of Corti was micro-dissected, dehydrated, embedded, sectioned, and observed using TEM (JEM-1400Plus; JEOL). Microtubules were identified by their properties, for example, around 25 nm diameter, tube-like high density structure 23 .
Culture of embryonic mouse cochlea and drug treatment. Cochlear organ culture was started from E17.5 mice. Briefly, cochleae were dissected in Leibobitz L-15 medium (Thermo Fisher Scientific) and established on coverslips coated with Matrigel matrix (Corning). Explants were then maintained for 1 h in vitro in DMEM/F-12 (Invitrogen) supplemented with FBS and ampicillin (Nacalai Tesque). Next, the medium was replaced with that containing DMSO only (control) or nocodazole (Sigma; 400 and 800 nM). After 2 days of culture in vitro, the explants were fixed with 4% PFA or methanol for immunostaining, or with 2% PFA plus 2.5% glutaraldehyde in 0.1 M HEPES (pH 7.4) for SEM.