Distinct roles of Eps8 in the maturation of cochlear and vestibular hair cells

Neuroscience 328 (2016) 80–91

DISTINCT ROLES OF EPS8 IN THE MATURATION OF COCHLEAR AND VESTIBULAR HAIR CELLS ELISA TAVAZZANI, a PAOLO SPAIARDI, a VALERIA ZAMPINI, ay DONATELLA CONTINI, aà MARCO MANCA, a GIANCARLO RUSSO, a IVO PRIGIONI, a WALTER MARCOTTI b AND SERGIO MASETTO a*

affects the hair bundle morphology but not the basolateral membrane currents. This difference is likely to explain the absence of obvious vestibular dysfunction in Eps8 KO mice. Ó 2016 IBRO. Published by Elsevier Ltd. All rights reserved.

a Department of Brain and Behavioral Sciences, University of Pavia, Via Forlanini 6, 27100 Pavia, Italy b Department of Biomedical Science, Sensory Neuroscience Group, Alfred Denny Building (B1 221), University of Sheffield, Western Bank, Sheffield S10 2TN, UK

Key words: Eps8, hair cell, vestibular, K+ channel, hair bundle, deafness.

Abstract—Several genetic mutations affecting the development and function of mammalian hair cells have been shown to cause deafness but not vestibular defects, most likely because vestibular deficits are sometimes centrally compensated. The study of hair cell physiology is thus a powerful direct approach to ascertain the functional status of the vestibular end organs. Deletion of Epidermal growth factor receptor pathway substrate 8 (Eps8), a gene involved in actin remodeling, has been shown to cause deafness in mice. While both inner and outer hair cells from Eps8 knockout (KO) mice showed abnormally short stereocilia, inner hair cells (IHCs) also failed to acquire mature-type ion channels. Despite the fact that Eps8 is also expressed in vestibular hair cells, Eps8 KO mice show no vestibular deficits. In the present study we have investigated the properties of vestibular Type I and Type II hair cells in Eps8-KO mice and compared them to those of cochlear IHCs. In the absence of Eps8, vestibular hair cells show normally long kinocilia, significantly shorter stereocilia and a normal pattern of basolateral voltage-dependent ion channels. We have also found that while vestibular hair cells from Eps8 KO mice show normal voltage responses to injected sinusoidal currents, which were used to mimic the mechanoelectrical transducer current, IHCs lose their ability to synchronize their responses to the stimulus. We conclude that the absence of Eps8 produces a weaker phenotype in vestibular hair cells compared to cochlear IHCs, since it

INTRODUCTION Hair cells are the sensory receptors of the auditoryvestibular system in all vertebrates. Sound or motion is transduced into electrical signals by the hair bundles, the mechanosensory actin-packed stereocilia protruding from the apical surface of hair cells. Stereocilia of different lengths are arranged in a staircase-like structure, the number of which changes depending on the inner ear organ. Hair bundle deflection, induced by sound or head motion, modulates the open probability of mechanoelectrical transducer channels localized at the tips of the stereocilia (Beurg et al., 2009), and as such generates a receptor potential. Cell voltage responses are then shaped by different types of voltage-dependent ion channels, among which Ca2+ channels are coupled to neurotransmitter (glutamate) exocytosis. Several genetic mutations that cause deafness in mice and humans have been shown to affect the development and/or function of cochlear hair cells (see Hereditary Hearing Loss Homepage http://hereditaryhearingloss. org). Given the common embryonic origins and biology of the auditory and vestibular hair cells, it is conceivable to postulate that single gene mutations known to cause inherited hearing loss would also lead to vestibular dysfunction. Instead, vestibular function is often retained even in the case of profound deafness (Jones and Jones, 2014). One possible explanation is that deficits in vestibular hair cells have sometimes gone undetected because of compensation or adaptation by the central nervous system. Deletion of Epidermal growth factor receptor pathway substrate 8 (Eps8), a gene involved in actin remodeling (Di Fiore and Scita, 2002), hampers normal stereocilia growth (Manor et al., 2011; Zampini et al., 2011) and ion channel expression in mouse cochlear inner hair cells (IHCs) (Zampini et al., 2011). Despite the similar expression profile of Eps8 in cochlear and vestibular hair cells (Manor et al., 2011; Zampini et al., 2011), Eps8 knockout (KO) mice are deaf but show no obvious vestibular

*Corresponding author. Tel: +39-0382987609; fax: +39-0382987527. E-mail addresses: [email protected] (V. Zampini), [email protected] (D. Contini), w.marcotti@sheffield.ac.uk (W. Marcotti), [email protected] (S. Masetto). y Present address: Paris Descartes University, Biomedical and Fundamental Science Faculty, Neurophotonics Laboratory, CNRS UMR8250, 45, rue des Saints Pe`res, 75270 Paris Cedex 06, France. à Present address: Department of Anatomy and Cell Biology, University of Illinois College of Medicine, 808 S. Wood St., Chicago, IL 60612, USA. Abbreviations: AP, action potential; d.c., direct current; EGTA, ethylene glycol tetraacetic acid; Eps8, Epidermal growth factor receptor pathway substrate 8; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; IHCs, inner hair cells; KO, knockout; OHCs, outer hair cells; P, postnatal day; Rm, cell membrane input resistance; Rs, series resistance; RT, room temperature; Vm, voltage membrane potential; Vrest, resting membrane potential; WT, wild type. http://dx.doi.org/10.1016/j.neuroscience.2016.04.038 0306-4522/Ó 2016 IBRO. Published by Elsevier Ltd. All rights reserved. 80

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

deficits. Similarly, it has been reported that a biallelic nonsense mutation of human EPS8, presumably coding a truncated non-functional protein or none (like the Eps8-KO mouse), results in deafness but no balance defects (Behlouli et al., 2014). In order to elucidate the importance of Eps8 in vestibular hair cells, we have investigated the bundle morphology and the biophysical properties of vestibular hair cells of the Eps8-KO mouse and compare them to those of WT mice. We provide evidence that the absence of Eps8 alters the growth of vestibular hair cells stereocilia. We have also found that, different from IHCs, the receptor potential of vestibular hair cells was not affected by the absence of Eps8. The above findings could explain why Eps8 deletion, and presumably EPS8 mutation, primarily affects the auditory function.

EXPERIMENTAL PROCEDURES All procedures used were approved by the Ministero Italiano della Salute (Rome, Italy) and animal experiments were carried out in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC). Eps8-KO mice were obtained by breeding heterozygote mice. All animals were genotyped as previously described (Offenha¨user et al., 2006). To increase the number of controls, and particularly for the hair cells dissociation protocol (see below), wild-type (WT) mice (C57 and Swiss CD1) were also obtained from Harlan Italy (Italy). As already reported (Contini et al., 2012), no differences in hair cells’ morphological and electrophysiological properties were found between C57 and Swiss CD1 mice. Kinocilia and stereocilia measurements were performed using enzymatically and mechanically dissociated hair cells from the semicircular canal cristae. Following brief anesthesia with 2-bromo-2-chloro-1,1,1-tri fluoroethane (Sigma Aldrich), mice were decapitated and the three ampullae with the sensory cristae dissected out. For the enzymatic treatment, the preparation was first placed in a Petri dish containing the following extracellular solution, for 7 min at room temperature (RT, 22–25 °C) (Extra_D, in mM): NaCl 138, CaCl2 0.1, KCl 5.8, MgCl2 0.9, HEPES 10, glucose 15, NaH2PO4 0.7, Na-pyruvate 2, plus vitamins (GIBCO Invitrogen, Grand Island, USA, #11130, 10 mL/L) and aminoacids (GIBCO Invitrogen, #11120, 20 mL/L); pH 7.4 with NaOH. Protease VIII (Sigma Aldrich; 0.05 mg/ml) was added to the above solution. Then, the cristae were incubated with crude papain (Calbiochem-Nova Biochem Corporation, USA; 0.5 mg/ml) plus L-cysteine (Sigma Aldrich; 0.3 mg/ml), dissolved in Extra_D, for 23 min at 37 °C. Finally, the ampullae were transferred to a Petri dish containing Extra_D plus bovine albumin serum (Sigma Aldrich; 1 mg/ml) for 40 min at RT to stop the enzymatic activity. Afterward, the cristae were transferred to the recording chamber filled with Extra_D. To mechanically dissociate the hair cells from the epithelium, each crista ampullaris was gently scraped with an eyelash. Hair cells were allowed to settle onto the glass-bottom of the recording chamber for 20 min,

81

and then visualized by an upright microscope (Olympus BX51WI, Japan) equipped with Nomarski optics and a 60x water immersion objective (Olympus LUMPlan FI, Tokyo, Japan). Images were acquired by a digital camera (ORCA-05G, Hamamatsu Photonics, Hamamatsu City, Japan) and digitally stored with cellSens Dimension software (version 1.6, Olympus, Japan); this software was also used for all measurements, calibration = 107.5 nm/pixel. Only hair cells with well-preserved hair bundles were considered. These were a small minority of the dissociated hair cells because hair bundles are closely embedded in the cupula, which is easily stressed during dissection. Since the morphology of hair cells can change following their dissociation (Zenner et al., 1990), we compared hair bundles without subdividing hair cells in Type I and Type II. For patch-clamp experiments, the organ of Corti or the whole cristae were dissected as previously described (Contini et al., 2012; Johnson et al., 2013). The preparations were placed in normal extracellular solution (in mM): 135 NaCl, 5.8 KCl, 1.3 CaCl2, 0.9 MgCl2, 0.7 NaH2PO4, 5.6 D-glucose, 10 Hepes-NaOH, 2 sodium pyruvate, amino acids (GIBCO Invitrogen), and vitamins (GIBCO Invitrogen, 10 mL/L) (pH 7.48; osmolality 310 mOsm/kg). The tissue and microelectrode were viewed using differential interference contrast optics employing an upright microscope (Zeiss Axioskop, Germany) equipped with X63 water immersion objective. Patch-clamp data were obtained from 18 WT IHCs, 13 Eps8-KO IHCs, 33 WT vestibular Type I hair cells, 17 Eps8-KO Type I hair cells, 26 WT vestibular Type II hair cells and 14 Eps8-KO Type II hair cells, from postnatal day (P) 6 to P29, where the day of birth is P0. All recordings were obtained at room temperature. Voltage- and current-clamp recordings were obtained using the following intra-pipette solution (in mM): 131 KCl, 3 MgCl2, 1 EGTA–KOH, 5 Na2ATP, 5 HEPES–KOH (pH 7.2; 293 mOsm/kg). Soda glass pipettes (Hilgenberg, Malsfeld, Germany) were pulled to tip diameters of about 2 lm, fire-polished and partially coated with Sylgard (Dow Corning 184, USA) to minimize the fast patch pipette capacitance transient. Electrophysiological recordings were made using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, USA). Data acquisition was controlled by pClamp software using a Digidata 1322A board (Molecular Devices, USA). The amplifier’s filter bandwidth was set at 2–5 kHz. Digital sampling frequency of voltage- and current-clamp protocols was at least three times the analog bandwidth of the signal recorded. Recordings were stored on computer for off-line analysis (Origin, OriginLab, USA). When filled with the intra-pipette solution, micropipettes had a resistance in the bath of 2–5 MX. Pipette capacitance and resistance were compensated in cell-attached configuration. The pipette resistance was kept as low as possible, despite the greater difficulty in obtaining a gigaseal, to minimize the series resistance (Rs). On-line Rs compensation in Type I hair cells may lead to substantial errors because no voltage range without active ion currents can be found in these cells. The low pipette resistance and good access to the whole-cell configuration gave acceptably low Rs (<10 MX). This was confirmed

82

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

in Type II hair cells, where Rs and cell membrane capacitance (Cm) were monitored during trains of 5 mV hyperpolarizing voltage steps delivered from the cell membrane holding potential of 63 mV. Current traces were not corrected for cell membrane capacitive transients; in a few traces however, as noted in the figure legends, they were partially blanked. Voltages in voltage-clamp experiments and cell membrane resting potential (Vrest) in current-clamp recordings were corrected a posteriori for the liquid junction potentials of intra-pipette solutions vs. extracellular solution (3 mV negative inside the pipette). Statistical comparisons of means were made by Student’s two-tailed t test. In the text, mean values are quoted ± standard deviation (S.D.), where P < 0.05 indicates statistical significance. In Figs. 2 and 3, data points refer to mean ± standard error (S.E.).

RESULTS Morphological features of Eps8-KO vestibular hair cells Eps8 is expressed at the stereocilia tips of both the cochlear and the vestibular hair cells (Manor et al., 2011). The hair bundles of Eps8-KO IHCs are shorter than normal, with the first (longest) row of stereocilia being the most affected (Zampini et al., 2011). Since no clear data are available concerning Eps8-KO vestibular hair bundles, we first investigated this aspect. Fig. 1A, B show the hair bundle populations in the intermediate region of the crista ampullaris from a WT and an Eps8KO mouse, respectively. WT and KO vestibular cristae presented similarly dense populations of hair bundles when visualized at their base, whereas the upper half (above the dashed line in Fig. 1B) appeared less packed. To elucidate the reason for the above difference, we investigated the morphology of the hair bundles in more detail by using dissociated hair cells. It is worth mentioning that while cochlear hair cells lose their kinocilium during maturation (at P10, Leibovici et al., 2005), vestibular hair cells keep it throughout their life. We found that vestibular hair bundles from both WT and Eps8-KO mice had a normal (eccentrically located and long) kinocilium and staircase organization (Fig. 1C, D). However, the longest row of stereocilia was found to be significantly shorter (about 50%; P < 0.0001) in Eps8-KO than in WT hair cells (Fig. 1E, F, G). The hair cell size was found to be not significantly different between Eps8-KO and WT mice, indicating that the absence of Eps8 was not affecting the normal growth of the cell body. The above data are summarized in Table 1. Electrophysiological properties from vestibular hair cells of WT and Eps8-KO mice Mammalian vestibular epithelia contain both Type I and Type II hair cells, which can be distinguished by the expression of specific K+ channels (Rennie and Correia, 1994; Eatock et al., 1998; Eatock and Songer, 2011). Fig. 2A, B show representative whole-cell current recordings from vestibular Type II hair cells of a WT and an Eps8-KO age-matched mice (P23). Cell membrane

hyperpolarization from 63 mV evoked an anomalous rectifying K+ current (IK,1) in both WT and Eps8-KO Type II hair cells. In some cells from both WT (Fig. 2A; see also Contini et al., 2012) and Eps8-KO (not shown) mice, membrane hyperpolarization additionally evoked a small slow inward rectifying cationic current (Ih). Cell membrane depolarization causes the activation of a rapid and transient outward rectifying K+ current (IKA) together with a delayed and sustained outward rectifying K+ current (IK,v) in both WT (Fig. 2A; see also Contini et al., 2012) and KO cells (Fig. 2B). The mean peak and steady-state current–voltage (I  V) relations are shown in Fig. 2C, D. The amplitude of the macroscopic inward and outward rectifying K+ currents in KO cells was not significantly different from that measured in WT cells. The above results indicate that the absence of Eps8 does not affect the K+ current profile of vestibular hair cells. The mean membrane input resistance (Rm), calculated between 63 mV and 53 mV, was also not significantly different between the two genotypes (721.07 ± 359.72 MX, n = 5 in WT cells vs. 978.18 ± 396.73 MX, n = 7 in KO cells). The resting membrane potential (Vrest) was found to be slightly (P = 0.029) less hyperpolarized in KO (63.64 ± 3.05 mV, n = 14) than in WT cells (67.38 ± 5.74 mV, n = 26). Since the amplitude of the macroscopic K+ currents was not significantly different, the more depolarized Vrest in Eps8-KO cells was likely to be due to a bias in hair cell sampling. As mentioned above, hair cells from both Eps8-KO and WT mice expressed Ih, a mixed Na+/K+ current expected to depolarize the Vrest. Depending on the relative contribution of Ih and different K+ currents in the voltage range close to 60 mV, Vrest is expected to be more or less depolarized. Fig. 3A, B show representative whole-cell currents recorded from vestibular Type I hair cells of a WT (P18) and an Eps8-KO (P16) mouse, respectively. Both cells expressed the low-voltage outward rectifying K+ current (IK,L), which is almost fully active at 60 mV, and the slow outward rectifying K+ current (IK,v) that activates for depolarization above 40 mV (Rennie and Correia, 1994; Ru¨sch and Eatock, 1996; Contini et al., 2012). Although the general features of the macroscopic current recorded in Type I hair cells appear to be similar during postnatal development (Ge´le´oc et al., 2004), it has been suggested that the K+ channel subunits contributing to IK,L are likely to change during the first three postnatal weeks (Hurley et al., 2006). Therefore, we investigated for possible differences in the K+ currents present in WT and Eps8-KO cells during the above timeframe. Fig. 3C, D show the mean peak and steady-state I  V relation at P7–9, whereas Fig. 3E, F show the I  Vs at P16-23. The size of the K+ currents was not significantly different between WT and Eps8-KO hair cells at both age ranges. Rm, calculated between 63 mV and 53 mV, was also not significantly different in either the younger (36.02 ± 11.95 MX, n = 4 in P7–9 WT cells vs. 32.61 ± 11.57 MX, n = 8 in P7–12 KO cells) or older hair cells (20.39 ± 8.43 MX, n = 7 in P16-23 WT cells vs. 16.50 ± 3.28 MX, n = 7 in P16–19 KO cells). The Vrest was 66.70 mV (±6.47; n = 17) in KO cells and 70.24 mV (±5.42; n = 33) in WT cells, i.e. slightly more hyperpolarized in WT than in KO cells,

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

83

Fig. 1. Stereocilia are abnormally short in Eps8-KO vestibular hair cells. (A, B) Representative photomicrographs of a small portion of the vestibular crista intermediate region from a P61 WT and a P65 Eps8-KO mouse, respectively. Several hair bundles are visible. Note that the hair bundle density seems similar at the base, but lower in the upper half of the Eps8-KO preparation (as delimited by the dashed line). The continuous lines indicate the maximal length of the hair bundle, which is dictated by the kinocilia (asterisks). (C, D) Representative photomicrographs of isolated vestibular hair cells from a P44 WT and a P44 Eps8-KO mouse, respectively. Several stereocilia and the kinocilium (asterisk) are visible. The arrowheads point at the longest stereocilium in the hair bundle. Note the staircase organization of the hair bundle. (E, F) Representative photomicrographs of isolated vestibular hair cells from a P42 WT and a P44 Eps8-KO mouse, respectively. Focus is on the kinocilium and the first row (longest) stereocilium. Note that the kinocilium has the same length in the two cells, whereas the longest stereocilium is much shorter in the Eps8-KO cell. (G) Histogram showing the average length of the kinocilia and the first stereociliary row in WT and KO mice. A statistically significant difference was found only for the stereocilia.

although the difference was not found to be significantly different (P = 0.055). Electrophysiological properties from IHCs of WT and Eps8-KO mice The onset of functional maturation in IHCs occurs at around P12, which corresponds to the onset of hearing. At birth, IHCs express IK,neo, a K+ current showing a 4-aminopyridine (4-AP)-sensitive and a TEA-sensitive component (Marcotti et al., 2003a), ISK2, an apaminsensitive Ca2+-activated K+ current (Marcotti et al., 2004), IK,1, an anomalous rectifying K+ current (Marcotti et al., 1999) and a Na+ and a Ca2+ current (INa and ICa; Marcotti et al., 2003b). At around P12, IHCs acquire IK, + current (Oliver et al., n, a low voltage-activated K 2003; Marcotti et al., 2003a), and IK,f, a fast-activating Ca2+-dependent K+ current (Kros et al., 1998), while ISK2, IK,1 and INa are down-regulated. It has been shown that in the absence of Eps8 IHCs fail to acquire both IK,n and IK,f (Zampini et al., 2011). In order to allow a direct comparison between the currents recorded from IHCs and vestibular hair cells, and for a better understanding of the underlying voltage response (see next section),

we performed some recordings from IHCs at different developmental ages (Fig. 4), matching those obtained in the vestibular hair cells (Figs. 2 and 3). Note that the current profile recorded in the KO P12 IHC (Fig. 4C: onset of cell maturation) resembled that obtained in the immature WT cell at P8 (Fig. 4A). The immature-type current profile observed in the IHC of the Eps8-KO mice was causing the negative slope in the I  V relation (Fig. 5B), which was not evident in age-matched WT cells (Fig. 5A). This leads to a net inward Na+/Ca2+ current at voltages around 30 mV (Marcotti et al., 2003a). The negative slope conductance near the action potential (AP) threshold (see Benson and Adams, 1987) determines the persistence of AP-like activity in adult Eps8-KO IHCs (compare Fig. 5C, D) (Zampini et al., 2011). To characterize the consequences of Eps8 deletion upon the receptor potential, we recorded voltage responses from vestibular and cochlear hair cells during sinusoidal current stimuli. Voltage response of vestibular and cochlear hair cells to sinusoidal currents Fig. 6 shows representative voltage responses to an injected sinusoidal current, which was used to mimic the

84

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

Fig. 2. Macroscopic currents are similar in WT and Eps8-KO mouse vestibular Type II hair cells. (A, B) Macroscopic currents from a control and a knockout adult P23 Type II hair cell, respectively. Currents were elicited by hyperpolarizing and depolarizing voltage steps from the holding potential of 63 mV to the various test potentials shown next to each trace. Residual transient capacitive artifacts have been partially erased. Hyperpolarizing steps elicited inward (anomalous) rectifying K+ currents (IK,1) and, in some cells, a small cationic current (Ih). Depolarizing steps evoked large transient outward K+ currents (IK,A) and a slow sustained K+ current (IK,v). No differences were found in the macroscopic currents between control and Eps8-KO mice. C, D: mean (± S.E.) peak and steady-state current (I) – voltage (V) relation obtained from 5 WT (P15–23) and 7 Eps8-KO Type II hair cells (P19–23), respectively. In this and the following figures, voltage commands (Vm) are shown as nominal values.

mechanoelectrical transducer (MET) current (0.2 Hz; 200 pA peak-to-peak; assuming a resting open probability of the MET channels of 10%; see Figure legend). Mature Type I and Type II vestibular hair cells from WT mice (P9: Fig. 6A, B, respectively) exhibited similar voltage responses to sinusoidal stimuli to those obtained from aged-matched Eps8-KO mice (not shown). Note that the extent of membrane depolarization in Type I hair cells (the ‘‘voltage gain” DV/DI) was much smaller than that observed in Type II hair cells, which was due to the much lower Rm produced by IK,L in Type I hair cells. We also found that the shape of the voltage response in Type I hair cells, but not that in Type II hair cells, overlapped to that of the stimulus (i.e. the voltage gain varies linearly with the amplitude of the MET current). This complete overlap was also due to the presence of IK,L since, by being almost fully activated at Vrest, it produces a ohmic (linear) change of the cell membrane voltage in response to the sinusoidal current stimulus. In Type II hair cells, IK,A and IK,v start activating close to Vrest, and therefore their kinetics interfere with the time course of the voltage response. This has two effects: the voltage gain is initially very large but, as the outward rectifying K+ currents activate,

it decreases and the voltage response flattens. The stimulus frequency (0.2 Hz) used for these experiments was selected to be in the vestibular frequency domain, i.e. below the cochlear frequency range (approximately 4–75 kHz in the mouse; Nyby, 2001). However, since fast depolarization in IHCs from Eps8-KO mice elicited Ca2+ dependent APs (see Fig. 5D), the slow depolarizing sinusoidal stimuli allowed to characterize the ‘‘analog” time course (direct current (d.c.) component) of the receptor potential. At immature developmental stages (P8), both the WT and Eps8-KO IHCs showed an intense firing activity (not shown; see Zampini et al., 2011). After the onset of hearing at P12, IHCs from WT mice stop firing APs and instead showed a voltage response that faithfully follows the sinusoidal stimulus (P22, Fig. 6C). In adult Eps8-KO IHCs (P23), APs could still be elicited and the shape of the voltage response did not match that of the stimulus (Fig. 6D), but resembled the responses observed in Type II hair cells (Fig. 6B). Our observations indicate that Eps8-KO IHCs fail to acquire the basolateral membrane property of mature graded receptors. Lack of expression of IK,n in IHCs from Eps8-KO mice caused their Rm to be significantly larger (718.19 ± 490.61 MX, n = 5, P23–29; P < 0.05) than that measured in WT

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

85

Fig. 3. Macroscopic currents are similar in WT and Eps8-KO mouse vestibular Type I hair cells. (A, B) Representative macroscopic currents recorded from a WT (P18) and an Eps8-KO (P16) Type I hair cell. Current recordings were obtained as described in Fig. 2. Note the presence of an outward current at 63 mV, which was due to the expression of IK,L. While hyperpolarizing steps deactivate IK,L, membrane depolarization produce the activation, from 43 mV, of an additional outward rectifying K+ current (IK,v). (C, D) Mean (± S.E.) peak and steady-state I  V relation obtained from WT and Eps8-KO Type I hair cells during early stages of developmental. Data refer to 4 Eps8-KO (P7-9) and 4 WT (P7-9) cells. (E, F) Mean (± S.E.) peak and steady-state I  V relation obtained from older WT and Eps8-KO Type I hair cells. Data refer to 7 Eps8-KO (P16-19) and 7 WT (P16–23) cells.

86

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

IHCs (66.00 ± 23.09 MX, n = 4, P15–28). Therefore, an additional indirect effect of Eps8 deletion is that the IHC membrane time constant is one order of magnitude greater than that in WT IHCs, which would preclude Eps8-KO IHCs to follow stimuli in the kHz (acoustic) range.

DISCUSSION We found that in the absence of Eps8 the normal growth of the stereociliary bundle was prevented in vestibular hair cells, which is consistent with previous observations from cochlear hair cells (Manor et al., 2011; Zampini et al., 2011). Despite this morphological defect, Eps8-KO mice are deaf but do not show any obvious vestibular defects (Manor et al., 2011; Zampini et al., 2011). One possible explanation for this discrepancy is that vestibular hair cells retain the kinocilium and have much longer stereocilia (Fig. 7) than those in cochlear hair cells, and as such vestibular accessory structures (i.e. the cupula in semicircular canal) are likely to provide sufficient coupling between the stimulus and the hair bundle even if the hair bundles of Eps8-KO mice are shorter. In the cochlea, IHC stereocilia are not connected to the overlaying tectorial membrane such that their reduced size in Eps8-KO mice is likely to compromise the normal deflection during sound stimulation. Moreover, different from cochlear IHCs, Eps8 deletion in vestibular hair cells does not impair the normal expression of mature-like basolateral membrane ion channels. Consistent with this, the receptor potential in response to an injected sinusoidal ‘‘MET” current was normal in vestibular hair cells, but substantially altered in IHCs. Therefore, even if the MET current could be elicited in Esp8-KO IHCs, the resulting receptor potential would be inadequate to signal the stimulus intensity and phase. Eps8 regulates stereocilia growth in cochlear and vestibular sensory cells Eps8 and Eps8L2, two members of the Eps8-like protein family, have been shown to regulate the growth (Eps8: Manor et al., 2011; Zampini et al., 2011) and maintenance (Eps8L2: Furness et al., 2013) of the stereociliary bundle in mammalian cochlear hair cells. Eps8 is highly expressed at the tips of the tallest stereocilia and in its absence the hair bundles are short and retain an immature phenotype. It is hypothesized that Eps8, by binding to the fast-growing ends of actin filaments (barbed end) at stereocilia tips (Schneider et al., 2002), might disadvantage the interaction with proteins inhibiting acting filament elongation (Manor et al., 2011), while favoring the action of those promoting it, such as espin (Rzadzinska et al., 2005). Here, we show that Eps8 deletion has the same impact on vestibular hair bundle, with the tallest stereocilia being about half of that present in WT hair cells, but without affecting their staircase-like organization (Manor et al., 2011; Zampini et al., 2011). The development of the hair bundle staircase-like structure is carefully orchestrated mainly during early stages of hair cell postnatal development (Frolenkov

Fig. 4. Typical voltage-clamp responses recorded from WT and Eps8-KO IHC. (A) Representative macroscopic ionic currents recorded from a P8 WT IHC in response to the voltage-clamp protocol shown on the top panel. Note the presence of an inward rectifying K+ current (IK,1) in response to hyperpolarizing steps from the holding voltage of 63 mV. Membrane depolarization elicited a fast inward current, presumably consisting of INa and ICa, and a slow outward rectifying K+ current, most likely carried out by ISK2, IK,v-like and IKA-like (Marcotti et al., 2003a; see text). (B) Current recordings from a P12 WT IHC in response to the voltage-clamp protocol shown above. Note the presence of IK,n (arrowheads), which is expressed in the mature-type cells, and IK,1, which is normally down-regulated at about P16. (C) Current recordings from a P12 Eps8-KO IHC. Hyperpolarizing steps elicited IK,1, while depolarizing steps elicited a small transient inward current (red trace), presumably consisting of ICa and INa, followed by a slow-activating outward rectifying K+ current clearly resembling the macroscopic current of a younger WT IHC (compare with panel A above). Note that ICa is present in WT IHCs at all ages (Marcotti et al., 2003b), but undetectable because of the larger and faster IK,f and IK,n in adult animals.

87

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91 Table 1. Morphological features of WT and Eps8-KO dissociated semicircular canal hair cells. All data are expressed as mean ± S.D. (lm)

Eps8-KO, n = 10 P37–56 WT, n = 14 P32–44 P value t-test

Kinocilium length

First stereocilium length

Cell body length

52.58 ± 5.31

22.32 ± 3.36

7.97 ± 0.79

53.53 ± 5.14

48.92 ± 4.66

8.16 ± 1.28

0.66

2.96E-13

0.68

et al., 2004). Initially, numerous short microvilli of uniform height appear on the apical surface of newly differentiated hair cells, at embryonic day 13.5 (E13.5) in the vestibule and at E15.5 in the cochlea, which surround a single centrally located kinocilium, a true cilium with a 9 + 2 arrangement of microtubules. During the following few days, the kinocilium migrates to the periphery of the cell surface and the short microvilli located near it begin to elongate, followed by those that are further and further away. This process generates rows of stereocilia that are graded in height. Stereociliary elongation in the shortest rows stops at around postnatal day 5 (P5) and that in the tallest row at about P15 (Manor and Kachar, 2008). Current results indicate that the final length of the hair bundle is unlikely to be determined by the kinocilium, since the first row of stereocilia in vestibular hair cells from Eps8-KO mice was about 50% shorter than normal, despite the fact that the kinocilium length was not affected

by the mutation. This observation agrees with previous studies showing that the growth of the tallest stereocilia row is not affected when the kinocilium either lacks its axonemal part (Jones et al., 2008) or it is entirely disconnected from the stereocilia (Lefe`vre et al., 2008). It is therefore conceivable that the main role of the kinocilium is to establish the polarity of the hair bundle (Jones et al., 2008) and possibly its initial growth, while the staircaselike structure of the hair bundle is mainly determined by the expression of the different proteins regulating actin polymerization. Indeed, it has recently been shown that twinfilin 2, which localizes to the tips of the shorter stereocilia of IHCs and outer hair cells (OHCs), causes a significant decrease of IHCs’ stereocilia length when overexpressed in vitro (Peng et al., 2009). Like Eps8, twinfilin 2 is an actin-binding protein, which inhibits actin polymerization presumably by sequestering G-actin monomers. It is therefore possible that while Eps8 allows

Fig. 5. Voltage- and current-clamp responses recorded from IHCs from WT and Eps8-KO mice. (A, B) Representative currents evoked by a ramp voltage protocol delivered from 123 mV to 37 mV. The red areas indicate the presumed voltage range of the hair cell receptor potential (from 63 mV to 23 mV), when considering that afferent neurotransmitter release occurs at rest, and ICa peaks at around 20 mV. Note that the Eps8KO IHC shows a negative slope in the receptor range, a condition required to generate repetitive action potentials (Benson and Adams, 1987). (C, D) Representative current-clamp responses evoked by a depolarizing current step of 50 pA delivered from the cell resting (zero-current) membrane potential (Vm; same cells as above). Note the persistence of AP-like activity in the Eps8-KO IHC.

88

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

Fig. 6. Voltage responses to sinusoidal injected currents in vestibular and cochlear hair cells. (A) Representative voltage response of a P9 vestibular Type I hair cell. Stimulus frequency and amplitude: 0.2 Hz and 200 pA peak-to-peak (it applies to all panels). Total number of cycles delivered: 3 – only the first and half the cycles of the sinusoidal stimulus/response are shown. Starting from the cell resting (zero-current) membrane potential, a current step of 20-pA amplitude and 1-s duration was delivered to simulate the resting open MET current (Johnson et al., 2012), before starting the sine-wave. Vm = cell membrane voltage response (blue trace). In this and the following panels, the trace of the current stimulus (black line) represents the true current delivered by the patch-clamp amplifier, but scaled in amplitude to maximize overlapping with the voltage response. The gray dashed horizontal line indicates the zero current level. (B) Representative voltage response of a P9 Type II hair cell. Note the different voltage range and shape of the cell membrane voltage response between Type II and Type I (panel A) hair cell. (C, D) Representative voltage responses (red trace) of a P22 WT and a P23 Eps8-KO cochlear IHC, respectively. Note that the voltage response from the WT IHC, but not that from the adult KO IHC, overlaps with the stimulus and APs were absent. The voltage response chosen for Fig. 6D had few APs in order to better show the d.c. component. The horizontal calibration bar in C (1 s) refers to all panels.

for the first stereociliary row to elongate by precluding the action of other capping proteins, twinfilin 2 limits the elongation of the shorter stereocilia to produce the mature staircase architecture of cochlear hair bundles (Peng et al., 2009). What causes the differential expression of actin-binding proteins by distinct stereocilia rows remains to be elucidated. Eps8 and ion channel expression In the absence of Eps8 IHCs, but not OHCs, fail to acquire their mature array of basolateral membrane channels (Zampini et al., 2011). However, similar to cochlear OHCs, the expression of ion channels in vestibular

Eps8-KO Type I and Type II hair cells was unaffected by the mutation. Therefore, the role of Eps8 on ion channel expression appears specific to IHCs. How this is attained, however, is not clear. The absence of IK,n and IK,f, and the persistence of IK,1 in mature Eps8-KO IHCs, indicate that Eps8 could be involved in regulating ion channel expression, possibly through multiple extracellular (BDNF-receptors; Menna et al., 2009) and intracellular (e.g. IRSp53 and Abi-1) partners (Vaggi et al., 2011). Alternatively, Eps8 could regulate the expression of voltage-gated ion channels indirectly, i.e. a secondary effect due to the altered MET apparatus observed in Eps8-KO mice. The stereocilia of IHCs are free-standing

89

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

embedded in the cupula, which might compensate for the shorter stereocilia present in Eps8-KO mice, which would result in a normal transducer current. Eps8 deletion affects auditory but not vestibular function

Fig. 7. Eps8 deletion in cochlear and vestibular hair cells. Schematic drawing showing the effect of Eps8 deletion upon vestibular and cochlear hair cells. A Type I hair cell is shown as representative of semicircular canal hair cells, since similar results were obtained for Type II hair cells. Note the difference in stereocilia height between WT and KO vestibular (present study) and cochlear (Zampini et al., 2011) hair cells. Arrows indicate the first (longest) stereocilia row of the vestibular hair bundle, which was significantly reduced in size in the absence of Eps8 (Fig. 1). The height of the shorter stereocilia has been scaled down to mimic the possible overall effect of the mutation, as seen in cochlear hair cells. The hair bundle of semicircular canal hair cells is almost completely embedded in the cupula (C), a gelatinous mass that bridges the width of the ampulla, forming a fluid barrier through which endolymph cannot circulate. It is the diaphragm-like displacement of the cupula that gives rise to hair bundle displacement during head rotation. At difference, IHCs stereocilia are freely standing (Weiss and Leong, 1985) and behave like hydrodynamic sensors which respond directly to endolymph motion in the space below the tectorial membrane (TM). Ion channel expression is the same in WT and KO vestibular hair cells, whereas KO IHCs lack IK,n and IK,f. IK,DR represents the delayed outward rectifying K+ current including IK,neo-4AP-sensitive and IK,neo-4AP-insensitive.

and behave as hydrodynamic sensors (Guinan, 2012). In Eps8-KO mice, IHCs have short and additional rows of stereocilia (Zampini et al., 2011), which could impact on the resting MET current, the role of which is crucial for the maintenance of the normal spontaneous APs in developing IHCs (Johnson et al., 2012). Alteration of this electrical activity in pre-hearing IHCs has been associated with defects in the synaptic machinery, basolateral membrane ion channels and formation of tonotopic maps in the brainstem (Roux et al., 2009; Johnson et al., 2013; Clause et al., 2014). The normal pattern of basolateral membrane ion channels found in mature vestibular hair cells from Eps8-KO mice suggests that, different from cochlear IHCs, their shorter stereociliary bundle is unlikely to play any developmental role. However, the hair bundle of semicircular canal hair cells is almost completely

Our work has provided some evidence explaining the reason for deafness but not vestibular phenotype in Eps8-KO mice. First of all, the normal endolymph movements produced by the acoustic stimuli is unlikely to be sufficient to stimulate the short hair bundles present in Eps8-KO IHCs. Secondly, even if some MET current is elicited, the altered ion channel expression in the basolateral membrane would preclude converting the stimulus intensity into the gradual receptor potential. Indeed, a mixed digital/analog response, with APs surmounting a d.c. depolarization, was elicited in Eps8KO IHCs when stimulated with a sinusoidal current (Fig. 6). Although neurotransmitter release could still occur in Eps8-KO IHCs, their ability to inform the brain about the timing of the natural stimulus (Palmer and Russell, 1986; Trussell, 2002; Magistretti et al., 2015) would be lost. As a final consideration, several gene mutations have been shown to affect the growth and maintenance of stereocilia in IHCs and produce deafness (Hilgert et al., 2009; Petit and Richardson, 2009). Whether the same mutations also altered vestibular hair bundles, however, has rarely been investigated, mainly because of the absence of overt vestibular dysfunction, e.g. abnormal posturing, imbalance, nystagmus, etc. (Jones and Jones, 2014). This might be due to different roles of particular genes in cochlear and vestibular end organs. Alternatively, it is possible that mild vestibular deficits are partially compensated by the central nervous system, and as such specific vestibular tests, such as evoked potentials, should be performed to detect them (see e.g. Street et al., 2008; Goodyear et al., 2012). Acknowledgments—Tavazzani Elisa, Spaiardi Paolo, Contini Donatella and Manca Marco performed and analyzed the experiments on vestibular hair cells; Zampini Valeria performed and analyzed the experiments on vestibular and cochlear hair cells and helped with the writing of the manuscript; Russo Giancarlo and Prigioni Ivo helped with the design of the experiments, the figures and the writing of the manuscript; Marcotti Walter helped with the experiments and the data analysis of experiments on cochlear hair cells, and the writing of the manuscript, Masetto Sergio designed the experiments, analyzed the data and wrote the manuscript. This work has been financially supported by PRIN, Italy (grant n° 2010599KBR_007 to Ivo Prigioni) and by Fondazione CARIPLO (grant n° 2011-0596 to Sergio Masetto). The funding sources had no involvement in study design, collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

REFERENCES Behlouli A, Bonnet C, Abdi S, Bouaita A, Lelli A, Hardelin JP, Schietroma C, Rous Y, Louha M, Cheknane A, Lebdi H, Boudjelida K, Makrelouf M, Zenati A, Petit C (2014) EPS8,

90

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91

encoding an actin-binding protein of cochlear hair cell stereocilia, is a new causal gene for autosomal recessive profound deafness. Orphanet J Rare Dis 9:55. Benson JA, Adams WB (1987) Significance of the steady-state current voltage relationship. In: Kaczmarek LK, Levitan IB, editors. Neuromodulation: the biochemical control of neuronal excitability. New York: Oxford University Press. p. 105–106. Beurg M, Fettiplace R, Nam JH, Ricci AJ (2009) Localization of inner hair cell mechanotransducer channels using high-speed calcium imaging. Nat Neurosci 12:553–558. }bsamen R, Clause A, Kim G, Sonntag M, Weisz CJ, Vetter DE, Ru Kandler K (2014) The precise temporal pattern of prehearing spontaneous activity is necessary for tonotopic map refinement. Neuron 82:822–835. Contini D, Zampini V, Tavazzani E, Magistretti J, Russo G, Prigioni I, Masetto S (2012) Intercellular K+ accumulation depolarizes Type I vestibular hair cells and their associated afferent nerve calyx. Neuroscience 227:232–246. Di Fiore PP, Scita G (2002) Eps8 in the midst of GTPases. Int J Biochem Cell Biol 34:1178–1183. Eatock RA, Songer JE (2011) Vestibular hair cells and afferents: two channels for head motion signals. Annu Rev Neurosci 34:501–534. Eatock RA, Ru¨sch A, Lysakowski A, Saeki M (1998) Hair cells in mammalian utricles. Otolaryngol Head Neck Surg 119:172–181. Frolenkov GI, Belyantseva IA, Friedman TB, Griffith AJ (2004) Genetic insights into the morphogenesis of inner ear hair cells. Nat Rev Genet 5:489–498. Furness DN, Johnson SL, Manor U, Ru¨ttiger L, Tocchetti A, Offenhauser N, Olt J, Goodyear RJ, Vijayakumar S, Dai Y, Hackney CM, Franz C, Di Fiore PP, Masetto S, Jones SM, Knipper M, Holley MC, Richardson GP, Kachar B, Marcotti W (2013) Progressive hearing loss and gradual deterioration of sensory hair bundles in the ears of mice lacking the actin-binding protein Eps8L2. Proc Natl Acad Sci USA 110:13898–13903. Ge´le´oc GS, Risner JR, Holt JR (2004) Developmental acquisition of voltage-dependent conductances and sensory signaling in hair cells of the embryonic mouse inner ear. J Neurosci 24:11148–11159. Goodyear RJ, Jones SM, Sharifi L, Forge A, Richardson GP (2012) Hair bundle defects and loss of function in the vestibular end organs of mice lacking the receptor-like inositol lipid phosphatase PTPRQ. J Neurosci 32:2762–2772. Guinan Jr JJ (2012) How are inner hair cells stimulated? Evidence for multiple mechanical drives. Hear Res 292:35–50. Hilgert N, Smith RJ, Van Camp G (2009) Function and expression pattern of nonsyndromic deafness genes. Curr Mol Med 9:546–564. Hurley KM, Gaboyard S, Zhong M, Price SD, Wooltorton JR, Lysakowski A, Eatock RA (2006) M-like K+ currents in type I hair cells and calyx afferent endings of the developing rat utricle. J Neurosci 26:10253–10269. Johnson SL, Kennedy HJ, Holley MC, Fettiplace R, Marcotti W (2012) The resting transducer current drives spontaneous activity in prehearing mammalian cochlear inner hair cells. J Neurosci 32:10479–10483. Johnson SL, Kuhn S, Franz C, Ingham N, Furness DN, Knipper M, Steel KP, Adelman JP, Holley MC, Marcotti W (2013) Presynaptic maturation in auditory hair cells requires a critical period of sensory-independent spiking activity. Proc Natl Acad Sci USA 110:8720–8725. Jones SM, Jones TA (2014) Genetics of peripheral vestibular dysfunction: lessons from mutant mouse strains. J Am Acad Audiol 25:289–301. Jones C, Roper VC, Foucher I, Qian D, Banizs B, Petit C, Yoder B, Chen P (2008) Ciliary proteins link basal body polarization to planar cell polarity regulation. Nat Genet 40:69–77. Kros CJ, Ruppersberg JP, Ru¨sch A (1998) Expression of a potassium current in inner hair cells during development of hearing in mice. Nature 394:281–284. Lefe`vre G, Michel V, Weil D, Lepelletier L, Bizard E, Wolfrum U, Hardelin JP, Petit C (2008) A core cochlear phenotype in USH1 mouse

mutants implicates fibrous links of the hair bundle in its cohesion, orientation and differential growth. Development 135:1427–1437. Leibovici M, Verpy E, Goodyear RJ, Zwaenepoel I, Blanchard S, Laine´ S, Richardson GP, Petit C (2005) Initial characterization of kinocilin, a protein of the hair cell kinocilium. Hear Res 203:144–153. Magistretti J, Spaiardi P, Johnson SL, Masetto S (2015) Elementary properties of Ca(2+) channels and their influence on multivesicular release and phase-locking at auditory hair cell ribbon synapses. Front Cell Neurosci 9:123. Manor U, Kachar B (2008) Dynamic length regulation of sensory stereocilia. Semin Cell Dev Biol 19:502–510. Manor U, Disanza A, Grati M, Andrade L, Lin H, Di Fiore PP, Scita G, Kachar B (2011) Regulation of stereocilia length by myosin XVa and whirlin depends on the actin-regulatory protein Eps8. Curr Biol 21:167–172. Marcotti W, Ge´le´oc GS, Lennan GW, Kros CJ (1999) Transient expression of an inwardly rectifying potassium conductance in developing inner and outer hair cells along the mouse cochlea. Pflugers Arch 439:113–122. Marcotti W, Johnson SL, Holley MC, Kros CJ (2003a) Developmental changes in the expression of potassium currents of embryonic, neonatal and mature mouse inner hair cells. J Physiol 548:383–400. Marcotti W, Johnson SL, Ru¨sch A, Kros CJ (2003b) Sodium and calcium currents shape action potentials in immature mouse inner hair cells. J Physiol 552:743–761. Marcotti W, Johnson SL, Kros CJ (2004) A transiently expressed SK current sustains and modulates action potential activity in immature mouse inner hair cells. J Physiol 560:691–708. Menna E, Disanza A, Cagnoli C, Schenk U, Gelsomino G, Frittoli E, Hertzog M, Offenhauser N, Sawallisch C, Kreienkamp HJ, Gertler FB, Di Fiore PP, Scita G, Matteoli M (2009) Eps8 regulates axonal filopodia in hippocampal neurons in response to brain-derived neurotrophic factor (BDNF). PLoS Biol 7(6):e1000138. Nyby JG (2001) Auditory communication among adults. In: Willott JF, editor. The handbook of mouse auditory research from behaviour to molecular biology. Florida: CRC Press. p. 3–18. Offenha¨user N, Castelletti D, Mapelli L, Soppo BE, Regondi MC, Rossi P, D’Angelo E, Frassoni C, Amadeo A, Tocchetti A, Pozzi B, Disanza A, Guarnieri D, Betsholtz C, Scita G, Heberlein U, Di Fiore P (2006) Increased ethanol resistance and consumption in Eps8 knockout mice correlates with altered actin dynamics. Cell 127:213–226. Oliver D, Knipper M, Derst C, Fakler B (2003) Resting potential and submembrane calcium concentration of inner hair cells in the isolated mouse cochlea are set by KCNQ-type potassium channels. J Neurosci 23:2141–2149. Palmer AR, Russell IJ (1986) Phase-locking in the cochlear nerve of the guinea-pig and its relation to the receptor potential of inner hair-cells. Hear Res 24:1–15. Peng AW, Belyantseva IA, Hsu PD, Friedman TB, Heller S (2009) Twinfilin 2 regulates actin filament lengths in cochlear stereocilia. J Neurosci 29:15083–15088. Petit C, Richardson GP (2009) Linking genes underlying deafness to hair-bundle development and function. Nat Neurosci 12:703–710. Rennie KJ, Correia MJ (1994) Potassium currents in mammalian and avian isolated type I semicircular canal hair cells. J Neurophysiol 71:317–329. Roux I, Hosie S, Johnson SL, Bahloul A, Cayet N, Nouaille S, Kros CJ, Petit C, Safieddine S (2009) Myosin VI is required for the proper maturation and function of inner hair cell ribbon synapses. Hum Mol Genet 18:4615–4628. Ru¨sch A, Eatock RA (1996) A delayed rectifier conductance in type I hair cells of the mouse utricle. J Neurophysiol 76:995–1004. Rzadzinska A, Schneider M, Noben-Trauth K, Bartles JR, Kachar B (2005) Balanced levels of Espin are critical for stereociliary growth and length maintenance. Cell Motil Cytoskeleton 62:157–165. Schneider ME, Belyantseva IA, Azevedo RB, Kachar B (2002). Nature 418(837):838. Street VA, Kallman JC, Strombom PD, Bramhall NF, Phillips JO (2008) Vestibular function in families with inherited autosomal dominant hearing loss. J Vestib Res 18:51–58.

E. Tavazzani et al. / Neuroscience 328 (2016) 80–91 Trussell LO (2002) Transmission at the hair cell synapse. Nat Neurosci 5:85–86. Vaggi F, Disanza A, Milanesi F, Di Fiore PP, Menna E, Matteoli M, Gov NS, Scita G, Ciliberto A (2011) The Eps8/IRSp53/VASP network differentially controls actin capping and bundling in filopodia formation. PLoS Comput Biol 7(7):e1002088. Weiss TF, Leong R (1985) A model for signal transmission in an ear having hair cells with free-standing stereocilia. III. Micromechanical stage. Hear Res 20:157–174.

91

Zampini V, Ru¨ttiger L, Johnson SL, Franz C, Furness DN, Waldhaus J, Xiong H, Hackney CM, Holley MC, Offenhauser N, Di Fiore PP, Knipper M, Masetto S, Marcotti W (2011) Eps8 regulates hair bundle length and functional maturation of mammalian auditory hair cells. PLoS Biol 9(4):e1001048. Zenner HP, Zimmermann U, Gitter AH (1990) Cell potential and motility of isolated mammalian vestibular sensory cells. Hear Res 50:289–293.

(Accepted 24 April 2016) (Available online 27 April 2016)