Ca_V1.3 Channels Are Essential for Development and Presynaptic Activity of Cochlear Inner Hair Cells

Impaired stimulus-secretion coupling in Ca_V1.3^-/- mice

Figure 1A displays Ca²⁺ currents of representative 3-week-old wt and Ca_V1.3^-/- IHCs elicited by depolarization to the peak Ca²⁺ current potential. The average current-voltage functions of the Ca²⁺ currents recorded from wt IHCs (n = 6 cells; hearing: p20-32) and Ca_V1.3^-/- IHCs (n = 7 cells; p14-32) are compared in Figure 1B. Together, the data demonstrate the dramatic reduction of the Ca²⁺ current after genetic ablation of Ca_V1.3 (∼92%). Exocytosis, recorded as membrane capacitance increments (ΔC_m) in response to depolarizations of different durations, was similarly reduced in Ca_V1.3^-/- cells (Fig. 1C,D). Two kinetic components of exocytosis can be discriminated in IHCs, with the fast one (up to ∼20-30 msec of stimulation) mainly representing exocytosis of the readily releasable pool of synaptic vesicles (RRP) (Moser and Beutner, 2000). Ca_V1.3^-/- IHCs still displayed a tiny fast secretory component, which can be appreciated from the example C_m trace in Figure 1C. When the C_m changes of Ca_V1.3^-/- IHCs were scaled by a factor of 12.8 (required to match peak Ca²⁺ currents of wt and Ca_V1.3^-/- IHCs) (Fig. 1B), the fast secretory components of IHCs from both genotypes overlapped (Fig. 1D, dashed line). This suggests that the residual Ca²⁺ channels of Ca_V1.3^-/- IHCs couple to fast exocytosis with comparable efficacy, provided that the Ca²⁺ current linearly relates to RRP exocytosis (A. Brandt and T. Moser, unpublished observations).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Strongly reduced Ca²⁺ current and exocytosis in Ca_V1.3^-/- IHCs. A plots representative Ca²⁺ current traces for wt (P31) and Ca_V1.3^-/- (P32) IHCs, which were elicited by step depolarization to the peak Ca²⁺ current potential (Solutions I_i and I_e were used for A-D). Data were leak-subtracted using a p/6 protocol. In B average current-voltage relationships (I-V) recorded from wt (n = 6 cells) and a Ca_V1.3^-/- (n = 7 cells) IHCs are displayed. The dashed line represents a scaled version of the Ca_V1.3^-/-I-V (scaling factor of 12.8 to match peak Ca²⁺ currents). C, D, Capacitance responses (ΔC_m) to paired depolarizations to the peak Ca²⁺ current potential, spaced 100 msec apart and having variable duration of 10, 20, 50, and 100 msec were recorded for wt (n = 6 cells; P20-P32) and Ca_V1.3^-/- (n = 9 cells; P14-P32) IHCs. C shows representative examples of wt and Ca_V1.3^-/- ΔC_m in response to a 2 ^* 20 msec paired pulse (low-pass filtered at 100 Hz). The bars indicate averages of baseline-subtracted Ca_V1.3^-/-C_m over 400 msec before and 40-100 msec after the first and second depolarization. Intervals between the pairs were at least 30 sec to allow for complete recovery of the readily releasable pool (Moser and Beutner, 2000). D plots mean wt and mutant ΔC_m responses to the first depolarization versus the respective duration of the stimuli. E shows the average ΔC_m response of Ca_V1.3^-/- IHCs (n = 9 recordings) to flash-photolysis of caged Ca²⁺. The mean Ca_V1.3^-/- ΔC_m is displayed along with its SE.

If the reduced depolarization-stimulated exocytosis of Ca_V1.3^-/- IHCs was solely attributable to the impaired Ca²⁺ influx, they should secrete normally when bypassing Ca²⁺ influx by direct elevation of Ca²⁺ at the release sites. Indeed, flash photolysis of caged Ca²⁺ stimulated robust ΔC_m in Ca_V1.3^-/- IHCs (Fig. 1E), which displayed two kinetic components as previously described for wt mouse IHCs (Beutner et al., 2001). The total amplitude of the average ΔC_m response (1.88 ± 0.21 pF; n = 9 recordings) was not statistically distinguishable from the mean amplitude of the wt flash responses obtained in a previous study (1.52 ± 0.09 pF; n = 68 recordings; p > 0.1; amplitude was independent of the flash-induced [Ca²⁺]_i change of at least >10 μM) (Beutner et al., 2001). Therefore, the secretory defect of Ca_V1.3^-/- IHCs is not caused by a lack of fusion competent vesicles, but is probably caused by a reduced number of Ca²⁺ channel-coupled release sites. The Ca²⁺ sensitivity of exocytosis in Ca_V1.3^-/- IHCs remains to be investigated in future experiments, including quantification of the flash-induced [Ca²⁺]_i change.

Developmental changes of Ca²⁺ current density: lack of postnatal electrical activity in Ca_V1.3^-/- IHCs

Here, we tested the Ca²⁺ current density of wt and Ca_V1.3^-/- IHCs at the end of the first week [postnatal day 6 (P6)], around and after the onset of hearing (P14-P32). Ca_V1.3^-/- IHCs did not display the developmental increase of membrane capacitance, which is observed in wt IHCs (Table 1) (Beutner and Moser, 2001). This, probably, indicates that Ca_V1.3^-/- IHCs do not increase in size during development. At any age, Ca²⁺ current density (Fig. 2A, top panel) and exocytosis (data not shown) of mutant cells were reduced to ∼5-10% of control. The remaining Ca²⁺ current density of Ca_V1.3^-/- IHCs showed a developmental decrease during the second postnatal week (Fig. 2A, bottom panel), as had been previously described for wt IHCs (NMRI mice) (Beutner and Moser, 2001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Developmental changes of Ca²⁺ current density: lack of Ca²⁺ spikes in Ca_V1.3^-/- IHCs. A shows mean absolute (top panel) and normalized (bottom panel, normalized to P6 values) Ca²⁺ current densities for Ca_V1.3^-/-, C57BL/6J, and NMRI IHCs (the NMRI data set was taken from Beutner and Moser, 2001) at P6 (n_NMRI = 10; n_C57BL/6J = 2; n_Cav1.3^-/- = 5 IHCs), P10 (n_NMRI = 9; n_Cav1.3^-/- = 7 IHCs), and P14-P32 (n_NMRI = 28; n_C57BL/6J = 8; n_Cav1.3^-/- = 8 IHCs). The top panel highlights the dramatic reduction in Ca²⁺ current density in Ca_V1.3^-/- IHCs. The bottom panel emphasizes the developmental current density reduction in all strains investigated. B, Whereas current injection readily triggered APs in wt IHCs (left panel, example is P5), even strong current injections failed to elicit APs in Ca_V1.3^-/- IHCs (right panel, example is P7). The inset compares the initial rise of membrane potential of representative wt and Ca_V1.3^-/- IHCs after injection of 300 pA.

AP firing takes place in immature IHCs of wt mice (Fig. 2B) but not in Ca_V1.3^-/- IHCs. Current-clamp experiments were performed at the end of the first postnatal week, when the Ca²⁺ channel density is largest (Fig. 2A). In contrast to wt IHCs, we observed only passive membrane charging from a slightly more depolarized resting potential (Table 1) followed by a repolarization even during strong depolarizing current injections (Fig. 2B). The depolarized V_m was not caused by increased membrane leak in Ca_V1.3^-/- IHCs (Table 1, R_in). Results similar to those of Figure 2B were obtained in 15 wt IHCs and 9 Ca_V1.3^-/- IHCs. We conclude that the remaining Ca²⁺ current density was too low to support spiking in Ca_V1.3^-/- IHCs.

Ca²⁺ channel types involved in the residual Ca²⁺ current of Ca_V1.3^-/- IHCs

Next, we investigated which channel types mediate the residual Ca²⁺ current in Ca_V1.3^-/- IHCs and how they contribute to stimulus-secretion coupling. To increase the current amplitude, Ba²⁺ (20 mm) was used as the permeating ion for the pharmacological dissection of the remaining Ca²⁺ channels. All experiments were performed between postnatal days 17 and 25. We first tested the presence of L-type Ca²⁺ channels using DHPs, because the voltage dependence and kinetics of the Ca²⁺ current were quite comparable between mutant and control IHCs (Fig. 1A,B). Figure 3A shows that the DHP agonist of L-type Ca²⁺ channels BayK8644 (5 μm; n = 4 IHCs), which prolongs the mean channel open time (Brown et al., 1984), strongly increased the Ca²⁺ current of mutant IHCs to ∼270%. As expected, BayK8644 also shifted the activation to more hyperpolarized potentials. The Ba²⁺ current of mutant IHCs was inhibited by the potent DHP antagonist isradipine (10 μm; n = 7 IHCs) to ∼40% of control (Fig. 3A). Application of the subtype-nonspecific blocker of voltage-gated Ca²⁺ channels Ni²⁺ (5 mm; n = 2 IHCs) (Fig. 3A) abolished DHP-insensitive Ba²⁺ current of Ca_V1.3^-/- IHCs, showing that it was mediated by Ca²⁺ channels.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Identity and function of the Ca²⁺ channels in Ca_V1.3^-/- IHCs. A plots average Ba²⁺ current IVs (mean ± SEM) of mutant IHCs under control conditions and during application of the DHP agonist BayK8644 (5 μm; n = 4 cells), the DHP antagonist isradipine (10 μm; n = 7 cells), or nickel (5 mm; n = 2 cells). Control I-V relationships before application of DHP agonists and antagonists were pooled. Individual I-V relationships were leak-corrected (see Materials and Methods). B compares leak-corrected Ca²⁺ currents and corresponding ΔC_m recorded in the perforated-patch configuration in response to pairs of 500 msec depolarizations to -6 mV during subsequent treatment with 200 nm and 5 μm of isradipine. C plots average Ba²⁺ current I-V relationships of a representative perforated-patch experiment in which application of 1 μm of ω-conotoxin (ctx) GVIA was followed by wash-in of 3 μm of ω-conotoxin MVIIC without any obvious effects on current amplitude or voltage dependence. D summarizes the mean effects of the specified drugs on the Ca_V1.3^-/- IHCs peak Ba²⁺ current. Each group of drugs was compared with their own control data acquired in perforated-patch experiments before the wash-in of the drug. The error probabilities for assuming statistically different mean amplitudes are provided underneath the plot.

At our holding potential (-86 mV), Ca_V1.2 channels are more sensitive to isradipine (complete inhibition by submicromolar isradipine) than Ca_V1.3 channels (micromolar isradipine required for full inhibition) (Koschak et al., 2001). To test a potential contribution of Ca_V1.2 channels, we subsequently applied 200 nm and 5 μm isradipine while recording the Ca²⁺ currents and exocytic C_m changes (Fig. 3B) (depolarization to -6 mV, experiment is representative of two further perforated-patch experiments). The residual Ca²⁺ current and ΔC_m of Ca_V1.3^-/- IHCs were only marginally affected by submicromolar isradipine (200 nm), arguing against a major contribution of Ca_V1.2 channels. In the presence of 5 μm isradipine ΔC_m were abolished, indicating that the remaining Ca²⁺ current did not induce exocytosis, exceeding our detection limit of ∼1 fF (∼30 fused vesicles, assuming a conversion factor of 37 aF per vesicle) (Lenzi et al., 1999).

We then used peptide toxins to probe the contributions of N- (Ca_V2.2), P/Q- (Ca_V2.1), and R-type (Ca_V2.3) Ca²⁺ channels to the Ba²⁺ current of Ca_V1.3^-/- IHCs. Experiments were performed in both the perforated-patch and whole-cell configuration for each drug. Figure 3C shows averaged I-V relationships of a representative perforated-patch experiment in which subsequent application of ω-conotoxin GVIA (1 μm, inhibitor of Ca_V2.2) and ω-conotoxin MVIIC (3 μm, blocking both Ca_V2.2 and Ca_V2.1) failed to reduce the Ba²⁺ current of Ca_V1.3^-/- IHCs. This observation was confirmed in three additional perforated-patch experiments (1 μm ω-conotoxin GVIA followed by 6 μm ω-conotoxin MVIIC).

Next, we investigated the presence of R-type (Ca_V2.3) Ca²⁺ channels by applying the specific toxin inhibitor SNX-482 (Tottene et al., 2000). No significant differences of peak Ba²⁺ currents were observed (1 μm SNX-482; n = 6 IHCs) (Fig. 3D). Interestingly, the inorganic Ca²⁺ channel blocker Ni²⁺ caused a significant inhibition of the Ca²⁺ current at a concentration of 50 μm (∼60%; n = 9 IHCs) (Fig. 3D), at which Ni²⁺ has been shown to selectively inhibit R-type Ca²⁺ channels (Gasparini et al., 2001). Ni²⁺ and isradipine blocked similar amounts of the Ba²⁺ currents in Ca_V1.3^-/- IHCs. Figure 3D summarizes the effects of DHPs, peptide toxins, and Ni²⁺ (50 μm and 5 mm) on the peak Ba²⁺ current of Ca_V1.3^-/- IHCs. We compared test and control responses for each drug. The control data were mainly obtained from perforated-patch experiments before drug application. Together, these experiments show that the remaining Ca²⁺ current in Ca_V1.3^-/- IHCs is mediated by L-type channels and, possibly, to a lesser extent by R-type channels.

Lack of functional large-conductance Ca²⁺-activated K⁺ channels

Mature wt IHCs displayed three distinct K⁺ currents (Fig. 4), which is in agreement with previous reports (Marcotti et al., 2003; Oliver et al., 2003). Depolarization recruited fast and slow outward currents (Fig. 4A). The fast current showed submillisecond activation and was selectively inhibited by low doses of TEA (5 mm) (Fig. 4B), as has been previously described for I_K,f in IHCs (Kros and Crawford, 1990). This current has been attributed to large-conductance Ca²⁺-activated K⁺ channels (BK channels; Langer et al., 2003). The slow outward current was less sensitive to TEA but primarily blocked by 4-aminopyridine (data not shown). This current has been termed I_K,s (Kros and Crawford, 1990) and is thought to be carried by delayed rectifier (K_V) K⁺ channels. When clamping mature IHCs close to their resting potential we observed a conductance, which inactivated during hyperpolarization (Fig. 4E) and was resistant to tetraethylammonium (TEA) and 4-aminopyridine (data not shown). This conductance was most likely mediated by KCNQ channels, which show low voltage activation and set the resting potential of IHCs (Marcotti et al., 2003; Oliver et al., 2003).

Ca_V1.3^-/- IHCs lacked fast outward currents at any age that we investigated (up to P35). Figure 4A demonstrates the kinetic and amplitude differences between representative outward currents of wt (P25) and Ca_V1.3^-/- (P30) IHCs. The lack of I_K,f in Ca_V1.3^-/- IHCs becomes even more obvious, when the currents are displayed normalized to their steady-state values (Fig. 4B). Currents of wt IHCs treated with 5 mm TEA and of Ca_V1.3^-/- IHCs greatly overlapped, except for a small, unblocked I_K,f component in the wt cells. We did not observe any effect of 5 mm TEA on the outward current of Ca_V1.3^-/- IHCs (data not shown).

One might argue that the apparent lack of functional BK channels from Ca_V1.3^-/- IHCs was attributable to the reduced voltage-gated Ca²⁺ influx, which, together with depolarization, normally coactivates the BK channels (Lewis and Hudspeth, 1983; Art and Fettiplace, 1987). However, we show here that elevation of [Ca²⁺]_i by loading Ca²⁺ through the patch pipette (1 mm), which was qualitatively confirmed by fura-2 [Ca²⁺]_i imaging (data not shown), did not reveal I_K,f in Ca_V1.3^-/- IHCs (Fig. 4C). In addition, we did not observe submillisecond current activation in Ca_V1.3^-/- IHCs even at very depolarized potentials (+24 mV), at which I_K,f in excised patches is also activated at low [Ca²⁺] (Oliver et al., 2003). We conclude that Ca_V1.3^-/- IHCs lack functional BK channels. I-V relationships from wt (n = 5 IHCs; P24-P29), wt + 5 mm TEA (n = 2 IHCs; P24, P29) and Ca_V1.3^-/- (high Ca^2+;n = 5 IHCs; P25) IHCs were obtained by time window analysis (Fig. 4D) and by exponential fitting (data not shown) to quantitatively confirm this finding. Traces of wt cells recorded at potentials positive to -45 mV were best fit by the sum of two exponential functions [activation time constants of the fast current component (I_K,f) were <1 msec; those of the slow current components (I_K,s) > 5 msec]. Traces of mutant hair cells were reasonably well described by single exponential functions, with time constants of activation similar to I_K,s of wt cells. Current amplitudes of Ca_V1.3^-/- IHCs and TEA-treated wt cells were similar (Fig. 4B,D), despite the much smaller C_m of Ca_V1.3^-/- IHCs (Table 1), arguing for a compensatory increase of K_V current density.

Ca_V1.3^-/- IHCs also showed KCNQ currents (Fig. 4E). Analysis of the deactivating current component measured after stepping from -65 to -125 mV (after 1.5 msec and at steady-state) revealed comparable mean current densities and deactivation time constants of wt (7.99 ± 0.79 pA/pF; 17.4 ± 1.3 msec; n = 6 cells; P25) and Ca_V1.3^-/- (9.55 ± 2.8 pA/pF; 22.1 ± 2.6 msec; n = 9 cells; P24-P25) IHCs. Together, these experiments indicate that Ca_V1.3 deficiency causes a selective lack of functional BK channels and a compensatory increase of the membrane density of K_V channels.

Persistence of cholinergic postsynaptic currents in Ca_V1.3^-/- IHCs

Here, we studied the developmental changes of efferent synaptic signaling in both Ca_V1.3^-/- and wt IHCs under resting conditions and during high [K⁺]_e stimulation. As shown in Figure 5A, spontaneous postsynaptic currents were readily observed in both groups at the end of the first postnatal week. At holding potentials negative to E_K⁺ (-80 mV), the postsynaptic signals were inward currents. At intermediate potentials (slightly positive to E_K⁺), these signals were composed of a brief, initial inward current followed by a long-lasting outward current (Fig. 5A,C). Figure 5B displays average current-voltage relationships of the slow current component recorded from P6 wt IHCs (n = 3 cells) as well as from P6 (n = 2 cells) and P35 (n = 3 cells) Ca_V1.3^-/- IHCs, with all currents reversing near E_K⁺. We did not obtain any postsynaptic currents in recordings from control IHCs at P25 despite high K⁺ stimulation (10 mm KCl; n = 3 cells; data not shown), which is in line with a previous study on rat IHCs (Glowatzki and Fuchs, 2000). On the contrary, every Ca_V1.3^-/- IHC tested (n = 14 cells; up to P35) displayed spontaneous postsynaptic currents (Fig. 5A). Figure 5, E and F, shows for a representative experiment that these persistent, postsynaptic currents were inhibited by 300 nm strychnine, blocker of the α9-receptor (Elgoyhen et al., 1994) or 1 μm dequalinium, blocker of small-conductance Ca²⁺-activated K⁺ channels (SK) (Strobaek et al., 2000). These findings were confirmed in three more P30-P35 Ca_V1.3^-/- IHCs. In addition, the single event characteristics were comparable among P6 wt IHCs and Ca_V1.3^-/- IHCs at P6 and P35 (Table 2). Thus, our data suggest that the postsynaptic currents of p35 Ca_V1.3^-/- IHCs are mediated by the same mechanism as in chick cochlear hair cells, mammalian outer hair cells, and immature wt IHCs (Fuchs and Murrow, 1992; Glowatzki and Fuchs, 2000; Oliver et al., 2000): transient cholinergic Ca²⁺ influx into the IHCs causing K⁺ current through SK channels. Figure 5D shows that these postsynaptic currents induced a brief depolarization followed by a long hyperpolarization in resting P35 Ca_V1.3^-/- IHCs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Continued presence of postsynaptic responses in Ca_V1.3^-/- IHCs. A compares representative spontaneous postsynaptic currents of wt and Ca_V1.3^-/- IHCs before and after the onset of hearing in wt (potentials are indicated next to the traces). I-V relationships of the slow component of the postsynaptic currents are plotted in B. C displays biphasic postsynaptic currents of a P30 Ca_V1.3^-/- IHC (at -75 mV) in more detail. D shows a stretch of a current-clamp V_m recording from the same cell as in C, illustrating the depolarizing onset and hyperpolarizing main component of these postsynaptic potentials. Extracellular application of 300 nm strychnine (P35; E) and dequalinium (1 μm; F; same cell as in E) abolished the postsynaptic currents of Ca_V1.3^-/- IHCs. The inhibitory effects of both drugs were reversible.

Mechanoelectrical transduction is functional in Ca_V1.3^-/- IHCs

Next, we investigated whether mechanoelectrical transduction (MET) was also impaired in the absence of Ca_V1.3. Stereociliar deflection was achieved by a fluid-jet that was driven by a sinusoidal voltage signal. Transduction currents were clearly present in Ca_V1.3^-/- IHCs. For each recorded cell, driving voltage and proximity of the fluid jet were optimized to gain the cells' maximum MET response. Mean peak transduction currents (P6, holding potential: -75 mV; 60 Hz stimulation) were -104 ± 16.4 pA for Ca_V1.3^-/- (n = 5 cells) and -99.9 ± 10.0 pA for wt (n = 7 cells) IHCs. Smaller amplitudes were observed in p30 wt and Ca_V1.3^-/- IHCs.

Role of Ca_V1.3 channels in stimulus-secretion coupling of IHCs

The Ca²⁺ current of Ca_V1.3^-/- IHCs amounted to only ∼8% of the total Ca²⁺ current measured in IHCs of hearing wt mice (Fig. 1A,B). This is similar to the current reduction observed in IHCs from early postnatal Ca_V1.3^-/- mice (Platzer et al., 2000). Depolarization-induced exocytosis of Ca_V1.3^-/- IHCs was strongly decreased (Fig. 1C,D), confirming the essential role of Ca²⁺ influx for hair cell transmitter release (Moser and Beutner, 2000; Spassova et al., 2001). Exocytosis was abolished when Ca_V1.3^-/- IHCs were treated with micromolar concentrations of the DHP antagonist isradipine (Fig. 3B), which also blocks most of the Ca²⁺ current in wt IHCs (Brandt and Moser, unpublished data). The presence of a small number of rapidly released vesicles indicates that L-type Ca²⁺ channels remaining after Ca_V1.3 ablation couple as efficiently to exocytosis as Ca_V1.3 itself. We suggest that the strong inhibition of IHC transmitter release, which itself is a sufficient reason for the deafness of the mutant, is caused by a reduction of presynaptic Ca²⁺ channel-coupled release sites.

Flash photolysis of caged Ca²⁺ elicited normal exocytic responses (Fig. 1E). This finding of a normal number of fusion competent vesicles is striking, given that recent morphological data suggest a major reduction of synaptic ribbons in Ca_V1.3^-/- IHCs (Glueckert et al., 2003). It has been suggested that synaptic ribbons are involved in sustained transmitter release from sensory synapses as well as in the exocytic responses to flash photolysis (Lenzi and von Gersdorff, 2001). Possible explanations of our finding include: (1) a selective defect of exocytosis caused by the reduction of Ca²⁺ channel-coupled release sites (see above) with a normal population of fusion competent vesicles (docked at and/or outside the active zone), (2) the presence of a large-scale exocytic process that is not related to synaptic transmission. Future measurements of glutamate release from Ca_V1.3^-/- IHCs (Glowatzki and Fuchs, 2002) will be needed to distinguish between these possibilities.

The Ca²⁺ channel-set of the cochlear inner hair cell

The incomplete blockade of hair cell Ca²⁺ currents by various DHP antagonists of L-type Ca²⁺ channels has left some uncertainty about the presence of non-L-type Ca²⁺ channels and their potential contribution to transmitter release (Fuchs et al., 1990; Moser and Beutner, 2000; Rodriguez-Contreras and Yamoah, 2001; Spassova et al., 2001). The initial analysis of Ca_V1.3^-/- mice (Platzer et al., 2000) had demonstrated that Ca_V1.3 is the predominant Ca²⁺ channel of immature mouse cochlear IHCs. Here, we took advantage of Ca_V1.3^-/- mice to demonstrate that Ca_V1.3 also dominates the Ca²⁺ current of IHCs in older mice (∼92%). The apparent discrepancy of the pharmacological and genetic approaches can be reconciled if the voltage dependence of DHP action on Ca_V1.3 is considered (Koschak et al., 2001).

We then used the Ca_V1.3^-/- mice to study the contribution of other Ca²⁺ channels in more detail (Fig. 3). From our pharmacological analysis we conclude that ∼60% of the residual Ca²⁺ current is mediated by L-type channels. Involvement of Ca_V1.1 channels is unlikely, because Ca_V1.1 activates extremely slowly. Also Ca_V1.2 seems not to contribute substantially, because we observed only marginal effects of 200 nm isradipine, whereas 300 nm completely blocks Ca_V1.2 channels (Koschak et al., 2001). Ca_V1.4 remains an attractive candidate for several reasons: (1) it is robustly stimulated by BayK 8644 and requires micromolar concentrations of isradipine for full blockade (Koschak et al., 2003); (2) Ca_V1.4 α1 subunits are expressed in sensory cells (retinal photoreceptor), where they must also be able to tightly couple to neurotransmitter release (Bech-Hansen et al., 1998; Strom et al., 1998); (3) Ca_V1.4 channels inactivate very slowly, similar to the L-type current component in Ca_V1.3^-/- mice reported here. This needs to be confirmed in future biochemical studies of Ca²⁺ channel subunit expression.

None of the subtype-specific peptide toxins, ω-conotoxin GVIA, ω-conotoxin MVIIC, and SNX-482 significantly blocked the DHP-insensitive Ca²⁺ current of Ca_V1.3^-/- IHCs. The inhibition by low micromolar Ni²⁺ could indicate that it was mediated by R-type channels (Gasparini et al., 2001), which then had to be insensitive to SNX-482. Highly Ni²⁺-sensitive R-type current components with varying SNX-482 sensitivity have previously been described (Sochivko et al., 2002). In conclusion, IHCs in addition to Ca_V1.3 contain a small number of further L-type channels of currently unknown molecular identity. A very small fraction of the Ca²⁺ current of the IHCs might be mediated by SNX-482-insensitive R-type channels. N- and P/Q-type Ca²⁺ channels do not contribute.

Impairment of hair cell development in the absence of Ca_V1.3^-/-

Ca_V1.3^-/- IHCs do not generate Ca²⁺ action potentials (Fig. 2B), which are a hallmark of normal IHC development (Kros et al., 1998; Glowatzki and Fuchs, 2000; Beutner and Moser, 2001). The lack of Ca²⁺ action potentials might be at the origin of some of the observed developmental defects. Ca²⁺ action potentials drive presensory hair cell exocytosis (Beutner and Moser, 2001) of glutamate (Glowatzki and Fuchs, 2002) and possibly neurotrophic factors (for review, see Rubel and Fritzsch, 2002), which is probably required for synaptogenesis and maintenance of afferent synapses and neurons during development. Indeed, the first signs of afferent dendrite degeneration are visible by electron microscopy of Ca_V1.3^-/- organs of Corti already at the end of the first postnatal week (Glueckert et al., 2003). It is likely that the defective afferent synaptic transmission and the subsequent degeneration of afferent dendrites lead to the persistent olivocochlear cholinergic input into Ca_V1.3^-/-IHCs (Glueckert et al., 2003) (Fig. 5). The signaling mechanism that causes efferent terminals to give up their axosomatic synaptic contacts with IHCs and to form axodendritic synapses with the afferent fibers during normal development remains to be characterized.

Postnatal Ca²⁺ spikes or the presence of Ca_V1.3 channels in the plasma membrane itself are essential for the acquisition of functional BK channels during hair cell development. It is unlikely that experimental conditions prevented the detection of BK currents in Ca_V1.3^-/- IHCs. We never observed submillisecond activation of outward currents in Ca_V1.3^-/- IHCs even at very depolarized potentials (+24 mV) and did not find any change in activation kinetics after including 1 mm Ca²⁺ into the pipette. Therefore, the absence of BK current in Ca_V1.3^-/- IHCs reflects a lack of functional BK channels rather than a defective BK channel activation caused by reduced Ca²⁺ influx.

The regenerative [Ca²⁺] changes of immature IHCs could coregulate the BK channel expression together with thyroid hormone (which is normal in Ca_V1.3^-/- mice; Platzer et al., 2000). However, it currently seems more likely that the lack of BK current in Ca_V1.3^-/- IHCs is not caused by defective gene expression. Preliminary data obtained by in situ hybridization suggests the presence of BK channel mRNA in Ca_V1.3^-/- IHCs (Waka et al., 2003). Alternatively, BK channels may not get sufficiently targeted or anchored at the hair cell plasma membrane in the absence of Ca_V1.3. Future experiments will be required to clarify the exact mechanism by which Ca_V1.3 Ca²⁺ channels regulate the abundance of functional BK channels in IHCs.

Clinical implications of the Ca_V1.3^-/- phenotype

So far, no human hereditary deafness has been reported to result from a channelopathy of Ca_V1.3. This is in contrast to the visual system, where mutations in the gene coding for the retinal L-type Ca²⁺ channel (Ca_V1.4) cause X-linked congenital stationary night blindness (Bech-Hansen et al., 1998; Strom et al., 1998). Defects of IHCs, their synapses or the auditory nerve are believed to cause a specific type of hearing impairment called auditory neuropathy. It includes a large variety of pathological conditions ranging from hereditary IHC lesions to demyelinating disease (Doyle et al., 1998; Brown and Dort, 2001; Lesinski-Schiedat et al., 2001; Varga et al., 2003). The “early” Ca_V1.3^-/- mouse resembles a synaptic model of auditory neuropathy, as the blockade of hair cell afferent synaptic transmission represents the primary defect before degeneration of afferent fibers and outer hair cells eventually occurs. Further genetic analysis of hearing-impaired families will be required to explore whether mutations in the gene coding for Ca_V1.3 cause human auditory neuropathy.