Ca_v 1.3 (α1D) Ca²⁺ currents in neonatal outer hair cells of mice

Abstract

Outer hair cells (OHC) serve as electromechanical amplifiers that guarantee the unique sensitivity and frequency selectivity of the mammalian cochlea. It is unknown whether the afferent fibres connected to adult OHCs are functional. If so, voltage-activated Ca²⁺ channels would be required for afferent synaptic transmission. In neonatal OHCs, Ca²⁺ channels seem to play a role in maturation since OHCs from Ca_v1.3-deficient (Ca_v1.3^−/−) mice degenerate shortly after the onset of hearing. We therefore studied whole-cell Ca²⁺ currents in outer hair cells aged between postnatal day 1 (P1) and P8. OHCs showed a rapidly activating inward current that was 1.8 times larger with 10 mm Ba²⁺ as charge carrier (I_Ba) than with equimolar Ca²⁺ (I_Ca). I_Ba started activating at −50 mV with V_max = −1.9 ± 6.9 mV, V_0.5 = −15.0 ± 7.1 mV and k = 8.2± 1.1 mV (n = 34). The peak I_Ba showed negligible inactivation (3.6 % after 300 ms) whereas the I_Ca (10 mm Ca²⁺) was inactivated by 50.7 %. OHC I_Ba was reduced by 33.5 ± 10.3 % (n = 14) with 10 μm nifedipine and increased to 178.5 ± 57.8 % (n = 14) by 5 μm Bay K 8644. A dose-response curve for nifedipine revealed an IC₅₀ of 2.3 μm, a Hill coefficient of 2.7 and a maximum block of 36 %. Average I_Ba density in OHCs was 24.4 ± 10.8 pA pF⁻¹ (n = 105) which is only 38 % of the value in inner hair cells. Single cell RT-PCR revealed expression of Ca_v1.3 in OHCs. In OHCs from Ca_v1.3^−/− mice, Ba²⁺ current density was reduced to 0.6 ± 0.5 pA pF⁻¹ (n = 9) indicating that > 97 % of the Ca²⁺ channel current in OHCs flows through Ca_v1.3.

Outer hair cells (OHCs) are specialised sensory cells of the mammalian cochlea that are part of the cochlear amplifier. In contrast to inner hair cells (IHCs) which produce graded receptor potentials and release excitatory transmitter onto afferent fibres in response to an acoustic stimulus, OHCs respond to depolarisation by excitatory transduction currents with a shortening of their cell body (Brownell et al. 1985). It is assumed that this length change feeds energy into the cochlear partition which is the basis for the extraordinary sensitivity and frequency selectivity of the mammalian inner ear (Davis, 1983; Dallos & Corey, 1991). Although each OHC is innervated by several afferent nerve fibres (type II afferents, for review see Pujol et al. 1997; Felix, 2002) it is not known whether OHCs do have functional presynapses and if so, what purpose the afferent fibres may serve. Small L-type Ca²⁺ currents have been found in guinea pig OHCs (Nakagawa et al. 1991; Chen et al. 1995) but their function remains unclear. Transmitter release in IHCs is dependent on Ca²⁺ influx through voltage-activated Ca²⁺ channels (Moser & Beutner, 2000). The underlying Ca²⁺ currents in neonatal mouse IHCs have fast activation and deactivation kinetics, show little inactivation, activate at negative voltages (around −65 mV in 1.3 mm external Ca²⁺) and have L-type channel pharmacology (Platzer et al. 2000). IHC Ca²⁺ currents thus resemble Ca²⁺ currents recorded in hair cells of auditory or vestibular organs of other vertebrates including the frog sacculus (Lewis & Hudspeth, 1983; Chabbert, 1997; Armstrong & Roberts, 1998; Rodriguez-Contreras & Yamoah, 2001); the frog crista ampullaris (Martini et al. 2000; Russo et al. 2003); the turtle papilla (Art et al. 1986; Art & Fettiplace, 1987; Ricci et al. 2000; Schnee & Ricci, 2003); and the chick papilla (Fuchs et al. 1990; Zidanic & Fuchs, 1995; Kollmar et al. 1997; Spassova et al. 2001). The molecular basis of the IHC L-type Ca²⁺ currents was investigated using a mouse model lacking Ca_v1.3 α1 (or α1D) subunits (Ca_v1.3^−/−, for nomenclature see Ertel et al. 2000). In IHCs of Ca_v1.3^−/− mice, the average Ba²⁺ inward current (I_Ba) density was reduced to less than 8 % of the wild-type value indicating that the majority of Ca²⁺ channels in mouse IHCs are Ca_v1.3 L-type Ca²⁺ channels (LTCCs, Platzer et al. 2000). The same channel type had been cloned before from hair cells of chick basilar papilla (Kollmar et al. 1997). Hair cell degeneration in Ca_v1.3^−/− mice starting in OHCs around postnatal day 14 (P14) and involving IHCs later on (Platzer et al. 2000; Glueckert et al. 2003) further suggested that Ca_v1.3 currents are normally present in wild-type mouse OHCs and are necessary for their differentiation and survival. We therefore investigated voltage-activated whole-cell Ca²⁺ currents in neonatal (P1-P8) OHCs. Single cell RT-PCR analysis and electrophysiological recordings in OHCs from Ca_v1.3^−/− mice were employed to determine the contribution of Ca_v1.3 channels for OHC Ca²⁺ currents.

METHODS

Tissue preparation

OHCs of NMRI mice (Charles River, Sulzfeld, Germany) and Ca_v1.3^−/− mice (Platzer et al. 2000) were studied either in organotypic cochlear cultures aged < P7 (Glowatzki et al. 1997) or following acute dissection of the organ of Corti (P1-P8). Animals were killed by decapitation and cochleae were removed according to national ethical guidelines. The organ of Corti was cut in three pieces: apical, medial and basal. Pieces were either mounted on coverslips with Cell Tak (Beckton Dickinson, Bedford, MA, USA) and maintained in culture for up to 24 h or a piece was fixed on a coverslip for immediate use by using Cell Tak or fixing it with two glass fibres. There was no difference regarding the current amplitudes or other current properties between OHCs from acutely isolated or cultured explants of the organ of Corti. For measuring I_Ba through Ca²⁺ channels, the recording chamber was constantly perfused with an extracellular solution composed of (mm): 160 Tris-(hydroxymethylamino)-methane HCl, 10 BaCl₂, 5.6 glucose, 1 MgCl₂, pH 7.4, 320 mosmol kg⁻¹ (‘10 mm Ba²⁺ solution’). Tris⁺ ions were used instead of Na⁺ and K⁺ to avoid any currents through voltage-activated Na⁺ channels present in rat OHCs (Oliver et al. 1997) and through K⁺ channels. When Ca²⁺ was used as charge carrier, Ba²⁺ ions were replaced by 10 mm Ca²⁺ (‘10 mm Ca²⁺ solution’) or 1.3 mm Ca²⁺ (‘1.3 mm Ca²⁺ solution’). In the latter case osmolarity was kept at 320 mosmol kg⁻¹ by adding Tris HCl. Ca²⁺- and Ba²⁺-free extracellular solution consisted of (mm) 2.3 MgCl₂, 173 TrisCl, 5.6 glucose, 2 EGTA, pH 7.4, 320 mosmol kg⁻¹. For some recordings of I_Ba in Ca_v1.3^−/− OHCs, the extracellular solution contained (mm): 139 TEACl, 15 4-aminopyridine (4-AP), 10 BaCl₂, 5.6 glucose, 1 MgCl₂, 0.0005 TTX (Alomone Labs, Israel; ‘TEA solution’). Test solutions were applied by a gravity-fed application system with a micropipette placed near the OHC of interest. The application pipette was made from theta glass (Hilgenberg, Malsfeld, Germany) allowing the subsequent application of two additional solutions in one experiment. The recording chamber was mounted on an upright microscope (Zeiss Axioskop, Oberkochen, Germany) and viewed with differential interference contrast and a ×40 water immersion objective. OHCs were exposed by removing the tectorial membrane and adjacent Hensen cells with cleaning pipettes made from borosilicate glass.

Electrical recording

Membrane currents from OHCs (third and second row) were recorded at 20–22 °C by the whole-cell patch clamp technique using an AXOPATCH 200B amplifier (Axon Instruments, Union City, CA, USA). Patch pipettes with a resistance of 4–6 MΩ were made from quartz glass. Using pipettes with a lower resistance resulted in very small or nondetectable Ca²⁺ inward current (I_Ca). Pipettes were filled with (mm): 110 cesium gluconate (prepared from CsOH and gluconic acid by repetitive re-crystallisation), 20 CsCl, 10 sodium phosphocreatine, 5 Hepes, 5 EGTA, 4 MgCl₂, 4 Na₂ATP, 0.1 CaCl₂, 0.3 GTP, pH 7.35, 305 mosmol kg⁻¹, with a free Ca²⁺ concentration of 1.7 nm (calculated according to Bers et al. 1994; see also http://www.stanford.edu/~cpatton/maxc.html). Series resistance (R_s) compensation was only used when R_s was larger than 12 MΩ. Raw currents were corrected for linear leak currrents and potentials were corrected by −20 mV for liquid junction potentials (LJP) measured according to Neher (1992). In detail, LJPs were 20.1 ± 0.3 mV (n = 3) for the 10 mm Ba²⁺ solution; 20.2 ± 0.2 mV (n = 3) for the 10 mm Ca²⁺ solution; 19.9 ± 0.5 mV (n = 3) for the 1.3 mm Ca²⁺ solution; and 20.4 ± 0.5 mV (n = 3) for the TEA solution. Calculated LJPs (Barry, 1994) were very similar and deviated by less than 1 mV from measured LJPs. Data were acquired using the software Pulse++ 1.7 (developed by Ulrich Rexhausen, Institute of Physiology II, Tuebingen, Germany, and Klaus Bauer, Max Planck Institute for Medical Research, Heidelberg, Germany) and an ITC16 interface (Instrutech, Greatneck, NY, USA) and were stored on a Macintosh computer. Current traces were sampled at 50 kHz and filtered at 10 kHz (fast protocols) or at 5 kHz/1.3 kHz (slow protocols). Data analysis and fitting was carried out using the Igor software package (Wavemetrics, Lake Oswego, OR, USA).

Current-voltage (I-V) curves of Ba²⁺ currents obtained by step depolarisations were fitted according to:

(1)

where I is the I_Ba at the time the I–V was determined, V is the membrane potential, V_rev is the extrapolated reversal potential, G_max is the maximum conductance of the cell, V_0.5 is the voltage for half-maximum activation and k is the slope factor of the Boltzmann function. The percentage of inactivation of the peak current after 300 ms was calculated as:

(2)

with I_peak being the current maximum of the peak current trace at t = t_peak and I_test the current determined at t = t_peak+ 300 ms. The inactivation time course of I_Ca in the 300 ms following the peak current was fitted with a double exponential function:

(3)

with A₀, A₁ and A₂ being constants and τ₁ and τ₂ being the fast and slow time constants of inactivation, respectively. For the dose-response curve for nifedipine, maximum I_Ba was elicited every 10 s with repetitive voltage pulses from −90 to 0 mV each lasting 40 ms. When a constant current level was reached - typically 3 min after establishing the whole-cell configuration when a run-up of I_Ba had saturated - the respective nifedipine concentration (dissolved in DMSO; concentration of DMSO was < 0.1 %) in the 10 mm Ba²⁺ solution was applied for 90 s and subsequently washed out. The percentage of I_Ba inhibition was determined for between five and ten OHCs, averaged, and a least-squares fit to the data was performed with a Hill equation:

(4)

with I_max,inh being the maximum percentage of I_Ba inhibition, IC₅₀ the nifedipine concentration for half-maximum inhibition, [nif] the nifedipine concentration and n_H the Hill coefficient.

Single cell RT-PCR analysis

Following acute dissection of the organ of Corti (P2-5), OHCs and IHCs were identified under the microscope using morphological criteria. After removing supporting cells with cleaning pipettes, between 1 and 15 target cells were gently sucked into a borosilicate glass collection micropipette with a diameter of 6–8 μm (Langer et al. 2003). The collected cells were immediately transferred to a prechilled microcentrifuge tube containing 20 μl of RT mix consisting of: 1 × reverse transcription buffer (Invitrogen), 10 mm dithiothreitol (Invitrogen), 0.5 mm each of dATP, dCTP, dGTP, dTTP (Roche Molecular Biochemicals), 2 U μl⁻¹ of RNase inhibitor (Peqlab), 25 μg ml⁻¹ of Oligo-dT primer (Roche Molecular Biochemicals) and 2.5 U μl⁻¹ of reverse transcriptase Superscript II (Invitrogen). Negative controls were also performed with perfusion solution to check for contamination by cell debris and with RT mix lacking the reverse transcriptase as control for contamination by genomic DNA. For the cDNA first strand synthesis, each microcentrifuge tube was incubated at 25 °C for 10 min followed by 60 min at 42 °C. The cDNA was stored at −20 °C until PCR analysis. PCR analysis was then performed with 9 μl of the reverse transcription product in a total volume of 50 μl of a PCR mix with final concentrations of 10 mm Tris, 50 mm KCl, 2 mm MgCl₂, 200 μm dNTPs, 0.3 μm of each primer, 1–2 U Taq Polymerase (Invitrogen) through 30 cycles (30” at 94 °C; 30” at 60 °C; 45” at 72 °C). The following primers were used to amplify a 421 bp stretch of the Ca_v1.3 subunit from exon 24 to exon 28 (Green et al. 1996) which is not subject to alternative splicing: sense primer: 5′-CTC GGT TGT GAA GAT TCT GAG-3′; antisense primer: 5′-GAT GAC GAA GCC CAC GAA GAT-3′. A second round of 40 cycles (same program) was necessary to amplify a detectable product using 5 μl of the first PCR product as a template with the following nested primers: sense: 5′-GAG GGT CTT GCG GCC TCT CAG AGC-3′; antisense: 5′-ATC TCC ACA CGG TAG TTG TAG ACA GGA-3′. PCR products were analysed by agarose gel electrophoresis. The Ca_v1.3 specificity of the PCR product was confirmed by DNA sequencing (MWG Biotech, Germany).

Statistics

Data are given as means ±s.d. unless otherwise indicated. Statistical comparison of means were made by Student's two-tailed t test with P < 0.05 as the criterion for statistical significance. All current traces shown in the figures are single traces.

RESULTS

Identity of OHC inward currents

With Tris⁺ ions in the external solution and Cs⁺ ions in the pipette solution, i.e. under conditions in which most of the currents through K⁺ and Na⁺ channels were blocked, depolarising voltage steps evoked inward currents (Fig. 1A) carried by either Ca²⁺ or Ba²⁺. I_Ca showed little (1.3 mm Ca²⁺) to moderate (10 mm Ca²⁺) inactivation whereas I_Ba (10 mm Ba²⁺) did not inactivate (Fig. 1A and C). Peak current-voltage (I-V) curves (Fig. 1B) were bell-shaped which is typical for currents through voltage-activated Ca²⁺ and also Na⁺ channels. They showed that the maximum current amplitude was small in 1.3 mm Ca²⁺, increased in 10 mm Ca²⁺ and further increased in 10 mm Ba²⁺ (Fig. 1B). Changing the charge carrier from 1.3 to 10 mm Ca²⁺ also shifted the voltage of maximum current (V_max) towards positive values, most probably due to a surface charge screening effect (Zhou & Jones, 1995). The I–V curve of I_Ba was typically shifted to negative values by about 5 mV as compared to the I–V of I_Ca obtained with equimolar Ca²⁺ (Fig. 1B) which is a known phenomenon of Ca²⁺ channel currents and has also been found in frog saccular hair cells (Rodriguez-Contreras & Yamoah, 2003). When the external solution contained neither Ca²⁺ nor Ba²⁺ and was supplemented with 2 mm EGTA only a tiny inward current of 4.3 pA (6.6 % of the control I_Ba in this OHC) could be evoked as shown by the peak current traces and I–V curves in Fig. 1C and D. Taken together these data demonstrate that neonatal OHCs possess voltage-activated Ca²⁺ currents.

Figure 1. Neonatal outer hair cells (OHCs) possess an inward current carried by Ca²⁺ or Ba²⁺.

A, depolarisation-evoked peak inward current traces of an OHC (postnatal day 6 (P6), holding potential: −70 mV) with 1.3 mm Ca²⁺, 10 mm Ca²⁺ or 10 mm Ba²⁺ in the extracellular solution. B, corresponding peak current-voltage (I–V) curves for the different charge carriers for the cell shown in A. C, I_Ba (10 mm Ba²⁺) in another OHC (P7) was greatly diminished in the absence of permeant divalent cations as shown by the current traces recorded at −10 mV and by the corresponding peak I–V curves (D). E, average I_Ba density ±s.d. as a function of age. The numbers of cells for a given age are indicated in the bars.

I_Ca and I_Ba were recorded in a number of cells and current densities (peak current divided by cell capacitance) were determined. Average peak current densities were 7.3 ± 2.2 pA pF⁻¹ (n = 6) in 1.3 mm Ca²⁺, 13.2 ± 5.7 pA pF⁻¹ (n = 14) in 10 mm Ca²⁺ and 24.4 ± 10.8 pA pF⁻¹ (n = 105) in 10 mm Ba²⁺ solution, yielding a ratio of current densities of 1:1.8:3.3 for the three different ion conditions. Because I_Ba was 1.8 times larger than equimolar I_Ca we mostly used 10 mm Ba²⁺ for characterising the Ca²⁺ channel currents in OHCs. Ba²⁺ current densities varied with age (Fig. 1E) but did not depend on the cochlear location. The average I_Ba density showed a maximum of 35.6 pA pF⁻¹ at P2, remained elevated above 19 pA pF⁻¹ until P7 and drastically decreased to 6.7 pA pF⁻¹ at P8. This contrasts to the developmental time course of the Ca²⁺ current density in IHCs which shows a pronounced maximum at P6 and only half of that maximum value at P2 (Beutner & Moser, 2001).

Activation properties of OHC Ca²⁺ channel currents

Figure 2 shows rapidly activating I_Ba evoked by depolarising voltage steps of 8 ms duration. Activation time constants were comparable to those of IHC I_Ba (τ_act≈0.3 ms for the peak current trace, Platzer et al. 2000). Figure 1B shows the corresponding I–V curve taken at the last millisecond of the depolarising voltage pulse. The current started activating at −50 mV and reached its maximum at −2.5 mV. Fitting the I–V curve with a Boltzmann function times driving force (see Methods) yielded the following parameters: V_0.5 = −15.5 mV, k = 7.2 mV and reversal potential (E_rev) = 50.6 mV.

Figure 2. The OHC I_Ba activates and deactivates rapidly.

A, selected current traces evoked by depolarisations from the holding potential of −90 mV to the potentials indicated show rapid activation and deactivation of I_Ba in an OHC (P4). B, the I–V curve was taken at the last millisecond of the voltage pulse and fitted by a first order Boltzmann equation times the driving force, see Methods.

Parameters such as maximum inward current, I_max; cell capacitance, C; maximum current density, I_Ba,D; potential of maximum inward current, V_max; potential of half-maximum activation, V_0.5; and the steepness of the Boltzmann fit, k, were determined for both outer and inner hair cells. As we aimed at comparing properties of I_Ba in both cell types, critical experimental conditions had to be identical. To compare maximum I_Ba, only those recordings from IHCs were included that had been carried out with the same intracellular solution containing purified cesium gluconate (see Methods) as was used for all OHC measurements since non-purified cesium gluconate reduced IHC peak currents (data not shown). For the comparison of voltages such as V_0.5 and V_max only those IHC recordings were taken into account that had been carried out with the same intra- and extracellular solutions as used in OHC experiments in order to exclude any possible voltage difference brought about by LJPs between the bath and the local perfusion solution used in most IHC measurements (Platzer et al. 2000). For these reasons, a relatively small number of IHC data was included in Table 1 although more data exist that had been acquired under slightly different conditions (Platzer et al. 2000).

Table 1.

Biophysical properties of Ba²⁺ currents in neonatal outer (P1–P7) and inner (P2–P9) hair cells

	Imax	C	IBa,D	Vmax	V0.5	k
Cell type	(pA)	(pF)	(pA pF−1)	(mV)	(mV)	(mV)
OHC	112.4 ± 48.7*	4.7 ± 0.8*	24.4 ± 10.8*	−1.9 ± 6.9	−15.0 ± 7.1*	8.2 ± 1.1*
	n = 105	n = 105	n = 105	n = 34	n = 34	n = 34
IHC	421 ± 206	6.5 ± 0.4	64.7 ± 29.4	−5.0 ± 4.1	−19.6 ± 4.6	7.6 ± 0.7
	n = 5	n = 5	n = 5	n = 13	n = 13	n = 13

P, postnatal day. Maximum current (I_max), cell capacitance (C), maximum current density (I_Ba,D), voltage of maximum current (V_max), half-maximal voltage for activation (V_0.5) and the slope for activation (k) for outer hair cells (OHC) and inner hair cells (IHC) are given as means ±s.d.V_0.5 and k were determined by fitting the data to a first order Boltzmann function as described in Methods.

^*Statistically significant difference (P < 0.05).

The most prominent difference between OHC and IHC I_Ba was the difference in average maximum current amplitude between P1 and P7 (112.4 pA or 27 % in OHCs compared to 421 pA in IHCs). To compensate for different cell sizes, current densities were determined and compared between OHCs and IHC. Average I_Ba density of OHCs was 24.4 pA pF⁻¹ which is only 38 % of that in IHCs. V_max and V_0.5 were slightly more positive in OHCs than in IHC (by 3.1 and 4.6 mV, respectively), and the steepness of the Boltzmann function was smaller in OHCs compared to IHCs.

Inactivation properties of OHC Ca²⁺ channel currents

OHC I_Ca showed little to moderate inactivation whereas in most cases I_Ba did not inactivate at all. OHCs were inhomogeneous with respect to inactivation in 10 mm Ca²⁺ solution: of 15 OHCs tested four cells did not show any inactivation like I_Ba, three cells showed weak inactivation (i.e. < 40 % reduction of the peak I_Ca during a 400 ms voltage pulse) and eight cells showed strong inactivation (> 40 % reduction of the peak I_Ca).

Figure 3 shows current responses of an OHC with strong inactivation in 10 mm Ca²⁺. To quantify the voltage-dependence of inactivation, the following inactivation protocol was used (Fig. 3A): starting from a holding potential of −70 mV OHCs were subjected to a conditioning pulse of 400 ms duration ranging from −70 to 30 mV before a test pulse to 0 mV was applied. Figure 3B shows that I_Ba did not inactivate during the conditioning pulse except at the highest potential tested (30 mV). In contrast, I_Ca significantly inactivated at potentials of −30 mV and above (Fig. 3C), consistent with the presence of Ca²⁺-dependent inactivation in OHC (but considerably less in IHC) Ca²⁺ channels. This is also demonstrated by the Ba²⁺ and Ca²⁺I–V curves (Fig. 3D and E) which were determined at the beginning (closed symbols) and at the end of the prepulse (open symbols). The I–V curves taken at the beginning of the prepulse further show that I_Ba was larger than I_Ca in the same cell. Figure 3F depicts the voltage dependence of I_Ba and I_Ca inactivation in terms of the normalised test pulse current (I_test/I_control) obtained by dividing the test current amplitudes (I_test) by the test current amplitude elicited without a depolarising prepulse (I_control). The normalised test pulse I_Ca was about 1 at prepulse potentials between −70 and −40 mV before it declined rapidly due to Ca²⁺-dependent inactivation. Interestingly, the normalised test pulse I_Ba showed an increase rather than a decrease with increasing prepulse potential. This is consistent with the known prepulse facilitation of LTCCs (Bourinet et al. 1994; Dolphin, 1996; Dai et al. 1999). When the inactivation curves were averaged for eight OHCs with strong inactivation of I_Ca, a slightly U-shaped inactivation curve with a minimum at 10 mV became apparent (Fig. 3G). This reflects that inactivation is maximal around V_max where Ca²⁺ influx is maximal and decreases when at more depolarised potentials I_Ca decreases due to the reduced driving force. On average, I_Ca inactivated by approximately 70 % during a 400 ms depolarising prepulse to 10 mV.

Figure 3. Inactivation behaviour of I_Ba and I_Ca in OHCs.

A, pulse protocol to determine voltage-dependent inactivation. Selected voltage-activated inward current traces are shown for Ba²⁺ (B) and Ca²⁺ (C) as charge carriers. D and E, I–V curves at the beginning (closed symbols) and at the end of the conditioning pulse (open symbols) for I_Ba (D) and I_Ca (E). F, normalised inactivation curves for I_Ba and I_Ca were obtained by dividing the test pulse current (I_test) by the current amplitudes without depolarising prepulse (I_control), respectively. G, average normalised I_Ca inactivation curve as a function of membrane potential (left ordinate, I_test/I_test,max; obtained from 8 OHCs). Its slight U-shape indicates Ca²⁺-dependent inactivation. For comparison, the average normalised I_Ca peak I–V curve of the prepulse (right ordinate, I_pre/I_pre,max) is shown to illustrate the voltage of peak I_Ca and the subsequent decline of I_Ca due to the reduced driving force.

To characterise the time course of inactivation, I_Ca traces were fitted with exponential functions which worked well with double exponentials (Fig. 4A). Average fast time constants of inactivation ranging between 10 and 20 ms did not vary with voltage whereas average slow time constants showed large variation between 105 and 200 ms and tended to become faster with increasing depolarisations (Fig. 4B). The average ratios A₁/A₂ of the two exponential inactivation terms (Methods, eqn (3)) amounted to 0.36-0.60 and did not show a systematic dependence on membrane potential. Ba²⁺ current traces with their weak inactivation could not be fitted systematically. To compare the inactivation time courses of I_Ba and I_Ca, the percentage of inactivation of the peak current after 300 ms was determined yielding 3.6 ± 1.5 % (n = 10) for I_Ba and 50.7 ± 23.7 % (n = 10) for I_Ca (10 mm Ca²⁺) which compares to 11.1 ± 4.8 % (n = 18) and 21.1 ± 9.3 % (n = 5) in IHCs, respectively (Platzer et al. 2000).

Figure 4. Kinetics of Ca²⁺-dependent inactivation during 400 ms depolarising voltage pulses.

A, peak I_Ca traces elicited by depolarisations to +0 mV from three OHCs are shown with respective double-exponential fits (black lines). From top to bottom trace, fast time constants of inactivation were 10.4, 12.0 and 34.6 ms; slow time constants were 91.8, 130.0 and 168.0 ms, and the ratio between the amplitudes of the fast and the slow inactivation term (A₁/A₂, see Methods, eqn (3)) was 0.36, 0.67 and 0.36, respectively. B, average fast and slow I_Ca inactivation time constants ±s.d. for 8 OHCs obtained from the double-exponential fits as a function of the membrane potential.

Pharmacological properties of OHC Ca²⁺ channels

To determine if DHP-sensitive LTCCs contribute to neonatal OHC Ca²⁺ channel currents we tested their modulation by the LTCC activator Bay K 8644. Channel activation by this DHP is a highly specific property of LTCCs. Superfusion with 5 μm Bay K 8644 increased the OHC I_Ba as shown by the peak I_Ba traces and I–V curves (Fig. 5). On average, Bay K 8644 increased the peak I_Ba from 92.4 ± 49.5 to 158.9 ± 76.5 pA (n = 14) which corresponds to an increase to 178.5 ± 57.8 % of the control. Because in IHCs the Bay K 8644-induced increase of I_Ba was much larger (to 292 %, Platzer et al. 2000) we tested different lots of Bay K 8644 to rule out that the low sensitivity of OHC I_Ba was due to a batch of poor quality. We also tested the same agonist solution with both OHCs and IHCs but the results did not change, confirming that IHC and OHC I_Ba indeed have different sensitivities to Bay K 8644.

Figure 5. The LTCC agonist Bay K 8644 increases the OHC I_Ba.

A, peak current traces before (control) and during the application of 5 μm Bay K 8644 in the bath in a 4-day-old OHC. B, corresponding I–V curves taken at the last millisecond of the voltage pulse are shown.

Next we determined the sensitivity of OHC currents to the DHP LTCC blocker nifedipine. Figure 6A and B shows peak I_Ba traces of an OHC before and during the application of nifedipine (10 μm) and corresponding I–V curves. Nifedipine reduced the peak current to 56.5 % of the control confirming that at least part of I_Ba was carried by LTCCs. After scaling the I–V curve of the nifedipine-resistant current to the control I–V curve (Fig. 6C) the two I–V curves matched remarkably. A dose-response curve for nifedipine was accomplished to define the sensitivity of the OHC I_Ba for nifedipine (Fig. 6D). I_Ba was partially blocked by nifedipine between 1 and 10 μm; higher concentrations (30 μm) did not further inhibit I_Ba. Fitting the data with a Hill equation (see Methods) gave a maximum inhibition of I_Ba of 36.3 ± 0.9 %, an IC₅₀ of 2.3 ± 0.1 μm and a Hill coefficient of 2.7 ± 0.4. The high IC₅₀ indicates that the LTCCs in OHCs have a low sensitivity to nifedipine. Incomplete block of I_Ba, however, could still result from a mixture of LTCCs and non-LTCCs. I_Ba inhibition at 10 μm nifedipine in OHCs (33.5 ± 10.3 % block of I_Ba, n = 14) was similar to IHC LTCCs (44.5 ± 11.3 % block of I_Ba, n = 4) but much lower than for Ca_v1.2 LTCCs (Koschak et al. 2001; Xu & Lipscombe, 2001). Taken together these pharmacological data suggest that Ca_v1.3 is also the major contributor to OHC L-type currents.

Figure 6. The LTCC antagonist nifedipine partially blocks the OHC I_Ba.

A, peak current traces before (control) and during the application of 10 μm nifedipine in the bath in an OHC (P4). B, corresponding I–V curves before and during nifedipine superfusion that were taken at the last millisecond of the voltage pulse. C, to compare I–V curves of the total current and the nifedipine-resistant current, I–V curves were scaled to their maximum. D, dose-response curve of the OHC I_Ba for nifedipine. For each nifedipine concentration, the percentage of I_Ba block was averaged for 5–10 OHCs. Fitting I_Ba block with a Hill equation yielded a maximum inhibition of 36.3 ± 0.9 %, an IC₅₀ of 2.3 ± 0.1 μm and a Hill coefficient of 2.7 ± 0.4.

Molecular identity of OHC Ca²⁺ channels

Single-cell RT-PCR experiments were performed to confirm that Ca_v1.3 α1 transcripts are present in OHCs. A 421 bp Ca_v1.3 α1-specific PCR product comprising a region not subject to alternative splicing (see Methods) was amplified (Fig. 7) in three independent samples of RNA harvested from IHCs or OHCs at P4.

Figure 7. Detection of Ca_v1.3 α1 subunit transcripts by single-cell RT-PCR.

RNA from IHCs, OHCs or epithelial (Hensen) cells was collected and subjected to RT-PCR as described in Methods. Three separate RNA samples obtained from pooled IHCs (6–7 cells per pool) or OHCs (21–35 cells per pool) were analysed. Lanes 1–4: control samples; lane 1: no template control; lane 2: Hepes-Hanks’ solution without cells; lane 3: OHC sample processed without reverse transcriptase; lane 4: epithelial (Hensen) cells; lanes 5–7: OHC preparations; lanes 8–10: IHC preparations.

To directly quantify the contribution of Ca_v 1.3 for total Ca²⁺ channel current in OHCs electrophysiological recordings were performed in Ca_v1.3^−/− mice. In Ca_v1.3^−/− OHCs, the Ba²⁺ current density was dramatically reduced (Fig. 8) as shown by the peak I_Ba trace (Fig. 8A, control) and the corresponding I–V curve (Fig. 8B, control). The large upward transients in the current traces that were prominent prior to the small inward currents in Ca_v1.3^−/− OHCs had a linear I–V curve with a reversal potential of approximately −90 mV and a conductance of about 0.6 nS and were unaffected by the K⁺ channel blockers TEA and 4-AP and the Na⁺ channel blocker TTX (data not shown). They might reflect outward currents carried by Cs⁺ or Na⁺ ions through a small fraction of the transduction channels, since it has been shown that 5–10 % of all transduction channels are open without mechanical stimulation (Geleoc et al. 1997). However, at present we do not understand why this outward current is transient and vanishes within the first 2 ms of the depolarising pulse.

Figure 8. The LTCC current is absent in OHCs from Ca_v1.3^−/− mice.

A, Ba²⁺ inward current traces at 0 mV in a 2-day-old Ca_v1.3^−/− OHC in the TEA solution before and during superfusion with 5 μm Bay K 8644. B, corresponding I–V curves taken at the last millisecond of the voltage pulse show that the maximum inward control current was as small as −13 pA and was not increased by Bay K 8644. C, mean I_Ba densities ±s.d. for Ca_v1.3^+/+ and Ca_v1.3^−/− mice obtained from the number of cells indicated.

The largest inward current of all Ca_v1.3^−/− OHCs tested had a peak amplitude of −13 pA that could not be increased by Bay K 8644 (Fig. 8A and B), and some OHCs showed no inward current at all. Average peak I_Ba was 2.8 ± 3.2 pA under control conditions and 3.0 ± 3.9 pA in the presence of 5 μm Bay K 8644 (n = 9). Cell capacitance as a measure of cell size was not affected by gene disruption (wild-type: 4.7 ± 0.8 pA pF⁻¹, n = 105; Ca_v1.3^−/−: 4.5 ± 1.1 pA pF⁻¹, n = 9). The average current density was reduced from 24.4 ± 10.8 pA pF⁻¹ (n = 105) in wild-type OHCs to 0.6 ± 0.5 pA pF⁻¹ (n = 9) in Ca_v1.3^−/− OHCs (Fig. 8C) leading us to conclude that more than 97 % of I_Ba in OHCs flows through Ca_v1.3 LTCCs.

DISCUSSION

Here we show the first analysis of Ca²⁺ currents in mouse OHCs aged from P1-P8. Neonatal OHCs seem to be the only mammalian cell type known so far that exclusively contains Ca_v1.3 Ca²⁺ channels. Ca_v1.3 channels are expressed in a variety of neurons, neuroendocrine, endocrine, cardiac pacemaker and sensory cells (Hell et al. 1993; Ihara et al. 1995; Platzer et al. 2000; Zhang et al. 2002; Mangoni et al. 2003). Their role for (neuro-)secretion and neurotransmission may be widely underestimated because no specific blocker exists that can differentiate between Ca_v1.3 and the widely distributed Ca_v1.2 (α1C) LTTCs. Interpretation of data regarding the role of different Ca²⁺ channel classes in synaptic transmission by sequentially blocking them with known Ca²⁺ channel blockers often disregards the fact that Ca_v1.3 channels are incompletely blocked by micromolar concentrations of LTCC blockers such as nifedipine and nimodipine. As a result, Ca_v1.3 currents may be misinterpreted as being non-LTCCs. Neonatal OHCs therefore seem to be a good model cell to study native Ca_v1.3 currents in isolation. Ca_v1.3 channels have been expressed in heterologous systems only recently (Scholze et al. 2001; Koschak et al. 2001; Xu & Lipscombe, 2001; Bell et al. 2001; Safa et al. 2001). The specific properties of native OHC Ca²⁺ channel currents show many similarities to currents from Xu & Lipscombe (2001) such as negative voltage of activation and lack of I_Ba inactivation. The differences in current properties of heterologously expressed Ca_v1.3 currents could be caused by different splice variants of the α1 subunit used in those studies.

Ca_v1.3-mediated currents in OHCs show many similarities with the Ca²⁺ currents in IHCs as well as with Ca²⁺ currents in auditory hair cells of chick (Zidanic & Fuchs, 1995) and turtle (Schnee & Ricci, 2003). With respect to hair cells of the mouse, the ratio I_Ba/I_Ca, the fast kinetics of activation, the lack of rapid inactivation of Ba²⁺ currents, and the partial block by nifedipine were similar in OHCs and IHCs. However, many differences between Ca²⁺ channel currents in the two hair cell types were found such as V_max, V_0.5 and k of I_Ba, the kinetics of inactivation of I_Ca and I_Ba, and the stimulation by Bay K 8644. Interestingly, Ca_v1.3 currents in most OHCs exhibited Ca²⁺-dependent inactivation similar to that in turtle auditory hair cells (Schnee & Ricci, 2003) whereas Ca²⁺-dependent inactivation was less pronounced in mouse IHCs (Platzer et al. 2000) and in guinea pig OHCs (Nakagawa et al. 1991; Chen et al. 1995). We have previously shown that recombinant Ca_v1.3 channels, like Ca_v1.3 in OHCs, also inactivate in a Ca²⁺-dependent manner (Koschak et al. 2001). Ca²⁺-dependent inactivation is largely determined by structural motifs in the C-terminal tail and involves calmodulin binding (Budde et al. 2002). Both rat or human C-terminal splice variants (Hui et al. 1991; Williams et al. 1992) identified so far contain the structural motifs required for Ca²⁺-dependent inactivation. It remains to be determined whether yet unknown C-terminal splice variants, accessory subunits (Dai et al. 1999) or other protein-protein interactions prevent Ca²⁺-dependent inactivation of Ca_v1.3 in IHCs. The recent findings on the modulation of heterologously expressed Ca_v1.3 channels by the presynaptic protein syntaxin that both increased the peak current density and slowed inactivation point in this direction (Song et al. 2003). The observed differences in current properties of inner and outer hair cells may result from: (i) IHC- and OHC-specific splicing of the α1 subunit, (ii) the association with different subtypes or splice variants of the auxiliary subunits β and α2δ or (iii) the interaction with different presynaptic molecules or scaffolding proteins. Single cell RT-PCR could clarify the prevalence of different Ca_v1.3 splice variants or different types of auxiliary subunits in IHCs and OHCs.

A major difference between outer and inner hair cells was the threefold lower I_Ba density in OHCs compared to IHCs. Before the onset of hearing around P12 (Rybak et al. 1992) IHCs already release transmitter onto afferent fibres in a Ca²⁺-dependent fashion (Beutner & Moser, 2001; Glowatzki & Fuchs, 2002). At this age, however, IHCs are not yet able to respond to an acoustic stimulus with a graded receptor potential but show spontaneous slow oscillations of the membrane potential (Kros et al. 1998, Marcotti et al. 1999, 2003) that depend on extracellular Ca²⁺. Ca²⁺-driven action potentials are a widespread phenomenon in the developing nervous system leading to the refinement of neuronal connectivity (for review see Mennerick & Zorumski, 2000; Spitzer, 2002). It is assumed that the periodic release of glutamate from the IHC caused by slow Ca²⁺ action potentials leads to refinement and maturation of synaptic connections in the auditory periphery and in the central auditory pathway. Oscillations of the membrane potential resembling the Ca²⁺ action potentials in IHCs have also been observed in OHCs but only after mild current injection (at least 40 pA, Marcotti & Kros, 1999). The threefold lower I_Ba density might be the reason for the lack of spontaneous Ca²⁺ action potentials in neonatal OHCs in vitro. It is not known, however, whether depolarising mechanisms exist in vivo that could drive the cell to the oscillation threshold. If Ca²⁺ action potentials indeed occurred in OHCs they could be involved in the formation of synaptic connections. Afferent and efferent innervation of OHCs are both built up during the first and second postnatal week, including: (i) innervation by collaterals of type I afferents at P0, (ii) innervation by type II afferents, (iii) retraction of type I afferent collaterals around P6 (Pujol et al. 1985, 1997) and (iv) efferent innervation by the medial olivocochlear bundle between P3 and P10 (Sobkowicz, 1992).

A general function of the Ca²⁺ channels in neonatal OHCs that has to be considered is their possible involvement in OHC-specific gene expression. Ca²⁺ currents, especially those through LTCCs, are involved in the regulation of gene expression in different cells and tissues (for review see Finkbeiner & Greenberg, 1998; Bradley & Finkbeiner, 2002). This could explain why the total lack of Ca²⁺ influx in Ca_v1.3^−/− mice prevents OHCs from adopting an adult phenotype. The following degeneration of OHCs observed in Ca_v1.3^−/− mice is probably the answer of the system to immature, non-functional OHCs around the time when hearing starts in wild-type animals.

The question remains if adult OHCs do have functional Ca²⁺ channels and, if so, for what purpose. We have shown that between P7 and P8, around the time when the OHC begins to adopt its mature phenotype by expressing the fast K⁺ current I_K,n (Marcotti & Kros, 1999) and the motor protein prestin (Weber et al. 2002), I_Ba density drops by 65 %. The presence of small, noninactivating L-type Ca²⁺ currents in guinea pig OHCs (Nakagawa et al. 1991; Chen et al. 1995) suggests that LTCCs are also of importance in mature OHCs of the mouse.