Biophysical and pharmacological characterization of nicotinic cholinergic receptors in rat cochlear inner hair cells

Abstract

Before the onset of hearing, a transient efferent innervation is found on inner hair cells (IHCs). This synapse is inhibitory and mediated by a nicotinic cholinergic receptor (nAChR) probably formed by the α9 and α10 subunits. We analysed the pharmacological and biophysical characteristics of the native nAChR using whole-cell recordings from IHCs in acutely excised apical turns of the rat organ of Corti. Nicotine did not activate but rather blocked the acetylcholine (ACh)-evoked currents with an IC₅₀ of 1 ± 0.1 μm. Antagonists of non-cholinergic receptors such as strychnine, bicuculline and ICS-205930 blocked ACh-evoked responses with an IC₅₀ of 8.6 ± 0.8 nm, 59 ± 4 nm and 0.30 ± 0.02 μm, respectively. The IHC nAChR was both permeable to (P_Ca/P_Na = 8 ± 0.9) and modulated by external Ca²⁺. ACh-evoked currents were potentiated by Ca²⁺ up to 500 μm but were reduced by higher concentrations of this cation. Ba²⁺ mimicked the effects of Ca²⁺ whereas Mg²⁺ only blocked these currents. In addition, elevation of extracellular Ca²⁺ reduced the amplitude of spontaneous synaptic currents without affecting their time course. The receptor had an EC₅₀ for ACh of 60.7 ± 2.8 μm in 0.5 mm Ca²⁺. In the absence of Ca²⁺, the EC₅₀ for ACh increased, suggesting that potentiation by Ca²⁺ involves changes in the apparent affinity for the agonist. These pharmacological and biophysical characteristics of the IHC nAChR closely resemble those of the recombinant α9α10 nAChR, reinforcing the hypothesis that the functional nAChR at the olivocochlear efferent–IHC synapse is composed of both the α9 and α10 subunits.

In mammals, sound waves are converted into electrical signals by two types of mechanotransducer hair cells, inner and outer hair cells (IHCs and OHCs, respectively), present in the sensory epithelium of the cochlea. Adult IHCs are largely innervated by afferent nerve fibres, whereas OHCs are the main target of descending olivocochlear efferent fibres (Guinan, 1996). However, before the onset of hearing (around P12 in rats), olivocochlear efferent fibres transiently innervate immature IHCs prior to reaching their final OHC targets (Simmons et al. 1996; Pujol et al. 1998). Synaptic release from efferent neurones modulates the firing frequency of IHCs (Glowatzki & Fuchs, 2000) and could play a role in the establishment of the adult innervation to the cochlea (Walsh et al. 1998). Cholinergic innervation of vertebrate hair cells is brought about by Ca²⁺ entering through a nicotinic acetylcholine receptor (nAChR) and the subsequent activation of small conductance calcium-dependent potassium channels (SK channels) (Fuchs & Murrow, 1992a; Doi & Ohmori, 1993; Blanchet et al. 1996; Dulon & Lenoir, 1996; Evans, 1996; Fuchs, 1996; Nenov et al. 1996b; Dulon et al. 1998; Glowatzki & Fuchs, 2000; Oliver et al. 2000). In rat OHCs, it has also been recently shown that the inhibitory response mediated by the SK channels is boosted by Ca²⁺ released from a postsynaptic calcium storage organelle ‘the synaptoplasmic cistern’, subsequent to the activation of the nAChRs (Lioudyno et al. 2004).

Pharmacological and biophysical studies performed with the native cholinergic receptors present in mammalian OHCs (Doi & Ohmori, 1993; Erostegui et al. 1994c; Blanchet et al. 1996; Chen et al. 1996; Dulon & Lenoir, 1996; Evans, 1996; Nenov et al. 1996b; Oliver et al. 2000) and in chick short hair cells, equivalent to mammalian OHCs (Fuchs & Murrow, 1992a; McNiven et al. 1996), as well as cellular localization data (Elgoyhen et al. 1994; Park et al. 1997; Luo et al. 1998; Morley et al. 1998; Elgoyhen et al. 2001; Morley & Simmons, 2002), strongly suggest that the native OHC receptor is composed of both the α9 and α10 nicotinic subunits (Elgoyhen et al. 1994; Elgoyhen et al. 2001; Weisstaub et al. 2002). Both the native OHC nAChR and the recombinant α9 and α9α10 nAChR have a peculiar pharmacological profile. A hallmark of these receptors is that they are not activated by nicotine, the prototypic agonist of the family (Elgoyhen et al. 1994, 2001; Sgard et al. 2002). Studies performed with the recombinant α9 and α9α10 nAChRs expressed in Xenopus laevis oocytes (Katz et al. 2000; Weisstaub et al. 2002; Sgard et al. 2002) demonstrated that both receptors have a high Ca²⁺ permeability. There are other biophysical features, however, in which the α9 and α9α10 nAChR differ significantly, namely, voltage sensitivity, modulation by Ca²⁺ and desensitization pattern (Katz et al. 2000; Elgoyhen et al. 2001; Weisstaub et al. 2002). Aside from the fact that the α9 and α10 subunits are the only nicotinic subunits expressed by cochlear OHCs and IHCs (Elgoyhen et al. 1994, 2001; Park et al. 1997; Luo et al. 1998; Morley et al. 1998; Morley & Simmons, 2002), studies with mice carrying a null mutation in the Alpha9 gene showed that these α9 knock-out mice are functionally de-efferented, thus indicating that the α9 subunit is a key component of the native receptor (Vetter et al. 1999). Moreover, we have recently shown that the developmental loss of functional nAChRs at the transient olivocochlear efferent–IHC synapse is strongly correlated with diminished expression of the α10 subunit, thus supporting a key role for this subunit in the functional expression of hair cell nAChRs as well (Katz et al. 2004).

In the present work we characterized pharmacological and biophysical features of the native nAChR present at the olivocochlear efferent–IHC synapse prior to the onset of hearing. We show that these characteristics of the IHC nAChR studied in isolation from the SK2 channels are very similar to those described for the recombinant α9α10 nAChR, providing further evidence that the native IHC functional receptor is composed of both the α9 and α10 subunits.

Methods

Animal procedures and isolation of the organ of Corti

Sprague-Dawley rats at postnatal ages 9–11 (P9–11; day of birth was considered P0) were anaesthetized with an i.p. injection of sodium pentobarbital. Animals were decapitated after assessing that a deep anaesthetic state had been obtained (by observing the lack of tail flick following tail pinch, and lack of eye blink response following a corneal touch). All experimental protocols were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publications no. 80-23) revised 1978. Apical turns of the organ of Corti were excised as previously described (Glowatzki & Fuchs, 2000) and used within 3 h. Cochlear preparations were mounted under an Axioskope microscope (Zeiss, Oberkochem, Germany) and viewed with differential interference contrast (DIC) using a 63× water immersion objective and a camera with contrast enhancement (Hamamatsu C2400-07, Hamamatsu City, Japan).

Electrophysiological recordings

Methods to record from IHCs were essentially as described (Glowatzki & Fuchs, 2000). Briefly, IHCs were identified visually, then by their whole-cell capacitance (7–12 pF) and by their characteristic voltage-dependent Na⁺ and K⁺ currents, including at older ages a fast-activating K⁺ conductance (Kros et al. 1998). Some cells were removed to access IHCs, but mostly the pipette moved through the tissue using positive fluid flow to clear the tip.

Currents in IHCs were recorded in the whole-cell patch-clamp mode using an Axopatch 200B amplifier (Axon Instruments, Union City, CA, USA), low-pass filtered at 2–10 kHz and digitized at 5–20 kHz with a Digidata 1200 board (Axon Instruments). Recordings were made at room temperature (22–25°C). Glass pipettes, 1.2 mm i.d., had resistances of 7–10 MΩ.

Solutions

The cochlear preparation was continuously superfused (∼0.5 ml min⁻¹) with an extracellular solution with an ionic composition similar to that of the perilymph (mm): 144 NaCl, 5.8 KCl, 1.3 CaCl₂, 0.9 MgCl₂, 0.7 NaH₂PO₄, 5.6 d-glucose and 10 Hepes buffer; pH was adjusted 7.4 with NaOH. For the Ca²⁺ permeability experiments and for the modulation of spontaneous synaptic currents by Ca²⁺, the extracellular solution was as follows (mm): 100 NaCl, 5.8 KCl, 1.3 mm CaCl₂, 0.9 MgCl₂, 5.6 d-glucose and 10 Hepes buffer; pH 7.4 (osmolality was adjusted to 300–320 mosmol kg⁻¹ with sucrose ∼26 g l⁻¹).

Solutions containing different divalent cation concentrations, with or without ACh, were applied by a gravity-fed multichannel glass pipette (∼150 μm tip diameter) positioned about 300 μm from the recorded IHC. All working solutions applied via the gravity-fed multi-channel glass pipette, except for those designed to study the effects of Mg²⁺ on the ACh-evoked responses, were made up in a saline containing 0.5 mm Ca²⁺ (unless otherwise stated) and without Mg²⁺, so as to optimize the experimental conditions for measuring currents flowing through the α9α10 receptors (Weisstaub et al. 2002). Working solutions contained (mm): 144 NaCl, 5.8 KCl, 0.5 CaCl₂, 0.7 NaH₂PO₄, 5.6 d-glucose, and 10 Hepes buffer; pH was adjusted 7.4 with NaOH. For the Ca²⁺ permeability experiments and for the modulation of spontaneous synaptic currents by Ca²⁺, the working solution was as follows (mm): 100 NaCl, 5.8 KCl, 0–10 mm CaCl₂, 5.6 d-glucose, and 10 Hepes buffer; pH 7.4 (osmolality was adjusted to 300–320 mosmol kg⁻¹ with sucrose ∼26 g l⁻¹).

The pipette solutions contained the following (mm). (1) KCl-EGTA saline (for recording the combined nAChR + SK response): 135 KCl, 3.5 MgCl₂, 0.1 CaCl₂, 5 ethyleneglycol-bis(β-aminoethyl ether)-N,N,N′,N′- tetraacetic acid (EGTA), 5 Hepes buffer, 2.5 Na₂ATP, pH adjusted to 7.2 with KOH. (2) KCl-BAPTA saline (for recording the isolated nAChR response): to preclude the activation of SK currents by Ca²⁺, in some experiments EGTA in the pipette solution was replaced by 10 mm of the fast calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA); 135 KCl, 3.5 MgCl₂, 0.1 CaCl₂, 10 BAPTA, 5 Hepes buffer, 2.5 Na₂ATP, pH adjusted to 7.2 with KOH. In order to minimize further the contribution of SK channel currents, in addition to using BAPTA in the pipette solution, the SK channel blocker apamin (1–10 nm) (Kohler et al. 1996) was added to the external working solutions.

In the experiments designed to study the voltage dependence of the nAChR currents and/or the changes in the current reversal potential, KCl in the KCl-BAPTA saline was replaced by either CsCl or Cs-gluconate, to eliminate voltage-dependent K⁺ currents (Kros et al. 1998). The osmolality of intracellular solutions ranged between 270 and 290 mosmol kg⁻¹. Spontaneous synaptic currents were recorded immediately after rupturing into the cell, in the extracellular saline containing 0.5 mm Ca²⁺ and with a recording pipette containing either a KCl-EGTA (combined nAChR + SK synaptic currents) or a KCl-BAPTA saline solution (isolated nAChR synaptic currents).

Evaluation of divalent cation permeability

For these experiments, ionic currents were elicited by ACh in IHCs held at different membrane potentials (−40 to +30 mV, in 5 mV steps). The relative calcium to sodium permeability (P_Ca/P_Na) was evaluated by computing the reversal potential (E_rev) of ACh-evoked currents upon changing the external Ca²⁺ concentration from 0.1 mm to 10 mm while keeping a fixed Na⁺ concentration of 100 mm. Data were then fitted by the constant field Goldman-Hodgkin-Katz (GHK) voltage equation extended to include divalent cations (Jan & Jan, 1976). Our assumptions were that Na⁺, K⁺ and Cs⁺ are equally permeant and that Cl⁻ is impermeant through this channel; in addition we made no allowance for the effect of surface charge on the value of P_Ca/P_Na (Mayer & Westbrook, 1987). Thus, taking into account the permeability to K⁺, Na⁺ and Ca²⁺, the E_rev was evaluated to be:

in which:

where P_K, P_Na and P_Ca are the membrane permeabilities to K⁺, Na⁺ and Ca²⁺, respectively (under our experimental conditions [K⁺]_i was replaced by [Cs⁺]_i); R is the molar gas constant, F is the Faraday constant, and T is absolute temperature (RT/F is 25.3 mV at 20°C). For comparison with previous data on the recombinant rat α9 and α9α10 nAChR (Katz et al. 2000; Weisstaub et al. 2002) and their recombinant human orthologues (Sgard et al. 2002), we used ionic concentrations and not ionic activities in our calculations.

Evaluation of the modulation of the native receptor by divalent cations

The effects of extracellular Ca²⁺, Ba²⁺ and Mg²⁺ on the ionic currents through the nAChR receptor were studied by measuring the amplitudes of the responses to 100 μm ACh upon varying the concentration of these cations from 0 to 10 mm. In all cases the IHC under study was preincubated for 1 min with a saline containing the concentration of divalent cations to be tested.

Amplitude values obtained at each concentration were normalized to that obtained in the same cell at a 0.5 mm concentration of the divalent cation being studied. Values from different IHCs were averaged and expressed as the mean ± s.e.m.

Evaluation of inhibition

The IC₅₀ values for the inhibition by Mg²⁺ and other antagonists (nicotine, strychnine, ICS-205930 and bicuculline) were evaluated in the presence of 0.5 mm Ca²⁺ and calculated by:

where I is the current obtained at the different concentrations of the antagonist, I_max is the current obtained in the absence of antagonist, X is the logarithm of the antagonist concentration and n_H is the Hill coefficient. IC₅₀ values were obtained for each IHC and then averaged. Data are presented as means ±s.e.m.

To study the inhibitory effects of Mg²⁺ or the other antagonists, in all cases the IHC under study was preincubated for 1 min with the same concentration of the antagonist to be tested.

Evaluation of the effect of divalent cations on the affinity of the nAChR for ACh

To evaluate changes in affinity and maximal response to ACh, concentration–response curves to ACh were carried out in the same IHCs at nominally 0 and 0.5 mm Ca²⁺. Values were normalized to the maximum obtained at 0.5 mm Ca²⁺. EC₅₀ values were calculated by:

Data analysis

For I–V curves and voltage dependence of divalent cation block, voltages were corrected for liquid junction potentials (−4 mV and −17 mV for the KCl (or CsCl) and Cs-gluconate intracellular solutions, respectively). Voltages were not corrected for the voltage drop across the uncompensated series resistance. Spontaneous synaptic currents were analysed with Minianalysis (Synaptosoft, Jaejin Software, Leonia, NJ, USA) and were identified using a search routine for event detection and confirmed by eye. τ_decay values were fitted with a monoexponential. Statistical analyses were carried by ANOVA followed by Tukey's test for multiple comparisons or by Student's two-tailed t test (when only two populations of data were compared). In both cases P < 0.05 was considered significant.

Materials

ACh-chloride, BAPTA free acid, Na₂ATP, strychnine-HCl (–)-bicuculline methbromide, ICS-205930-HCl, (–)-nicotine-di-d-tartrate, apamin and all other reagents were from Sigma Chemical Co. (St Louis, MO, USA). BAPTA and Na₂ATP were dissolved at the moment of preparing the intracellular solutions. All other drugs were dissolved in distilled water as 1–10 or 100 mm stocks and stored in aliquots at −20°C.

Results

Voltage sensitivity, pharmacological profile and desensitization pattern of the isolated nAChR current

Application of 100 μm ACh, in the presence of a low Ca²⁺–Mg²⁺ free saline solution in the bath and using a KCl-EGTA-based pipette solution, caused inward currents in postnatal (P9–11) IHCs voltage clamped at −90 mV and outward currents at −50 and −30 mV. The current–voltage relation (Fig. 1A, nAChR + SK) was bell shaped and the average reversal potential, −67.6 ± 4.2 mV (n = 5), was near the potassium equilibrium potential (E_K, −82 mV). Under these conditions, ACh-activated currents were mediated by a combined flux of cations through the α9α10-containing nAChR and a K⁺ flux through calcium-dependent SK channels. Negative to E_K, at −90 mV, the ACh-activated inward current is mediated by current flow through both the nAChR channel and the SK channel whereas at voltages between E_K and 0 mV, the outward current flows through SK channels. In order to isolate the nAChR current, the calcium-activated SK current was blocked by using the calcium chelator BAPTA (10 mm) instead of EGTA in the pipette solution plus the addition of 1 nm apamin, a specific SK channel blocker (Köler et al. 1996), to the bath solution. Under these conditions the residual currents reversed nearer to 0 mV (E_rev =−10.7 ± 1.6 mV; n = 3), and the I–V relation was more linear (Fig. 1A, nAChR).

Figure 1. Voltage sensitivity and pharmacological profile of the native IHC nAChR.

A, representative I–V relationships illustrating the voltage dependence of the combined (nAChR + SK, KCl-EGTA saline solution in the recording pipette) and isolated responses (nAChR; CsCl-BAPTA saline solution in the recording pipette plus 1 nm apamin in the extracellular solution) evoked by exogenously applied ACh (100 μm). B, concentration–response curves in the presence of different antagonists: strychnine (IC₅₀: 8.6 ± 0.8 nm; n_H: 0.65 ± 0.1); ICS-205930 (IC₅₀: 59 ± 4 nm; n_H:1.2 ± 0.2), bicuculline (IC50: 0.3 ± 0.02 μm; n_H: 0.9 ± 0.1) and nicotine (IC₅₀: 1 ± 0.1 μm; n_H: 0.5 ± 0.1). Responses were normalized to the response obtained with 60 μm ACh in the absence of antagonist. C, representative records of currents evoked by 60 μm ACh in the absence or presence of different concentrations of only two of the antagonists used (strychnine and nicotine). In all cases the effects of the antagonists were completely reversed by washing the cells in a saline solution without antagonist. In A and B, the records and data points for I–V and concentration–response curves are representative of results obtained in 3–6 IHCs. In B and C all ACh responses were elicited in IHCs voltage clamped at –90 mV.

We characterized the effects of drugs reported to block both the recombinant α9α10 nAChRs and OHC native receptors (Elgoyhen et al. 1994; Erostegui et al. 1994c; Chen et al. 1996; Rothlin et al. 1999; Verbitsky et al. 2000; Elgoyhen et al. 2001), on the isolated nAChR-mediated current in IHCs. Nicotine is the prototypic cholinergic agonist of nicotinic receptors, except for the recombinant α9 and α9α10 receptors on which it behaves as an antagonist (Elgoyhen et al. 1994, 2001). When tested in IHCs, nicotine did not evoke currents itself but rather blocked the ACh-evoked responses in a reversible manner with an IC₅₀ of 1 ± 0.1 μm (n = 3–5 cells; Fig. 1B). Another pharmacological hallmark of the recombinant α9 and α9α10 nAChRs is that they are potently blocked by antagonists of non-cholinergic receptors such as strychnine (glycine receptors), bicuculline (γ-aminobutyric acid type A receptors), and ICS-205930 (ligand gated serotonin receptors) (Rothlin et al. 1999; Elgoyhen et al. 2001). As illustrated by the concentration–response curves in Fig. 1B, these three compounds blocked ACh-evoked responses in IHCs with an IC₅₀ of 8.6 ± 0.8 nm for strychnine (n = 3–5 cells), 59 ± 4 nm for ICS-205930 (n = 4 cells) and 0.3 ± 0.02 μm for bicuculline (n = 3–5 cells). The effects of all four drugs were readily reversed by superfusing the preparation with normal saline solution (Fig. 1C).

We have previously shown that homomeric α9 and heteromeric α9α10 channels differed in their desensitization kinetics. Whereas responses to ACh in α9α10-expressing oocytes are markedly reduced upon the continuous (1 min) or intermittent (10 s at 0.025 Hz) application of the agonist, in α9-expressing oocytes, ACh-evoked responses do not desensitize with either protocol (Elgoyhen et al. 2001). Figure 2 shows that the native nAChR in IHCs also desensitized upon either the continuous (1 min) or intermittent (10 s at 0.05 Hz) application of ACh. In the continuous presence of 1 mm ACh, currents through the nAChR decayed with a time constant of 13.8 ± 0.5 s (n = 4, Fig. 2A and C). In the case of intermittent application of the agonist, peak amplitudes decayed with a time constant of 37.9 ± 4.2 s (n = 3, Fig. 2B and D). It must be borne in mind that the decay time constant values obtained for desensitization might be flawed by the limitation imposed by the slow perfusion of the intact cochlear coil. Solution exchange around the cell, as evaluated by the change in holding current when changing form 5.8 mm to 40 mm KCl, is complete in around 1–3 s according to the position of the perfusion pipette in relation to the cell under study (data not shown, see Glowatzki & Fuchs, 2000; Katz et al. 2004). This slow ACh application, which is similar to the slow perfusion method used to study the recombinant α9 and α9α10 in Xenopus oocytes (see Elgoyhen et al. 2001), should lead to an underestimation of the peak current as desensitization would start before the peak is reached, and therefore to an underestimation of desensitization itself.

Figure 2. Desensitization pattern of the native IHC nAChR.

Representative records illustrating that ACh-evoked responses desensitize both for the continued (A) and intermittent application of ACh (B), obtained in 3–4 IHCs voltage clamped at –90 mV. C, plot of I/I_maxversus time during the continuous application of 1 mm ACh. Illustrated is the average trace obtained from 4 IHCs. The grey line is a monoexponential fit to the data (τ_decay = 13.8 ± 0.5 s). D, plot of I/I_maxversus time during the intermittent application of ACh. Data points are the average current amplitudes (n = 3 IHCs) at each subsequent ACh pulse. Data were fitted by a monoexponential decay equation; τ_decay = 37.9 ± 4.2 s.

Ca²⁺ permeability of the IHC nAChR

The inhibitory sign of the cholinergic olivocochlear synapse depends upon Ca²⁺ influx through the nAChR and the subsequent activation of SK channels (Fuchs & Murrow, 1992a; Evans, 1996; Dulon et al. 1998; Blanchet et al. 1996, Glowatzki & Fuchs, 2000; Oliver et al. 2000; Gomez-Casati et al. 2004). It has been previously shown that both the α9 and α9α10 recombinant receptors are highly permeable to Ca²⁺ (Katz et al. 2000; Weisstaub et al. 2002; Sgard et al. 2002). In order to know the extent to which calcium permeates the native IHC nAChR, we studied the variation in the reversal potential of the ACh-evoked currents upon variation of the external Ca²⁺ concentration while keeping constant the Na⁺ concentration. Figure 3A shows representative I–V curves in the presence of different Ca²⁺ concentrations. As expected for a Ca²⁺-permeable channel, increasing the extracellular Ca²⁺ concentration caused a significant positive shift in E_rev. In Fig. 3B, averaged E_rev data from different IHCs are plotted against the different Ca²⁺ concentrations employed. P_Ca/P_Na was 8 ± 0.9 (n = 6 cells) as estimated by the extended constant field equation (see Methods). In 10 mm external Ca²⁺, we found that there was some residual activation of SK channels even though we used 10 mm BAPTA in the pipette and 1 nm apamin in the external solution (data not shown). Therefore, for the permeability experiments we used 10 nm apamin in all the Ca²⁺ concentrations tested. Notwithstanding, there still could be a small contribution of SK channels to the total ACh-evoked current in 10 mm Ca²⁺. If this were the case, we would be underestimating the shift in E_rev upon changing to this highest Ca²⁺ concentration and thus underestimating P_Ca/P_Na.

Figure 3. Ca²⁺ permeability of the native nAChR.

A, representative I–V curves obtained in the presence of different external Ca²⁺ concentrations (0.5, 5 and 10 mm) and a fixed extracellular Na⁺ concentration (100 mm). Note that as expected for a Ca²⁺ permeable channel, E_rev shifted to more positive potentials as the Ca²⁺ concentration was increased. B, plot of E_rev (each data point is the mean ± s.e.m., n = 6 IHCs) versus the external Ca²⁺ concentration. Data were fitted with the GHK voltage equation extended to include divalent cations (see Methods) and the estimated P_Ca/P_Na was 8 ± 0.9.

Modulation of the native IHC nAChR by Ca²⁺, Ba²⁺ and Mg²⁺

We have previously shown that the homomeric α9 nAChR is strongly blocked by micromolar concentrations of either Ca²⁺ or Ba²⁺ in the extracellular solution (Katz et al. 2000), whereas the heteromeric α9α10 receptor shows a more complex response upon variation of these divalent cations. At micromolar extracellular Ca²⁺ or Ba²⁺ concentrations, ACh-evoked responses through this receptor are potentiated, but at concentrations above 0.5 mm responses are strongly reduced (Elgoyhen et al. 2001; Weisstaub et al. 2002). As the α9 and α9α10 nAChRs are blocked by Mg²⁺ (Katz et al. 2000; Weisstaub et al. 2002), in order to investigate the modulation by Ca²⁺ and Ba²⁺ of the native IHC nAChR we chose to vary their concentration without compensation by other divalent cations.

Figure 4A shows representative responses of IHCs to the exogenous application of 100 μm ACh in the presence of increasing concentrations of Ca²⁺ (0, 0.1, 0.5, 1 and 3 mm) in the extracellular solution. Currents were negligible in a Ca²⁺-free solution, were potentiated by micromolar concentrations of Ca²⁺ and were inhibited by Ca²⁺ concentrations above 0.5 mm. The bar graph in Fig. 4B illustrates the averaged data obtained from different IHCs. Current amplitudes at different Ca²⁺ concentrations in each IHC were normalized with respect to those obtained at 0.5 mm in the same cell. A one-way ANOVA indicated that differences in mean current amplitudes in 0, 0.1, 1 and 3 mm Ca²⁺ were significant with respect to the mean amplitude obtained at 0.5 mm Ca²⁺ (P < 0.001). When Ba²⁺ substituted for Ca²⁺ in the external solution (Figs 4C and D), the same dual effect, potentiation and block, was observed. The one-way ANOVA indicated that differences in mean current amplitudes between 0, 0.1, 3 and 5 mm Ba²⁺ were significant with respect to the mean amplitude obtained at 0.5 mm Ba²⁺ (P < 0.001).

Figure 4. Modulation of the isolated nAChR by Ca²⁺ and Ba²⁺.

A, representative records of currents evoked by 100 μm ACh in the presence of different concentrations of Ca²⁺. B, bar graph illustrating that in the physiological range Ca²⁺ both potentiates and diminishes ionic currents through the nAChR studied in isolation from the SK channels. C and D, the same as A and B but in the presence of Ba²⁺ instead of Ca²⁺. Asterisks indicate that the amplitudes obtained at a given Ca²⁺ or Ba²⁺ concentration were significantly different from that obtained at 0.5 mm for either cation. All experiments were carried out in IHCs voltage clamped at −90 mV (n = 3–7 cells). Mean current amplitudes used to normalize the data in the bar graphs were: −752.4 ± 49.6 pA (n = 7 IHCs) and −480 ± 99 pA (n = 7 IHCs) in 0.5 mm Ca²⁺ and 0.5 mm Ba²⁺, respectively. Slope conductances at −90 mV were 8.4 ± 0.5 nS and 5.3 ± 1.1 nS in 0.5 mm Ca²⁺ and 0.5 mm Ba²⁺, respectively.

In order to evaluate the effects of Mg²⁺ on the native nAChR, the same type of experiments as those shown in Fig. 4 were carried out on IHCs superfused with an external solution in which Ca²⁺ was replaced by Mg²⁺. Figure 5 shows representative traces of the responses obtained in a nominally Ca²⁺-free solution (no added buffer) containing increasing concentrations of Mg²⁺. As illustrated in Fig. 5A, upon increasing the concentration of Mg²⁺ to 0.1 and 0.3 mm, responses to 100 μm ACh were not potentiated but strongly reduced. Even at concentrations < 0.1 mm, no potentiation of ACh-evoked responses were observed (data not illustrated). The reduction in the current amplitude was rapidly reversed upon superfusing the cells with a saline solution without Mg²⁺. The concentration–response curve in Fig. 5B, performed in a solution containing 0.5 mm Ca²⁺, shows that Mg²⁺ blocked the IHC nAChR with an IC₅₀ of 0.2 ± 0.01 mm (data were pooled from n = 3–11 IHCs).

Figure 5. Effects of Mg²⁺ on the isolated nAChR currents.

A, representative records of currents evoked by 100 μm ACh in the absence of Ca²⁺, with (0.1 and 0.3 mm) or without Mg²⁺. Note that micromolar concentrations of Mg²⁺ do not potentiate but strongly reduce these currents. B, concentration–response curve obtained in the presence of 100 μm ACh, 0.5 mm Ca²⁺ and increasing concentrations of Mg²⁺. The amplitudes of the ACh-evoked currents in the presence of the different Mg²⁺ concentrations were normalized to the values obtained in the absence of this cation. Mg²⁺ blocked the nAChR with an IC₅₀ of 0.21 ± 0.01 mm, n_H = 0.53 ± 0.06 (data were pooled from 3 to 11 IHCs). All experiments were carried out in IHCs voltage clamped at −90 mV.

Mechanism of Ca²⁺ potentiation

To evaluate whether changes in the apparent affinity of ACh for the native IHC receptor could account for the potentiation by Ca²⁺, concentration–response curves to ACh were performed at two different Ca²⁺ concentrations. Figure 6 illustrates these results for IHCs superfused with either nominally calcium-free or 0.5 mm Ca²⁺ saline. Under these conditions the EC₅₀ values were 160 ± 6 μm (n = 4 IHCs) and 60.7 ± 2.8 μm (n = 3–7 IHCs) for 0 and 0.5 mm Ca²⁺ saline, respectively. These differences in the EC₅₀ show that in a nominally Ca²⁺-free saline solution the apparent affinity of the IHC nAChR receptor for the agonist was lower than that observed in the presence of 0.5 mm Ca²⁺. The maximal response obtained in zero Ca²⁺ was not significantly different from that obtained in 0.5 mm Ca²⁺ (for example, in the same IHC the amplitude of currents evoked by 1 mm ACh was −746 pA in 0 Ca²⁺ and −772 pA in 0.5 mm Ca²⁺; see concentration–response curves in Fig. 6 for average responses in all IHCs tested under both conditions), indicating that at high ACh concentrations, the response of the native receptor becomes independent of Ca²⁺ in the external medium.

Figure 6. Mechanism of Ca²⁺ potentiation.

Concentration–response curves performed in 0 and 0.5 mm Ca²⁺ are shown. A significant leftward shift in the EC₅₀ upon increasing the Ca²⁺ concentration to 0.5 mm (EC₅₀: 160 ± 6 μm and 60.7 ± 2.8 μm for 0 and 0.5 mm Ca²⁺, respectively) was observed, thus suggesting that potentiation by Ca²⁺ involves changes in the apparent affinity of the native receptor for ACh. Responses were normalized to the maximal value obtained in 0.5 mm Ca²⁺. Data were pooled from 3 to 7 IHCs. All experiments were carried out in IHCs voltage clamped at −90 mV.

Mechanism of block by divalent cations

We studied the mechanism of divalent cation block using Ba²⁺ instead of Ca²⁺ to preclude the contribution of SK channels. Representative I–V curves in the presence of 0.5 and 5 mm Ba²⁺ (Fig. 7A) show that block by Ba²⁺ was sensitive to membrane voltage being more pronounced at hyperpolarized than at depolarized potentials. This effect can be better appreciated by the conductance versus voltage plot (ratios of conductance in 5 mm Ba²⁺/conductance in 0.5 mm Ba²⁺) where it can be observed that conductance in 5 mm Ba²⁺ is significantly lower than that in 0.5 mm Ba²⁺ at hyperpolarized potentials (Fig. 7B). In agreement with this, the bar graph in Fig. 7C shows that rectification of the I–V curve was significantly greater in higher barium concentrations. The ratio of current amplitudes at +60 and −90 mV (+ 43 and −107 mV when corrected for liquid junction potential) was 2.3 ± 0.1 (n = 14 IHCs) and 4.2 ± 0.4 (n = 12 IHCs) for 0.5 mm and 5 mm Ba²⁺, respectively.

Figure 7. Voltage dependence of divalent cation block.

A, representative I–V curves obtained at 0.5 and 5 mm Ba²⁺ in the extracellular solution. Note that in 5 mm Ba²⁺ the nAChR passes much less inward than outward current. B, the ratio of chord conductance in 5 mm Ba²⁺ with respect to that in 0.5 mm Ba²⁺ plotted against membrane voltage, shows that block by Ba²⁺ is much more effective at hyperpolarized than at depolarized potentials. The break in the x-axis corresponds to the data points that are close to or at the reversal potentials for the currents through the nAChR in 0.5 and 5 mm Ba²⁺. C, the ratio of current amplitudes at depolarized potentials (+43 mV) with respect to that at hyperpolarized potentials (−107 mV) shows that outward rectification is significantly higher in ACh-evoked currents elicited in the presence of 5 mm Ba²⁺ than those obtained in 0.5 mm Ba²⁺. Data were pooled from 12 to 14 IHCs.

Ca²⁺ modulation of isolated nAChR synaptic currents

It has been reported that changes in extracellular Ca²⁺ acting on nAChRs modify spontaneous cholinergic synaptic currents in sympathetic ganglion neurones (Amador & Dani, 1995). In the cochlear preparation used in the present study, spontaneous inhibitory postsynaptic currents (IPSCs) due to efferent synaptic transmission were observed in IHCs superfused with a saline solution containing 0.5 mm Ca²⁺ and recorded with a patch pipette containing a KCl-EGTA based saline solution (see Methods). Under this condition, synaptic currents are composed of ions flowing through both the nAChRs and the SK channels. As previously described (Glowatzki & Fuchs, 2000), at voltages between −70 and −50 mV these IPSCs have two components (Fig. 8A, upper record and see also Fig. 1 in Glowatzki & Fuchs, 2000), a fast inward component due to the inward flux of Na⁺ and Ca²⁺ through the nAChR and a more delayed and longer lasting outward component due to the opening of the calcium activated SK channels. The resting potential of IHCs is positive to E_K and therefore the functional significance of efferent activity is to hyperpolarize the IHC. These spontaneous synaptic currents were inward at −90 mV with mean amplitude of −28.4 ± 1 pA. IPSCs rose rapidly (time to peak 9.6 ± 0.7 ms) and decayed more slowly (τ_decay 30.3 ± 1.9 ms, n = 52 events, 6 cells, Fig. 8A and B). At −30 mV, IPSCs were outward with a mean amplitude of 33.2 ± 1.7 pA, time to peak = 12 ± 0.6 ms and τ_decay = 31 ± 1.6 ms, n = 40 events, 2 cells (Fig. 8A). The E_rev of IPSCs was −66 ± 1 mV, a value close to E_K and similar to that obtained for ACh-evoked currents (see Fig. 1A).

Figure 8. Ca²⁺ modulation of isolated nAChR spontaneous synaptic currents.

A, representative recordings of combined (nAChR + SK) and isolated (nAChR) spontaneous synaptic currents at different voltages. Note that at −70 mV, the trace is biphasic (arrows) in the combined current records. B, superimposed averages of combined (black) and isolated (grey) synaptic currents. C and D, bar graphs showing τ of decay and amplitudes, respectively, of the combined nAChR + SK and isolated nAChR synaptic currents. Note that amplitudes are not significantly different whereas τ of decay is significantly faster in the isolated nAChR synaptic currents. E, superimposed averages of isolated synaptic currents recorded in the presence of 0.5 (black) and 10 mm (grey) Ca²⁺ in the extracellular solution. F and G, bar graphs showing that synaptic current amplitudes (F) are significantly reduced in the presence of 1.3 and 10 mm Ca²⁺ with respect to the amplitudes obtained in the presence of 0.5 mm Ca²⁺ and that no changes were found in τ of decay (G) upon varying the external Ca²⁺ concentration. All experiments were carried out in IHCs voltage clamped at −90 mV.

When synaptic currents were recorded with a patch pipette containing a CsCl-BAPTA based saline solution and superfused with an external solution containing 0.5 mm Ca²⁺ plus 1 nm apamin, synaptic currents were inward at −90 mV and outward at positive voltages (E_rev≈−18 mV). This is in agreement with observations on the isolated nAChR currents evoked by exogenously applied ACh (see Fig. 1A). Note that under this condition, no synaptic currents could be detected at −30 mV (Fig. 1A). This is due to the fact that this membrane voltage is close to E_rev for the nAChR and also shows that outward currents through SK channels have been effectively blocked. Consistent with the notion that the temporal course of IPSCs is governed by SK channel gating (Oliver et al. 2000), the kinetics of the synaptic currents studied in isolation from the SK channels at −90 mV were faster (time to peak = 4.5 ± 0.5 ms, τ_decay = 20.4 ± 1.5 ms, n = 41 events, 5 cells; Figs 8B and C) than those of the combined current. The average waveform for the combined response nAChR + SK (see Fig. 8B) seems to have a fast and a slow component; however, it can be better fitted by one exponential and not by two. Therefore, even though in the presence of SK channels, the slowing of the synaptic current kinetics is significant, one can only speculate that the faster component seen in Fig. 8B, is due to the activation of the nAChR and the slower due to the opening of SK channels. Considering the effect of the presence of SK channels in the kinetics of the response, it is puzzling that no significant differences were found in the amplitude of the isolated nAChR synaptic currents with respect to that of the combined nAChR + SK IPSCs (Fig. 8B and D). At −90 mV, ionic currents through both nAChRs and SK channels are inward, but the driving force for K⁺ entry is quite low (E_K =−82 mV), whereas the driving force for Na⁺ entry through the nAChR is high (E_Na =+84 mV). Therefore, at –90 mV, under this ionic condition, the contribution of SK channels to the peak synaptic current amplitude could be very small. Nevertheless, to be able to clearly separate both components, it would be necessary to measure synaptic currents in the same cell with and without SK channels.

To evaluate whether Ca²⁺ modulates the time course and/or the amplitude of the isolated cholinergic synaptic currents through native nAChRs, we measured these parameters in IHCs voltage clamped at −82 mV (using BAPTA in the pipette solution and 1 nm apamin in the external solution) and superfused with saline solutions containing different concentrations of Ca²⁺. We could only analyse Ca²⁺ concentrations ≥ 0.5 mm, as in lower Ca²⁺ concentrations the frequency of spontaneous currents was too low for evaluation. As illustrated by the records (Fig. 8E) and the bar graph (Fig. 8F), increasing the concentration of Ca²⁺ to 1.3 and 10 mm, caused a significant reduction (12% and 36%, respectively) in the amplitude of the nAChR synaptic currents with respect to the values obtained at 0.5 mm Ca²⁺. Amplitudes were −13.3 ± 0.3, −11.8 ± 0.3 and −8.6 ± 0.1 pA in 0.5 (n = 95 events, 5 cells), 1.3 (n = 106 events, 5 cells, P < 0.05) and 10 mm (n = 791 events 5 cells, P < 0.001) Ca²⁺, respectively. No significant differences were found in the decay time constants at the different Ca²⁺ concentrations: τ_decay = 19 ± 0.8, 22 ± 1 and 21 ± 0.3 ms for 0.5, 1.3 and 10 mm Ca²⁺ (Fig. 8G). These results are consistent with the earlier interpretation (Katz et al. 2000; Weisstaub et al. 2002) that divalent cations reduce ionic current by acting as blocking particles in the open channel. This hypothesis also accords with the observation of voltage-dependent rectification produced by Ba²⁺ (see Fig. 7).

Discussion

The main goal of the present work was to characterize, from a pharmacological and biophysical standpoint, the receptor mediating cholinergic synaptic effects on IHCs prior to the onset of hearing. Even though this is a transient synapse present in IHCs at an immature stage of cochlear development, it is functional as of postnatal day 3 (P3) to P13–14 (Glowatzki & Fuchs, 2000; Katz et al. 2004) and, as in other hair cells (Fuchs & Murrow, 1992a, b; Doi & Ohmori, 1993; Blanchet et al. 1996; Dulon & Lenoir, 1996; Evans, 1996; Fuchs, 1996; Nenov et al. 1996b; Dulon et al. 1998; Oliver et al. 2000), it is mediated by a nAChR probably composed of both the α9 and α10 subunits (see Elgoyhen et al. 1994, 2001). The activation of the α9α10-containing nAChR has been shown to be always coupled to the activation of the SK channels that hyperpolarize the hair cell (Glowatzki & Fuchs, 2000; Katz et al. 2004; Marcotti et al. 2004). In addition, it has been demonstrated that the presence of a functional nAChR channel depends on the presence of both the α9 and α10 subunits and that the expression of both the α10 subunit and the associated SK channels are down-regulated in IHCs after the onset of hearing, a time at which direct efferent contacts to IHCs start to retract, and cholinergic sensitivity disappears from these cells (Katz et al. 2004).

The characterization of the native nAChR was done in isolation from the associated SK channel in order to compare its characteristics to those of the recombinant α9 and α9α10 nAChRs expressed in Xenopus oocytes. The pharmacological profile and Ca²⁺ permeability of the native IHC nAChR were found to be identical to those described for both the α9 and α9α10 recombinant nAChRs. However, when comparing other biophysical characteristics, the native IHC nAChR resembles more closely the heteromeric α9α10 nAChR. Of particular interest is the observation that the native nAChRs, like heteromeric α9α10 receptors, possess two calcium binding sites. An extracellular site that facilitates ligand gating of the channel is present in both native nAChRs and heteromeric α9α10 receptors, but absent from homomeric α9 receptors expressed in Xenopus oocytes. However, native nAChRs IHC, heteromeric α9α10 and homomeric α9 nAChRs alike demonstrate open channel block by divalent cations (Katz et al. 2000; Elgoyhen et al. 2001; Weisstaub et al. 2002). Thus, comparisons of conserved and varying sequences in the α9 and α10 subunits should offer important clues to the structural bases for these functional differences.

Voltage sensitivity, pharmacological profile and desensitization pattern of the native IHC nAChR

The reversal potential and voltage sensitivity of the isolated nAChR current are consistent with those reported for the recombinant heteromeric α9α10 nAChRs expressed in Xenopus oocytes (Elgoyhen et al. 2001; Weisstaub et al. 2002) and for the ACh-evoked current in isolated adult guinea pig OHCs (Blanchet et al. 1996), developing rat OHCs (Dulon & Lenoir, 1996), short chick hair cells (McNiven et al. 1996) and developing rat IHCs (Glowatzki & Fuchs, 2000). The I–V relationship of the IHC nAChR presents, as those of the α9 and α9α10 recombinant nAChRs (Elgoyhen et al. 2001), considerable rectification around the reversal potential. However, this I–V differs from that of the recombinant α9 mainly in the fact that the native nAChR, as shown for the recombinant α9α10 nAChR (Elgoyhen et al. 2001; Weisstaub et al. 2002; Sgard et al. 2002) can pass a considerable amount of current in the hyperpolarized direction. Conversely, currents through the α9 nAChR at hyperpolarized potentials are potently blocked by Ca²⁺ (IC₅₀ 100 μm) (Katz et al. 2000). This difference could be accounted for by the fact that Ca²⁺ only blocks the α9 nAChR whereas it both potentiates and blocks the native and the heteromeric α9α10 nAChR receptors within the same range of physiological concentrations: at a concentration of 0.5 mm potentiation predominates whereas at higher Ca²⁺ concentrations, blockade starts to be apparent (see discussion below). This hypothesis is reinforced by the fact that the I–V curve in the presence of 5 mm Ba²⁺ has the same shape (outwardly rectifying) as that reported for α9 in 1.8 mm Ca²⁺ (Katz et al. 2000; Elgoyhen et al. 2001). A similar outward rectification was obtained with the recombinant α9α10 nAChR upon elevating the concentration of Ca²⁺ to 3 mm (Weisstaub et al. 2002).

The pharmacological characteristics of the native IHC nAChR are indistinguishable from those of the recombinant α9 and α9α10 nAChRs and also similar to those reported for the native OHCs cholinergic receptor (Elgoyhen et al. 1994, 2001; Erostegui et al. 1994b; Chen et al. 1996; Rothlin et al. 1999; Verbitsky et al. 2000). Namely, the IHC cholinergic receptor is blocked by nicotine as well as by antagonists of non-cholinergic receptors such as strychnine (glycine receptors), bicuculline (γ-aminobutyric acid type A receptors), and ICS-205930 (ligand gated serotonin receptors).

The apparent affinity of the IHC nAChR for ACh (EC₅₀ = 60.7 μm) is somewhat lower than that reported (EC₅₀ 7–30 μm) for both the recombinant α9 and α9α10 (Elgoyhen et al. 1994, 2001; Weisstaub et al. 2002; Sgard et al. 2002) and the native hair cell nAChRs (Blanchet et al. 1996; Chen et al. 1996; Dulon & Lenoir, 1996; Erostegui et al. 1994c; Fuchs & Murrow, 1992a, b; Housley & Ashmore, 1991; McNiven et al. 1996). However, it is higher than the apparent affinity (EC₅₀ = 122 μm) reported by Blanchet et al. (2000) for the nAChR of guinea pig OHCs. Several factors could account for this rather wide range of EC₅₀ values: species differences, the type of preparation used (oocytes, isolated hair cells, cochlear coil), combined nAChR + SK versus isolated nAChR (see Blanchet et al. 2000) and also a very important source of variations could be the ionic composition of the external and internal solutions used to evaluate this parameter. Resolution of these issues must await further work with the isolated nAChR in IHCs and OHCs from different species, studied under the same ionic conditions.

Desensitization in the presence of the agonist is a widespread feature of ligand-gated channels (Jones & Westbrook, 1996). In the case of nAChRs, it has been shown that subunit composition alters the degree to which desensitization takes place (Fenster et al. 1997; Gerzanich et al. 1998). In agreement with this, we have shown that the homomeric α9 nAChR does not significantly desensitize either upon the continued or intermittent presence of ACh, whereas currents through the α9α10 recombinant receptor are strongly diminished in both experimental situations (Elgoyhen et al. 2001). We now show that the IHC nAChR markedly desensitizes when exposed to ACh over the same time periods, thus resembling the α9α10 nAChRs. Cholinergic responses in mammalian OHCs were also shown to desensitize in the presence of ACh (Blanchet et al. 1996; Dulon & Lenoir, 1996; Evans, 1996; Nenov et al. 1996a).

Calcium permeability of the native IHC nAChR

A key step for the cholinergic inhibition of cochlear hair cells is the activation of calcium-dependent SK channels by an increase in cytoplasmic calcium (Fuchs, 1996; Glowatzki & Fuchs, 2000; Oliver et al. 2000). The high calcium permeability of the native IHC receptor found in the present work is consistent with this physiological role. This result is also in line with previous work on both chick short hair cells (Fuchs & Murrow, 1992a; McNiven et al. 1996) and rodent OHCs (Blanchet et al. 1996) that has shown evidence that the hair cell receptor must have a considerable Ca²⁺ permeability. It should be remembered, however, that Ca²⁺ blocks the total inward current, composed of Na⁺ and Ca²⁺, through the IHC nAChR. This interaction violates the principle of ionic independence assumed by the Goldman-Hodgkin-Katz formalism and therefore the values obtained for relative permeabilities should be taken only as estimates to compare to the values obtained for the α9 and α9α10 recombinant receptors (Katz et al. 2000; Elgoyhen et al. 2001; Weisstaub et al. 2002; Sgard et al. 2002) and to other recombinant and native ligand-gated channels.

The present P_Ca/P_Na estimated for the IHC nAChR, however, is of about one order of magnitude lower than that obtained by Jagger et al. (2000) for the nAChR present in hair cell precursor lines from the immortomouse. The source of this discrepancy is not clear. However, aside from the fact that channels could behave differently in a cell line, one possible explanation could be that the nAChR of the immortomouse hair cells might have a different nAChR subunit composition and/or stoichiometry from those of rat cochlear hair cells.

The high Ca²⁺ permeability of the native IHC nAChR estimated in the present study (P_Ca/P_Na∼8) is similar to that reported for the homomeric α9 and the heteromeric α9α10 nAChRs (Katz et al. 2000; Weisstaub et al. 2002) and is also in the same range as that of the homomeric α7 nAChR (P_Ca/P_Na∼10–20) (Bertrand et al. 1990; Sands & Barish, 1991; Séguéla et al. 1993; Katz et al. 2000), highest among all nAChRs. The P_Ca/P_Na of this group of nAChRs is similar to that of other ligand-gated channels with high Ca²⁺ permeability, such as the NMDA receptors and cyclic nucleotide-gated channels (Mayer & Westbrook, 1987; Frings, 1997). As for the hair cell nAChR, in many cases the main role of these calcium-permeable receptors is not that of depolarizing the membrane, but rather to trigger cellular functions through Ca²⁺ entry into the cell (Burnashev, 1998).

Effects of divalent cations on the isolated IHC nAChR currents

Modulation by external Ca²⁺, which can vary dramatically in concentration during high synaptic activity (Brown et al. 1995), is a widespread feature of both recombinant and native nAChRs. The tight modulation of the IHC nAChR by extracellular Ca²⁺ resembles what has been described for the α9α10 receptor where Ca²⁺ potentiates ACh responses in the micromolar range and blocks them at concentrations > 0.5 mm (Weisstaub et al. 2002). Studies performed in chick short hair cells show that the native avian cholinergic receptor is also biphasically modulated by Ca²⁺, but the range of concentrations at which modulation is observed in that preparation was significantly higher (McNiven et al. 1996). In isolated mammalian OHCs loaded with internal BAPTA, ACh-evoked currents at +32 mV were strongly reduced, and almost completely abolished at –81 mV, by removal of external Ca²⁺ (Blanchet et al. 1996). However, in those experiments Ca²⁺ was replaced with Mg²⁺, which potently blocks the recombinant α9α10 nAChR (Weisstaub et al. 2002) and, as we show here, the native IHC nAChR. A similar biphasic effect of Ca²⁺ on the IHC nAChR was recently reported (Marcotti et al. 2004), although that involved both cation current through the nAChR and the accompanying SK current. Therefore, although the results were qualitatively similar, the interpretation was restricted by that additional complexity.

As reported for the α9α10 nAChR (Weisstaub et al. 2002) and for other nAChRs (Ifune & Steinbach, 1991; Mulle et al. 1992b; Booker et al. 1998; Liu & Berg, 1999), micromolar Ba²⁺, but not Mg²⁺, mimicked the enhancing effects of Ca²⁺ on the IHC nAChR. The effects of Ba²⁺ on ACh-evoked currents have been previously studied in guinea pig OHCs (Erostegui et al. 1994a). In that case, however, the authors did not see any ACh-activated current when they replaced extracellular Ca²⁺ by Ba²⁺. The discrepancy with the present results might be accounted for by the fact that: (1) they were looking at the combined nAChR + SK currents and evaluated the effects of Ba²⁺ at 0 mV, near the reversal potential for current through the nAChR, and (2) Ba²⁺ is both a poor activator and a blocker of SK channels (Soh & Park, 2001), making the outward currents through SK negligible. As reported for OHCs (Nenov et al. 1996b), Mg²⁺ also blocked currents through the native IHC nAChR at physiological concentrations. Therefore, the effects of OC efferent activation on IHCs will depend not only on the external Ca²⁺ concentration, but on the Ca²⁺/Mg²⁺ ratio in the perilymph. It remains uncertain why Mg²⁺ does not potentiate responses of ACh on IHCs. Mg²⁺ has a smaller metallic radius (150 pm) than Ca²⁺ and Ba²⁺ (180 and 215 pm, respectively). As charge density is greater in Mg²⁺ than in Ba²⁺ and Ca²⁺, the interaction between either Ca²⁺ or Ba²⁺ with a H₂O molecule would be weaker than that of Mg²⁺. Thus, as postulated for α7 nAChRs, the difference in the actions of these divalent cations could possibly be due to the fact that Mg²⁺ has a higher charge density than Ca²⁺ and Ba²⁺ resulting in a larger hydration radio and a slower dehydration rate, which may prevent this divalent cation from interacting with the sites involved in potentiation (Eddins et al. 2002).

Mechanism of Ca²⁺ potentiation and block

Calcium potentiates nAChRs through allosteric effects at an extracellular site (Decker & Dani, 1990; Mulle et al. 1992b; Vernino et al. 1992; Galzi et al. 1996) and is independent both from Ca²⁺ influx and membrane potential (Mulle et al. 1992b; Vernino et al. 1992; Weisstaub et al. 2002). Likewise, gating of the hair cell's nAChR was enhanced even when cytoplasmic calcium was heavily buffered with BAPTA. The fact that Ca²⁺ caused a leftward shift in the ACh concentration–response curve, indicates that potentiation by Ca²⁺ involves changes in the apparent affinity of the native IHC receptor for ACh. A change in the apparent affinity could imply changes in agonist binding, in the cooperativity of the system, in the coupling between agonist binding and channel gating and/or changes in the kinetic properties of these receptors. Increased agonist binding seems less likely as it has been reported that Ca²⁺ is a competitive inhibitor of Torpedo nAChRs in the 0.1–1 mm range, and therefore increasing Ca²⁺ from 0 to 0.5 mm would decrease ACh occupancy of the binding site (Chang & Neumann, 1976). Cooperativity of ACh binding in hippocampal nAChRs is altered by extracellular Ca²⁺ levels (Bonfante-Cabarcas et al. 1996), but we did not find significant differences in the Hill coefficients of the concentration–response curves to ACh obtained in the presence or absence of Ca²⁺. Changes in the kinetic properties of the receptors by Ca²⁺, leading to increases in the probability or the frequency of channel opening, have been previously shown for other nAChRs (Mulle et al. 1992b; Amador & Dani, 1995). However, further analysis would be needed to prove this hypothesis.

The voltage dependence of Ba²⁺ block suggests that, as reported for the recombinant homomeric α9 and heteromeric α9α10 receptors (Katz et al. 2000; Weisstaub et al. 2002), the site of action of either Ca²⁺ or Ba²⁺ might lie within the channel pore and might involve hindering the passage of Na⁺ through the channel in the presence of divalent cations. This was also suggested for neuronal and muscle nAChRs (Decker & Dani, 1990; Mulle et al. 1992a, b) and for other ligand-gated and mechanosensitive channels (Frings et al. 1995; Ricci & Fettiplace, 1998).

Ca²⁺ modulates nAChR spontaneous synaptic currents

The modulatory effects of Ca²⁺ were also seen on spontaneous synaptic currents which result from ACh released by efferent synaptic terminals contacting the IHCs (Glowatzki & Fuchs, 2000; Katz et al. 2004). Elevating Ca²⁺ from 0.5 mm to 1.3 and 10 mm significantly decreased the amplitude of synaptic currents without affecting their time course. The reduction in amplitude is in agreement with, although less pronounced than, the blocking effects of Ca²⁺ on the amplitude of currents evoked by exogenous ACh. The lack of effect of external Ca²⁺ on the time course of synaptic currents is in line with the idea that Ca²⁺ potentiation of nAChRs might be due to an effect on channel gating (Mulle et al. 1992b; Amador & Dani, 1995) whereas Ca²⁺ block is most likely due to the hindering of the passage of monovalent ions through the pore (Decker & Dani, 1990; Mulle et al. 1992a, b; Katz et al. 2000; Weisstaub et al. 2002). Therefore, Ca²⁺ block would affect the conductance of the channels without affecting channel kinetics and hence it should not have any effect on the time course of synaptic currents. In the IHC nAChR, as well as in the recombinant α9α10 receptor, the potentiating effects of external Ca²⁺ are only evident at very low Ca²⁺ concentrations (0–0.5 mm). Therefore, we could not study the potentiating effects of Ca²⁺ on synaptic currents due to the extremely low number of events we found in the absence of external Ca²⁺.

Physiological implications

In view of the high Ca²⁺ influx the native IHC nAChR can generate and the critical role Ca²⁺ plays in the inhibitory sign of this fast cholinergic synapse, it is not surprising that its activity is tightly regulated by external Ca²⁺ within physiological ranges of concentration. In vivo, during periods of intense synaptic activity, extracellular Ca²⁺ at the synaptic cleft could be significantly reduced (Brown et al. 1995) and thus provide a type of negative feedback control by decreasing receptor function. Conversely, during normal synaptic activity, with physiological concentrations of Ca²⁺ and Mg²⁺ in the perilymph (1.3 and 0.9 mm, respectively) the total current through this receptor would be reduced due to the hindering effects divalent cations have on the Na⁺ inward current. This would minimize the depolarization of the hair cell membrane upon activation of the α9α10 nAChR and maximize Ca²⁺ influx through the activated receptor, thus leading to the activation of SK channels that hyperpolarize the cell.