Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The adamantane derivative 3,5-dimethyl-1-adamantanamine (memantine) is a well established blocker of ligand-gated ion channels permeable for Ca²⁺ such as the NMDA-type glutamate receptor (Chen et al., 1992; Parsons et al., 1993; Bresink et al., 1996) or nicotinic acetylcholine receptors (Buisson and Bertrand, 1998). In either case, memantine acts as an open channel blocker, that enters the channel pore and sterically occludes the ion pathway (Chen et al., 1992; Buisson and Bertrand, 1998). For NMDA-receptors, this pore-block strongly depends on the transmembrane voltage, with highest efficacy at hyperpolarized potentials (Bresink et al., 1996); for nAchRs, the voltage-dependence is less pronounced (Buisson and Bertrand, 1998). Therapeutically, memantine is used to prevent excitotoxic neuronal cell death associated with neurodegeneration and stroke and is induced by massive influx of Ca²⁺ through overactivated NMDA receptors (Chen et al., 1992; Parsons et al., 1999).

Cochlear outer hair cells (OHCs) are the central element of the active mechanical amplification mechanism that is crucial for the exquisite sensitivity and frequency-resolving capacity of the mammalian hearing organ (Dallos, 1992; Dallos and Evans, 1995). Mechanistically, this amplification is based on the ability of OHCs to alter their cell length in response to changes in membrane potential, a process that works at frequencies of up to at least 70 kHz (Brownell et al., 1985; Gale and Ashmore, 1997; Frank et al., 1999). Cochlear amplification is controlled by the cholinergic medial olivocochlear system that synapses onto OHCs (reviewed by Guinan, 1996). This inhibitory synapse uses an excitatory nAchR to supply the postsynaptic OHC with Ca²⁺ that initiates an inhibitory hyperpolarization by opening Ca²⁺-activated SK2 channels (Fig. 1A) (Fuchs and Murrow, 1992; Blanchet et al., 1996; Evans, 1996; Oliver et al., 2000).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Memantine blocks IPSCs in OHCs. A, diagram illustrating efferent innervation of OHCs and postsynaptic Ca2+-signaling that generates fast IPSCs carried by SK2 channels (ET, efferent synaptic terminals). B, whole -cell voltage-clamp recording from an OHC at 64 mV. Increase of the extracellular K+ concentration induced IPSCs, which were reversibly inhibited by 100 µM memantine. The baseline shift induced by high K+ is caused by an unblocked fraction of the hair cell's resting K+ conductance. C, dose-response curve for inhibition of IPSCs by memantine. Postsynaptic currents were normalized to the response in the absence of memantine (see Materials and Methods). Line indicates the best fit of the Hill equation to data obtained from seven OHCs (nH = 1.8; IC50 = 16.1 µM).

The OHC nAChR contains the 9 subunit (Elgoyhen et al., 1994; Glowatzki et al., 1995; Vetter et al., 1999) and exhibits a high permeability for Ca²⁺ (Jagger et al., 2000; Katz et al., 2000). This high permeability allows for significant Ca²⁺ influx into the OHC that may lead to globally increased cytoplasmic [Ca²⁺] (Doi and Ohmori, 1993). In analogy to excitotoxic processes in CNS neurons, abnormally high Ca²⁺ levels may initiate or support degradation of OHC; loss of functional OHCs is well known to be a major cause of sensorineural hearing loss induced by various harmful stimuli including noise or ototoxic agents (Cody and Russell, 1985; Patuzzi et al., 1989). So far, it has been shown that Ca²⁺ entering through the nicotinic receptor can trigger structural alterations of OHCs, resulting in a decreased axial stiffness (Dallos et al., 1997; Sridhar et al., 1997), and elevated intracellular Ca²⁺ has been implicated in the impairment of OHC function after acoustic overstimulation (Fridberger et al., 1998). Here we test the effect of memantine on efferent synaptic transmission and Ca²⁺-influx in OHCs and investigate block of the nAChR as the mechanism underlying the observed inhibitory effects.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Patch-Clamp Recordings on Outer Hair Cells. The apical turn of the organ of Corti was dissected from cochleae of three- to six-week-old Wistar rats as described previously (Oliver et al., 2000). The preparation was performed in a solution containing 144 mM NaCl, 5.8 mM KCl, 0.1 mM CaCl₂, 2.1 mM MgCl₂, 10 mM HEPES, 0.7 mM Na₂HPO₄, and 5.6 mM glucose, pH adjusted to 7.3 with NaOH. For recordings, OHCs located between half and one turn from the apex of the cochlea were chosen. If necessary, supporting cells were removed with gentle suction from a cleaning pipette carefully avoiding mechanical disturbance of the efferent nerve terminals.

Electrophysiology on Heterologously Expressed Channels. AChR subunits and SK2 channels were heterologously expressed in Xenopus laevis oocytes. Oocytes were surgically removed from adult females and dissected manually. Four to 5 days before electrophysiological recordings, Dumont stage VI oocytes were injected with about 50 ng RNA. For coexpression experiments, the total amount of RNA was kept constant and the different RNAs were coinjected in equal concentrations.

Molecular Biology. The coding region of the rat 10 gene (GenBank accession no. AF196344) was amplified from rat brain cDNA by PCR using 5'- and 3'- adapter-primers containing suitable restriction sites (GAGACCCGGGAGCTCCACC, ATGGGGACAAGGAGCCACTACC, and GAGTCTAGATTACAGGGCTTGCACCAGTACAATG). The amplified fragments were subcloned into the X. laevis oocyte expression vector pGEM-HE (gift of Dr. J. Tytgat), yielding pGEM-HE-nAchR-10, verified by sequencing. Capped mRNAs for 9, 10, and SK2 were synthesized in vitro using the mMESSAGE mMACHINE kit (Ambion, Austin, TX). To detect 10 and 9 transcripts, PCR was performed using reverse transcribed RNA isolated from either OHCs (containing some Deiters cells) or supporting cells (Deiters and Hensen's cells) as template. Cells were collected from rat organs of Corti using suction glass micropipettes (diameter 10 µm). RNA was prepared from ~100 cells of each fraction using the QIAGEN RNeasy kit (QIAGEN, Hilden, Germany) according to the manufacturers' instructions. The oligonucleotides used as primers in the PCR reactions were chosen to span an intron in the human 9 and 10 genes to allow differentiation between products originating from cDNA and products originating from contaminating genomic DNA (9 sense, CGTCCTCATATCGTTCCTCGCTCCG; 9 antisense, TGGTAAGGGCTGTGGAGGCAGTGA; 10 sense, GCAGCCTACGTGTGCAACCTCCTGC; and 10 antisense, AGGTGTCCCAGCAGGAGAACCCGAG). For each PCR reaction, RNA corresponding to ~3 cells was used as template; the number of cycles was 40 for amplification of 9 and 10, and 30 for GAPDH used as a control.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

IPSCs were recorded from whole-cell voltage-clamped OHCs in acutely isolated organs of Corti during depolarization of the efferent presynaptic terminal with elevated extracellular [K⁺]. As shown previously, these IPSCs are K⁺-currents through SK2 channels activated by brief elevation of subsynaptic [Ca²⁺] (Oliver et al., 2000). Ca²⁺-transients are generated by the opening of postsynaptic nAChRs in response to presynaptically released ACh (Glowatzki and Fuchs, 2000; Oliver et al., 2000). In the presence of memantine, IPSCs were reversibly reduced in a concentration dependent manner (Fig. 1B). Data obtained from seven OHCs yielded a half-blocking concentration for memantine of 16.1 µM (Fig. 1C). Based on the known effect of memantine on ligand-gated ion channels, it seemed most likely that inhibition of the complex IPSCs resulted from block of Ca²⁺-entry via the nAChR. We therefore aimed at testing the action of memantine on the OHC nAChR expressed heterologously in X. laevis oocytes. However, application of 100 µM ACh to oocytes injected with 9-specific cRNA yielded very small currents (9.3 + 5.0 nA at 80 mV; Fig. 2A), consistent with previous reports (Elgoyhen et al., 1994; Katz et al., 2000). No currents exceeding background levels were observed with the rat homolog of the 10 subunit, a member of the nAChR family recently identified by Boulter and colleagues (GenBank accession no. AF196344; Fig. 2, A and D). As shown in Fig. 2B by RT-PCR on OHCs isolated from the rat organ of Corti (see Materials and Methods), 10 mRNA is indeed present in these sensory cells, although it was not detected in the supporting cell fraction containing Hensen and Deiters cells. In a control experiment with RNA from OHCs that was not reverse transcribed, PCR amplified a fragment of ~900 bp (Fig. 2B, lane 4) which most likely resulted from contamination with genomic DNA as the length of this fragment is in good agreement with the sequence defined by the primer pair in the human genome (bacterial artificial chromosome from chromosome 11; GenBank accession no. AC060812).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   9 and 10 nAChR subunits are expressed in OHCs and coassemble to functional channels. A, ACh (100 µM) induced inward currents in oocytes injected with RNA coding for 9 (lower trace) but not in oocytes injected with RNA coding for 10 (upper trace). Holding potential was 80 mV. Fast downward spikes in the traces are artifacts resulting from switching between solutions. B, detection of 10 and 9 mRNA in OHCs by RT-PCR. Fragments of the expected length were amplified for 10 and 9 from OHCs (lanes 1 and 2) but not from supporting cells (lane 3, data for 9 not shown). Controls were GAPDH amplified from supporting cells (lane 4), OHC-RNA without RT added (lane 5), and water (lane 6) as a template for PCR. C, same experiment as in A, but for coexpression of both subunits. Note the different current scaling. The recording from A (9) was added for comparison. D, current amplitudes from experiments as in A and C summarized for oocytes injected with RNA coding for a nonconducting SK2 channel mutant (control), 9, 10, and 9/10 (values are mean ± standard error of 5, 13, 13, and 18 oocytes, respectively).

When both 9 and 10 were coexpressed in oocytes, large inward currents with peak amplitudes of up to 35 µA (at 80 mV) were recorded upon application of ACh (Fig. 2C, D). Similar to 9-mediated currents, the time-course was characterized by an initial transient declining to a smaller plateau of variable amplitude with respect to the peak current. The increase in current amplitude of more than 3 orders of magnitude (compared with homomeric 9 expression) together with the coexpression of 9 and 10 in OHCs suggest that heteromultimerization of both subunits is essential to give fully functional receptor channels. Moreover, the absence of any other of the known nAChR subunits (Morley et al., 1998) strongly suggests that the OHC nAChR is a heteromer composed of 9 and 10 subunits.

A characteristic feature of homomeric 9 channels is their exceptionally high Ca²⁺-permeability (Jagger et al., 2000; Katz et al., 2000). This Ca²⁺-permeability is thought to be essential for the OHC nAChR, since it allows for a Ca²⁺ influx sufficiently high to effectively activate SK-type potassium channels. However, high endogenous expression levels of Ca²⁺-activated Cl-channels characterize the oocyte expression system (Stühmer and Parekh, 1995). Therefore, opening of Ca²⁺-permeable channels in an external medium containing Ca²⁺ leads to coactivation of a Cl conductance. When external Ca²⁺ was substituted for Mg²⁺, currents induced by ACh application onto 9/10 heteromeric channels were reduced by a factor of roughly 10 (Fig. 3A). Thus, a large fraction of the ACh-induced current measured in CaNFR was attributable to opening of Ca²⁺-activated Cl-currents. This was also supported by the reversal potential of the current in CaNFR of about 25 mV (data not shown), close to the estimated Cl equilibrium potential in X. laevis oocytes (Stühmer and Parekh, 1995). Consequently, 9/10 heteromeric channels had a significant Ca²⁺-permeability, similar to what is known from homomeric 9 receptors. Application of ACh in the absence of extracellular Ca²⁺ allowed recording of 9/10 currents in isolation. Heteromeric channels yielded currents that were 100-fold larger than currents recorded from 9 channels under the same conditions, confirming the large gain of receptor conductance by coexpression that was observed in the presence of external Ca²⁺ (Fig. 3, B and C). In the absence of Ca²⁺, 9/10 also showed consistent kinetics characterized by slow desensitization on a time scale of seconds. Desensitization was not observed with 9 channels within the limits of the speed of solution exchange (Fig. 3B).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Memantine blocks heteromeric 9/10 receptors with mild voltage dependence. A, currents evoked by activation of 9/10 nAChRs are dependent on external Ca2+. Traces show subsequent applications of 100 µM ACh to the same oocyte in CaNFR (Ca2+) and MgNFR (Mg2+) at 80 mV. B, currents from homomeric 9 (upper trace) and heteromeric 9/10 nAChRs (lower trace), recorded by application of 100 µM ACh in nominally Ca2+-free external solution (MgNFR; 80 mV). C, current amplitudes from oocytes expressing 9 and 9/10, recorded as in Fig. 3B (mean ± S.E. from 12 and 15 oocytes, respectively). D, ACh-induced (100 µM) currents recorded from 9/10 nAChRs in the presence of the memantine concentrations indicated. E, memantine dose-inhibition curves of 9/10 nAChR. Peak currents were recorded as in D and normalized to the current amplitude after washout of the blocker. Continuous lines show fits of the Hill equation to data from 7 oocytes measured in MgNFR (solid) and CaNFR (gray), yielding IC50 values of 1.6 and 1.3 µM and nH of 1.2 and 1.0, respectively. F, current-voltage relations of 9/10 nAChR were recorded in the absence (black) and presence (gray trace) of 1 µM memantine in response to 2-s voltage ramps from 100 to +50 mV. Currents were recorded in MgNFR and were leak-corrected by subtracting the response to the same voltage ramp preceding the application of 100 µm ACh.

The action of memantine was tested on the isolated 9/10 current in MgNFR. Memantine blocked this current in a completely reversible manner with an IC₅₀ value of 1.6 µM and a Hill coefficient of 1.2 (Fig. 3, D and E). Channel block by memantine has been analyzed most extensively at NMDA-type glutamate receptors (Chen et al., 1992; Bresink et al., 1996), where block of the open pore is strongly voltage-dependent. To address the issue of blocking mechanism and voltage dependence at 9/10 channels, current-voltage relations were determined from voltage ramps in the absence and presence of memantine (Fig. 3F). In either case, current-voltage curves were highly nonlinear and showed considerable rectification at negative and positive potentials. The reversal potential was 6.4 + 1.0 mV (n = 3), consistent with a nonspecific cation channel. As shown in Fig. 3F, memantine block was observed over the whole voltage range tested. It increased by a factor of 1.6 from +50 mV to 80 mV and thus exhibited only mild voltage-dependence. We also measured the impact of memantine on the chloride current, activated secondarily in the presence of external Ca²⁺. Memantine inhibited the peak chloride current with an efficiency not significantly different from block of the isolated 9/10 current (Fig. 3E).

In hair cells, Ca²⁺ influx via nAChRs activates SK2 channels to give rise to IPSCs. Figure 4 shows, that this activation cascade may be reconstituted in X. laevis oocytes by coexpression of the 9/10 nAChR with SK2 channels. In oocytes expressing both channel species, application of ACh evoked a biphasic response at 70 mV. An initial inward current carried mainly by chloride (see above) was followed by an outward current due to the activation of SK2 channels (Fig. 4A). With the Cl driving force largely abolished and an increased driving force for K⁺ at a membrane potential of 30 mV, ACh induced a monophasic potassium outward current, similar to the response of isolated OHCs to ACh application (Blanchet et al., 1996; Evans, 1996). However, in coinjected oocytes, SK channel activation occurred considerably more slowly than in OHCs (20-80% rise time of 1.8 ± 0.2 s, n = 3). Memantine block of the SK2 current showed a dose-response relation that was characterized by an IC₅₀ value of 0.7 µM and a Hill coefficient of 1.1 (Fig. 4, B and C), very similar to the values obtained for the memantine block of 9/10 receptors (see Fig. 3E).

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Activation of SK2 channels by coexpressed 9/10 receptors is blocked by memantine. A, application of 100 µM ACh for the time indicated to an oocyte coexpressing 9, 10, and SK2. Traces are subsequent recordings from the same oocyte at the holding potentials indicated. B, K+ current induced by ACh (100 µM) in an oocyte as in A was reversibly blocked by memantine at the concentrations indicated. Holding potential was 30 mV. C, memantine dose-response curve for inhibition of the nAChR-induced SK2 current (n = 5 oocytes). Peak outward currents were recorded at 30 mV and normalized to the values preceding memantine application. Continuous line shows fit of the Hill equation (IC50 = 0.7 µM; nH= 1.1). Inhibition curve of IPSCs by memantine () is replotted from Fig. 1C.

However, these values were considerably different from those obtained for memantine-induced inhibition of IPSCs in OHCs (Fig. 1C). Thus, inhibition of IPSCs required 10-fold higher concentrations of memantine than inhibition of SK2 currents in the oocyte system (Fig. 4C). This difference might suggest that the nAChR underlying the generation of IPSCs is less susceptible to memantine block than 9/10 and may thus be indicative for a different subunit composition of OHC nAChRs. To test this possibility, we examined the effect of memantine on the nAChR of OHCs directly. AChR currents can be measured in isolation by uncoupling SK channel activation from Ca²⁺ influx via the nAChR with the fast Ca²⁺-chelator BAPTA (Fuchs and Murrow, 1992; Blanchet et al., 1996). Application of ACh to OHCs dialyzed with 10 mM BAPTA from the recording pipette induced inward currents of 110 ± 39 pA at 94 mV (n = 4). These currents were blocked by memantine with an IC₅₀ value of 1.1 µM and a Hill coefficient of 0.9 (Fig. 5A, B). This affinity is in close agreement with the values obtained for the block of 9/10 receptors and strongly suggests that the observed differences in blocking potency of memantine are not caused by a different receptor; instead, they may result from the specific mechanism of coupling between nAChRs and SK channels in OHCs.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Steady-state block of nAChR currents in OHCs by memantine. A, currents through nAChRs were recorded from an OHC at a holding potential of 94 mV. Inward currents induced by application of 200 µM ACh were reversibly blocked by coapplication of memantine as indicated by horizontal bars. The intracellular solution contained 10 mM BAPTA to prevent activation of SK currents. B, dose-inhibition relation for memantine block of nAChR currents (), recorded from four OHCs as in A. Steady-state amplitudes were normalized to the value recorded in the absence of memantine. The fit to the Hill equation (continuous line) yielded IC50 = 1.1 µM and nH = 0.9. Dose-inhibition curves for memantine block of 9/10 currents in oocytes () and of IPSCs () are replotted for better comparison.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have examined the action of memantine, a pore blocker of ligand gated ion channels, on inhibitory synaptic input to cochlear OHCs. It is shown that memantine abolishes IPSCs in the micromolar range by blocking Ca²⁺ entry via the OHC nAChR. To be able to analyze this block at the molecular level, we aimed to reconstitute the nAChR in a heterologous expression system.

Identity of the nAChR of OHCs. The nAChR of OHCs has been shown to contain the 9 subunit by a variety of methods including in situ hybridization (Elgoyhen et al., 1994; Morley et al., 1998), single-cell RT-PCR (Glowatzki et al., 1995), transgenic expression of green fluorescent protein controlled by the 9 promoter (Zuo et al., 1999), and inactivation of the 9 gene (Vetter et al., 1999). Homomeric 9 receptors, however, yield remarkably small currents when heterologously expressed in X. laevis oocytes (Figs. 2A and 3B; see also Elgoyhen et al., 1994), raising the possibility that an additional subunit is needed to yield the fully functional OHC receptor. However, OHCs lack any other known nAChR subtypes (Morley et al., 1998).

Memantine Block of 9/10 Receptors. Heteromeric 9/10 channels are reversibly blocked by memantine. This adamantane derivative is a well-characterized open-channel blocker of NMDA-type glutamate receptors with an IC₅₀ value of around 1 µM (at 70 mV) and also blocks neuronal 4/2 nAChRs with an IC₅₀ value of 7 µM (at 100 mV) (Chen et al., 1992; Parsons et al., 1993; Buisson and Bertrand, 1998). Thus, the affinity for memantine of 9/10 receptors ( 1 µM, see Figs. 3E and 5B) is considerably higher than of neuronal nAChRs and equals the affinity of NMDA receptors.

Action of Memantine on Cochlear Efferent Transmission. The block of OHC synaptic transmission by memantine establishes an impact on cochlear physiology by a drug that is used therapeutically (Parsons et al., 1999). The olivocochlear system acts on the cochlea by reducing its response to acoustic stimulation on two time scales. A "fast effect" caused by the quick opening of Ca²⁺-activated K⁺ channels by the nAChRs is followed by a slower effect, which is thought to involve second-messenger systems (Sridhar et al., 1995, 1997). Because the most obvious effect of memantine is the block of fast IPSCs, it can be expected to block the fast efferent effect, based on SK2 channel opening. Moreover, because this inhibition is caused by abolishment of Ca²⁺-influx via the OHC nAChR, the second, slower efferent effect that is thought to be triggered by the Ca²⁺ influx in the same synaptic events (Sridhar et al., 1997) is also expected to be blocked by memantine.