INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The endolymph in the cochlea is unique as an extracellular fluid because its predominant cation is K⁺ and its electrical potential (endocochlear DC potential, EP) is positive by 80 to 90 mV relative to the perilymph. Both the EP and the high K⁺ concentration in the endolymph are essential for the transduction of sound by hair cells. Transduction of sound begins at the cilia of the hair cells, which bear mechano-electrical transducer channels that are in contact with the endolymph. When the transducer channels open, K⁺ ions flow into hair cells. The main driving force for the K⁺ influx through the transducer channels is the electrical gradient across the ciliary membrane produced by the sum of the EP and the resting membrane potential of the hair cells (Dallos, 1996).

It has been accepted that the stria vascularis in the cochlea produces the endolymph and the EP, but a general agreement on the mechanism involved in EP production has not been reached. Within the stria vascularis are intermediate cells, which are melanocytes and migrate from the neural crest during ontogeny (Hilding and Ginzberg, 1977). It is known that a congenital deficiency in these cells causes low EP and an increase in the threshold of sound pressure levels required to elicit compound action potentials (Cable et al., 1994; Steel and Barkway, 1989). Recent physiological models for the mechanism of EP generation suggest that K⁺ channels may play an important role (Salt et al., 1987; Wangemann and Schacht, 1996), and we have proposed that the K⁺ channels in intermediate cells may in particular be critical in the generation of EP and K⁺ transport (Takeuchi and Ando, 1998b; Takeuchi and Ando, 1999). Our proposal is based on the anatomical structure of the stria vascularis and these electrophysiological observations: (i) the EP is reduced by some K⁺ channel blockers (Marcus et al., 1985; Takeuchi et al., 1996), and (ii) intermediate cells have a relatively large K⁺ conductance (Takeuchi and Ando, 1998b, 1999).

In this study, we examined the possibility that the membrane potential of intermediate cells (E_m) plays a critical role in the EP generation. This hypothesis can be tested by determining if agents that alter the E_m also alter the EP in similar measure. To accomplish this, the effect of K⁺ channel blockers on the E_m were compared with their effect on the EP. We studied the effect of the blockers on the E_m rather than on currents because (i) we hypothesize that the E_m of the intermediate cell generates the EP directly, and (ii) drug effects on the E_m cannot be determined precisely from those on currents. The E_m and the EP were studied using the same animal species (guinea pigs) to exclude species-dependent differences in drug sensitivity.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Measurement of membrane potentials of dissociated intermediate cells

Dissociated cells were prepared in essentially the same manner as reported previously (Takeuchi and Ando, 1998b). Cochleae of albino guinea pigs were obtained under deep anesthesia with pentobarbital sodium (75-100 mg/kg, i.p.). Tissue strips of the stria vascularis were incubated for 30 min at 24-26°C in a control bath solution containing 0.2% trypsin and then kept at 4°C for up to 4 h in the same solution until use. The strips were dissected with fine needles under visual control. Though intermediate cells of albino guinea pigs do not contain pigmented inclusions, single intermediate cells can be separated from other cells on the basis of their morphological characteristics (i.e., octopus-like shape and dendrite-like projections; Fig. 1), which are similar to those seen in gerbil intermediate cells (Takeuchi and Ando, 1998b). The E_m of these cells was measured in the zero-current clamp mode of the whole cell patch clamp technique with an amplifier (3900A with 3911A, Dagan, Minneapolis, MN) at 35 ± 1°C. The above technique allowed stable measurements of the E_m for up to 30 min. Pipette resistance was 2.2-3.6 M when filled with the pipette solution. Data storage and analyses were performed using pCLAMP (6.0.3, Axon, Foster City, CA). Liquid junction potentials were measured against a 3-M KCl electrode and corrected.

View larger version (112K): [in this window] [in a new window]   FIGURE 1   Light micrograph of an intermediate cell dissociated from albino guinea pig. p, patch pipette. Bar, 10 µm.

Two control bath solutions with different K⁺ concentrations ([K⁺]_bath) were prepared for the testing of the effect of K⁺ channel blockers on the E_m because: (i) effects of K⁺ channel blockers on the E_m might be dependent on [K⁺]_bath (see Results), and (ii) the K⁺ concentration in the intercellular space in the stria vascularis, to which the intermediate cell is exposed in vivo, might differ from that in the blood (see Discussion). The bath solution with 3.6 mM K⁺ contained (in mM): NaCl 110, Na-aspartate 40, KCl 3.6, HEPES 6, Tris 2.6, CaCl₂ 0.7, MgCl₂ 1, and glucose 5 (pH 7.4). The bath solution containing 1.2 mM K⁺ was prepared by replacing K⁺ with Na⁺. The pipette solution contained (in mM): KCl 110, K-aspartate 15, KOH 10, EGTA 5, HEPES 6, Tris 6, MgCl₂ 1.1, and Mg-ATP 2 (pH 7.2). E_m values were recorded for 2 min in each experimental condition, and steady state values were pooled for further analysis. When E_m values flickered even after a 2-min trial (Fig. 3, E and F), means for the last 30 s were used for analysis.

Vascular perfusion and measurement of the endocochlear potential

EP levels were maintained by vascular perfusion and K⁺ channel blockers were administered to the artificial blood in essentially the same manner as reported previously (Takeuchi et al., 1996), which is a modification of the technique described by others (Wada et al., 1979b; Kobayashi et al., 1984). A schematic drawing of this method is presented in Fig. 5 G. The vascular perfusion technique has at least two advantages for the study of the effect of K⁺-channel blockers on the EP, namely, blockers are quickly delivered to the stria vascularis, which has a dense capillary network, and the artificial blood flows constantly and is independent of the blood circulation system of animals. Albino guinea pigs weighing 260 to 450 g were anesthetized with ketamine (70 mg/kg, i.m.) and xylazine (10 mg/kg, i.m.) and artificially ventilated with room air. A pharyngo-laryngectomy was then performed and the bony wall of the skull base was removed. The basilar artery at both ends and all of the identifiable branches from the basilar artery except for the right anterior inferior cerebellar artery were occluded and a polyethylene catheter with a tip diameter of approximately 250 µm was inserted into the basilar artery. Perfusion was performed at a constant flow rate (1.2-1.6 ml/min) using a syringe pump. The temperature of the artificial blood at the catheter tip was 35 ± 1°C. Immediately after starting vascular perfusion, the animal was severed at the neck to secure venous routes for drainage. The artificial blood contained (in mM): NaCl 126, KCl 3.6, NaHCO₃ 24, CaCl₂ 1.3, MgCl₂ 1.2, KH₂PO₄ 0.5 and glucose 5. The artificial blood was bubbled with a mixture of 95% O₂ and 5% CO₂ for 60 min at 35 ± 1°C (pH, 7.4 ± 0.1) and stored in syringes.

EP levels were recorded from the scala media of the basal turn using a glass microelectrode (tip diameter approximately 1 µm) filled with 150 mM KCl and an electrometer (FD223, WPI, Sarasota, FL). The electrode was inserted into the scala media of the basal cochlear turn through the lateral wall of the cochlear duct. An Ag-AgCl wire placed in the subcutaneous tissue of the head served as reference. The recording system was zeroed when the pipette tip was placed on the spiral ligament before insertion into the scala media. Voltage drifts were within ±3 mV when measured with the pipette tip pulled back to the spiral ligament after experiment. Data from animals whose EP (>76 mV) was stable for 6 min before the administration of K⁺ channel blockers were pooled for further analysis.

The care and use of animals used in this study were approved by the Kochi Medical School Animal Care and Use Committee.

K⁺ channel blockers

We examined the effect of five K⁺ channel blockers on both the E_m of the intermediate cell and the EP. The blockers used were BaCl₂, CsCl, quinine-HCl, tetraethylammonium (TEA), and 4-aminopyridine (4-AP). These blockers were directly dissolved in the control bath solution and the artificial blood. As 4-AP is strongly basic, it was neutralized by equimolar HCl. When the concentration of blockers exceeded 5 mM, equimolar NaCl was removed.

Data presentation and statistics

Data were presented as mean ± SE. N indicates number of either cells (E_m) or animals (EP). Data obtained in the presence of blockers were compared with those under control conditions by paired Student's t-test, and changes were regarded as significant when P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of K⁺ channel blockers on the membrane potential of intermediate cells

Preliminary experiments revealed that currents in guinea pig intermediate cells were composed of two distinct K⁺ currents similar to those found in gerbil intermediate cells (Takeuchi and Ando 1999). These currents were (i) inwardly rectifying K⁺ currents, which were blocked almost completely by 0.5 mM Ba²⁺, and (ii) depolarization-activated outward K⁺ currents, which were resistant to 0.5 mM Ba²⁺ (Fig. 2 A). The chord conductances determined at the end of 500-ms voltage pulses were 152.1 ± 31.4 nS between 160mV and the reversal potential (E_rev, 88 mV) and 104.9 ± 9.3 nS between E_rev and 0 mV (N = 9). The membrane potential (E_m) recorded in the zero-current clamp mode was dependent on [K⁺]_bath (Fig. 2 B). The E_m changed by 55.1 mV per 10-fold change in [K⁺]_bath (Fig. 2 C), indicating that the major determinant of E_m was a K⁺ conductance.

View larger version (31K): [in this window] [in a new window]   FIGURE 2   K+ conductance in the intermediate cell. (A) K+ currents in the intermediate cell. Depolarization-activated currents could be separated from inward rectifier currents by 0.5 mM Ba2+. Bath K+ concentration ([K+]bath) was 3.6 mM. Membrane capacitance, 30 pF. Series resistance (6 M) was compensated by 80%. Arrowheads, zero-current level. Bar, 5 nA. (B) Trace of membrane potential (Em) recorded in zero-current clamp mode showing effect of changes in [K+]bath. (C) Em recorded in zero-current clamp mode plotted against [K+]bath (N = 6). The slope of the straight line obtained after applying linear regression was 55.1 mV/10-fold change in [K+]bath.

Representative E_m traces showing the effects of the K⁺ channel blockers are presented in Fig. 3, and dose-response relations are summarized in Fig. 4. As the E_m in control conditions was dependent on [K⁺]_bath, we calculated changes in E_m (E_m) to evaluate the effect of [K⁺]_bath on drug sensitivity (Fig. 4, B and D). When extracellular K⁺ concentration was set at the level of blood plasma ([K⁺]_bath = 3.6 mM), the effect of Ba²⁺ on the E_m began to appear at 0.15 mM, and E_m increased in a dose-dependent manner. Ba²⁺ showed stronger effects when [K⁺]_bath was reduced to 1.2 mM; the E_m was statistically significant even at 0.05 mM, and increased in a dose-dependent manner. The dose-response curve of E_m was shifted to the left at 10-fold less Ba²⁺ concentration when [K⁺]_bath was reduced from 3.6 mM to 1.2 mM (Fig. 4 B). Thus, when 5 mM Ba²⁺ is used, the resulting E_m was 47.2 ± 2.6 mV when [K⁺]_bath was 3.6 mM, and 77.8 ± 2.3 mV when [K⁺]_bath was 1.2 mM. The E_m recovered almost completely upon washout of 5 or 15 mM Ba²⁺ to 85.3 ± 2.7 mV when [K⁺]_bath was 3.6 mM, and to 105.8 ± 2.7 mV when [K⁺]_bath was 1.2 mM (Fig. 3, A and B). When Cs⁺ was present, the concentrations needed to obtain the same responses observed with Ba²⁺ were higher, and, as observed for Ba²⁺, its effects were dependent on [K⁺]_bath (i.e., larger E_m at lower [K⁺]_bath). When 30 mM Cs⁺ was present, the E_m was 7.0 ± 2.6 mV when [K⁺]_bath was 3.6 mM, and 23.0 ± 2.1 mV when [K⁺]_bath was 1.2 mM. Upon washout of 50 mM Cs⁺, the E_m recovered almost completely to 89.0 ± 1.1 mV when [K⁺]_bath was 3.6 mM and to 106.4 ± 2.0 mV when [K⁺]_bath was 1.2 mM (Fig. 3, C and D).

View larger version (41K): [in this window] [in a new window]   FIGURE 3   Traces of membrane potential (Em) recorded in zero-current clamp mode showing effects of K+ channel blockers under two [K+]bath conditions. (A and B) Ba2+, (C and D) Cs+, (E and F) quinine, (G and H) 4-aminopyridine, (I and J) tetraethylammonium. In A, C, E, G, and I, [K+]bath = 3.6 mM. In B, D, F, H, and J, [K+]bath = 1.2 mM.

View larger version (40K): [in this window] [in a new window]   FIGURE 4   Dose-response relations of membrane potential (Em). (A and C) Effect of blockers on Em (N = 8 for each data set). Em values under control conditions are also presented (control). 3.6K+ and 1.2K+ indicate [K+]bath in mM. (B and D) Dose-response relations of changes in Em from control values (Em) calculated from data presented in A and C.

Quinine at 0.1 mM did not cause significant changes in the E_m. When 0.3 to 1 mM quinine was used, flickering of the E_m, due most likely to interactions between quinine and K⁺ channels, was observed (Fig. 3, E and F). When the concentration of quinine was raised from 0.3 mM to 1 mM, a large change in the E_m appeared (Fig. 4 C). The E_m caused by 1 mM quinine was 82.2 ± 1.9 mV when [K⁺]_bath was 3.6 mM and 85.4 ± 6.8 mV when [K⁺]_bath was 1.2 mM. [K⁺]_bath-dependent differences in E_m like those observed for Ba²⁺ and Cs⁺, were not apparent at concentrations up to 1 mM (Fig. 4 D). Upon washout of 3 mM quinine, the E_m settled down to approximately 0 mV after a small transient recovery (Fig. 3, E and F). Neither 10 mM 4-AP nor 30 mM TEA resulted in apparent changes in the E_m regardless of [K⁺]_bath (Figs. 3, G-J, and 4, C and D).

Effect of blockers and [K⁺]_blood on the EP

The EP was depressed by Ba²⁺ in a dose-dependent manner, and it took 3 to 8 min for the EP to reach its minimum (Figs. 5 A, 6 A, and 7 A). When 5 mM Ba²⁺ was administered, the EP decreased from the control value of 83.5 ± 4.7 mV to a minimum value of 5.1 ± 5.4 mV, thereafter recovering over 20-30 min to a maximum of 13.3 ± 4.2 mV. Upon washout of 5 mM Ba²⁺, the EP recovered completely to 83.6 ± 4.9 mV in 5 to 8 min. These effects of Ba²⁺ confirm those reported previously (Marcus et al., 1985).

View larger version (32K): [in this window] [in a new window]   FIGURE 5   Recordings of EP showing the effect of blockers and the K+ concentration in artificial blood ([K+]blood). (A) Ba2+, (B) Cs+, (C) quinine, (D) 4-aminopyridine, (E) tetraethylammonium, (F) [K+]blood. Arrowheads, data points for comparison with Em (G) Method for vascular perfusion and recording of EP.

Cs⁺ had less intense effects on the EP than Ba²⁺ (Figs. 5 B, 6 A, and 7 C). When 10 mM Cs⁺ was administered, the EP decreased from a control value of 88.9 ± 1.9 mV to 81.1 ± 2.3 mV within 5 min, and remained at this level for up to 10 min. At 30 mM Cs⁺, the EP decreased to 72.1 ± 1.2 mV in 3 min and remained at this level for 3 to 5 min. Subsequently, the EP decreased to 40.0 ± 6.5 mV in the next 30 min. The EP recovered incompletely to 66.7 ± 6.0 mV in 10 min upon washout of 30 mM Cs⁺.

The EP was also depressed by quinine in a dose-dependent manner (Figs. 5 C, 6 A, and 7 B). Quinine at 0.3 mM caused a decrease in the EP from the control value of 85.7 ± 2.6 mV to 81.5 ± 3.0 mV in 3 to 6 min, whereas 1 mM quinine elicited a large biphasic change in the EP (Fig. 5 C). During the first phase, which continued for 8 to 10 min, the EP declined to 12.8 ± 4.3 mV. During the second phase, which began before the first phase reached a steady level, the EP declined to 6.9 ± 2.1 mV. Upon washout of 1 mM quinine, the EP recovered only partially to 17.9 ± 4.9 mV. Neither TEA at 30 mM nor 4-AP at 10 mM caused statistically significant changes in the EP (Figs. 5, D and E, and 6 B).

Manipulations of the K⁺ concentration in the artificial blood ([K⁺]_blood) altered the EP only relatively weakly (Figs. 5 F and 7 D). Upon elevation of [K⁺]_blood from 3.6 mM to 36 mM, the EP declined from 83.0 ± 2.1 mV to 72.0 ± 2.2 mV in 3 min and remained at a quasi-steady level for 10 min. A reduction in [K⁺]_blood from 3.6 mM to 1.2 mM did not cause significant changes in the EP over 10 min of observation. These observations confirm previous reports (Wada et al., 1979a; Marcus et al., 1985).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

E_m measurement and effect of K⁺ channel blockers on E_m

E_m was measured in the zero-current clamp mode of the whole-cell patch clamp configuration. Though it cannot be excluded that the dialysis of cytosol by the pipette solution might suppress ion channels that are dependent on native cytosolic substances, it is unlikely that the observed K⁺ conductance is silent in native cells and is artificially activated under the experimental condition. At least two distinct K⁺ channels (i.e., the inward rectifier K⁺ and the depolarization-activated K⁺ channels) contribute to the E_m of the intermediate cell and the sensitivities of these ion channels to blockers vary (Takeuchi and Ando, 1998b, 1999).

The reduction of the E_m by Ba²⁺ and Cs⁺ was enhanced by lowering the extracellular K⁺ concentration from 3.6 mM to 1.2 mM (Fig. 4 B). Two mechanisms may be responsible for the enhanced effect of blockers at lower [K⁺]_bath: (i) reduced competition between K⁺ ions and blockers, and (ii) larger E_m values under control conditions. The first mechanism may apply to competitive blockers such as Ba²⁺ and Cs⁺, as these blockers are known to compete with K⁺ ions, most probably in the selective filter of K⁺ channels (Armstrong and Taylor, 1980; Hagiwara et al., 1976). The second mechanism may apply to all the effective blockers.

Mechanisms underlying EP inhibition by K⁺ channel blockers

Ba²⁺, Cs⁺, and quinine depressed the EP in a dose-dependent manner, whereas TEA at 30 mM and 4-AP at 10 mM did not change the EP (Fig. 6). It is likely that the blockers in the artificial blood exerted most of their effects on the EP by affecting cells facing the intercellular space of the stria vascularis (intrastrial space) because: (i) a dense capillary network is present in the stria vascularis, and (ii) it is known that when Ba²⁺ and quinine are administered to the perilymph, these blockers cause an increase in the EP (Marcus, 1984; Wang et al., 1993), which is opposite to what we observed when they were administered to the artificial blood. Because the stria vascularis is one of the most densely vascularized tissues in the inner ear, it is expected that the effects of blockers will appear quickly.

View larger version (23K): [in this window] [in a new window]   FIGURE 6   Dose-response relationships of EP. EPs under control conditions are also presented. (A) Ba2+, Cs+, and quinine. N = 4 for each data set. (B) 4-aminopyridine (4-AP) and tetraethylammonium (TEA). N = 3 for each data set.

With regard to the effect of Ba²⁺ on the EP, the relatively fast onset of EP decline, and the rapid and complete recovery upon washout suggest first, that the decline in EP was due to blockage of K⁺ channels, and second, that Ba²⁺ exerts its effect extracellularly. The recovery of the EP observed during prolonged administration of 5 mM Ba²⁺ might be due to blockage of pathways which mediate K⁺-diffusion from the endolymph to the perilymph. It is likely that such pathways reside in the basolateral membrane of hair cells, which are not well vascularized. Because administration of Ba²⁺ to the perilymph in the scala tympani causes an increase in the EP (Marcus, 1984), Ba²⁺ might be able to reach the perilymph and induce a slow increase in the EP when administered to the vascular perfusate at high concentrations.

Prolonged administration of 30 mM Cs⁺ and 1 mM quinine results in a late and gradual decrease in the EP (Fig. 5, B and C). It is known that Cs⁺ is taken up via Na⁺-K⁺-ATPase (Baker et al., 1969) and that Cs⁺ blocks K⁺ channels from within the cell. The late onset of the effect of 30 mM Cs⁺ on the EP might be due to the gradual accumulation of Cs⁺ in the cell. That the recovery of the EP upon washout of 30 mM Cs⁺ is incomplete might be due to remaining Cs⁺ within the cell. With regard to quinine, this molecule is relatively lipophilic and is thus expected to permeate through the plasma membrane to some extent. The second phase of the EP decline caused by 1 mM quinine might be caused by the effects of quinine on intracellular structures, such as the intracellular domains of K⁺ channels and K⁺ transporters in mitochondria (Diwan, 1986). Another possibility might be such that quinine inhibited a certain type of Ca²⁺ channels, whose block disturbed the normal operation of other mechanisms that were responsible for the generation of the EP. The limited recovery of the EP upon washout of 1 mM quinine (Fig. 5 C) may be due to the irreversible binding of quinine to K⁺ channels since the effect of 3 mM quinine on the E_m of the intermediate cell was irreversible (Fig. 3, E and F). Not only intermediate cells, but also other cells in the stria vascularis, including marginal cells, might be affected by prolonged administration of 30 mM Cs⁺ and 1 mM quinine.

The basilar artery and the anterior inferior cerebellar artery were expanded throughout perfusion and did not show apparent changes in their widths upon administration of blockers (data not shown). Thus, it is likely that the vascular perfusion at a constant flow rate was able to overcome any vasoconstriction caused by blockers, although possible effects of blockers on perfused vessels peripheral to the anterior inferior cerebellar artery cannot be excluded.

Comparison between the effect of K⁺ channel blockers on E_m and EP

With regard to the method used to measure EP, it is possible for the blockers to accumulate in native cells and thus cause effects not observed in patch-clamped cells. To enable us to compare the changes in the EP (EP) and E_m values directly, we used those EP values that were most likely to be affected only minimally by the intracellular presence of the blockers. Thus, because Ba²⁺ is likely to cause the decline in the EP by predominantly inhibiting K⁺ channels from the extracellular side (see previous subsection) and the maximal effects of Ba²⁺ appeared in 3 to 8 min, we used EPs at the quasi steady-states observed within 3 to 8 min of blocker administration to compare EP with E_m (Fig. 5). With regard to diffusion of blockers, we assume that the time course of diffusion from capillaries to effective sites in the stria vascularis for Ba²⁺, Cs⁺, and quinine is similar. As 1 mM quinine caused biphasic changes in the EP and the second phase began before the first phase reached a steady state, we decided to use the EP value obtained just before the onset of the second phase for comparison with E_m (Fig. 5 C).

E_m values were compared with EPs to test our hypothesis that the E_m of the intermediate cell directly influences the EP (Fig. 7). The polarity of E_m is reversed in Fig. 7 to compare with EP because the voltage jump across the plasma membrane of intermediate cells and/or basal cells in the direction from the perilymph to the endolymph (i.e., the extracellular potential in the intrastrial space relative to the interior of intermediate cells) is likely to be the major source of the EP (Salt et al., 1987; Wangemann and Schacht, 1996; Takeuchi and Ando, 1999). The relationship between the Ba²⁺ concentration ([Ba²⁺]) and EP is similar to the relationship between [Ba²⁺] and E_m obtained at the lower [K⁺]_bath, whereas this relationship is not as clear at the higher [K⁺]_bath (Fig. 7 A). The relationship between [Cs⁺] and EP is also closer to the relationship between [Cs⁺] and E_m obtained at the lower [K⁺]_bath (Fig. 7 C). With regard to the effect of quinine, [K⁺]_bath-dependent differences in E_m were not observed at quinine concentrations up to 1 mM (Fig. 4 D), and the relationship between [quinine] and EP is similar to the relationship between [quinine] and E_m (Fig. 7 B).

View larger version (29K): [in this window] [in a new window]   FIGURE 7   Comparison between changes in EP (EP) and Em (Em). (AC) Effect of blockers. 3.6K+ and 1.2K+ indicate [K+]bath in mM. Data presented in Figs. 4, C and D, and 6 A are rearranged. (D) Effect of K+ concentration. Control values were obtained when [K+]blood and [K+]bath were 3.6 mM. Data presented in Fig. 2 C are rearranged to obtain Em. N = 3 for EP.

Comparison of the effect of [K⁺]_bath on E_m and the effect of [K⁺]_blood on EP

The E_m responses to changes in [K⁺]_bath were close to those expected from the Nernst equation for a K⁺ selective membrane, whereas the EP responses to changes in [K⁺]_blood were relatively small (Fig. 7 D). If, as we hypothesize, the EP is directly dependent on the membrane potential of intermediate cells determined by K⁺ channels, then the resistance of EP to changes in [K⁺]_blood should be explained. In view of the close relationships between E_m obtained at [K⁺]_bath of 1.2 mM and EP (Fig. 7, A and C), we suggest that the K⁺ concentration in the intrastrial space ([K⁺]_IS) is actively controlled at a concentration lower than [K⁺]_blood, and that because of this [K⁺]_IS will not rise to [K⁺]_blood levels even if [K⁺]_blood is raised substantially. That[K⁺]_IS is actively controlled is likely when the following well known facts are considered: (i) the intrastrial space is quite small as plasma membranes facing the intrastrial space are associated very closely, and (ii) the basolateral membrane of marginal cells has strong immunoreactivity for Na⁺-K⁺-ATPase and Na⁺-K⁺-Cl cotransporter (Iwano et al., 1989; Crouch et al., 1997). Both transporters act to take up K⁺ from the intrastrial space (Wangemann and Schacht, 1996). The enzyme activity of Na⁺-K⁺-ATPase in the stria vascularis is dependent on K⁺ concentration (K_m, 0.87 mM) and responds sensitively to changes in K⁺ concentration in the range between 0 and 5 mM (Kuijpers and Bonting, 1969). Similarly, the transport activity of Na⁺-K⁺-Cl cotransporter measured by furosemide-sensitive Rb⁺ uptake is also dependent on K⁺ (Rb⁺) concentration (K_m, 1.80 mM), and responds sensitively to changes in K⁺ (Rb⁺) concentration in the range between 0 and 5 mM (Duhm, 1987). The above properties of Na⁺-K⁺-ATPase and Na⁺-K⁺-Cl cotransporter, in association with their abundant expression, seem to be favorable to the regulation of [K⁺]_IS at relatively low revels. In addition, it has been reported that radioactive K⁺ administered to the perilymph accumulates in the endolymph and that within 40 min the concentration of radioactive K⁺ in the endolymph is 12 times that in the perilymph (Konishi et al., 1978). The latter suggests that the K⁺ taken up by the marginal cell is mainly from the perilymph, not the blood plasma. Though the [K⁺]_IS has been examined using double-barreled ion selective microelectrodes (Salt et al., 1987; Ikeda and Morizono, 1989), the reported values (3.6-22.1 mM) should be considered preliminary because of technical difficulties related to the small volume of this space (Wangemann and Schacht, 1996). It is thus possible that changing the [K⁺]_blood has only limited effects on the [K⁺]_IS, explaining the relatively small effect of [K⁺]_blood on the EP.

If the [K⁺]_IS is maintained at approximately 4 mM, there should be a mechanism other than K⁺ channels in intermediate cells because the E_m of the intermediate cells is less sensitive to Ba²⁺ and Cs⁺ than the EP when the [K⁺]_bath is 3.6 mM (Fig. 7, A and C). This putative mechanism, if any, should be insensitive to TEA and 4-AP, as sensitive to quinine as K⁺ channels in intermediate cells, and more sensitive to Ba²⁺ and Cs⁺ than K⁺ channels in intermediate cells. Such a mechanism has not been found in the stria vascularis until now, and cannot be mediated by the large conductance K⁺ channel in basal cells, as this channel is completely blocked by extracellular TEA at 1 mM (Takeuchi and Irimajiri, 1996b).

Five-compartment (two-cell) model of the stria vascularis

On the basis of the results presented in this and previous studies (Salt et al., 1987; Wangemann and Schacht, 1996; Takeuchi and Ando, 1999), we propose an extended version of the electrophysiological model of the stria vascularis (Fig. 8). The interior of the stria vascularis is isolated by two distinct cell sheets connected by tight junctions; one sheet is the marginal cell layer, and the other is the basal cell layer. Capillary endothelial cells are also connected by tight junctions. The junctions are expected to be electrically tight so as to minimize electrical shunting between the endolymph and the perilymph. The intermediate cells and the capillary network are located between the two cell sheets mentioned above. The marginal cell layer serves as a barrier between the endolymph and the intrastrial space. The basal cell layer functions as a barrier between the intrastrial space and the perilymph, and the capillary endothelial cell layer functions as a barrier between the intrastrial space and the blood in the capillary. Intermediate cells, basal cells, pericytes, endothelial cells, and fibrocytes in the spiral ligament constitute an electrical syncytium, because gap junctions connect these cells (Kikuchi et al., 1995; Takeuchi and Ando, 1998a). Accordingly, the stria vascularis, including the perilymph and the endolymph, can be regarded to be composed of five compartments separated by four distinctive membranes (m1-m4 in Fig. 8). In this model, the electrical potential of the blood in capillaries is assumed to be the same as that of the perilymph. When E_m1 through E_m4 are defined as potential differences across each membrane relative to the extracellular side in vivo, the EP would be the sum of E_m1, E_m2, E_m3, and E_m4.

View larger version (57K): [in this window] [in a new window]   FIGURE 8   Five-compartment (two-cell) model of the stria vascularis. m1-m4 denote membranes bordering adjacent compartments. Cells constituting each membrane are indicated in parentheses. ma, marginal cell; i, intermediate cell; p, pericyte; e, endothelial cell; b, basal cell; f, fibrocyte in the spiral ligament. Electrical potentials are referred to the perilymph. KIR, inward rectifier K+ channel. KV, depolarization-activated K+ channel. Asterisk, unspecified ion channel(s) that determine the membrane potential of m1. In intrastrial space, both electrical potential and [K+] are hypothetical values based on this study.

Measurements in vivo using intracellular microelectrodes indicate that the intracellular potential of marginal cells relative to the endolymph, which corresponds to E_m4, is 0 to 10 mV (Melichar and Syka, 1987; Offner et al. 1987; Salt et al., 1987; Ikeda and Morizono, 1989). As the basolateral membrane of marginal cells (m3) has a large Cl conductance (Takeuchi et al., 1995; Takeuchi and Irimajiri, 1996a) and the transepithelial voltage recorded from cell sheets of the stria vascularis shows a large change when the Cl concentration on the basolateral side is reduced (Wangemann et al., 1995), E_m3 is most likely to be determined predominantly by the Cl conductance. The E_m3 may be estimated to be approximately 10 mV from the intracellular potential of dissociated marginal cells, which depends mostly on the Cl conductance in the basolateral membrane (Takeuchi et al., 1997). The intracellular potential of cells connected by gap junctions relative to the perilymph (E_m1) is estimated to be in the range between 10 and 0 mV, as potentials in this range have been recorded in vivo (Melichar and Syka, 1987; Salt et al., 1987; Ikeda and Morizono, 1989). The above discussion suggests that neither m1, m3, nor m4 are likely to be the location of the positive voltage jump, which is directly related to the EP.

We consider here the possibility that m2 is the location of the positive voltage jump directly related to the EP. The m2 is composed of intermediate cells, pericytes, inner membranes (i.e., membranes facing the intrastrial space) of basal cells, and inner membranes of endothelial cells. With regard to the surface area of cells constituting m2, intermediate cells are likely to be a major component of m2 in view of the fine network composed of these cells in situ (Ando and Takeuchi, 1999). The relatively large K⁺ conductance in these cells also suggests that intermediate cells may be major determinants of the conductive properties of m2 (Takeuchi and Ando 1998b, 1999). Assuming that (i) m2 is a K⁺-selective membrane, (ii) intracellular K⁺ concentration is 140 mM, and (iii) [K⁺]_IS is actively controlled at 1-2 mM, then E_m2 could be approximately 110 mV (the potential in the cell relative to the intrastrial space), and the electrical potential in the intrastrial space referred to the perilymph could be approximately 100 to 110 mV when E_m1 is 10 mV. The presence of a space with both a positive potential and a relatively low K⁺ concentration in the stria vascularis has been proposed previously (Salt et al., 1987; Ikeda and Morizono, 1989). This space may correspond to the intrastrial space. Using the estimated E_m3 of 10 mV, the intracellular potential of marginal cells in vivo (relative to the perilymph) may be 90 to 100 mV. This value is in conformity with the intracellular potential of marginal cells measured in vivo (Melichar and Syka, 1987; Offner et al., 1987; Salt et al., 1987; Ikeda and Morizono, 1989). In the above discussion, we have assumed that the transepithelial voltage across the marginal cell layer (i.e., the voltage jump from the intrastrial space to the endolymph) is negative (Fig. 8). With regard to this transepithelial voltage, it has been reported that the transepithelial voltage across vestibular dark cells, which are analogous to marginal cells, changes from positive to negative when the K⁺ concentration on the apical side is maintained at 145 mM and the basolateral K⁺ concentration is reduced from 3.5 to 2 mM (Wangemann et al., 1996). The above report supports the model (Fig. 8), as it is well known that marginal cells and vestibular dark cells have very similar electrophysiological properties and functions (Wangemann, 1995).

Thus, in summary, it is the most likely that m2 is the location of the voltage jump that is directly related to the EP generation, as suggested previously (Salt et al. 1987; Wangemann and Schacht, 1996; Takeuchi and Ando, 1999), and that the conductive property of m2 depends largely on K⁺ channels in the intermediate cell. Further studies are needed to clarify the electrophysiological properties of m1 (asterisk in Fig. 8) and the non-intermediate cells constituting m2.

Finally, we would like to stress the important role of marginal cells in the five-compartment model. Although the plasma membranes of marginal cells (m3 and m4) are not likely to be the location of the positive voltage jump as discussed above, marginal cells may contribute to EP generation by taking up K⁺ from the intrastrial space, thus maintaining the low [K⁺]_IS, which is essential for a large K⁺ diffusion potential across m2. Marginal cells also play a major role in the K⁺ secretion into the endolymph (Wangemann et al., 1995). It is known that K⁺ is secreted via the KVLQT1/I_sK channel in the apical membrane of marginal cells, and that the K⁺ secretion is impaired by a null mutation of the I_sK gene (Vetter et al. 1996).