Lack of pendrin HCO₃^- transport elevates vestibular endolymphatic [Ca²⁺] by inhibition of acid-sensitive TRPV5 and TRPV6 channels

Abstract

The low Ca²⁺ concentration of mammalian endolymph in the inner ear is required for normal hearing and balance. We reported [Yamauchi et al. Biochem Biophys Res Commun, 2005] that the epithelial Ca²⁺ channels TRPV5 and TRPV6 are expressed in the vestibular system and that TRPV5 expression is stimulated by 1,25-dihydroxyvitamin D₃ (1,25(OH)₂D₃), as also reported in kidney. TRPV5/6 channels are known to be inhibited by extracellular acidic pH. Endolymphatic pH, [Ca²⁺] and transepithelial potential of the utricle (UP) were measured in Cl^-/ ${\mathrm{HCO}}_3^{-}$ exchanger pendrin (SLC26A4) knockout mice in vivo. Slc26a4^-/- mice exhibit reduced pH and UP and increased [Ca²⁺]. Monolayers of primary cultures of rat semicircular canal duct (SCCD) cells were grown on permeable supports and cellular uptake of ⁴⁵Ca²⁺ was measured individually from the apical and basolateral sides. Net uptake of ⁴⁵Ca²⁺ was greater after incubation with 1,25(OH)₂D₃. Net ⁴⁵Ca²⁺ absorption was dramatically inhibited by low apical pH and was stimulated by apical alkaline pH. Gadolinium, lanthanum and ruthenium red reduced apical uptake. These observations support the notion that one aspect of vestibular dysfunction in Pendred syndrome is a pathological elevation of endolymphatic [Ca²⁺] due to luminal acidification and consequent inhibition of TRPV5/6-mediated Ca²⁺ absorption.

INTRODUCTION

Pendrin (PDS or SLC26A4) is an anion exchanger that is capable of transporting iodide, chloride, formate, nitrate and bicarbonate and functions as a ${\mathrm{HCO}}_3^{-}$ -secreting mechanism in kidney (30; 31; 33). SLC26A4 is expressed in kidney (29), the inner ear, thyroid (7), mammary gland (26), uterus (34), testis (16), vas deferens (3) and placenta (2). Mutation or deletion of the SLC26A4 gene leads to acidification of the urine (15) and to Pendred syndrome, which is an autosomal recessive disorder characterized by sensorineural hearing loss and goiter (21; 25).

The availability of SLC26A4 knockout mice (Slc26a4^-/-) makes it possible to perform a direct study of the inner ear defects that occur in the absence of SLC26A4 (6). Slc26a4^-/- mice are completely deaf and also display signs of vestibular dysfunction (6). In the vestibular system, pendrin is expressed in the apical membrane of vestibular transitional cells in the utricle and ampullae (40). If SLC26A4 secretes ${\mathrm{HCO}}_3^{-}$ in the vestibule, and if there are no strong compensatory mechanisms, it can be predicted that there may be an acidification of endolymph when SLC26A4 is deleted. An altered endolymphatic pH can be expected to affect other ion transport processes in the luminal membrane of epithelial cells bordering the lumen, since several ion channels are known to be highly sensitive to extracellular pH.

The [Ca²⁺] of vestibular endolymph (ca. 250 μM) is lower than perilymph (ca. 1 mM) and it has a critical role in sensory transduction through hair cells (20). A Ca²⁺ absorption system in inner ear epithelial cells must be present to maintain the low [Ca²⁺] of vestibular endolymph. Recently, we reported that the epithelial Ca²⁺ channels TRPV5 and TRPV6 are expressed in the semicircular canal duct (SCCD) of the vestibular system and expression of TRPV5 is regulated by 1,25(OH)₂D₃, as in some other systems (9; 36; 44). TRPV5 and TRPV6 belong to the transient receptor potential (TRP) family of channels and are the only two highly Ca²⁺-selective TRP channels (12; 22). TRPV5 is expressed in specific tissues such as kidney, placenta and bone and plays a major role in Ca²⁺ transport and is localized at the apical membrane of epithelial cells or at the ruffled border membrane of osteoclasts (10; 11; 37). Recently, the molecular mechanisms of TRPV5/6 inhibition by both intra- and extracellular acidic pH were reported (45).

Due to the presence of the TRPV5/6 Ca²⁺-absorptive pathway in the vestibular system and its known inhibition by extracellular acid in other systems, we hypothesized that part of the vestibular dysfunction observed in Pendred syndrome may be due to an acidification of endolymph (loss of ${\mathrm{HCO}}_3^{-}$ secretion), which in turn would lead to an elevation of luminal [Ca²⁺]. We therefore sought to measure the luminal pH and [Ca²⁺] in wild type (Slc26a4^+/+), heterozygous (Slc26a4^+/-) and knockout (Slc26a4^-/-) mice and to determine if Ca²⁺ uptake across the apical and basolateral membranes of vestibular Ca²⁺-absorbing epithelia were sensitive to apical acidification and to known TRPV5/6 blockers.

We indeed observed a reduced pH and increased [Ca²⁺] in the vestibular lumen of Slc26a4^-/- mice. Further, 1,25(OH)₂D₃ increased ⁴⁵Ca²⁺ absorption and TRPV5/6 inhibitors reduced apical uptake of ⁴⁵Ca²⁺. ⁴⁵Ca²⁺ absorption was inhibited by apical acid pH and was stimulated by apical alkaline pH, consistent with the notion that one aspect of vestibular dysfunction in Pendred syndrome is a pathological elevation of endolymphatic [Ca²⁺] due to luminal acidification and consequent inhibition of TRPV5/6-mediated Ca²⁺ absorption.

METHODS

Utricular endolymphatic potential (UP), pH and [Ca²⁺]

Adult Slc26a4^-/- and Slc26a4^+/- mice were obtained from a colony at Kansas State University that was established with breeders kindly provided by Dr. Susan Wall. The mouse strain 129Sv/Ev (Taconic, Germantown, NY) was used as the source of Slc26a4^+/+ mice, since Slc26a4^-/- mice were generated in this background. Young adult mice 30-142 days old were deeply anesthetized with Tribromoethanol (640 mg/kg i.p.; Fluka 90710) in 0.9% NaCl. The Institutional Animal Care and Use Committee of Kansas State University approved all experimental protocols.

UP, pH and [Ca²⁺] were measured with double barreled microelectrodes (one side pH- or Ca²⁺- sensitive and the other for voltage; see below for details) using procedures developed by modifying previously described protocols (18; 19). For both pH and Ca²⁺ electrodes, two pieces of glass tubing containing a glass filament (World Precision Instruments 1B100F-4, Sarasota, FL) were cut to 81 mm and 60 mm and pulled using a micropipette puller (Narishige PD-5, Tokyo, Japan). After heating pulled capillaries at 180 °C for 2 h, the ion-selective barrel was silanized by placing the open end of that barrel through a hole in the lid of a beaker at 210 °C in which we put 0.08 ml Dimethyldichlorosilane (Fluka 40136) for 90 s. The reference barrel was protected from silanization by sealing the open end with Parafilm (Alcan Packaging). After heating electrodes at 180 °C for 3 h, the tips were broken to ~3 μm OD.

For pH electrodes, the reference barrel was filled with 1 M KCl and the ion-selective barrel was filled at the tip with Hydrogen ionophore II - cocktail A (Fluka 95297) and back-filled with buffer solution (KCl 500 mM, HEPES 20 mM, pH7.34). For Ca²⁺ electrodes, the reference barrel was filled with 150 mM KCl and the ion-selective barrel was filled at the tip with Calcium ionophore I - cocktail A (Fluka 21048) and back-filled with 500 mM CaCl₂. Connection to each barrel was made with a Ag-AgCl wire. Each electrode was connected to an input of a dual electrometer (World Precision Instruments FD223) and buffered outputs were led to a Data Acquisition System (Axon Instruments DIGIDATA 1322A) and recorded with AxoScope 9 (Axon Instruments). A pulse generator was used to inject current pulses (1 nA) through the reference barrel to monitor the resistance of the electrode.

pH electrodes were calibrated at three different pH values (composition in mM): pH 6 (130 NaCl, 20 MES), pH 7 (130 NaCl, 20 HEPES) and pH 8 (130 NaCl, 20 Tricine) and had an average slope of 56.6 ± 0.7 mV/pH (n=16). Ca²⁺ electrodes were calibrated at three Ca²⁺ concentrations (composition in mM): 10 μM Ca²⁺ (0.121 CaCl₂, 150 KCl, 10 HEPES, 1 C₆H₇NO₆Na₂ (Nitrilotriacetic acid; Sigma N-0128), pH 7.4), 100 μM Ca²⁺ (0.1 CaCl₂, 150 KCl, 10 HEPES, pH7.4) and 1 mM Ca²⁺ (1 CaCl₂, 150 KCl, 10 HEPES, pH 7.4) and had an average slope of 26.2 ± 0.8 mV/decade concentration (n=13). Data were analyzed using custom software written by P.W. in LabTalk (Origin 6.0, OriginLab, Northampton, MA).

To reduce the diffusion of CO₂ through the exposed surface of the perilymph to the ambient air, which leads to higher pH compared to unexposed perilymph, we put liquid Sylgard 184 (Dow Corning Corporation, Midland, MI) at the fluid surface after placing the electrodes in the perilymph. The electrode was maintained below the surface of the perilymph solution after the measurement of inner ear fluids and the first calibration point was taken in situ in the bulla pocket of the temporal bone. The pH 7 buffer for calibration was placed on the Sylgard layer and the electrode retracted up into this buffer from the perilymph surrounding the utricle. Approximately 85% of the electrodes survived the travel through the Sylgard layer; outcomes from experiments in which electrodes failed this maneuver were excluded from the data set. The remaining calibration points to determine slope sensitivity were obtained by placing calibration solutions in a small conducting Ringer- agarose (Fisher Scientific, BP164) cup that rested on the exposed neck muscles. This procedure was used to preclude voltage offsets introduced through movement of the reference electrode. In some cases the electrode failed after moving to the calibration cup. In these few cases, the pre-experiment electrode slope was used in conjunction with the in situ calibration point. Calibration of Ca²⁺ electrodes was identical except that the first point was taken by flushing the perilymph space with 1 mM Ca²⁺ solution and raising the electrode into a region of the pool remote from the tissues.

Perilymph pH, [Ca²⁺] and UP were corrected for the liquid junction potentials of 1.3 mV (pH electrode) and 4 mV (Ca²⁺ electrode) between the voltage barrels in calibrating solution and in perilymph and between the voltage electrode in perilymph and endolymph.

Cellular uptake measurements

Primary cultures of Wistar rat SCCD were prepared and incubated in DMEM/F-12 (Invitrogen, #12500-062) supplemented with 5% fetal bovine serum, penicillin (100 U/ml), streptomycin (100 μg/ml) as described previously (44). Primary cultures of SCCD cells from rat were used to study transport function since murine SCCD cells have been observed to not proliferate in culture. Briefly, SCCD epithelial cells from neonatal rats were cultured on 12 mm diameter Snapwell permeable supports (0.4 μm pore, Costar 3801, Corning, NY). Confluence of primary cultures (10-14 days after seeding) was verified visually and by measurement of transepithelial electrical resistance using an Endohm meter (World Precision Instruments). Cultures were incubated further for 24 h in the presence or absence of 1,25(OH)₂D₃ (DM-200, Biomol, Plymouth Meeting, PA) (44).

The method for cellular ⁴⁵Ca²⁺ uptake determination was modified from Den Dekker et al. (5). Cells were washed twice on both sides with pre-warmed Ca²⁺-free solution (in mM: 145 NaCl, 4 KCl, 1 MgCl₂, 20 HEPES, 5 glucose, pH 7.4). Cells were then incubated in this medium at 37 °C and 5% CO₂ for 45 – 60 min prior to treatment. Cells tested with Ca²⁺ channel blockers or different pH buffers had the apical medium changed after 45 minutes of pre-treatment, allowing 15 min incubation with the appropriate blocker or pH. The blocker medium was Ca²⁺-free solution with an added mixture of known inhibitors of TRPV5/6 (10 μM Ruthenium Red, 100 μM LaCl₃ and 100 μM GdCl₃) or each blocker in the same concentration by itself. The various pH buffers were similar to Ca²⁺-free solution except the pH 5.5 contained 20 mM MES as a buffer in place of HEPES and the pH of each was adjusted with HCl or NaOH as appropriate. The pre-treatment buffer on either the apical or basolateral side was replaced with ⁴⁵Ca²⁺ uptake buffer (in mM: 145 NaCl, 4KCl, 1 MgCl₂, 0.1 CaCl₂, 20 HEPES, 5 glucose with 10 μM Verapamil, pH 7.4 and at least 1 μCi ⁴⁵Ca²⁺/ml) for 15 min. Verapamil was utilized to preclude measuring Ca²⁺ uptake via L-type Ca²⁺ channels, which play important roles in some epithelia (27). Verapamil at the concentration used here is sufficient for a virtually complete block of Ca²⁺ current through these channels (13). The radiotracer was purchased as 1 mCi (37 MBq) ⁴⁵Ca²⁺ as CaCl₂ in 17 μl H₂O (Perkin Elmer Life Sciences, Boston, #NEZ013001MC).

The cells were then rinsed 3 times on both sides with ice cold wash buffer (in mM: 145 NaCl, 4KCl, 1 MgCl₂, 0.5 CaCl₂, 1.5 LaCl₃, 20 HEPES, 5 glucose, pH 7.4). The wash buffer was removed from both compartments and 200 μL of Lysis buffer (Lysis Reagent 1 from cAMP kit RPN225, Amersham Biosciences, Piscataway NJ) was added to the apical compartment. After 5 minutes the lysate was collected and radioactivity was determined with a liquid scintillation counter (Packard Tri-Carb 2100TR, Meriden, CT). The difference in uptake of ⁴⁵Ca²⁺ from the apical and basolateral sides was taken as a relative measure of net flux.

Statistics

Data are presented as means ± SE from n observations. Net uptakes are the difference of the mean apical and basolateral uptakes; the SE was pooled from the variance of the apical and basolateral uptake data, as calculated from the variances with the Java applet at http://home.ubalt.edu/ntsbarsh/Business-stat/otherapplets/Pooled.htm. Significance between the UP, pH and Ca²⁺ control and knockout mouse data and the significance of Ca²⁺ uptake data were calculated with an unpaired t-test. The comparison of endolymph values to the corresponding perilymph values (* within bars, Fig. 2) were performed with a paired t-test. P < 0.05 represents a significant difference.

Figure 2.

Endolymphatic pH and [Ca²⁺]. (Top panel) endolymphatic pH of control and knockout (-/-) genotypes compared to control perilymphatic pH (dotted line). (Bottom panel) endolymphatic [Ca²⁺] of control and knockout (-/-) genotypes compared to control perilymphatic pH (dotted line). * inside bars, P < 0.05 versus perilymphatic pH or [Ca²⁺]; * above bars, P < 0.05. Number of measurements at each bar.

RESULTS

UP, pH and [Ca²⁺] in the utricle

Across all measurements, Slc26a4^+/- mice showed no significant differences in UP, pH and [Ca²⁺] compared with Slc26a4^+/+ mice, consistent with an autosomal recessive trait. There were no differences between Slc26a4^+/+ mice compared with Slc26a4^+/- mice in the transepithelial potential (UP; -1.5 ± 0.5 mV (n=8) vs -1.5 ± 0.3 mV (n=9)), perilymphatic pH (7.14 ± 0.02 (n=5) vs 7.19 ± 0.02 (n=5)), endolymphatic pH (7.42 ± 0.02 (n=5) vs 7.40 ± 0.06 (n=5)), perilymphatic [Ca²⁺] (1.62 ± 0.10 mM (n=3), vs 1.51 ± 0.28 mM (n=4)), endolymphatic [Ca²⁺] (266 ± 51 μM (n=3) vs 230 ± 28 μM (n=4)). The data and the SE were pooled between the two genotypes and taken as “control” reference values against which the knockout mice were evaluated.

There was a small, but statistically significant, decrease of about 3 mV in the UP of slc26a4^-/- mice compared to control mice (Fig. 1). No significant differences between the control and knockout mice were observed in perilymphatic pH and [Ca²⁺] (Fig. 1). The pH was higher in endolymph of control mice than in the perilymph, consistent with secretion of ${\mathrm{HCO}}_3^{-}$ by pendrin. Endolymphatic pH was significantly decreased in Slc26a4^-/- mice, compared with control mice (Fig. 2). Further, endolymphatic [Ca²⁺] was less than in perilymph of control mice and was significantly increased in Slc26a4^-/- mice to a level higher than in perilymph (Fig. 2). These abnormalities of endolymphatic potential, pH and [Ca²⁺] in Slc26a4^-/- mice are compatible with their vestibular dysfunction, exhibited as circling behavior and head-tilting (6).

Figure 1.

Utricular endolymphatic potential (UP), perilymphatic pH and [Ca²⁺]. (Top panel) UP of pooled (Control) wild-type (Slc26a4^+/+; n=8) and heterozygous (Slc26a4^+/-; n=9) and of knockout (Slc26a4^-/-) mice. *, P < 0.05 . (Middle panel) perilymphatic pH of control and knockout mice; no significant difference (P > 0.05). (Bottom panel) perilymphatic [Ca²⁺] of control and knockout mice; no significant difference (P > 0.05). Number of measurements at each bar.

The high endolymphatic [Ca²⁺] in Slc26a4^-/- mice would be consistent with altered otoconia, as reported earlier (41). Indeed we consistently observed an absence of the normal otoconia but observed a single giant crystal (Fig. 3), assumed to be CaCO₃, which is the compound of calcium observed in otoconia.

Figure 3.

Otoconia from Control (Slc26a4^+/-) and knockout (Slc26a4^-/-) mice. A: Normal otoconia situated in Utricle from Slc26a4^+/- mouse. B: Normal otoconia isolated from Slc26a4^+/- utricle at higher magnification. C: Utricle from Slc26a4^-/- mouse with giant crystal (arrow). D: Giant crystal isolated from Slc26a4^-/- utricle at higher magnification.

Increased Ca²⁺ absorption by 1,25(OH)₂D₃

We tested whether Ca²⁺ absorption is stimulated by exposure to vitamin D since transcript expression of two genes involved in epithelial Ca²⁺ absorption, TRPV5 and calbindin, is upregulated (44). Stimulation of transport would therefore serve as part of the fingerprint for the involvement of this Ca²⁺ transport system in vestibular function.

Apical-to-cell ⁴⁵Ca²⁺ uptake and basolateral-to-cell ⁴⁵Ca²⁺ uptake in primary SCCD cells were measured separately. Net uptake was obtained by subtraction of the mean basolateral-to-cell ⁴⁵Ca²⁺ uptake from the mean apical-to-cell ⁴⁵Ca²⁺ uptake. Without any added inhibitors of TRPV5/6, cultured cells exhibited a net ⁴⁵Ca²⁺ uptake consistent with an absorptive flux from the apical to the basolateral side (Fig. 4). This direction corresponds to endolymph (normally, low [Ca²⁺]) to perilymph (normally, high [Ca²⁺]). Incubation with 1,25(OH)₂D₃ for 24 h significantly increased net Ca²⁺ uptake by 71% (Fig. 4).

Figure 4.

Upregulation of ⁴⁵Ca²⁺ uptake by 1,25(OH)₂D₃ across SCCD cells grown on permeable supports. (left) apical-to-cell ⁴⁵Ca²⁺ uptake (Apical) and basolateral-to-cell ⁴⁵Ca²⁺ uptake (Basolateral) of cells incubated with and without 1,25(OH)₂D₃ (Vit D, 100 nM, 24 hr). (right) net ⁴⁵Ca²⁺ uptake was calculated by subtraction of the mean basolateral-to-cell ⁴⁵Ca²⁺ uptake from the mean apical-to-basolateral ⁴⁵Ca²⁺ uptake; SEM was pooled. Net ⁴⁵Ca²⁺ uptake of cells with 1,25(OH)₂D₃ was regarded as 100% (n = 4). Uptakes are expressed as count per minute (cpm). * P < 0.05.

Regulation by pH

The pH of the apical solution was varied over a wide range to test the dependence of Ca²⁺ uptake on apical pH of the SCCD cells. Decreases of the apical pH from 8.2 to 7.4 and 5.5 each significantly reduced net ⁴⁵Ca²⁺ uptake by 37% and 96%, respectively (Fig. 5).

Figure 5.

Net ⁴⁵Ca²⁺ uptake under the different apical pH. Measured after a 15 minute incubation with the pH of 5.5, 7.4, 8.2. All cells were incubated with 1,25(OH)₂D₃ (100 nM, 24 hr) before the measurement (n = 6). Uptakes are expressed as count per minute (cpm).

Inhibition of Ca²⁺ uptake by inhibitors

Adding a mixture of agents known to block TRPV5 and TRPV6 Ca²⁺ channels (in μM: 10 ruthenium red, 100 LaCl₃, 100 GdCl₃) 15 min after 1,25(OH)₂D₃ treatment stopped the net uptake completely (Fig. 6-A). Both apical-to-cell and basolateral-to-cell uptakes were significantly reduced (Fig. 6-A). Since the mixture of Ca²⁺ inhibitors completely blocked Ca²⁺ net uptake, individual blockers were examined. 10 μM ruthenium red significantly decreased net Ca²⁺ uptake by 31% (Fig. 6-B). 100 μM LaCl₃ and 100 μM GdCl₃ showed stronger effects than ruthenium red (Fig. 6-C).

Figure 6.

Inhibition of ⁴⁵Ca²⁺ uptake into SCCD cells by TRPV5/6 inhibitors SCCD cells were incubated with 1,25(OH)₂D₃ (100 nM, 24 hr) before the measurement. A: (left) apical-to-cell ⁴⁵Ca²⁺ uptake (Apical) and basolateral-to-cell ⁴⁵Ca²⁺ uptake (Basolat) of cells with and without the mixture of Ca²⁺ inhibitors (in μM: 10 ruthenium red, 100 GdCl₃, 100 LaCl₃). (right) net ⁴⁵Ca²⁺ uptake was calculated by subtraction of the mean basolateral-to-cell ⁴⁵Ca²⁺ uptake from the mean apical-to-basolateral ⁴⁵Ca²⁺ uptake; SEM was pooled. Net ⁴⁵Ca²⁺ uptake of cells without the mixture of Ca²⁺ inhibitors was taken as 100% (n = 6). B: Net ⁴⁵Ca²⁺ uptake of control and ruthenium red-treated cells (n = 4). ⁴⁵Ca²⁺ uptake of cells without inhibitor was taken as 100%.C: Apical-to-cell ⁴⁵Ca²⁺ uptake of control, gadolinium (GdCl₃) and lanthanum (LaCl₃) treated cells (n = 4). ⁴⁵Ca²⁺ uptake of cells without inhibitor was taken as 100%. Concentrations for individual inhibitors are same as described in A. cpm, count per minute; Ctrl, control; Ruth Red, ruthenium red. * P < 0.05 ; ns, not significant.

DISCUSSION

We report for the first time that Slc26a4^-/- mice posses a markedly lower pH and higher [Ca²⁺] in the luminal fluid, endolymph, than in wild-type and heterozygous mice. We also present the first evidence of net Ca²⁺ absorption by SCCD epithelium in the vestibular system.

SLC26A4 is expressed in several discrete areas in the inner ear, including the apical membrane of vestibular transitional cells of the utricle and ampullae of the semicircular canals (8; 40). These observations, along with the absence of pendrin transcript in the SCCD (Raveendran & Marcus, unpublished observation) suggest that pH homeostatic contributions by the Cl^-/ ${\mathrm{HCO}}_3^{-}$ exchanger pendrin occurs mainly outside of the SCCD.

In spite of the physical separation between the sites of ${\mathrm{HCO}}_3^{-}$ secretion and of calcium absorption, it is expected that there is no significant gradient of pH among the utricle, ampullae and canal ducts since the lumen has a relatively large cross-sectional area that favors good diffusion, unlike most smaller tubular epithelial structures such as the nephron. The volume of the utricle is almost as large as the SCCD (32) and convection of endolymph through the SCCD is one of the mechanisms of vestibular perception (14; 39). It is thus reasonable to assume that both the utricle and the SCCD have similar ion concentrations and pH by diffusion and convection. The TRPV5/6 channels in the SCCD would therefore be exposed on their extracellular face to the relatively acidic conditions in the Slc26a4^-/- mice.

Regulation of luminal pH by SLC26A4 could conceivably influence transport of a number of ions that are actively moved by vestibular epithelial cells. However, the deletion of SLC26A4 had no effect on luminal (endolymphatic) [K⁺] in the utricle (28), suggesting that neither the K⁺ secretory channel in the apical membrane of vestibular dark cells nor K⁺ exit pathways are strongly sensitive to luminal pH. However, the negative shift of the UP in Slc26a4^-/- mice is consistent with an increase in apical membrane conductance of highly polarized epithelial cells, such as the hair cells (17), as opposed to low-voltage K⁺ -secretory dark cells (41). Increased apical conductance would increase the fractional contribution of the negative intracellular potential to the observed transepithelial potential, the UP (17). A candidate for the apical acid-activated conductance is the ASIC1b channel identified in the apical membrane of vestibular hair cells (35), although this channel in an oocyte expression system was only transiently activated by acid; ASIC1b inactivated during exposure to a step-change in acid with a time constant on the order of 1 sec (1).

Previously, we reported the expression of the epithelial Ca²⁺ channels TRPV5 and TRPV6 in primary cultures of SCCD epithelial cells (44). Because of the known sensitivity of the TRPV5/6 channels to external pH (46), we hypothesized that low pH of vestibular endolymph due to absence of the Cl^-/ ${\mathrm{HCO}}_3^{-}$ exchanger pendrin inhibits TRPV5 and TRPV6, and consequently leads to higher [Ca²⁺] in endolymph.

The normal [Ca²⁺] of vestibular endolymph (ca. 250 μM; (42)) is known to be necessary for optimal hair cell function (20) and is likely important in the utricle for proper otoconial formation, which contains precipitated CaCO₃. The otoconial crystals comprise the inertial mass that is coupled to the sensory stereocilia of the utricular hair cells and are relatively fine “stones” under normal conditions, but form giant crystals in Slc26a4^-/- mice (Fig. 3, (40)). The low [Ca²⁺] of normal endolymph points to the existence of one or more Ca²⁺ absorption pathways in the vestibular system. However, in addition to absorption it is known that there are also Ca²⁺ secretory processes in the inner ear. The plasma membrane Ca²⁺ ATPase PMCA2 is expressed in the apical membrane (stereocilia) of hair cells and mutation of PMCA2 leads to an absence of otoconia, a lowered [Ca²⁺] and to vestibular dysfunction (43).

Our results here of elevated [Ca²⁺] in the endolymph of Slc26a4^-/- mice compared to the endolymph of wild-type mice is consistent with inhibition of the Ca²⁺ absorbing function of the TRPV5/6 channels by acidic pH in those mice. Interestingly, endolymphatic [Ca²⁺] was also higher than the abluminal fluid, perilymph, in Slc26a4^-/- mice. That elevation of endolymphatic [Ca²⁺] (2.6 mM compared to 1.1 mM in perilymph) is even higher than the equilibrium concentration of 1.5 mM that could be accounted for by passive distribution of Ca²⁺ at the UP of -4.4 mV in knockout mice. These observations point to continued active secretion of Ca²⁺ in the vestibular system in the face of inhibited Ca²⁺ absorption.

TRPV5 is predominantly expressed and controlled by 1,25(OH)₂D₃ in the kidney, while TRPV6 is predominantly expressed and controlled by 1,25(OH)₂D₃ in the intestine (9; 36). Although the SCCD epithelium expresses transcripts for both TRPV5 and TRPV6, only TRPV5 is responsive to 1,25(OH)₂D₃ in SCCD primary cultures (44) and expression of TRPV6 is about 10-fold less in the native epithelium (Yamauchi, Raveendran & Marcus; unpublished observations). Our observation of increased Ca²⁺ absorption by SCCD is consistent with observations in kidney and with the increased expression of TRPV5 transcript by 1,25(OH)₂D₃ in SCCD.

Both intra- and extra-cellular acid pH inhibits TRPV5 and they each enhance the other’s effect (45). Binding of extracellular protons to glutamate-522 in the pore region of TRPV5 causes conformational changes and closure of the channel (46). TRPV6 has a histidine at the position equivalent to glutamate-522 in TRPV5, which predicts similar pH sensitivity for both TRPV5 and TRPV6 (46). The very steep dependence of activity of these channels and of SCCD Ca²⁺ uptake on pH is strong evidence for the functional expression of the TRPV5/6 channels in the apical membrane of SCCD epithelium.

In addition, TRPV5 is known to be inhibited by ruthenium red, gadolinium and lanthanum (12; 38). TRPV6 is also inhibited by ruthenium red but with less potency [IC₅₀ 0.12 μM rabbit TRPV5 vs 9 μM mouse TRPV6 (12)]. Each of the multivalent blockers have poor specificity, but the effectiveness of all of these conditions to decrease Ca²⁺ absorption in SCCD epithelium is strong support for the notion that a major part of the Ca²⁺ absorption from the semicircular canal is via the TRPV5/6 channels. Ruthenium red is an inhibitor of a wide range of Ca²⁺ channels, including voltage-gated and epithelial Ca²⁺ channels (4; 12). The net Ca²⁺ uptake in the SCCD was inhibited only 31% by 10 μM ruthenium red in spite of the 100x lower IC₅₀ mentioned above for TRPV5 and IC₅₀ for TRPV6 similar to the concentration used. This result suggests either that there is a significant contribution of another Ca²⁺-permeable channel in this epithelium or, more likely, points to a combination of species and condition differences to the cited studies.

The trivalent cations La³⁺ and Gd³⁺ are also effective blockers of TRPV5 and TRPV6 (23; 38), but are also not specific enough to discriminate TRPV5 and TRPV6 from other Ca²⁺ channels. In our study, a mixture of ruthenium red, La³⁺ and Gd³⁺ produced a complete block of net Ca²⁺ uptake.

Vectorial Ca²⁺ absorption occurs by well-characterized processes (10). Ca²⁺ enters the cell from the lumen down an electrochemical gradient through TRPV5/6 in the apical cell membrane. Cytotoxic accumulation of Ca²⁺ in the cytosol is prevented by immediate binding of Ca²⁺ to Ca²⁺-binding proteins, calbindin-9K and/or calbindin-28K, which carry the Ca²⁺ in a bound state by diffusion to the basolateral cell membrane. Ca²⁺ is then extruded from the cell by an ATP-dependent Ca²⁺-ATPase (PMCA) and a Na⁺/Ca²⁺ exchanger (NCX). In this way, influx at the apical membrane and apical to basolateral flux are correlated in a 1:1 fashion (24). The radiotracer accumulation in the cells from the basolateral side predominantly indicates the basolateral-to-cell unidirectional cycle of the NCX. As the net flux is reduced, this coupled basolateral uptake will also be reduced, as seen in Figure 6.

In conclusion, we have demonstrated an acid-sensitive vectorial Ca²⁺ absorption by the SCCD that can account for the pathological endolymph composition observed in mice with the SLC26A4 mutation. These observations increase our understanding of the etiology of Pendred syndrome.