Loss of cochlear HCO3 secretion causes deafness via endolymphatic acidification and inhibition of Ca2+ reabsorption in a Pendred syndrome mouse model

Philine Wangemann, Kazuhiro Nakaya, Tao Wu, Rajanikanth J. Maganti, Erin M. Itza, Joel D. Sanneman, Donald G. Harbidge, Sara Billings, Daniel C. Marcus

Abstract

Pendred syndrome, characterized by childhood deafness and postpuberty goiter, is caused by mutations of SLC26A4, which codes for the anion exchanger pendrin. The goal of the present study was to determine how loss of pendrin leads to hair cell degeneration and deafness. We evaluated pendrin function by ratiometric microfluorometry, hearing by auditory brain stem recordings, and expression of K+ and Ca2+ channels by confocal immunohistochemistry. Cochlear pH and Ca2+ concentrations and endocochlear potential (EP) were measured with double-barreled ion-selective microelectrodes. Pendrin in the cochlea was characterized as a formate-permeable and DIDS-sensitive anion exchanger that is likely to mediate HCO3 secretion into endolymph. Hence endolymph in Slc26a4+/− mice was more alkaline than perilymph, and the loss of pendrin in Slc26a4−/− mice led to an acidification of endolymph. The stria vascularis of Slc26a4−/− mice expressed the K+ channel Kcnj10 and generated a small endocochlear potential before the normal onset of hearing at postnatal day 12. This small potential and the expression of Kcnj10 were lost during further development, and Slc26a4−/− mice did not acquire hearing. Endolymphatic acidification may be responsible for inhibition of Ca2+ reabsorption from endolymph via the acid-sensitive epithelial Ca2+ channels Trpv5 and Trpv6. Hence the endolymphatic Ca2+ concentration was found elevated in Slc26a4−/− mice. This elevation may inhibit sensory transduction necessary for hearing and promote the degeneration of the sensory hair cells. Degeneration of the hair cells closes a window of opportunity to restore the normal development of hearing in Slc26a4−/− mice and possibly human patients suffering from Pendred syndrome.

  • pendrin
  • stria vascularis
  • Slc26a4
  • Kcnj10
  • Trpv5

pendred syndrome is an autosomal-recessive disorder caused by mutations of the SLC26A4 gene that codes for the protein pendrin (10). Childhood deafness, postpuberty goiter, and an enlarged endolymphatic duct are the hallmarks of Pendred syndrome (12, 28, 30, 32). Although deafness is generally profound, it is variable and sometimes late in onset (6, 41). Studies of Pendred syndrome have recently been facilitated by the development of an animal model (Slc26a4−/− mice) and pendrin-specific polyclonal antibodies (9, 33). Expression has been found in the inner ear and thyroid gland consistent with the clinical manifestations of deafness and goiter (10, 11, 32, 44). In addition, pendrin expression has been found in the kidney (33), mammary gland (31), uterus (39), testes (20), and placenta (2). No expression was found in fetal or adult brain, consistent with a peripheral cause of deafness (10, 39).

Pendrin belongs to the gene family SLC26A, which contains sulfate transporters, and therefore was initially thought to be a transporter for sulfate (10). Subsequent expression studies in Xenopus laevis oocytes, Sf9 insect, and HEK-293 cells have shown that pendrin functions as an exchanger that transports anions such as Cl, I, HCO3, and formate but not sulfate or oxalate (34–36). Functional studies in native tissues have only been performed in the murine renal collecting duct, where pendrin is expressed in β- and non-α-, non-β-intercalated cells (33). Isolated, perfused collecting ducts from NaHCO3-fed, deoxycorticosterone-treated Slc26a4−/− mice reabsorbed HCO3, whereas ducts from similarly treated Slc26a4+/+ mice secreted HCO3, suggesting that pendrin is involved in HCO3 secretion (33). It is conceivable that pendrin is involved in HCO3 secretion in the inner ear. Evidence for HCO3 secretion into endolymph comes from the observations that endolymphatic pH and HCO3 concentrations are relatively high (15, 16). If pendrin would be involved in HCO3 secretion, it would be expected that loss of pendrin leads to an acidification of pH. The first aim of the present study was to determine whether pendrin in the inner ear is functional and whether loss of pendrin affects endolymphatic pH.

Pendred syndrome can be associated with deafness at birth or with progressive hearing loss during childhood (6, 41). This implies that hearing develops in at least some cases but is lost during childhood. Compared with humans, the life span of mice is much compressed. Furthermore, mice are born developmentally more immature than humans. Degeneration of the organ of Corti was found in Slc26a4−/− mice to begin around the time of the normal onset of hearing. Adult mice were found deaf due to a loss of Kcnj10 protein expression in the stria vascularis and an ensuing loss of the endocochlear potential, which is the major driving force for sensory transduction (9, 44). The stria vascularis of adult mice was found to degenerate, and this degeneration was found to be associated with an invasion of macrophages (18). Taken together, these findings raised the question of whether Kcnj10 expression and endocochlear potential are first developed and then lost, which implies the existence of a window of opportunity to intervene and prevent loss of function. Alternatively, it was conceivable that Kcnj10 expression and the endocochlear potential failed to develop, which may point to a prenatal defect that would be difficult to correct during postnatal life. Thus the second aim of the present study was to determine whether Slc26a4−/− mice develop hearing, express Kcnj10 in the stria vascularis, and generate an endocochlear potential.

Endolymphatic Ca2+ concentration is unusually low for an extracellular fluid, which is critically important for sensory transduction. Reabsorption of Ca2+ from vestibular endolymph has recently been shown to involve the epithelial Ca2+ channels Trpv5 and Trpv6 (Trpv5 synonyms: ECaC1 and CaT2; Trpv6 synonyms: ECaC2, CaT1) (46). The observations that Trpv5 and Trpv6 are inhibited by an acidification of extracellular pH (29, 42) and that transepithelial Ca2+ absorption in Trpv5- and Trpv6-expressing rat semicircular canal epithelium was inhibited by extracellular acidification (24) raised the hypothesis that loss of pendrin leads to inhibition of sensory transduction via inhibition of Ca2+ absorption and an increase in endolymphatic Ca2+ concentration. Thus the third aim of the present study was to determine whether Trpv5 and Trpv6 channel proteins are expressed in the cochlea and whether loss of pendrin leads to a change in endolymphatic Ca2+ concentration that may present a key event in the etiology of deafness.

METHODS

Animal use.

Mongolian gerbils were obtained from a commercial source (Charles River, Wilmington, MA) and housed at Kansas State University (KSU). Adult Slc26a4−/− and Slc26a4+/+ mice were obtained either from the colony of Dr. Susan Wall (Emory University) or from a colony at KSU that was established with breeders kindly provided by Dr. Wall. Preweaning mice and Slc26a4+/− mice were solely obtained from the KSU colony. Gerbils were deeply anesthetized with pentobarbital sodium (100 mg/kg ip) and killed by decapitation. Preweaning mice were deeply anesthetized with 4% tribromoethanol (0.013 ml/g body wt ip). Adult mice were deeply anesthetized either with 4% tribromoethanol (0.014 ml/g body wt ip) or with pentobarbital sodium (100 mg/kg body wt ip) and killed by decapitation or transcardial perfusion. All procedures involving animals were approved by the Institutional Animal Care and Use Committee of Kansas State University.

Ratiometric pH measurements.

Temporal bones were extracted from gerbils after death, and lateral wall tissues were obtained by microdissection in Cl-free solution. Cl-free solution contained (in mM) 150 Na-gluconate, 1.6 K2HPO4, 0.4 KH2PO4, 4 Ca-gluconate2, 1 MgSO4, and 5 glucose, pH 7.4. Great care was taken to not strip stria vascularis from the underlying spiral ligament. Lateral wall tissues were loaded with BCECF by incubation with 5 μM BCECF-AM for 45 min at 37°C. Tissues were mounted with fine glass needles in a bath chamber installed on the stage of a confocal microscope (PASCAL 5, Carl Zeiss, Jena). For optimal fluorescence imaging, the epithelial side of the tissue faced toward the coverslip. Tissues were initially superfused with 150 Cl solution that contained (in mM) 150 NaCl, 3.5 KCl, 1 CaCl2, 1 MgCl2, and 5 glucose, pH 7.4. Cl steps from 150 to 15 mM were performed by replacement of 135 mM NaCl with an equimolar amount of Na-gluconate, while increasing the amounts of CaCl2 and MgCl2 from 1 to 4 mM to compensate for chelation. Formate (10 mM) and DIDS (1 mM, predissolved in DMSO) were simply added to solutions. Fluorescence originating from surface epithelial cells in the spiral prominence region was recorded without interference of underlying connective tissue or capillary networks. Contributions of connective tissue were excluded since connective tissue cells did not load BCECF under the chosen conditions. Furthermore, contributions of capillary endothelial cells that loaded BCECF were avoided by choice of an appropriate optical section. Fluorescence was imaged in an alternating manner in response to 458- and 488-nm excitation using the same detector operated at a fixed gain, offset, and amplification. The fluorescence ratio was converted to pH according to a calibration performed in droplets of BCECF acid dissolved in HEPES-buffered solutions of varying pH (Fig. 1).

Fig. 1.

Calibration of ratiometric pH measurements. The ratio of fluorescent emissions of BCECF is plotted against various pH values. Fluorescence emissions were detected in response to excitation using the 458- and 488-nm lines of an argon laser. The Kd of 7.06 ± 0.03 is close to the theoretical value of 7.0.

Electrophysiological pH and Ca2+ measurements.

The endolymphatic pH and Ca2+ concentrations and the endocochlear potential were measured in situ with double-barreled microelectrodes. Procedures were developed by modifying previously described protocols (21). Measurements were made in the basal turn of the cochlea by a round-window approach through the basilar membrane of the first turn. After electrodes were placed in the perilymph, the surgical cavity was covered with liquid Sylgard 184 (Dow Corning, Midland, MI). This maneuver was designed to prevent the measurement of artificially elevated perilymphatic pH values due to the loss of tissue CO2 into ambient air. Indeed, perilymph values in a preliminary series of experiments that lacked this precaution were significantly higher (7.67 ± 0.04, n = 8 vs. 7.33 ± 0.04, n = 11).

Calibration consisted of taking a reference value and obtaining the slope of the electrode in an agar cup in situ. This method was devised to minimize the contribution of electrode drift and differences between reference electrodes. After cochlear measurements of endolymph and perilymph, the surgical cavity was flooded with pH 7 or 1 mM Ca2+ calibration solutions and the electrode was lifted through the Sylgard layer to rapidly obtain a reference measurement. Subsequently, the slope of the electrode was obtained by placing an agar cup on the exposed neck muscles and connecting it to three inflow lines for calibration solutions and one suction line. The electrode was moved into the cup, and the tip was perfused with calibration solutions (1 min/solution). Agar cups holding ∼100 μl were prepared with weakly buffered Ringer solution. Slopes of Ca2+- and pH-sensitive electrodes were 25.0 ± 0.6 mV/decade concentration (n = 19) and 57.1 ± 0.3 mV/pH unit (n = 24), respectively.

Double-barreled glass microelectrodes were manufactured from filament-containing glass tubing (1B100F-4, World Precision Instruments, Sarasota, FL) using a micropipette puller (Narishige PD-5, Tokyo, Japan). Before silanization, microelectrodes were baked at 180°C for 2 h to ensure dryness. The longer ion-selective barrel was mounted in the lid of a beaker. The beaker was heated to 210°C and silanized by a 90-s exposure to 0.08 ml dimethyldichlorosilane (Fluka 40136) at room temperature. The shorter reference barrel was protected from silanization by sealing the open end with Parafilm (Alcan Packaging, Chicago, IL). After silanization, microelectrodes were baked at 180°C for 3 h and tips were broken to a final outer diameter of ∼3 μm. For pH electrodes, the reference barrel was filled with 1 M KCl and the ion-selective barrel was filled at the tip with liquid ion exchanger (Fluka 95297, Hydrogen ionophore II-Cocktail A) and back-filled with buffer solution (500 mM KCl, 20 mM HEPES, pH 7.34). For Ca2+ electrodes, the reference barrel was filled with 150 mM KCl and the ion-selective barrel was filled at the tip with liquid ion exchanger (Fluka 21048, Calcium ionophore I-Cocktail A) and back-filled with 500 mM CaCl2. Each barrel was connected to an input of a grounded dual electrometer (FD223, World Precision Instruments) via Ag-AgCl wire electrodes. The animal was grounded via a Ag-AgCl wire electrode inserted into the neck musculature. Current pulses (1 nA) were injected via the reference barrel to monitor electrode resistance.

pH-sensitive electrodes were calibrated using solutions of three different pH values (in mM): pH 6: 130 NaCl, 20 MES; pH 7: 130 NaCl, 20 HEPES; and pH 8: 130 NaCl, 20 tricine. Ca2+-sensitive electrodes were calibrated using solutions of three different Ca2+ concentrations (mM): 0.01 mM Ca2+: 150 KCl, 0.121 CaCl2, 10 HEPES, 1 nitrilotriacetic acid (Sigma N-0128, Ca2+ buffer), pH 7.4; 0.1 mM Ca2+: 150 KCl, 0.1 CaCl2, 10 HEPES, pH 7.4; and 1 mM Ca2+: 150 KCl, 1 CaCl2, 10 HEPES, pH 7.4.

Data were recorded analog (Flat Bed Chart recorder, Kipp & Zonen) for convenient annotation and digital for presentation and data analysis (DIGIDATA 1322A and AxoScope 9, Axon Instruments). Data were analyzed using custom software written by P. Wangemann in LabTalk (Origin 6.0, The Origin, Northampton, MA).

Measurements of the endocochlear potential were reduced by 1.3 and 4.2 mV to correct for the liquid junction potentials that occur by advancing the electrode filled with 1 M KCl or 150 mM KCl from perilymph (150 NaCl) to endolymph (150 mM KCl). Similar corrections were also made to the measurements of the endolymphatic pH and the perilymphatic Ca2+ concentration since pH and Ca2+ measurements were calibrated with NaCl- and KCl-containing solutions, respectively.

Blood electrolyte measurements.

Plasma profiles were obtained using a blood analyzer (Stat Profile M, Nova Biomedical, Waltham, MA). No significant differences were found between Slc26a4−/− and Slc26a4+/+ mice (Table 1).

View this table:
Table 1.

Analysis of blood plasma from Slc26a4+/+ and Slc26a4−/− mice

Auditory brain stem recordings.

Mice were deeply anesthetized with 4% tribromoethanol (0.013–0.014 ml/g body wt ip) and placed on a thermal pad to maintain normal body temperature. The mastoid, vertex, and ventral neck region of the animal were connected via subdermal platinum needle electrodes (F-E2, Astro-Med, West Warwick, RI) and short (31 cm) leads to the main channel, reference channel, and ground of the preamplifier, respectively. Auditory brain stem recordings were performed in a custom constructed, electrically shielded, and sound-attenuated chamber (inner dimensions 23 × 23 × 23 cm) using a digital data-acquisition system (BioSig32 software, RA4LI Preamplifier, RP2.1 Enhanced Real Time Processor, PA5 Programmable Attenuator, ED1 Electrostatic Speaker Driver, Tucker-Davis Technologies, Alachua, FL). Tone burst and click stimuli were presented (21/s) via a free-field electrostatic speaker (SigGen software, ES1 speaker, Tucker-Davis). Acoustic stimuli were calibrated using a 1/4-in. condenser microphone (SigCal IRP4.2 software, Tucker-Davis; PS9200 microphone, Acoustical Interface, Belmont, CA) that replaced the mouse's head. Clicks (1-ms duration) and tone bursts (2-ms duration, 0.5-ms gate time; 8, 16, and 32 kHz) were presented with alternating phase (0 and 180°). Responses, recorded over 10 ms, were filtered (300-Hz high pass, 3,000-Hz low pass, and 60-Hz notch), and 1,000 recordings were averaged. Click and tone burst stimuli were presented at intensities varying between 90 and 10 dB SPL in 10-dB intervals. Auditory thresholds were obtained by a visual comparison of waveforms.

Immunocytochemistry.

Mice were deeply anesthetized and killed by transcardial perfusion with Cl-free or NaCl solution (6 ml, 1 min) followed by Cl-free or NaCl solution with 4% paraformaldehyde (24 ml, 4 min). Cl-free solution contained (mM) 150 Na-gluconate, 1.6 K2HPO4, 0.4 KH2PO4, 4 Ca-gluconate2, 1 MgSO4, and 5 glucose, pH 7.4. NaCl solution contained (mM) 150 NaCl, 1.6 K2HPO4, 0.4 KH2PO4, 0.7 CaCl2, 1 MgCl2, and 5 glucose, pH 7.4. Cryosection and whole mounts of stria vascularis were prepared.

For cryosections, temporal bones were decalcified in 10% EDTA, processed through a sucrose gradient, and infiltrated with polyethylene glycol. Midmodiolar cryosections (12 μm, CM3050S, Leica, Nussloch, Germany) were blocked in PBS-TX (137 mM NaCl, 10. 1 mM Na2HPO4, 1.8 mM KH2PO4, 2.7 mM KCl, pH 7.4, with 0.2% Triton X-100) and 5% bovine serum albumin. Slides were incubated overnight at 4°C with primary antibody in PBS-TX with 1–3% BSA. Primary antibodies included rabbit anti-pendrin (1:500, h766–780, a kind gift from Dr. Ines Royaux, National Institutes of Health), goat anti-Kcnq1 (1:200, C20, Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-Kcnj10 antibody (1:300, Alomone, Jerusalem, Israel), rabbit anti-rat Trpv5 (1:100, CAT21-A, Alpha Diagnostics, San Antonio, TX), and rabbit anti-rat Trpv6 (1:100, Alpha Diagnostics). Slides were washed in PBS-TX and incubated for 1 h at room temperature with secondary antibodies at a 1:1,000 dilution in PBS-TX with 1–3% BSA. Secondary antibodies included donkey anti-rabbit Alexa 488 and chicken anti-goat Alexa 594 (Molecular Probes, Eugene, OR). After incubation, slides were washed with PBS-TX and mounted with FluorSave (Calbiochem, La Jolla, CA).

For whole mounts, the stria vascularis was isolated by microdissection and postfixed for 2 h at 4°C in Cl-free solution containing 4% paraformaldehyde. Fixed stria vascularis was washed 2× in Cl-free solution and 1× in PBS-TX and blocked with 5% BSA in PBS-TX for 45 min at room temperature and again washed 3× in PBS-TX. The tissue was then incubated overnight at 4°C with anti-Kcnj10 antibody (1:300; see above) in PBS-TX with 1–3% BSA, washed with PBS-TX, and incubated for 1 h at room temperature with donkey anti-rabbit Alexa 488 secondary antibody (1:1,000, Molecular Probes) in PBS-TX with 1–3% BSA. After antibody incubation, the stria vascularis was washed 3× in PBS-TX and mounted with FluorSave between two coverslips.

Cryosections and whole mounts were viewed by confocal microscopy (LSM 510 Meta, Carl Zeiss, Göttingen, Germany). Laser-scanning brightfield images were collected to aid orientation and document structural preservation.

Statistics.

Data are given as average ± SE. Differences were determined by paired and unpaired t-test, as appropriate. Significance was assumed at P < 0.05.

RESULTS

Pendrin protein expression in the cochlea is functional.

Pendrin has been shown to be expressed in the apical membrane of endolymph-facing, spiral prominence epithelial cells (44). When expressed in heterologous expression systems, pendrin has been shown to mediate formate-enhanced and DIDS-sensitive cytosolic alkalinizations in response to reductions in the extracellular Cl concentration (34, 36). Similar protocols were used here to evaluate whether pendrin expressed in the cochlea is functional. Spiral prominence regions of the gerbil cochlea were isolated, and surface epithelial cells were loaded with BCECF (Fig. 2). Reductions in extracellular Cl concentration from 150 to 15 mM in the absence of formate resulted in a cytosolic alkalinization by 0.13 ± 0.02 pH units (29 cells in 8 preparations). Addition of 1 or 10 mM formate resulted in an acidification by 0.03 ± 0.01 (9 cells in 2 preparations) and 0.12 ± 0.02 pH units (14 cells in 4 preparations), respectively. Addition of 1 mM DIDS had no significant effect (−0.03 ± 0.02 pH units, 13 cells in 2 preparations). Paired experiments revealed that Cl-induced alkalinizations were enhanced by 10 mM formate and inhibited by 1 mM DIDS. These observations are consistent with the presence of functional pendrin protein in the cochlea.

Fig. 2.

Ratiometric pH measurements in pendrin-expressing spiral prominence epithelial cells from a gerbil. A: location and microdissection of spiral prominence (SP), outer sulcus (OS), and stria vascularis (SV). RM, Reissner's membrane. B and C: expression of Kcnj10, Kcnq1, and Pendrin in the SP region of the lateral wall. Confocal immunohistochemistry of cryosections is shown. Kcnj10 was expressed in intermediate cells of the stria vascularis, and Kcnq1 in the apical membrane marginal cells of the SV and pendrin in the apical membrane of SP epithelial cells. Pendrin-expressing cells were clearly distinct from marginal and intermediate cells of the SV. D: SP epithelial cells were loaded with BCECF. OS cells suffered from damage received during microdissection, and marginal cells of the SV did not load BCECF under the conditions chosen. E: cellular pH measurements. Cl concentration steps from 150 to 15 mM (Cl15) caused an alkalinization that was enhanced in the presence of 10 mM formate. F: data summary. Cl step-induced cellular alkalinization was enhanced by 10 mM formate and inhibited by 1 mM DIDS.

Mice lacking pendrin do not develop hearing.

Inner and outer hair cells in the organ of Corti have been shown to develop normally in Slc26a4−/− mice but to begin to degenerate between postnatal day 7 (P7) and P15, which encompasses the onset of hearing at P12 (9). These observations raised the hypothesis that Slc26a4−/− mice may develop hearing for at least a brief period. Hearing was evaluated by auditory brain stem response thresholds in Slc26a4+/− and Slc26a4−/− mice. Slc26a4+/− mice began hearing at P12, with the threshold improving daily (Fig. 3). In contrast, Slc26a4−/− lacked hearing at all ages tested. The equivalency of development was evaluated by observing the time of eye opening. No difference in eye opening was found among Slc26a4+/+, Slc26a4+/−, and Slc26a4−/− mice (11.3 ± 0.2, n = 7; 11.5 ± 0.1, n = 41; and 11.5 ± 0.1 postnatal days, n = 26, respectively).

Fig. 3.

Evaluation of hearing in Slc26a4+/− and Slc26a4−/− mice. Hearing thresholds were determined by auditory brain stem recordings using click and tone-burst stimuli at 8, 16, and 32 kHz. A: example of auditory brain stem recordings made in a postnatal day 15 (P15) Slc26a4+/− mouse using 16-kHz tone-burst stimuli at different amplitudes. B: data summary. The onset of hearing occurred in Slc26a4+/− mice at P12. Hearing in Slc26a4+/− mice improved daily, reaching nearly adult levels by P15. Slc26a4−/− mice did not develop hearing.

Lack of pendrin leads to loss of Kcnj10 and endocochlear potential.

We have shown previously that adult Slc26a4−/− mice lack protein expression of the K+ channel Kcnj10 in the stria vascularis and consequently do not generate an endocochlear potential (44). The finding raised the question of whether Slc26a4−/− mice never express Kcnj10 in the stria vascularis or whether these mice first express Kcnj10 and then lose expression. This question was addressed by measuring the endocochlear potential and Kcnj10 expression in stria vascularis before and after the onset of hearing. Slc26a4−/− mice developed a small endocochlear potential at P10 that was progressively lost during further development (Fig. 4). Consistent with this observation, Kcnj10 was expressed at P10 and was progressively lost during further development (Fig. 5).

Fig. 4.

Measurements of endocochlear potential (EP) in Slc26a4+/− and Slc26a4−/− mice before and after the onset of hearing. At P10, Slc26a4+/− and Slc26a4−/− mice generated an endocochlear potential. This potential grew with further development in Slc26a4+/− mice but was lost in Slc26a4−/− mice. *P < 0.05.

Fig. 5.

Development of expression of K+ channel Kcnj10 in SV of Slc26a4+/− and Slc26a4−/− mice. Confocal immunohistochemistry of whole mounts of SV is shown. Top: images from Slc26a4+/− mice of various ages. Bottom: images from Slc26a4−/− mice. The scale bar shown in the bottom right image represents 10 μm and pertains to all images in this figure.

Lack of pendrin causes acidification of endolymphatic pH.

The observations that pendrin is expressed in the apical membrane of spiral prominence epithelial cells and that pendrin is an anion exchanger that accepts pH equivalents raised the hypothesis that pendrin controls endolymphatic pH. pH was measured with double-barreled ion-selective electrodes in endolymph and perilymph of Slc26a4+/− and Slc26a4−/− mice. Measurements were made before and after the onset of hearing. At all ages, endolymph of Slc26a4+/− mice was more alkaline than perilymph. In contrast, endolymph of Slc26a4−/− mice was more acidic than perilymph (Fig. 6). No difference was found in the pH of perilymph or blood between Slc26a4+/+ and Slc26a4−/− mice (Fig. 6, Table 1).

Fig. 6.

Measurements of the endolymphatic and perilymphatic pH and Ca2+ concentration in Slc26a4+/− and Slc26a4−/− mice before and after the onset of hearing. At all ages, the endolymphatic pH compared with the perilymphatic pH was more alkaline in Slc26a4+/− mice and more acidic in Slc26a4−/− mice. With development, the endolymphatic Ca2+ concentration decreased in Slc26a4+/− and increased in Slc26a4−/− mice. Significant differences between endolymph and perilymph are marked (H, higher in endolymph; L, lower in endolymph). *Significant differences between Slc26a4+/− and Slc26a4−/− (*P < 0.05). Insignificant differences are marked (ns). Numbers next to error bars represent the number of animals.

Cochlear epithelial cells express Ca2+ channels Trpv5 and Trpv6.

Epithelial Ca2+ channels Trpv5 and Trpv6 have recently been found to contribute to Ca2+ reabsorption from vestibular endolymph (24, 46). This finding raised the question of whether these channels are also expressed in cochlear epithelial cells and contribute to Ca2+ absorption in the cochlea. Expression of Trpv5 was mainly found in marginal cells of stria vascularis, and expression of Trpv6 was found mainly in inner and outer sulcus epithelial cells (Fig. 7). Staining for Trpv5 and Trpv6 was similar in Slc26a4+/− and Slc26a4−/− mice.

Fig. 7.

Expression of the Ca2+ channels Trpv5 (AC) and Trpv6 (D and E) in the cochlea of Slc26a4+/− and Slc26a4−/− mice at P16. Confocal immunohistochemistry of cryosections is shown. Trpv5 was mainly expressed in SV and less in inner sulcus (IS) and OS epithelial cells. Trpv6 was expressed mainly expressed in IS and OS epithelial cells and less in SV. Expression found in Slc26a4+/− mice was also found in Slc26a4−/− mice. Phalloidin-based staining of F-actin was used to provide orientation within the sections.

Lack of pendrin causes elevation of endolymphatic Ca2+ concentration.

Ca2+ concentration was measured with double-barreled ion-selective electrodes in endolymph and perilymph of Slc26a4+/− and Slc26a4−/− mice before and after the onset of hearing. At P10, Ca2+ concentration in endolymph of Slc26a4+/− mice was lower than Ca2+ concentration in perilymph. During further development, endolymphatic Ca2+ concentration progressively decreased to adult levels (Fig. 6). In contrast, endolymphatic Ca2+ concentration in Slc26a4−/− mice at P10 was similar to Ca2+ concentration in perilymph. During further development, endolymphatic Ca2+ concentration progressively increased. No difference was found in Ca2+ concentration of perilymph or blood between Slc26a4+/+ or Slc26a4−/− mice (Fig. 6, Table 1).

DISCUSSION

The most salient findings of the present study are as follows. 1) Pendrin in spiral prominence epithelial cells of the cochlea is a functional anion exchanger. 2) At P10, before the onset of hearing, Slc26a4−/− mice express Kcnj10 in the stria vascularis and generate a small endocochlear potential. During further development, Kcnj10 expression and endocochlear potential are lost and Slc26a4−/− mice fail to develop hearing. 3) Lack of pendrin leads to an acidification and an increase in the Ca2+ concentration of endolymph. 4) Epithelial cells enclosing endolymph express the acid-sensitive Ca2+ channels Trpv5 and Trpv6 in the apical membrane.

Pendrin mediates HCO3 secretion into endolymph.

The observation that pendrin is a functional anion exchanger and that loss of pendrin leads to an acidification of endolymph suggests that the main anions transported are alkaline equivalents such as HCO3. HCO3 is a likely substrate for pendrin since the stria vascularis generates CO2 and the spiral prominence is heavily expressing carbonic anhydrase, which converts CO2 to HCO3 (14, 26). The stria vascularis is a source of CO2 because it has a higher metabolic rate than neighboring tissues and because it has a respiratory quotient of 1.2, which means that it generates 1.2 CO2 for every O2 molecule consumed (22). In addition, spiral ligament and spiral prominence fibrocytes express Slc4a7, which is a Na+/HCO3 cotransporter likely involved in the uptake of HCO3 into cells (3). Consistent with pendrin-mediated HCO3 secretion is the observation that the HCO3 concentration in endolymph is higher than in perilymph and that inhibition of carbonic anhydrase leads to an acidification of endolymph and a reduction in HCO3 (16, 38). Furthermore, acoustic stimulation, which increases metabolism and CO2 production, has been shown to cause an alkalization of endolymph (15).

Coincidentally, pendrin-mediated HCO3 secretion has recently been demonstrated in the cortical collecting duct of the kidney (33). A systemic effect as origin of the acidified endolymphatic pH, however, is unlikely since no difference in plasma or perilymphatic pH was observed.

In the absence of pendrin, endolymph was more acidic than perilymph. Acid secretion into endolymph, which was uncovered in the absence of pendrin, may occur in several endolymph-facing epithelial cells, including cochlear interdental cells, known to express the subunits E and B1 of the vH+-ATPases in the apical membrane and the Cl/HCO3 exchanger Slc4a2 (AE2) in the basolateral membrane (19, 37). Additional acid secretion sites may include strial marginal cells, which express the E subunit of the vH+-ATPases in their apical membrane (37).

The possibility that the endocochlear potential would provide a driving force for pH equivalents and that loss of the endocochlear potential in Slc26a4−/− mice would be the sole cause for endolymphatic acidification is unlikely since a similar acidification of endolymph was found in the vestibular labyrinth, although there is no difference in the transepithelial voltage between Slc26a4+/− and Slc26a4−/− mice (24). HCO3 secretion into vestibular endolymph is likely mediated by pendrin expressed in vestibular transitional cells (44).

Early loss of Kcnj10 prevents the development of hearing.

The endocochlear potential, which is generated by the K+ channel Kcnj10, is crucial for hearing since it drives sensory transduction (23, 43). Adult Slc26a4−/− mice lack Kcnj10 protein expression and fail to generate an endocochlear potential, which is consistent with the fact that adult Slc26a4−/− mice are deaf (9, 44).

In general, mice are born blind and deaf. Eyes and ear canals open at P11, and hearing begins at P12. Expression of the K+ channel Kcnj10 in the stria vascularis begins in mice around P8, and the onset of expression is paralleled by the onset of an endocochlear potential (13). At P10, Slc26a4−/− mice expressed Kcnj10 in the stria vascularis and generated a small endocochlear potential. This finding demonstrates that the stria vascularis of Slc26a4−/− mice can express Kcnj10 protein and can generate an endocochlear potential.

A window of opportunity may exist to prevent loss of Kcnj10 expression and to enable Slc26a4−/− mice to develop a normal endocochlear potential and normal hearing. This narrow window of opportunity appears to close with the degeneration of the organ of Corti that begins between P7 and P15 (9). Pendrin and Kcnj10 are expressed in different cells. The link between loss of pendrin function and loss of Kcnj10 expression remains unclear.

Endolymphatic acidification inhibits Ca2+ absorption.

Under physiological conditions, endolymphatic Ca2+ concentration is 20–30 μM, which is very low for an extracellular compartment (4). Higher or lower concentrations have been shown to suppress transduction currents and microphonic potentials (25, 40). The endolymphatic Ca2+ concentration is likely controlled by secretory and absorptive processes. Ca2+ secretion may involve basolaterally expressed Ca2+ influx mechanisms in conjunction with apically expressed Ca2+-ATPases such as PMCA2 (45). Support for active Ca2+ secretion into endolymph comes from the finding that pharmacological inhibition of Ca2+-ATPases leads to a fall in endolymphatic Ca2+ concentration (17) and that deaf-waddler mice, which bear a loss-of-function mutation in the apically expressed Ca2+-ATPase PMCA2, have a very low endolymphatic Ca2+ concentration (45).

Reabsorption of Ca2+ may entail apically expressed Ca2+ channels Trpv5 and Trpv6 in conjunction with basolaterally expressed Ca2+-ATPases and Na+/Ca2+ exchangers (27, 45, 46). The finding that Trpv5 is expressed in the apical membrane of stria marginal cells in conjunction with the finding of the Ca2+-ATPase PMCA1 in their basolateral membrane (45) suggests that the stria vascularis is involved in Ca2+ reabsorption. Expression of Trpv5 and Trpv6 is of great interest in view that these channels are inhibited by extracellular acidification (29, 42). Endolymphatic acidification may inhibit Ca2+ absorption and lead to the observed elevation in endolymphatic Ca2+ concentration. The majority of Ca2+ that enters the hair cell bundle through the transduction channel under physiological conditions is extruded by Ca2+-ATPases that are expressed in the hair cell bundle (47). Although Ca2+ is necessary to maintain the sensitivity of the bundle through adaptation and motility (1, 5, 8), an entry of excess Ca2+ into the sensory cells leads to Ca2+ overloading and cell death. The capacity of cochlear hair cells to extrude Ca2+ may be limited, especially in outer hair cells that appear to not express plasma membrane Ca2+-ATPases in their basolateral membrane (7). Ca2+ overload may be the cause of cellular degeneration that was observed to begin with outer hair cells between P7 and P15 (9).

In conclusion, our data demonstrate that pendrin is a functional formate-permeable and DIDS-sensitive anion exchanger that likely mediates HCO3 secretion into endolymph. Hence endolymph is alkaline, and loss of the pendrin leads to acidification. Endolymphatic acidification may be responsible for inhibition of Ca2+ reabsorption via the acid-sensitive Ca2+ channels Trpv5 and Trpv6. Failure to lower endolymphatic Ca2+ may inhibit sensory transduction necessary for hearing and promote the degeneration of the sensory cells. Degeneration of the sensory cells closes a window of opportunity to restore the normal development of hearing in Slc26a4−/− mice and possibly human patients suffering from Pendred syndrome.

GRANTS

The support by National Institutes of Health (NIH) research Grants R01-DC-01098 to P. Wangemann and R01-DC-00212 to D. C. Marcus and of the confocal and the molecular biology core facilities through KSU-COBRE NIH Grant P20-RR-017686 is gratefully acknowledged.

Acknowledgments

The authors thank Dr. Alexander Gow (Wayne State University), Dr. Alec Salt (Washington University), and Andrew Hoyard (Tucker Davis Technologies, Alachua, FL) for valuable help in setting up auditory brain stem recordings at Kansas State University. This work would not have been possible without the dedication of Susan Rose, James Dille, Britt Neely, and Dr. Bart Carter from the Animal Resource Department at the Kansas State University College of Veterinary Medicine.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract