Endolymphatic Na+ and K+ Concentrations during Cochlear Growth and Enlargement in Mice Lacking Slc26a4/pendrin

Slc26a4 Δ/Δ mice are deaf, develop an enlarged membranous labyrinth, and thereby largely resemble the human phenotype where mutations of SLC26A4 cause an enlarged vestibular aqueduct and sensorineural hearing loss. The enlargement is likely caused by abnormal ion and fluid transport during the time of embryonic development, however, neither the mechanisms of ion transport nor the ionic composition of the luminal fluid during this time of development are known. Here we determine the ionic composition of inner ear fluids at the time at which the enlargement develops and the onset of expression of selected ion transporters. Concentrations of Na+ and K+ were measured with double-barreled ion-selective electrodes in the cochlea and the endolymphatic sac of Slc26a4 Δ/+, which develop normal hearing, and of Slc26a4 Δ/Δ mice, which fail to develop hearing. The expression of specific ion transporters was examined by quantitative RT-PCR and immunohistochemistry. High Na+ (∼141 mM) and low K+ concentrations (∼11 mM) were found at embryonic day (E) 16.5 in cochlear endolymph of Slc26a4 Δ/+ and Slc26a4 Δ/Δ mice. Shortly before birth the K+ concentration began to rise. Immediately after birth (postnatal day 0), the Na+ and K+ concentrations in cochlear endolymph were each ∼80 mM. In Slc26a4 Δ/Δ mice, the rise in the K+ concentration occurred with a ∼3 day delay. K+ concentrations were also found to be low (∼15 mM) in the embryonic endolymphatic sac. The onset of expression of the K+ channel KCNQ1 and the Na+/2Cl−/K+ cotransporter SLC12A2 occurred in the cochlea at E19.5 in Slc26a4 Δ/+ and Slc26a4 Δ/Δ mice. These data demonstrate that endolymph, at the time at which the enlargement develops, is a Na+-rich fluid, which transitions into a K+-rich fluid before birth. The data suggest that the endolymphatic enlargement caused by a loss of Slc26a4 is a consequence of disrupted Na+ transport.


Introduction
Enlargement of vestibular aqueduct (EVA) is a common inner ear malformation found in children with non-syndromic sensorineural hearing loss and hearing loss associated with Pendred syndrome [1]. Mutations of SLC26A4 are a prevalent cause of EVA and of progressive and often fluctuating hearing loss with an onset before or around speech acquisition [2,3,4,5,6]. SLC26A4 codes for pendrin, an anion exchanger that functions as a Cl 2 / HCO 3 2 exchanger in the inner ear [7]. In some populations, mutations in SLC26A4 occur in 13-14% of deaf subjects [8,9].
Slc26a4 D/D mice closely resemble the human phenotype, since Slc26a4 D/D mice acquire an enlargement of the vestibular aqueduct during embryonic development. The vestibular aqueduct is a bony tunnel that forms around the epithelia-lined endolymphatic duct. Swelling of the fluid-filled epithelium of the inner ear in Slc26a4 D/D mice leads to an enlargement of the vestibular aqueduct. Slc26a4 D/D mice fail to acquire hearing during the postnatal phase of development [7,10]. Studies on Slc26a4 D/D mice suggest that the endolymphatic enlargement entails a ,10-fold larger luminal volume in the cochlea, that the endolymphatic acidification entails a ,2-fold higher H + ion concentration, and that the enlargement and acidification are key events that mark the onset of the pathology [11,12].
The development of the murine inner ear begins with an ectodermal invagination that forms an otocyst at embryonic day (E) 9.5. The otocyst is filled with amniotic fluid, which is a NaClrich fluid [13]. The otocyst extends two protrusions: one to form scala media of the cochlea and the other to form the endolymphatic duct and sac, while the middle of the otocyst is reorganized into the vestibular labyrinth [14]. The lumen of the endolymphatic sac opens at E10.5 and the lumen of the cochlea begins to open at E14.5. The fluid inside the lumen is called endolymph and the fluid surrounding the epithelium is called perilymph. In Slc26a4 D/D mice, the enlargement of the endolymphatic lumen begins at E14.5, which coincides with cochlear growth and lumen formation between E14.5 and E18.5 [1,15].
The mesenchymal tissue surrounding the cochlea, the vestibular labyrinth and the endolymphatic duct forms cartilage during embryonic development that is later converted into bone (Fig. 1A). In addition, the cochlea and the vestibular labyrinth are enclosed in the temporal bone. Access to scala media of the cochlea is available via the round window that is located near the base of the cochlea. The endolymphatic sac is the only part of the inner ear that is not enclosed in cartilage or bone.
The mechanisms that lead to cochlear enlargement in Slc26a4 D/D mice are not known. It is conceivable that fluid accumulation leading to the enlargement occurs in the wake of an abnormally low rate of ion absorption or an abnormally high rate of ion secretion. Ion transport in the inner ear has so far mainly been studied in adult animals, when endolymph consists of a K + -rich fluid [16]. The enlargement of the endolymphatic lumen in Slc26a4 D/D mice, however, occurs during embryonic development. Thus, it is of great interest to determine whether endolymph in Slc26a4 D/D and Slc26a4 D/+ mice at the time of enlargement is a K + -rich fluid, as observed in the adult inner ear, or whether endolymph at this time is a Na + -rich fluid similar in composition to amniotic fluid. The determination of the ion composition during the phase of enlargement is the very first step for understanding how the enlargement occurs. Thus the main goal of the present study was to determine the Na + and K + concentrations in cochlear endolymph during embryonic development.

Ethics Statement
All procedures involving animals were approved by the Institutional Animal Care and Use Committee at Kansas State University (IACUC#: 2961, 2925 and 3245).

Animals
A colony of Slc26a4 D/+ and Slc26a4 D/D mice was maintained at Kansas State University. Slc26a4 D/+ dams and Slc26a4 D/D sires were housed in monogamous pairs. Litter sizes averaged 4.3 pups with Slc26a4 D/+ and Slc26a4 D/D offspring in the near Mendelian ratio of 50.1 to 49.9. The average gestational period was 21 days. Gestational age was counted from the day when a vaginal plug was detected. This day was set to embryonic (E) day 0.5 (E0.5). Pregnancies were verified by ultrasound (Terason t3000, Universal bone-enclosed inner ear consisting of the cochlea, the vestibular labyrinth and the endolymphatic duct and sac. Note, that scala media of the cochlea is accessible via the round window and that the endolymphatic sac is the only structure not enclosed in bone. B-E) Experimental configurations for the measurement of voltages and of Na+ and K+ concentrations. Voltage electrodes were used to record the voltage (V) and ion-selective electrodes, which were filled at the tip with liquid ion exchanger, recorded a signal that consisted of the sum of the voltage (V) and the ion-voltage (VK+ or VNa+). B) Configuration for in situ measurements of voltage and K+ concentrations in the cochlea. C-E) Configuration for in vitro measurements. Temporal bones were isolated and superfused in a bath chamber that was outfitted with an inflow and a suction outlet. C) Configuration for in vitro measurements of voltage and K+ concentrations in the cochlea. D) Configuration for in vitro measurements of voltage and K+ concentrations in the endolymphatic sac. E) Configuration for in vitro measurements of Na+ and K+ concentrations in the cochlea. doi:10.1371/journal.pone.0065977.g001 Medical Systems, Bedford Hills, NY). The colony was maintained free of known and suspected murine pathogens, as reported earlier [15]. Slc26a4 D/+ and Slc26a4 D/D mice ranging in age from E14.5 to P72 were used in the present study. Time-pregnant dams were deeply anesthetized with 4% tri-bromo-ethanol and sacrificed by decapitation after harvesting embryos by sterile laparotomy. Embryos were sacrificed by decapitation. Neonatal mice (P0) were anesthetized by a combination of intraperitoneal (i.p.) injection of 0.014 ml/g body weight of 4% tri-bromo-ethanol and rapid cooling on a slush of ice. Older mice (.P3) were anesthetized solely by i.p. injection of 0.014 ml/g body weight of 4% tri-bromo-ethanol. Neonatal and older mice were sacrificed by decapitation under deep anesthesia.
Measurement of voltage, Na + and K + concentrations Voltage, Na + and K + concentrations were measured with double-barreled ion-selective electrodes. Measurements in prenatal and neonatal mice were carried out using a novel in vitro approach. This novel approach was developed since the wellestablished in situ approach is not suitable for measurements in prenatal mice. Limitations of the in vitro approach were established by comparing results from postnatal mice obtained in vitro to data obtained in situ.
For in situ measurements of K + concentrations, one barrel of the double-barreled electrode was used to measure the K + concentration and the other barrel was used to measure the transepithelial voltage (Fig. 1B). Electrodes were manufactured using established protocols [17]. Briefly, the K + selective barrel was silanized prior to being filled at the tip with liquid ion exchanger (Ionophore I -cocktail B, Cat# 60398, Sigma-Aldrich) and backfilled with 1 M KCl solution. The voltage barrel was filled with 500 mM NaCl solution. Each barrel was connected via a Ag/ AgCl electrode to a dual electrometer with high input-resistance (FD223, World Precision Instruments, Sarasota, FL).
Mice between ages P62 and P72 were deeply anesthetized and maintained at 37uC body temperature using a heated platform. Surgery consisted of a tracheotomy to maintain unobstructed ventilation, the subdermal installation of a flowing 1 M KCl reference electrode in the chest region, and of a ventral approach to the temporal bone, using established methods [17]. Briefly, a small opening in the temporal bone was made and measurements were obtained by guiding the electrode through the round window into the perilymph-filled scala tympani and then into the endolymph-filled scala media. Anoxia was induced by i.p. injection of succinylcholine chloride (1 mg/g body weight) after obtaining stable readings of the K + concentration and the endocochlear potential.
For in vitro measurements of K + and Na + concentrations, temporal bones were isolated from pre-and postnatal mice and placed into a bath chamber superfused with warm (37uC) artificial perilymph (in-line heater SHM-8 and controller TC-344B, Warner instruments, Hamden, CT). Artificial perilymph contained (in mM): 135 NaCl, 25 NaHCO 3 , 4 KCl, 1.5 CaCl 2 , 1 MgCl 2 and 5 glucose and was bubbled with a humidified gas mixture consisting of 5% CO 2 and 95% O 2 to set the pH to pH 7.3 -pH 7.4. The bath chamber was grounded via a Ag/ AgCl electrode that was bathed in a vial filled with 1 M KCl solution and connected to the bath chamber via an agar-bridge made from 150 mM NaCl solution ( Fig. 1C-E). For measurements in the cochlea, an approach via the round-window was used. For in vitro measurements of K + concentrations, one barrel of the doublebarreled electrode was used to measure the K + concentration and the other barrel was used to measure the transepithelial voltage (Fig. 1C). For in vitro measurement of Na + concentrations, one barrel of the double-barreled electrode was used to measure the Na + concentration and the other barrel was used to measure the K + concentration as an indicator for successful penetration into the endolymph-filled scala media (Fig. 1E). This unusual configuration was found necessary since at the ages of interest the transepithelial voltage was found to be near zero and the Na + concentration in endolymph was found to be similar to the concentration in perilymph such that neither the recording of the voltage nor the recording of the Na + concentration could serve as an indicator for successful penetration into the endolymph-filled scala media.
Calibration of ion selective electrodes in warm (37uC) solutions was performed immediately after withdrawal of the electrodes from the temporal bone. All measurements were corrected for calculated liquid junction potentials (JPCalc in Clampex, Molecular Devices, Silicon Valley, CA [19]).
Data were recorded via a flat-bed chart recorder (BD12E, Kipp & Zonen, The Netherlands) for easy annotation and also were recorded digitally (Digidata 1322A and AxoScope 9, Molecular Devices) for easy analysis using custom software written by P.W. (Origin 6.0, The Origin Company, Northhampton, MA).

Quantitative RT-PCR
Total RNA was isolated from the cochlea of Slc26a4 D/+ and Slc26a4 D/D mice between E14.5 and P8 (RNeasy micro, Qiagen, Valencia, CA). Quantity and quality of total RNA were evaluated by microfluidic electrophoresis (BioAnalyzer, Agilent, Santa Clara, CA) and by microliter absorption photometry (Nanodrop, Wilmington, DE). RNA samples were accepted only when they were free of contamination and when RIN was .7.0, which indicates excellent quality. RIN is a quality indicator on a scale from 0 (worst) to 10 (best) that is computed from microfluidic electrophoresis runs (BioAnalyzer, Agilent).
Total RNA, primers, enzymes and buffers necessary for quantitative RT-PCR reactions were assembled with the assistance of an automatic pipetting station (Biomek NX P , Beckman Coulter, Fullerton, CA) with hardware modifications and software programming by P.W. Reactions were carried out in 96-well plates with each well containing ,10 ng of total RNA, gene specific primers, SYBR-green, and an enzyme mix containing reverse transcriptase and DNA polymerase (iScript, BioRad, Hercules, CA) in a total volume of 25 ml. Reverse transcription was performed for 10 min at 50uC and terminated by heating to 95uC for 5 min (OneStepPlus, Applied Biosystems, Foster City, CA). PCR consisted of 40 cycles of 10 s melting at 95uC, 30 s annealing and elongation at 58uC, and 15 s fluorescence detection at 78uC (OneStepPlus, Applied Biosystems).
Gene specific primers were designed using software (Primer 3.0 [20]). Primer pairs spanned introns to discourage amplification of genomic DNA ( Table 1). Amplification of a single product of the appropriate size was verified by microfluidic electrophoresis (BioAnalyzer, Agilent).
The number of template molecules (cDNA templates ) was estimated according to where 6.02610 23 molecules/mol represents Avogadro's number, Product threshold is the weight of the PCR-product at threshold (0.49610 29 g) that was obtained from calibration experiments, Product size is the size of the product in base pairs (bp), Weight bp is average weight of one bp (660 g/mol), Efficiency is the PCRefficiency obtained from the slope of the log-linear phase of the growth curve [21] and C t is the cycle at which the fluorescence of the product molecules reaches a common threshold chosen in the middle of the log-linear part of the growth curve.

Statistics
Numerical data were presented as average 6 sem with n being the number of replicates. Statistical significance was determined based on unpaired t-tests. Significance was assumed at p,0.05.

Selectivity and sensitivity of K + -and Na + -selective electrodes
The sensitivity and selectivity of K + electrodes was evaluated by making recordings in solutions containing various concentrations of KCl and NaCl. Recordings were corrected for calculated liquid junction potentials and corrected measurements were subjected to a three-dimensional fit to the Nicolski equation: where V represents the measured voltage, V i is an offset term, S K+ is the slope representing the sensitivity for K + , [K + ] and [Na + ] represent the K + and Na + concentrations, and A is the selectivity coefficient. K + electrodes were found to be highly sensitive with S K+ near the theoretical value of 60 (52.960.5 mV/decade, n = 4)  Fig. 2A). The sensitivity and selectivity of Na + electrodes was evaluated in a similar fashion using solutions containing various concentrations of NaCl and KCl. Recorded voltages were corrected for calculated liquid junction potentials and corrected voltages were subjected to a three-dimensional fit to the Nicolski equation: where S Na+ is the slope representing the sensitivity for Na + and all other symbols have the same meaning as above. Na + electrodes were found to be highly sensitive with S Na+ near the theoretical value of 60 (57.860.9 mV/decade, n = 12) and moderately selective with A being different from zero (0.1360.03, n = 12). Accordingly, steep voltage changes were recorded when the NaCl concentration was changed at constant KCl concentration of 3 or 13 mM and less steep changes were observed when the KCl concentration was altered in the presence of a constant NaCl concentration of 10 mM (Fig. 2B).

Simplified calibration procedure for K + -selective electrodes
The high selectivity of K + electrodes encouraged us to simplify the calibration process for K + electrodes by using solutions composed of KCl and NaCl, where the total cation concentration was kept constant at 150 mM. Recorded voltages were corrected for calculated liquid junction potentials and corrected voltages were subjected to a two-dimensional fit to a simplified Nicolski equation: The KCl concentration was varied between 10 to 150 mM in the presence of a constant NaCl concentration of 10 mM. Note, that Na + selective electrodes were highly selective for Na + but, although to a lesser degree, also detected K + , which is evident from the steep relationships between the Voltage and the NaCl concentration (Curve 1 and 2) and the less steep relationship between the voltage and the KCl concentration (Curve 3). C) Two-dimensional calibration of K + selective electrodes. Calibration procedures were simplified by using solutions in which the sum of NaCl and KCl was maintained constant at 150 mM. doi:10.1371/journal.pone.0065977.g002 where symbols have the same meaning as above. Again, K + electrodes were found to be highly sensitive with S K+ near the theoretical value of 60 (54.960.4, n = 137) and highly selective with A near the ideal value of zero (0.01860.001, n = 137). Accordingly, the relationship between voltage and K + was found to be non-linear (Fig. 2C).

K + concentration and voltage measurements in situ and in vitro in the cochlea of adult mice
Measurements of the endocochlear potential and of ion concentrations in cochlear endolymph are usually performed in situ using adult animals, although some investigators have succeeded to obtain measurements in pre-weaning and even neonatal mice [7,22]. The in situ technique, however, cannot be extended to the prenatal phase of development. Therefore, we established an in vitro method where measurements of the transepithelial voltage and of endolymphatic ion concentrations are conducted in isolated superfused temporal bones. This method was evaluated by comparing data from adult Slc26a4 D/+ mice (P62-P72) that were obtained with the in situ technique and with the in vitro method (Fig. 3). Measurements in situ yielded an endolymphatic K + concentration of 157614 (n = 4) mM and an endocochlear potential at normoxic conditions of 9562 (n = 6) mV. Upon induction of anoxia by injection of succinylcholine, the endocochlear potential declined within 2-3 min to 22962 (n = 6) mV, while the endolymphatic K + concentration remained stable (15369 mM, n = 4). Measurements in vitro were obtained within 10-15 min after sacrifice and yielded an endolymphatic K + concentration of 144612 (n = 6) mM, which is similar to results obtained in situ, and an endocochlear potential of 21663 (n = 6) mV, which is similar to the anoxia potential found in situ. These data suggest that the in vitro method is suitable to evaluate the K + concentration, and presumably the Na + concentration, and that in vitro measurements of the transepithelial potential more closely approximate anoxic rather than normoxic conditions.

K + concentration and voltage measurements in the cochlea and endolymphatic sac
The in vitro method was used to determine the endolymphatic K + concentration and the transepithelial potential in the cochlea and the endolymphatic sac during development. Measurements were made with double-barreled K + selective electrodes in prenatal and neonatal Slc26a4 D/+ and Slc26a4 D/D mice (Fig. 4).
In the cochlea, the endolymphatic K + concentration between age E14.5 and E17.5 was 10.760.7 (n = 20) mM with no differences between Slc26a4 D/+ and Slc26a4 D/D mice (Fig. 4A). At E19.5, the endolymphatic K + concentration began to rise in both genotypes. At P0 and P3, the endolymphatic K + concentrations in the cochlea of Slc26a4 D/D mice was significantly lower compared to in Slc26a4 D/+ mice, however, this difference disappeared with further development. The perilymphatic K + concentration was 3.2960.07 (n = 72) mM and did not change with age and displayed no differences between genotypes. Slc26a4 D/+ mice developed a positive transepithelial voltage of 4.461.7 (n = 6) mV at E19.5 (Fig. 4B). With further development, the transepithelial voltage became negative. Transepithelial voltages in Slc26a4 D/D mice remained close to zero till age P5, but was found to be negative in adult mice (23.960.5 mV, n = 5).
In the endolymphatic sac, the endolymphatic K + concentration between ages E14.5 and E17.5 was 15.461.6 (n = 21) mM with no differences between Slc26a4 D/+ and Slc26a4 D/D mice (Fig. 4C) and the transepithelial potential was 25.160.7 (n = 21) mM, again with no differences between Slc26a4 D/+ and Slc26a4 D/D mice (Fig. 4D). Between ages E14.5 and E17.5, no difference was found between the endolymphatic K + concentration in the cochlea and the endolymphatic sac.
Na + and K + concentration measurements in the cochlea The in vitro method was further used to determine the endolymphatic Na + and K + concentrations in the cochlea of Slc26a4 D/+ and Slc26a4 D/D mice at ages E16.5 and P0 (Fig. 5). At both ages the endolymphatic K + concentration had been found in both genotypes to be significantly higher than the perilymphatic K + concentration and the transepithelial voltage had been found to be zero, see above. Measurements were made with doublebarreled Na + and K + selective electrodes. Recordings with the K + electrode were used as indicator of a successful penetration into scala media ( Fig. 5A and 5B).
At age E16.5, the endolymphatic Na + and K + concentrations in the cochlea were 14163 (n = 8) mM and 9.660.3 (n = 8) mM, respectively. No differences were detected between Slc26a4 D/+ and Slc26a4 D/D mice. At age P0, the endolymphatic Na + dropped in Slc26a4 D/+ mice to 80612 (n = 3) mM and the endolymphatic K + concentrations rose to 80614 mM (n = 3). Smaller changes were observed in Slc26a4 D/D mice where the endolymphatic Na + concentration dropped to 13164 (n = 3) mM and the K + concentration rose to 3162 (n = 3) mM. No difference was found between the genotypes in the sum of the Na + and the K + concentration (Fig. 5C and D). At E16.5 the sum was 15163 (n = 8) and at P0 the sum was 16065 (n = 6). The perilymphatic Na + and K + concentrations did not vary with genotypes or age and were 14763 (n = 14) mM and 4.260.3 (n = 14) mM, respectively. The sum of the perilymphatic Na + and K + concentrations was 15163 mM (n = 14).

HCO 3 2 and Cl 2 concentration estimations in the cochlea
The HCO 3 2 concentrations in cochlear endolymph were estimated under the assumptions that CO 2 , HCO 3 2 and pH are in equilibrium such that the HCO 3 2 concentration can be calculated based on measurements of the endolymphatic pH. In the cochlea at age ,E16.5, the endolymphatic pH had been reported to be pH,7.44 in Slc26a4 D/+ mice and pH,7.11 in Slc26a4 D/D mice and the perilymphatic pH had been reported to be pH 7.30 in both genotypes [11]. According to the Henderson-Hasselbalch equation, where pK a is the logarithmic acid dissociation constant (pK a = 6.1), '0.03' is the solubility of CO 2 in water and P CO2 is the partial pressure of CO 2 (P CO2 = 5%), the endolymphatic HCO 3 2 concentration was calculated to be 25 mM in Slc26a4 D/+ mice and 11.7 mM in Slc26a4 D/D mice and 18 mM in perilymph of both genotypes. With further development, based on endolymphatic pH measurements at P10, the endolymphatic HCO 3 2 concentration was calculated to be 35 mM in Slc26a4 D/+ mice and 10.5 mM in Slc26a4 D/D mice [7].
The Cl 2 concentration in cochlear endolymph was estimated under the assumptions that Na + , K + , Cl 2 and HCO 3 2 are the major ions in cochlear endolymph, and that the sum of anions matches the sum of cations. Accordingly, the endolymphatic Cl 2 concentration at age ,E16.5 was estimated to be 126 mM ( = 151-25) in Slc26a4 D/+ and 139 mM ( = 151-11.7) in Slc26a4 D/D mice and the perilymphatic Cl 2 concentration was estimated to be 133 mM ( = 151-18) in both genotypes.
Expression of Atp1a1, Slc12a2, Kcnq1, Scnn1a, Scnn1b and Scnn1g in the cochlea The molecular mechanism for K + secretion into endolymph has been established in the adult cochlea where K + is taken up across the basolateral membrane of stria marginal cells via the Na + /K + ATPase that contains the a-subunit ATP1A1 and the Na + /2Cl 2 / K + cotransporter SLC12A2. K + is then secreted into endolymph via the apical K + channel KCNQ1 [23,24]. Whether this mechanism is responsible for the prenatal rise in the endolymphatic K + concentration is not unknown.
Cochlear mRNA expression of Atp1a1, Slc12a2 and Kcnq1 was quantified between ages E14.5 and P8 and the onset and localization of SLC12A2 and KCNQ1 protein expression was determined in the cochlea of Slc26a4 D/+ and Slc26a4 D/D mice. Expression levels of Atp1a1 mRNA rose nearly in parallel between age E14.5 and P8 in Slc26a4 D/+ and Slc26a4 D/D mice (Fig. 6A) with the exception that between E17.5 and P0 a steeper rise was observed in Slc26a4 D/D mice compared to Slc26a4 D/+ mice. In contrast, expression levels of Slc12a2 and Kcnq1 were stable between E14.5 and E17.5, increased between E17.5 and P0, where stable between P0 and P4 and then increased between P4 and P8 with no differences between genotypes ( Fig. 6B and 6C).
Cryosections were prepared from the cochlea of Slc26a4 D/+ and Slc26a4 D/D mice and revealed in Slc26a4 D/D mice the dramatic enlargement of scala media (Fig. 7). At E17.5, no protein expression of SLC12A2 or KCNQ1 was detected, however, expression of both proteins was found at E19.5 in Slc26a4 D/+ and Slc26a4 D/D mice. Expression of SLC12A2 was detected in the basolateral membrane of strial marginal cells and in the basolateral membranes of epithelial cells in Kölliker's organ and expression of KCNQ1 was found in the apical membrane of strial marginal cells. No difference in the location of expression was detected between Slc26a4 D/+ and Slc26a4 D/D mice. These results are based on the evaluation of 2-3 pairs of Slc26a4 D/+ and Slc26a4 D/D littermates that were collected from two timedpregnant dams per gestational age.
The molecular mechanism for Na + absorption has been established for the adult cochlea in Reissner's membrane epithelial cells as well as in outer sulcus epithelial cells and Claudius' cells [25,26,27]. Outer sulcus epithelial cells absorb cations from endolymph including Na + and K + , however, Reissner's membrane and Claudius' epithelial cells selectively absorb Na + across the apical membrane via the Na + channel ENaC which consists of the subunits Scnn1a, Scnn1b and Scnn1g. Cochlear mRNA expression of the three ENaC subunits was quantified in Slc26a4 D/+ and Slc26a4 D/D mice between the ages E14.5 and P8. Expression levels of Scnn1a were stable during the time of observation and did not differ between Slc26a4 D/+ and Slc26a4 D/D mice (Fig. 6D). In contrast, expression levels of Scnn1b rose in Slc26a4 D/+ mice between P4 and P8 and in Slc26a4 D/D mice between E14.5 and E17.5, which is 9-15 days earlier (Fig. 6E). Similarly, expression levels of Scnn1g rose between E17.5 and P0 in Slc26a4 D/+ mice and between E14.5 and E17.5 in Slc26a4 D/D mice, which is 1-5 days earlier (Fig. 6F). With the exception of the measurement at E17.5, no difference in expression levels of Scnn1g were observed between Slc26a4 D/+ and Slc26a4 D/D mice.

Discussion
The most salient findings of the present study are that endolymph in the cochlea and the endolymphatic sac of Slc26a4 D/+ and Slc26a4 D/D mice is a NaCl rich fluid during the phase of cochlear growth and enlargement (Figs. 4 and 5), that the onset of the rise in the K + concentration in the cochlea occurs prenatally and coincides with a decline in the Na + concentration (Fig. 5), and that the rise in the K + concentration in the cochlea coincides with the onset of expression of the Na + /2Cl 2 /K + cotransporter SLC12A2 and the K + channel KCNQ1 (Fig. 7).
The molecular mechanisms for K + secretion into endolymph have been well established in the adult cochlea, where K + secretion is mediated by marginal cells of stria vascularis [16,24].
Marginal cells take up K + across the basolateral membrane via the Na + /K + -ATPase and the Na + /2Cl 2 /K + -cotransporter SLC12A2 and secrete K + into endolymph across the apical membrane via the K + channel KCNQ1/KCNE1 (formerly called IsK and MinK). The presented data suggest that this mechanism becomes operational with the onset of expression of SLC12A2 and KCNQ1 between E17.5 and E19.5 resulting in a rise of endolymphatic K + concentration in Slc26a4 D/+ and Slc26a4 D/D mice (Figs. 4 and 7). Notably, the onset of the rise in the endolymphatic K + concentration was found to occur prenatally, at E19.5 (Fig. 4A), which is ,5 days earlier than the previously reported onset at P3 that was based on serially-sectioned freeze-dried labyrinths and Xray analysis of elemental composition [28,29]. The expression of SLC12A2 in the cochlea, however, was not limited to strial marginal cells but was also found in the basolateral membrane of epithelial cells in Köllikers organ (Fig. 7). These cells, which are lost during early postnatal development, have previously been shown to express Na + /K + -ATPase, which suggest they may be engaged in ion secretion [15]. Whether and what ions are transported across the apical membrane is not known.
The molecular mechanism for Na + absorption is well established in Reissner's membrane and Claudius' epithelial cells of the adult cochlea, where Na + absorption from endolymph in mediated via the Na + selective channel ENaC and Na + is extruded across the basolateral membrane via the Na + /K + -ATPase [25,27]. Other epithelial cells, such as outer sulcus cells and sensory hair cells, also absorb Na + but do not distinguish between Na + and K + absorption since their major apical cation channel is a cation non-selective channel [26,30]. The presented quantification of mRNA expression suggest that regulation of Na + absorption, if present, does not occur at the transcriptional level of the channel-forming a-subunit Scnn1a or the accessory c-subunit Scnn1g, but occurs possibly at the transcriptional level of the b-subunit Scnn1b and at translational or regulatory levels. Interestingly, our observation of upregulation of the b-subunit between P4 and P8 coincides with the onset of amiloride-sensitive Na + transport in Reissner's membrane [31].
The onset of the rise in the K + concentration coincided with the observation of a positive transepithelial potential at E19.5 in Slc26a4 D/+ mice (Fig. 4B). This transepithelial potential may be generated by marginal cells of stria vascularis, which have been shown to generate a positive transepithelial potential when bathed at the apical membrane with a Na + -rich solution [24]. This positive potential, however, is different from the endocochlear potential that is generated across the basal cell barrier of stria vascularis, begins to rise at P5, develops within ,10 days to levels of 90 -100 mV (Fig. 3), and is exquisitely sensitive to anoxia [22]. The positive endocochlear potential is essentially a K + diffusion potential that is generated by the K + channel KCNJ10 that is located in intermediate cells of stria vascularis, which are part of the basal cell barrier [16].
From P3 onward, a negative transepithelial potential was found in the cochlea (Fig. 4B). Negative potentials of similar magnitude have been recorded in situ under anoxic conditions [22]. The negative potentials recorded in vitro are most likely anoxia potentials that were recorded since our experimental conditions did not maintain normoxic conditions, which are necessary for the positive endocochlear potential. This shortcoming, however, has been shown to not affect the outcome of the ion measurements (Fig. 3C). The negative endocochlear potential found under anoxic conditions (Fig. 3D) is a K + diffusion potential that is thought to be generated by K + channels in the basolateral membrane of the sensory hair cells and that depends on the openness of the nonselective transduction channel in the apical membrane of the sensory cells. Since the present measurements of the transepithelial Figure 5. Measurements of the endolymphatic Na + and K + concentrations in vitro. A) Representative traces of the Na + (red) and K + (black) concentration in the cochlea of an Slc26a4 D/+ mouse at age E16.5. B) Representative traces of the Na + (red) and K + (black) concentration in the cochlea of an Slc26a4 D/+ mouse at age P0. C-D) Summary of Na + and K + concentration measurements in cochlear endolymph in Slc26a4 D/+ and Slc26a4 D/D mice at age E16.5 and P0. Numbers next to bars represent the number of animals. doi:10.1371/journal.pone.0065977.g005 potential were conducted in vitro, which mimics anoxic conditions, the presence of the negative transepithelial potential observed from P3 onward indicates the presence of mature sensory hair cells. This observation is consistent with previous findings on the opening of the transduction channels in murine cochlear hair cells at neonatal stage [32]. The observation that up to P5 neither a positive nor a negative potential was observed in the cochlea of Slc26a4 D/D mice may be related to a retardation of development [12].
The rise in the K + concentration during the neonatal phase of development was found to be delayed by ,3 days in Slc26a4 D/D mice since K + concentrations reached similar levels at P0 in Slc26a4 D/+ mice and P3 in Slc26a4 D/D mice. This delay in the concentration change may in part be due to the ,10 fold larger volume of scala media in Slc26a4 D/D mice compared to Slc26a4 D/+ mice that would require a ,10 fold higher rate of K + secretion and Na + absorption [15]. This delay, however, may also, at least in part, be due to a delayed vascularization of stria vascularis in Slc26a4 D/D mice that had been observed at P3 [11].
The endolymphatic sac has been shown to reabsorb fluid and thereby drain the cochlea during the phase of cochlear lumen formation in Slc26a4 D/+ mice. Failure to drain the cochlea leads to enlargement of the cochlea in Slc26a4 D/D mice [15]. It is conceivable that loss of Slc26a4 disrupts NaCl absorption and thereby leads to the enlargement. The negative transepithelial potential found in the prenatal endolymphatic sac (Fig. 4D) is consistent with electrogenic Na + absorption mediated by a channel such as ENaC, which has been found to be expressed in the Figure 6. Expression of mRNAs that code for selected Na + and K + channels and transporters in the cochlea of Slc26a4 D/+ and Slc26a4 D/D mice. A) The a-subunit of the Na + /K + ATPase Atp1a1. B) The Na + /2Cl 2 /K + cotransporter Slc12a2. C) The a-subunit of the K + channel Kcnq1. D-F) The a-, band c-subunit of the Na + channel ENaC. Measurements in Slc26a4 D/D mice that differed significantly from measurements in Slc26a4 D/+ mice are marked (*). doi:10.1371/journal.pone.0065977.g006 endolymphatic sac [33]. Whether ENaC or other ion transporters mediate fluid absorption in the endolymphatic sac is beyond the scope of the present study.
In summary, we have determined 1) that endolymph in the cochlea and the endolymphatic sac of Slc26a4 D/+ and Slc26a4 D/D mice is a NaCl rich fluid during the phase of cochlear growth and enlargement, 2) that the onset of K + secretion occurs in the cochlea prenatally, and 3) that the rise in the K + concentration coincides with the onset of expression of the K + channel KCNQ1 and the Na + /2Cl 2 /K + cotransporter SLC12A2. The determination of the ionic composition of endolymph during the phase of cochlear growth in Slc26a4 D/+ mice and enlargement in Slc26a4 D/ D mice presents an important first step toward elucidating mechanisms that cause the enlargement. The elucidation of these mechanisms is important for the identification of drug targets that can be exploited to prevent enlargement and restore normal cochlear development in patients carrying mutations of SLC26A4.