Pendred syndrome is due to loss-of-function mutations of Slc26a4, which codes for the HCO3− transporter pendrin. Loss of pendrin causes deafness via a loss of the K+ channel Kcnj10 in stria vascularis and consequent loss of the endocochlear potential. Pendrin and Kcnj10 are expressed in different cell types. Here, we report that free radical stress provides a link between the loss of Kcnj10 and the loss of pendrin. Studies were performed using native and cultured stria vascularis from Slc26a4+/− and Slc26a4−/− mice as well as Chinese hamster ovary (CHO)-K1 cells. Kcnj10, oxidized proteins, and proteins involved in iron metabolism were quantified by Western blotting. Nitrated proteins were quantified by ELISA. Total iron was measured by ferrozine spectrophotometry and gene expression was quantified by qRT-PCR. At postnatal day 10 (P10), stria vascularis from Slc26a4+/− and Slc26a4−/− mice expressed similar amounts of Kcnj10. Slc26a4−/− mice lost Kcnj10 expression during the next 5 days of development. In contrast, stria vascularis, obtained from P10 Slc26a4−/− mice and kept in culture for 5 days, maintained Kcnj10 expression. Stria vascularis from Slc26a4−/− mice was found to suffer from free radical stress evident by elevated amounts of oxidized and nitrated proteins and other changes in protein and gene expression. Free radical stress induced by 3-morpholinosydnonimine-N-ethylcarbamide was found to be sufficient to reduce Kcnj10 expression in CHO-K1 cells. These data demonstrate that free radical stress provides a link between loss of pendrin and loss of Kcnj10 in Slc26a4−/− mice and possibly in human patients suffering from Pendred syndrome.
pendred syndrome, the most frequent hereditary form of syndromic deafness, is an autosomal recessive disease that affects the inner ear and the thyroid and causes childhood deafness and postpuberty goiter (9, 21, 29). It is caused by loss-of-function mutations of the gene SLC26A4, which encodes the anion exchanger pendrin (13, 16, 32). Pendrin has been characterized in heterologous expression systems to be an anion exchanger that accepts Cl−, HCO3−, I−, and formate (37).
Investigations of the cause of deafness in Pendred syndrome have recently been facilitated by the creation of mouse model (12). The following observations have been made in the Pendred syndrome mouse model: 1) enlargement of endolymphatic spaces prenatally, 2) acidification of endolymph at postnatal day 10 (P10), 3) increased Ca2+ concentration in the endolymph at P10, 4) reduction in endocochlear potential at P10, 5) loss of Kcnj10 protein expression at P15, 6) macrophage invasion in stria vascularis at ∼P35, 7) no changes in endolymphatic K+ concentration in adult mice, and 8) degeneration of strial marginal cells in adult mice (12, 20, 42, 44).
In the cochlea, pendrin is mainly expressed in the apical membrane of outer sulcus and spiral prominence epithelial cells that border endolymph (Fig. 1) (14, 42). The observation that loss of pendrin leads to acidification of endolymph suggests that pendrin mediates HCO3− secretion (28, 44).
Sensory transduction in the cochlea depends on the endocochlear potential, which is generated by the K+ channel Kcnj10 in stria vascularis (26, 41). The generation of a small endocochlear potential at P10 is consistent with the expression of K+ channel Kcnj10. Both Kcnj10 expression and endocochlear potential are lost during further development and consequently Slc26a4−/− mice never develop hearing (42, 44).
Loss of pendrin cannot directly affect Kcnj10 protein expression, since Kcnj10 and pendrin are expressed in different cells in the cochlear lateral wall (Fig. 1A) (42, 44). In this study, we hypothesize that free radical stress could be the link between loss of pendrin and reduction of Kcnj10 protein expression.
Increased oxidative stress in the Pendred syndrome mouse model may be a consequence of enlarged endolymphatic spaces and/or the acidic endolymph. The K+ concentration is maintained at normal levels in the enlarged Slc26a4−/− endolymph volume, which implies increased rates of K+ secretion by stria vascularis to compensate for K+ leakage (42). The rate of ion transport correlates with ATP production and metabolism in stria vascularis (23, 25). Elevated rates of K+ secretion in the stria vascularis of Slc26a4−/− mice could enhance free radical stress as a result of increased energy metabolism, which is a major source of oxygen free radicals. Oxidative stress in stria vascularis may also be a consequence of endolymph acidification. The acidic pH of endolymph could reduce the cytosolic pH in strial marginal cells that border endolymph. Cytosolic acidification is a condition favorable to release of iron from proteins and in the presence of superoxide anion (O2•−), nitric oxide radical (·NO) leads to the formation of peroxynitrite (ONOO−) and hydroxyl radical (•OH) (4, 10).
Oxygen free radicals can originate from a leaky electron transfer in the mitochondrial electron transport chain during production of ATP. Incomplete reduction of O2 during metabolism generates the free radical (O2•−), which spontaneously dismutates to hydrogen peroxide (H2O2). These two reactive oxygen species can be involved in further reactions that generate more aggressive, damaging radicals. In the presence of Fe2+, H2O2 can enter into the Fenton reaction that yields the extremely aggressive •OH. •NO synthesized by nitric oxide synthase can react with O2•− to yield peroxynitrite (ONOO−) which decomposes to form •OH and nitrate radical (·NO2) (5, 40). Transition metal ions bound to lipids and proteins can react with H2O2, •OH, and •NO2 to generate intermediates that subsequently cause oxidation and nitration of proteins and lipids, impairing the ability of the cell to function normally (30, 38).
Defense mechanisms against free radical stress entail sequestration of catalysts that promote the formation of free radicals and detoxification by chemical conversion. The free radical O2•− can be converted to O2 and H2O by superoxide dismutase in conjunction with catalase or glutathione peroxidase. Catalysts that generate free radicals, such as Fe, Zn, and Cu, are sequestered by ferritin, metallothionein, and ceruloplasmin (34, 36, 46). The cellular availability of the catalyst Fe is controlled via transcriptional and translational mechanisms. The expression of the Fe chelator transferrin (Trf) and the uptake mechanism transferrin receptor (Tfrc) are controlled by an iron-response protein, aconitase (Aco1) (8, 11). When the cytosolic free Fe concentration is high, retardation of Tfrc mRNA degradation is lifted and translation of Trf is promoted leading to an increased expression of the chelator Trf and a decreased expression of the Fe uptake mechanism Tfrc. Together, these changes in the expression of Fe-related proteins can mediate a reduction in the cytosolic free Fe concentration. Decreases in Tfrc transcript and increases in Trf protein expression, however, are often not sufficient to control oxidative stress and can be taken as an indication for the presence of Fe-mediated oxidative stress.
In the present study, we obtained direct and indirect measures of oxidative and nitrative stress before and after the onset of hearing in stria vascularis and a preparation of spiral ligament that included pendrin-expressing outer sulcus and spiral prominence epithelial cells. The levels of oxidized and nitrated proteins in conjunction with the mRNA expression of genes involved in antioxidant defenses were used to assess oxidative and nitrative stress. We evaluated the impact of oxidative and nitrative stress on the protein levels of Kcnj10 using an expression system model. To ascertain whether the loss of Kcnj10 was a consequence of the conditions prevalent in the stria vascularis of Slc26a4−/− mice, stria vascularis from P10 organ cultures and the corresponding time points in vivo were compared for the protein level expression of Kcnj10.
Animal use and tissue preparations.
The Slc26a4−/− mouse strain was created in the laboratory of Dr. E. Green at the National Institutes of Health (12). Slc26a4−/− and Slc26a4+/− mice were raised in a colony at Kansas State University that was established with breeders kindly provided by Dr. S. Wall (Emory University). Mice were deeply anesthetized with tribromoethanol (560 mg/kg ip) and killed by transcardial perfusion with Cl−-free solution. Cl−-free solution contained (in mM) 150 Na-gluconate, 1.6 K2HPO4, 0.4 KH2PO4, 4 Ca2+-gluconate, 1 MgSO4, and 5 glucose, pH 7.4. Temporal bones housing the cochleae were removed and separate fractions of stria vascularis and spiral ligament were obtained by microdissection (Fig. 1B). All procedures concerning animals were approved by the Institutional Animal Care and Use Committee of Kansas State University.
Real-time RT-PCR in the presence of SYBR green (Molecular Probes, Eugene, OR) was carried out in 96-well plates (OneStep RT-PCR Kit, Qiagen Valencia, CA; iCycler, Bio-Rad, Hercules, CA) using gene-specific primers (Table 1). Transcripts were quantified in paired experiments using one master mix and total RNA isolated from pairs of age- and sex-matched Slc26a4−/− and Slc26a4+/− mice. RT was performed for 30 min at 50°C and terminated by heating to 95°C for 15 min. PCR consisted of 40 cycles of 1-min annealing at 60°C, 1-min elongation at 72°C, 20-s hot measurement at 78°C, and 1-min denaturation at 94°C. Amplification of a single product was verified by gel electrophoresis and identity of the product was verified once by sequencing. The number of template molecules (T) was calculated according to T = 10^ log (number of molecules at Ct)/(PCR efficiency^Ct). Ct was defined as the cycle at which the fluorescence of the product molecules reached a set threshold. The number of molecules at Ct was calibrated by amplifying known numbers of 18S rRNA molecules. The content of 18S in total RNA was estimated under the assumption that total RNA consists of 100% of 18S and 28S rRNA. PCR efficiency was obtained from the slope of the log-linear phase of the growth curve (31).
Isolation of protein.
Tissue fractions were transferred from the dissection dish into 0.5-ml Eppendorf tubes and Cl−-free solution was removed after a pulse spin. Four different methods for protein isolation were used. In method 1, proteins were extracted by first adding 20 μl protein extraction buffer 1 (ReadyPrep Sequential Extraction Kit, cat. no. 163–2101 and 163–2102, Bio-Rad) and vortexing for 3 min. The homogenized tissue was centrifuged at 16,000 rpm for 10 min at room temperature. Subsequently, extraction buffer 2 was added and the process was repeated. Proteins in the supernatant were transferred into a new tube and either used immediately or stored at −80°C. In method 2, proteins were extracted by adding 30 μl of Tris-Triton buffer (50 mM Tris, 150 mM NaCl, 1% Triton-X) and incubating for 30 min in a sonicating water bath at 0°C (Fisher Scientific FS20). Proteins in the supernatant were transferred to a new tube and either used immediately or stored at −80°C. In method 3, proteins were extracted by adding 30 μl of Tris-SDS buffer (50 mM Tris, 150 mM NaCl, 0.5% SDS) and incubating for 30 min in a sonicating water bath at 0°C. Proteins in the supernatant were transferred to a new tube and either used immediately or stored at −80°C. In method 4, the tissue was fixed in 0.6 N trichloroacetic acid on ice, centrifuged at 15,000 rpm for 10 min at 4°C, and the supernatant was removed. The proteins were extracted by adding 30 μl of urea buffer (9 M urea, 0.065 mM DTT, 2% Triton X-100) and incubating for 30 min in a sonicating water bath at 0°C. The protein-containing solution was neutralized by adding 1 M Tris. To this solution, 5 μl of 1% SDS were added and resulting mixture was incubated for 30 min in a sonicating water bath at 0°C. Proteins in the supernatant were transferred to a new tube and either used immediately or stored at −80°C.
Detection of oxidized proteins.
Carbonyl groups of oxidized proteins were derivatized with dinitrophenylhydrazine (Oxyblot Kit, cat. no. S7150, Millipore, Billerica) to form denitrophenylhydrazone (DNP). DNP-labeled proteins were separated by SDS gel electrophoresis (5 μl/lane) and detected in Western blots. Procedures were carried out according to the manufacturer's recommendations. Differences in the presence of oxidized proteins were evaluated by comparison to the expression of actin or tubulin.
Quantitation of nitrated proteins by ELISA.
Protein was isolated using the Tris-Triton method (method 2) and the level of nitration was quantified by two different immunoabsorbent assays (cat. no. HK501, Cell Sciences, Canton, MA; cat. no. 17-376, Millipore) according to the manufacturer's recommendation. The total protein content of the sample was measured (NanoOrange protein quantitation kit, cat. no. N666, Invitrogen, Carlsbad, CA). Differences in the amount of nitrated proteins were evaluated by comparison of nitrated protein content to the total protein content.
Quantitative Western blotting.
An equal volume of Laemelli buffer containing 5% β-mercaptoethanol was added to the protein samples and incubated at 75°C for 10 min. Protein samples (15 μl) were separated by SDS-PAGE gel electrophoresis (4–15% Tris-SDS Polyacrylamide Precast Gels, Bio-Rad). After separation, proteins were transferred to either nitrocellulose or PVDF membranes (0.2-μm-pore size, Bio-Rad), blocked with 5% dry milk in TBS-Tween (137 mM NaCl, 20 mM Tris·HCl, 0.1% Tween 20, pH 7.6), and probed with primary antibodies (rabbit anti-Kcnj10, 1:1,000, cat. no. APC-035, Alomone Labs, Jerusalam, Israel; mouse anti-Kcnj10, cat. no. H00003766-M01, Novus Biologicals, Littleton, CO; rabbit anti-trf, 1:750, cat. no. A76, Biomeda, Foster City, CA; mouse anti-tfrc 1:2,000, cat. no. 13-6800, Zymed, San Francisco, CA; rabbit anti-actin, 1:1,000, cat. no. A2066, Sigma, St. Louis, MO; rabbit anti-tubulin, 1:500, cat. no. ab6046, Cambridge, MA). Membranes were washed 4× for 15 min in TBS-Tween and incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (anti-rabbit, 1:10,000, cat. no. 1858415; anti-mouse, 1:10,000, cat. no. 1858416, Pierce, Rockford, IL). After being washed 4× for 15 min in TBS-Tween, HRP was detected by chemiluminescence (SuperSignal West Femto Maximum Sensitivity Substrate, cat. no. 34095, Pierce) using a camera-based imaging workstation (4000MM, Kodak). Quantification of oxidized proteins and of Kcnj10 protein was carried out by integration of signals arising from specific staining (Fig. 2). Differences in Kcnj10 expression were evaluated by comparison to the expression of actin or tubulin.
Protein was isolated using the Tris-Triton method (method 2). An equal volume of sepharose protein G beads (cat. no. P3296,Sigma) was added to protein lysate followed by incubation on ice for 60 min. This mixture was spun at 10,000 g for 10 min at 4°C and the supernatant (precleared lysate) was transferred to a fresh Eppendorf tube. Ten micrograms of mouse monoclonal antibody for Kcnj10 were added to the cold precleared lysate and the mixture was incubated at 4°C for 1 h. Fifty microliters of protein G beads washed in Tris-Triton buffer were added to the antibody-lysate mixture followed by incubation for 3 h at 4°C on a shaker. The supernatant was removed after a quick spin (10,000 g for 30 s) and the beads were washed 4× in Tris-Triton buffer. Fifty microliters of 1× Laemelli buffer were added to the bead pellet and this mixture was vortexed and heated at 95°C for 15 min followed by centrifugation at 10,000 g for 5 min. The supernatant was collected in a fresh Eppendorf tube and used as the sample for Western blotting.
Quantification of total tissue Fe content.
Total tissue Fe content (tissue-Fe) was measured using a modified ferrozine-based assay (15). Freshly isolated tissue fractions of stria vascularis and spiral ligament were transferred into Eppendorf tubes and Cl−-free solution was removed. Guanidine hydrochloride (2 μl) was added to the tissue and mixing was facilitated by a pulse spin. After incubation for 45 min at 60°C, 2 μl of FAT solution were added. FAT solution contained 0.5 M ferrozine, 0.5 M ascorbic acid, and 1 M Tris·HCl, pH 7. Mixing was facilitated by a pulse spin and the tissue was incubated for 45 min at 60°C. Absorbance at 562 nm was measured in the supernatant of the 4-μl sample (ND1000, Nanodrop Technologies) and standards containing FeCl2 that were processed in parallel. Following tissue-Fe measurements, the total protein content of samples was determined (NanoOrange Protein Quantitation Kit). Measurements of the tissue-Fe are reported per milligram of protein.
Freshly microdissected stria vascularis from P10 Slc26a4−/− and Slc26a4+/− mice was cultured in 35-mm flat-bottom culture dishes (cat. no. 08-757-100A, Fisher) containing 2 ml media for 1, 3, and 5 days. The media consisted of DMEM (cat. no. 30-2002, ATCC) supplemented with 10% FBS (cat. no. 30-2020, ATCC), 500 μM furosemide (cat. no. F4381, Sigma), and 1% pencillin-streptomycin antibiotic mixture. The media was changed every day. Cultured stria vascularis at the end of 1, 3, and 5 days was transferred from the culture dish to a 0.5-ml Eppendorf tube and washed twice with Cl−-free solution to remove excess media. Protein was isolated using the Tris-Triton buffer (method 3). Protein was either used immediately or stored at −80°C.
Chinese hamster ovary cells (CHO)-K1 (cat. no. CRL-9618, ATCC, Manassas, VA) were cultured as monolayers in DMEM supplemented with 10% FBS and 1% pencillin-streptomycin antibiotic mixture (cat. no. 30-2300, ATCC) in a cell culture incubator maintained at 37°C and 5% CO2. The CHO-K1 cells were transfected in a T25 flask (cat. no. 10-126-39, Fisher, St. Louis, MO) with a GFP-Kcnj10-Export sequence plasmid (a generous gift from Dr. N. Klocker, Dept. of Physiology, University of Freiburg, Germany) using Lipofectamine 2000 (cat. no. 11668-019, Invitrogen) according to the manufacturer's instructions. The CHO-K1 cells containing the GFP-Kcnj10-Export plasmid were split 24 h posttransfection and seeded in three wells. To replicate the conditions of organ culture, 500 μM furosemide was added to each well 48 h posttransfection and wells were subjected to different levels of oxidative and nitrative stress. The first of the three wells was exposed to oxidative stress by adding 10 μM H2O2 (cat. no. H1009, Sigma). The second well was exposed to oxidative and nitrative stress by adding 1 mM SIN-1 hydrochloride (cat. no. M5793, Sigma). The third well served as a control. This protocol was devised to ensure that differences in transfection efficiency did not impact the results. After 12 h, cells were lysed and protein was extracted using the Tris-Triton buffer. Protein was either used immediately or stored at −80°C.
Data are expressed as means ± SE. Data were compared by paired and unpaired t-test as appropriate. Significance was assumed at P < 0.05.
Stria vascularis contains elevated levels of nitrated proteins.
The amount of nitrated proteins in relationship to the total protein was quantified at ages P15 and P35 in stria vascularis and the spiral ligament preparation of Slc26a4−/− and Slc26a4+/− mice (Fig. 3A). Increased levels of nitration were observed in stria vascularis but not in the spiral ligament preparation of Slc26a4−/− compared with Slc26a4+/− mice. The increased levels of nitrated proteins suggest that stria vascularis of Slc26a4−/− mice is under increased nitrative stress.
Stria vascularis contains elevated levels of oxidized proteins.
The presence of oxidized proteins was evaluated at P10 and P15 in the stria vascularis and the spiral ligament preparation of Slc26a4−/− and Slc26a4+/− mice (Fig. 3B). Detection of oxidized proteins involved DNP-labeling of carbonyl groups in oxidized proteins by derivatization with DNPH. DNP-labeled proteins were detected by Western blot. No labeling was observed in the absence of derivatization. The levels of oxidized proteins were increased in stria vascularis but not in the spiral ligament preparation of Slc26a4−/− compared with Slc26a4+/− mice. The increased levels of oxidized proteins indicate that stria vascularis of Slc26a4−/− mice is under increased oxidative stress.
Antioxidant defenses are reduced in stria vascularis.
Transcripts coding for superoxide dismutase (Sod2) and metallothionein (Mt3) were quantified in stria vascularis and in the spiral ligament preparation of Slc26a4−/− and Slc26a4+/− mice (Fig. 4). Expression of Sod2 was reduced in stria vascularis of Slc26a4−/− mice at ages P10 and P30. Furthermore, transcripts for Mt3 were reduced in stria vascularis of Slc26a4−/− mice at P15. No changes in expression of Sod2 and Mt3 were found in spiral ligament. These observations suggest that defenses against free radical stress are weakened in stria vascularis of Slc26a4−/− mice before and after the onset of hearing.
Altered iron metabolism in stria vascularis is consistent with oxidative stress.
Transcripts were quantified in stria vascularis and in the spiral ligament preparation of Slc26a4+/− and Slc26a4−/− mice (Fig. 5). Transcripts included the Fe uptake mechanism transferrin receptor (Tfrc), the Fe chelator transferrin (Trf), the Fe storage protein ferritin (Ftl1), and the expression regulator iron-responsive protein (Aco1). Quantifications were performed before (P10) and after the onset of hearing (P15), after weaning (P30–P35), and in adulthood (P75). The numbers of transcripts coding for Tfrc were found to be reduced in Slc26a4−/− mice and this reduction was observed in stria vascularis at all ages including P10, and in the spiral ligament preparation at P75. The number of transcripts coding for Trf was elevated in Slc26a4−/− mice compared with Slc26a4+/− mice and this elevation was observed in stria vascularis from P35 onward but not in the spiral ligament preparation. Transcripts for Ftl1 were found reduced in number and this reduction was seen in stria vascularis from P15 onward and in the spiral ligament preparation at P10 and at P75. In addition, transcripts for Aco1 were present in stria vascularis and spiral ligament of Slc26a4+/− and Slc26a4−/− mice. No difference in the number of transcripts was seen in stria vascularis of Slc26a4−/− compared with Slc26a4+/− mice; however, transcripts in the spiral ligament preparation were reduced at P15 and at P75.
Protein expression of Trf and Tfrc was quantified at age P35 in stria vascularis and the spiral ligament preparation of Slc26a4+/− and Slc26a4−/− mice (Fig. 6). Protein expression of Trf was increased and protein expression of Tfrc was decreased in stria vascularis but not in the spiral ligament preparation of Slc26a4−/− mice compared with Slc26a4+/− mice. The increase in Trf expression is consistent with the increase in Trf transcripts (Fig. 5) and with Aco1-mediated translational expression regulation via an elevated free Fe pool. The decrease in Tfrc expression is consistent with the observed decrease in transcripts and Aco1-mediated transcriptional regulation via an elevated free Fe pool.
Tissue-Fe was measured at age P10, before the onset of hearing, in stria vascularis and the spiral ligament preparation of Slc26a4+/− and Slc26a4−/− mice (Fig. 6). No difference in tissue-Fe was found in stria vascularis and spiral ligament of Slc26a4−/− compared with Slc26a4+/− mice.
The changes in the transcript and protein level expression of Trf and Tfrc indicate Fe-mediated oxidative stress in the stria vascularis of Slc26a4−/− mice. The reduction in Ftl1 transcripts is consistent with a reduced capacity to sequester Fe and weakened defenses against oxidative stress.
Western blot for Kcnj10 protein in stria vascularis.
Before evaluating the effect of oxidative and nitrative stress on Kcnj10 protein expression, the specificity of Kcnj10 antibodies was evaluated. Western blots revealed a considerable interexperimental variability in the pattern of bands and smears detected by anti-Kcnj10 antibodies. Reprobing individual blots with two different anti-Kcnj10 antibodies, one of which was a monoclonal, revealed highly similar patterns of bands and smears (Fig. 7, A vs. C and D vs. F). Differences in the pattern of bands and smears were a function of the protein extraction method. The interexperimental variability, however, was encountered with all methods of protein extraction (compare Fig. 7, C and D). Immunoprecipitation with the mouse monoclonal and detection with the rabbit polyclonal antibody revealed monomers of Kcnj10 at ∼40 kDa as well as higher molecular weight bands and smears (Fig. 7G). This observation confirmed that the mouse monoclonal and the rabbit polyclonal antibody recognized the same protein in samples from stria vascularis.
Furthermore, incubating the primary antibody with a blocking peptide (KLEESLREQAEKEGSALSVR) resulted in the loss of all bands and smears (Fig. 7, H and I). In this experiment, a single lane of a blot was probed for Kcnj10 protein using the rabbit polyclonal antibody and then stripped and cut into two pieces that were reprobed for Kcnj10 in the absence and presence of blocking peptide. The cut pieces were rejoined for detection. This experimental protocol was chosen to unambiguously demonstrate that the addition of blocking peptide is sufficient to prevent detection of Kcnj10. Taken together, these data verify that the rabbit polyclonal and the mouse monoclonal antibody detect Kcnj10 and that Kcnj10 in stria vascularis exists as monomers (∼40 kDa), as well as in variable higher molecular weight complexes. In subsequent experiments both smears and bands were regarded as Kcnj10 protein and quantification of expression was based on integrals of staining as shown in Fig. 2.
The variability in the pattern of Kcnj10 expression is likely due to the interaction with the dystrophin glycoprotein complex. This complex contains proteins that greatly vary in their mobility patterns due to variable glycosylations and stoichiometries. Kcnj10 protein has been shown to be associated with the dystrophin glycoprotein complex in brain and retina and with aquaporin 4 in retina (6, 7). Moreover, ubiquitination may contribute to the observed smears.
Progressive loss of Kcnj10 expression in the stria vascularis of Slc26a4−/− mice.
Protein expression of Kcnj10 was quantified in stria vascularis before the onset of hearing (P10 and P11) and after the onset of hearing (P13 and P15; Fig. 8). As expected, stria vascularis of Slc26a4+/− mice maintained and stria vascularis of Slc26a4−/− mice progressively lost the expression of Kcnj10 protein. These results are consistent with previous studies showing loss of Kcnj10 protein expression at P15 by immunohistochemistry and the loss of endocochlear potential (44). The loss of Kcnj10 protein expression occurs at a time point when the stria vascularis of Slc26a4−/− mice is under increased oxidative and nitrative stress. This finding suggests that increased oxidative and/or nitrative stress contributes to the loss of Kcnj10 protein expression in the stria vascularis of Slc26a4−/− mice.
Protocol development for organ culture of stria vascularis.
A previous study showed that stria vascularis in organ culture is poorly preserved and shows signs of degeneration in strial marginal and intermediate cells within 5 days (2). The degeneration of stria vascularis could be due to low apical concentration of K+ in organ culture that would lead to elevated rates of K+ secretion from strial marginal cells (43). Increased ion transport rates would increase ATP consumption, ATP production, and thereby enhance mitochondrial stress (25). Furosemide and furosemide analogs have been shown to reduce ATP consumption and inhibit K+ secretion by strial marginal cells (23, 43). Reduced ATP consumption would reduce ATP production and thereby lower oxidative stress. Therefore, furosemide supplementation might be useful in preserving the integrity of stria vascularis in organ culture. Stria vascularis was cultured in media with/without furosemide for 0, 1, 3, and 5 days and evaluated for viability by quantifying the levels of tubulin. In the absence of furosemide, tubulin expression was gradually lost. In the presence of furosemide, stria vascularis could be cultured for at least 5 days (Fig. 9). Therefore, to evaluate the expression of Kcnj10 in organ culture, stria vascularis was cultured in media supplemented with furosemide.
Similar expression of Kcnj10 in organ culture of Slc26a4−/− and Slc26a4+/− mice.
Stria vascularis from P10 Slc26a4−/− and Slc26a4+/− mice were cultured. Protein expression of Kcnj10 was quantified in organ culture at 1, 3, and 5 days, corresponding to the time points of measurements in native tissue at P11, P13, and P15. Under conditions of equal oxidative stress in organ culture, stria vascularis of Slc26a4−/− mice was able to maintain Kcnj10 expression at a level similar to Slc26a4+/− mice (Fig. 10). These observations suggest that oxidative stress mediates the loss of Kcnj10 expression in the native tissue.
Loss of Kcnj10 protein expression under oxidative and nitrative stress.
The effect of oxidative and nitrative stress on protein expression of Kcnj10 was evaluated in an heterologous expression system (Fig. 11). Cells were transfected with a vector that contains Kcnj10-GFP-export-sequence. This plasmid was used in our study because the localization of Kcnj10 to plasma membrane in native tissues is lost under cell culture. However, insertion of the ER export sequence for Kir2.1 into the COOH terminal of Kcnj10 enables it to translocate to the plasma membrane (39). Cells were treated with 500 μM furosemide to replicate the conditions present in organ culture. Oxidative stress was induced by incubation with 10 μM H2O2. Furthermore, oxidative and nitrative stress was induced by incubation with 1 mM SIN-1. SIN-1 releases equimolar amounts of O2•− and •NO that react to produce peroxynitrite (27). Kcnj10 expression was reduced in cells treated with SIN-1 but not in cells treated with H2O2. No significant change was found in the expression levels of tubulin suggesting that the observed differences in Kcnj10 expression are not due to cell death or degeneration. SIN-1-mediated loss of Kcnj10 suggests that free radical stress comprising both oxygen and nitrogen radicals is capable of reducing Kcnj10 expression. It is conceivable that observed loss of Kcnj10 expression in stria vascularis of Slc26a4−/− mice is due to the combined effect of oxidative and nitrative stress.
The most salient findings of the present study are that 1) free radical stress is increased in stria vascularis of Slc26a4−/− mice as early as P10, which is before the onset of hearing; 2) Kcnj10 expression is lost progressively in the stria vascularis of Slc26a4−/− mice; 3) under conditions of equal oxidative stress, stria vascularis of P10 Slc26a4−/− and Slc26a4+/− mice have a similar pattern of expression in organ culture; and 4) Kcnj10 expression is reduced under free radical stress in cell culture. These results suggest that free radical stress is involved in the etiology of hearing loss in Pendred syndrome.
Free radical stress that exceeds the capacity of defense mechanisms has been implicated in inner ear pathogenesis including acoustic trauma and drug-induced ototoxicity (17, 33). Stria vascularis is prone to free radical stress due to its high metabolic activity and dense vascular system (25). Oxygen free radical (O2•−) is a product of energy metabolism in the mitochondria and the nitrogen free radical (·NO) is generated by nitric oxide synthase in the endothelial cells of stria vascularis. The reaction between O2•− and •NO produces peroxynitrite (ONOO−). ONOO− decomposes to form the very aggressive nitrate radical (•NO2), which causes nitration of proteins. The formation of •NO2 from ONOO− is facilitated under acidic conditions or in the presence of CO2 (5, 24). The sensitivity of Kcnj10 protein expression to free radical stress and the presence of oxidative and nitrative stress in stria vascularis suggest that free radical stress is responsible for the loss of Kcnj10 in Pendred syndrome.
Reduction of Kcnj10 protein expression under free radical stress is a pathobiologic mechanism that is not limited to Pendred syndrome. Loss of Kcnj10 under free radical stress is also implicated in other disease models. Kcnj10 protein expression, unlike the expression of other K+ channels, is reduced in the retina after ischemia-reperfusion injury, an insult that is associated with oxidative stress (19). Furthermore, mice expressing a mutant superoxide dismutase (SOD1-G93A) have increased oxidative stress and reduced expression of Kcnj10 protein in spinal cord tissues (1, 22).
The origin of increased free radical stress in stria vascularis may be related to the enlarged endolymphatic volume or the acidified endolymphatic pH of Slc26a4−/− mice. Enlargement of the cochlear and vestibular endolymphatic compartments develop in Slc26a4−/− mice prenatally after embryonic day 13 (E13), when the onset of pendrin expression fails to occur in Slc26a4−/− mice (12). Shortly after birth, the composition of endolymph changes from a Na+-rich fluid to a K+-rich fluid. The onset of K+ secretion occurs at P3 (3, 45). K+ is secreted by stria vascularis and the rate of K+ secretion is controlled by the endolymphatic K+ concentration (41, 43). Enlarged endolymphatic spaces with an inherently less favorable surface-to-volume ratio may require larger rates of K+ secretion to offset leakage and maintain the normal high endolymphatic K+ concentration found in Slc26a4−/− mice (42). Higher rates of K+ secretion can be associated with higher rates of oxidative stress since metabolism and ATP production in stria vascularis are tightly linked to the rate of ion transport (23, 25). It is conceivable that the onset of oxidative stress in stria vascularis coincides with the onset of K+ secretion at P3, which is before the onset of Kcnj10 expression and rise of endocochlear potential at P8 and the onset of hearing at P12 (3, 18, 35, 44).
Acidification of endolymph was observed in Slc26a4−/− mice at P10 and older ages (28, 44). Acidification could contribute to free radical stress in stria vascularis if the lowered pH of endolymph would decrease the cytosolic pH of the adjacent stria marginal cells that are the most likely source of O2•−. The cytosolic pH could induce oxidative stress by enhancing the stability of •OH and by promoting the formation of the aggressive •OH, •NO2 from O2•− and •NO (5, 10).
In conclusion, the data demonstrate that loss of pendrin leads to free radical stress in stria vascularis and that free radical stress is sufficient to cause the loss Kcnj10 expression. The loss of Kcnj10 expression in stria vascularis leads to the loss of the endocochlear potential and the failure to develop hearing in the mouse model of Pendred syndrome.
The support by National Institutes of Health Research Grants R01-DC-01098 and P20-RR-017686 (molecular biology and biochemistry core facility) is gratefully acknowledged. N. Gunhanlar from Bilkent University (Ankara, Turkey) participated in the development of Western blots during her 2006 summer internship in the laboratory.
The authors are grateful to C. Linsenmeyer and S. Billings for genotyping animals and proofreading drafts of the manuscript.
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.
- Copyright © 2008 the American Physiological Society