The semicircular canal duct (SCCD) epithelium is a vestibular epithelial domain that was recently shown to actively contribute to endolymph homeostasis by Cl− secretion under control of β2-adrenergic stimulation. By analogy to other Cl− secretory epithelia, we hypothesized that SCCD also provides an active absorptive pathway for Na+ under corticosteroid control. Measurements of short-circuit current (Isc) demonstrated stimulation (7–24 h) by the glucocorticoids hydrocortisone (EC50 13 nM), corticosterone (33 nM), prednisolone (70 nM), and dexamethasone (13 nM) over physiologically and therapeutically relevant concentrations and its block by amiloride (IC50 470 nM) and benzamil (57 nM), inhibitors of the epithelial sodium channel (ENaC). Isc was also partially inhibited by basolateral ouabain and Ba2+, indicating the participation of Na+-K+-ATPase and a K+ channel in Na+ transport. By contrast, aldosterone stimulated Isc only at unphysiologically high concentrations (EC50 102 nM). The action of all steroids was blocked by mifepristone (RU-486; Kd ∼0.3 nM) but not by spironolactone (Kd ∼0.7 μM). Expression of mRNA for the α-, β-, and γ-subunits of ENaC was demonstrated in the presence and absence of glucocorticoids. These findings are the first to identify SCCD in the vestibular labyrinth as a site of physiologically significant ENaC-mediated Na+ absorption and osmotically coupled water flux. They further demonstrate regulation of Na+ transport by natural and therapeutic glucocorticoids. The results provide for the first time an understanding of the therapeutic benefit of glucocorticoids in the treatment of Meniere's disease, a condition that is associated with increased luminal fluid volume.
- inner ear
- vestibular labyrinth
- cortisol (hydrocortisone)
the semicircular canal duct (SCCD) is an epithelial domain in the vestibular labyrinth that has a high ratio of surface area to endolymph volume and would therefore be a strong candidate for a site of ion homeostasis. Indeed, it was recently shown that these cells secrete Cl− and that Cl− transport is controlled by β2-adrenergic receptors via a cAMP pathway (27). Interestingly, several Cl−-secreting epithelia have been found to also be capable of absorbing Na+ (23), suggesting an additional role for the SCCD in controlling ion homeostasis and thereby endolymph volume.
Meniere's disease is a debilitating condition that originates from the auditory and vestibular periphery and that has a prevalence of ∼2.8 million people in the United States. Symptoms include severe vertigo and the disease is associated with pronounced swelling of the luminal compartment of the inner ear (endolymphatic hydrops). This swelling can be due either to hypersecretion or hypoabsorption. One clinical approach to management of Meniere's disease is the administration of glucocorticoids, such as dexamethasone or prednisone (19, 36). This treatment has been effective in controlling the vertigo in ∼52 to 76% of cases of diagnosed Meniere's disease (19, 36). The underlying mechanism of action in the inner ear was previously unidentified, although it was considered likely that ion transport processes are affected.
Glucocorticoids can modify cell function via genomic and nongenomic pathways, including regulation of ion transport (10, 47). One distinguishing feature between genomic and nongenomic action is the time course of the effect. Genomic processes typically take substantially longer, on the order of hours, than the nongenomic effects, which occur within seconds or minutes (47). There is considerable evidence in many epithelia that sodium transport is stimulated by corticosteroids and that a common part of the transport pathway is the epithelial sodium channel (ENaC) in the apical membrane (7, 39).
ENaC encompasses a cation channel family composed of heteromultimeric subunits. A channel with high selectivity for Na+ over K+ results from the association of α-, β-, and γ-subunits (1). Recently, a δ-subunit of ENaC was also cloned and found to produce a highly Na+-selective channel (43). A highly Na+-selective ENaC mediates Na+ absorption in many salt-reabsorbing epithelia with high electrical resistance and is present in tissues such as cortical collecting tubules of the kidney, urinary bladder, colon, glandular ducts, respiratory airways, and amphibian skin (11, 33). ENaC is defined pharmacologically by its inhibitor sensitivity to the K+-sparing diuretic amiloride and benzamil.
The goals of the present study were to determine 1) whether SCCD epithelium engages in cation absorption via ENaC, 2) whether this absorption is regulated by glucocorticoid hormones via activation of glucocorticoid receptors (GR), and 3) whether other ion transport pathways are involved in glucocorticoid-mediated cation absorption. Our findings establish the SCCD as a newly discovered site for cation homeostasis in the inner ear and provide a basis of action for treatment of Meniere's disease with glucocorticoids.
MATERIALS AND METHODS
Primary cultures of SCCD epithelium.
Epithelial cells from the semicircular canals of neonatal (3–5 days) Wistar rats, excluding common crus, were dispersed and seeded on 6.5-mm diameter Transwell permeable supports (Costar 3470, Corning, NY) and cultured as described previously (27) with DMEM (Invitrogen 31600–034, Carlsbad, CA) or DMEM/F-12 (Invitrogen 12500–062) medium. Confluence of primary cultures (4–6 days after seeding) was verified by measurement of transepithelial resistance (RT; >2,600 Ω/cm2 at room temperature before treatment with steroid) using an Endohm meter (World Precision Instruments, Sarasota, FL). Cultures treated with steroids in the presence and absence of antagonist were exposed for 24 h unless otherwise stated. Epithelia cultured from a common dispersion of canals had a similar short-circuit current (Isc) after incubation with a maximal concentration of agonist, but occasionally there were marked differences between some batches (see Figs. 4E and 5D). Each experimental series was performed on cultures from a single batch.
Epithelia were bathed for most experiments in symmetric bicarbonate-buffered physiological saline. The composition of the solution was (in mM) 120 NaCl, 25 NaHCO3, 3.3 KH2PO4, 0.8 K2HPO4, 1.2 MgCl2, 1.2 CaCl2, and 5 glucose, pH 7.4, equilibrated with 95% O2-5% CO2. A HEPES-buffered solution equilibrated with air was used for the Ba2+ experimental series; HEPES buffer and the absence of bicarbonate/CO2 did not alter the response of Isc to experimental reagents. The composition of the HEPES-buffered solution was (in mM) 150 NaCl, 3.6 KCl, 1 MgCl2, 0.7 CaCl2, 5 glucose, and 10 HEPES, pH 7.4.
Transepithelial voltage (VT) and RT were measured from confluent monolayers of SCCD in an Ussing chamber (AH 66–0001; Harvard Apparatus, Holliston, MA) maintained at 37°C and connected to a voltage/current clamp amplifier (model VCC600, Physiologic Instruments, San Diego, CA) via Ag/AgCl electrodes and HEPES-buffered bath solution bridges with 3% agar. Cell monolayers were bathed on both sides with solutions warmed to 37°C. VT and RT were measured under current clamp, and the equivalent Isc was calculated from Isc = VT/RT. Apical and basolateral side solutions were stirred by bubbling of the appropriate gas, and drugs were added to the baths cumulatively.
RNA isolation and RT-PCR.
Total RNA was extracted both from dexamethasone-treated and nontreated SCCD primary culture cells using RNeasy Micro Kit following the manufacturer's protocol (74004, Qiagen, Valencia, CA). Rat kidney total RNA (7926, Ambion, Austin, TX) was used as positive control.
Qualitative expression at the level of mRNA of ENaC channel subunits in SCCD epithelia was determined using a One-Step RT-PCR Kit following the manufacturer's protocol (210210, Qiagen) with 50 ng total RNA in each reaction. Total RNA quantity and quality were determined with an Agilent BioAnalyzer (model 2100, Palo Alto, CA). Each of the 35 PCR cycles was composed of 95°C for 1 min, 55°C for 1 min, and 72°C for 1 min. The primers used for RT-PCR of α-, β-, and γ-ENaC subunits are specified in Table 1. To exclude the possibility of genomic DNA amplification during the PCR reaction, RT-negative controls were performed. PCR products were run on 2% agarose gels and detected by ethidium bromide. PCR products were then purified by using either a gel extraction kit (28704, Qiagen) or PCR purification kit (28104, Qiagen), and purified PCR products were sequenced to verify the identity of the RT-PCR products.
Experimental agents were added to the bath as 1,000× concentrates. Amiloride (A-7410, Sigma), forskolin (F-6886, Sigma), benzamil (B-2417, Sigma), EIPA (A-3085, Sigma), spironolactone (S-3378, Sigma), mifepristone (M-8046, Sigma), corticosterone (C-2505, Sigma), prednisolone (P-6004, Sigma), ouabain (O-3125, Sigma), actinomycin D (A-9415, Sigma), cycloheximide (239763, Calbiochem), and bumetanide (B-3023, Sigma) were dissolved in DMSO. Cyclodextrin-encapsulated dexamethasone (D-2915, Sigma), hydrocortisone (H-0396, Sigma), tetra-ethyl-ammonium (TEA; T-2265, Sigma), 4-amino pyridine (4-AP; A-0152, Sigma), iberiotoxin (IbTX; I-2141, Sigma), and BaCl2 dihydrate (11760, Fluka Chemica) were dissolved in water. Progesterone (P-8783, Sigma) and aldosterone (215360050, Acros Organics) were dissolved in absolute ethanol.
Data are presented as original recordings and as mean values ± SE from n observations. Student's t-test was used to determine statistical significance of paired and unpaired samples. Differences were considered significant for P < 0.05. The Hill equation was fitted to concentration-response curves by using individual data points to retain appropriate weighting and presented here plotted with the mean and SE. Kd for mifepristone and spironolactone was determined as previously described (44).
Increased Isc from genomic action of corticosteroids.
The glucocorticoids dexamethasone (100 nM) and hydrocortisone (1 μM) and the mineralocorticoid aldosterone (1 μM) were added to both sides of SCCD monolayers. Glucocorticoids increased Isc over a period of 10 h (Fig. 1A), reaching significance at 7 h [control 0.6 ± 0.1 μA/cm2 (n = 3); dexamethasone 2.4 ± 0.6 μA/cm2 (n = 3); hydrocortisone 2.6 ± 0.2 μA/cm2 (n = 3)] and increased further by 10 h [control 0.6 ± 0.3 μA/cm2 (n = 3); dexamethasone 3.6 ± 1.1 μA/cm2 (n = 3); hydrocortisone 3.0 ± 0.2 μA/cm2 (n = 3)]. The high concentration of aldosterone (1 μM) also led to a monotonic increase in Isc that became significant at 10 h (3.0 ± 0.4 μA/cm2, n = 3–5; Fig. 1A). Isc stabilized by 20–24 h of exposure (dexamethasone 20–24 h: 4.3 ± 0.4 μA/cm2, n = 27; 48 h: 4.9 ± 0.3 μA/cm2, n = 3), and exposures for 24 h were subsequently used. Dexamethasone significantly increased Isc and VT but decreased RT; hydrocortisone, corticosterone, prednisolone, and aldosterone significantly increased Isc and VT (Table 2). In contrast to short-term nongenomic action which takes place within seconds to minutes, long-term induction of Isc by corticosteroids is consistent with their genomic action.
The notion of genomic action was tested further by incubation with actinomycin D (50 nM), an inhibitor of transcription, or cycloheximide (1 μM), an inhibitor of translation. Epithelia were incubated for 9 h in the absence or presence of dexamethasone (100 nM), with or without the inhibitors. The stimulation of Isc by dexamethasone was significantly decreased by each of the inhibitors (Fig. 1B). Neither inhibitor prevented stimulation of Isc by forskolin (10 μM; data not shown), suggesting that these inhibitors were specific to the Na+ pathway and that there was no substantive turnover of proteins in the cAMP-dependent Cl− secretory pathway (27).
Effects of ion transport inhibitors.
Incubation of confluent monolayers with dexamethasone (100 nM) and hydrocortisone (1 μM) led to an increase in VT and Isc with a decrease in RT (Table 2 and Fig. 2, A and B). Two well-established blockers of ENaC, amiloride and benzamil, were tested for their effects on Isc. Amiloride (10 μM) decreased Isc from the apical side (Table 2 and Fig. 2, B and C) and had no effect from the basolateral side (Fig. 2C). Both amiloride and benzamil decreased Isc monotonically over the concentration range of 0.1 nM to 10 μM in epithelia treated with dexamethasone (100 nM; Fig. 2D). The concentration-response curves for amiloride and benzamil were fitted with the Hill equation. Amiloride had an IC50 of 470 ± 140 nM (n = 7) and a Hill coefficient of 0.7 ± 0.1 (Fig. 2D). The more potent amiloride analog benzamil decreased Isc from the apical side with an IC50 of 57 ± 17 nM (n = 5) and a Hill coefficient of 0.7 ± 0.1 (Fig. 2D). By contrast, an inhibitor of Na+/H+ exchange, EIPA, had almost no effect on Isc from both the apical and basolateral side (Fig. 2D), indicating that a Na+/H+ exchanger is not involved.
We investigated whether forskolin stimulates Cl− secretion in dexamethasone-treated epithelia in the continued presence of apical amiloride. Indeed, increasing intracellular cAMP by exposure to apical and basolateral forskolin (10 μM) in the presence of apical amiloride (10 μM) led to an increased Isc in dexamethasone (100 nM)-treated epithelia (Fig. 2B) as previously shown in untreated epithelia (Fig. 2A) (27). Interestingly, the stimulation by forskolin was significantly greater after treatment with dexamethasone [without dexamethasone (Fig. 2A): 0.58 ± 0.06 μA/cm2, n = 38; with dexamethasone (Fig. 2B): 1.30 ± 0.19 μA/cm2, n = 15].
The block by benzamil was 10 times more potent than amiloride and this potency is consistent with the block of ENaC (26). In addition to inhibiting Isc, amiloride and benzamil also increased RT in dexamethasone-treated SCCD monolayers (Table 2); this is consistent with the block of ENaC by amiloride and benzamil. Taken together, these data are consistent with the pharmacological fingerprint of highly Na+-selective ENaC (15, 21).
We investigated whether a ouabain-sensitive Na+-K+-ATPase, basolateral K+ channels, and a bumetanide-sensitive Na+-K+-2Cl− cotransporter could be involved in glucocorticoid-mediated Na+ absorption. It was found that basolateral, but not apical, ouabain (1 mM; Fig. 3A) decreased Isc in epithelia exposed to dexamethasone (100 nM). By contrast, neither basolateral nor apical bumetanide (10 μM; Fig. 3B) changed Isc in epithelia exposed to dexamethasone (100 nM). These results demonstrate the involvement of Na+-K+-ATPase but not Na+-K+-2Cl− cotransporter in Na+ transport.
The potassium channel blocker Ba2+ (1 mM) caused a significant decrease in Isc (Fig. 3, C and D) in epithelia exposed to dexamethasone (100 nM) when applied to the basolateral but not the apical side. By contrast, TEA (20 mM), IbTX (100 nM), and 4-AP (1 mM) had no significant effect on Isc from either side (Fig. 3, C and D) in epithelia exposed to dexamethasone. This result suggests the involvement of Ba2+-sensitive basolateral potassium channels in Na+ transport.
Concentration dependence of corticosteroid stimulation of Isc.
Epithelia were incubated with the corticosteroids dexamethasone, hydrocortisone, corticosterone, prednisolone, and aldosterone for 24 h at four to six concentrations each. Isc increased monotonically with increasing concentrations, and the results were fitted with the Hill equation (Fig. 4 and Table 3). The steep part of each curve for the glucocorticoids was within the therapeutic or physiological range, whereas aldosterone was only effective at concentrations much higher than found under normal physiological conditions (Table 4).
These results are not limited to an early developmental stage because Na+ transport via ENaC [amiloride- and benzamil-sensitive Isc during dexamethasone (100 nM) stimulation] was also observed in SCCD monolayers cultured from adult rats (data not shown).
Effect of corticosteroid receptor antagonists.
We investigated whether dexamethasone, hydrocortisone, and aldosterone stimulate Isc by activation of GR and/or mineralocorticoid receptors (MR). SCCD monolayers were incubated in the presence of dexamethasone (100 nM), hydrocortisone (1 μM), or aldosterone (1 μM) and in the presence or absence of receptor antagonists. The GR antagonist mifepristone (a.k.a. RU-486 and RU-38486) was used at concentrations of 0.1 and 1 μM. The Kd for glucocorticoid agonist binding to GR (4, 40) has been reported to be similar to that for mifepristone (20). There is virtually no binding of mifepristone to MR (14).
Mifepristone significantly reduced the effects of dexamethasone, hydrocortisone, and aldosterone (Fig. 5, A–C), consistent with action of all of these corticosteroids at GR. We determined the effective Kd for mifepristone on dexamethasone-stimulated Isc in SCCD epithelium to be ∼0.3 nM (Figs. 5D and 6A). The Kd of mifepristone for GR from our functional studies is in agreement with the reported binding studies in rodents (20). Mifepristone is also known to be an antagonist of the progesterone receptor. However, progesterone (10–1,000 nM) had no effect on Isc (Fig. 4F).
The MR antagonist spironolactone had no significant effect on dexamethasone-, hydrocortisone-, or aldosterone-treated epithelia (Fig. 5, A–C), consistent with MR being not functional in regulating Isc in these cells. We estimated the effective Kd for spironolactone on dexamethasone-stimulated Isc in SCCD epithelium to be ∼0.7 μM (Figs. 5D and 6A) from these measurements. These findings suggest that both glucocorticoids and mineralocorticoid stimulate Isc solely by activation of GR.
Transcripts encoding α-, β-, and γ-ENaC subunits were detected by RT-PCR in SCCD epithelium in both the absence (Fig. 7A) and presence (Fig. 7B) of dexamethasone (100 nM; 24 h) treatment (n = 3 each). Bands corresponding to the expected lengths for each subunit fragment were observed, and the identities were confirmed by sequencing of the cDNA. No contamination with genomic DNA was observed in reverse transcription-negative controls (−RT) amplified by PCR. All three ENaC subunits were detected in rat kidney RNA under identical conditions (data not shown). Even though SCCD epithelial cells have transcripts for ENaC in the absence of dexamethasone exposure, little functional (amiloride sensitive) Isc was observed (Table 2).
These findings are the first to identify the SCCD epithelium as a site of physiologically significant Na+ absorption in the vestibular labyrinth and to demonstrate its regulation by glucocorticoids. Much interest has been focused on epithelia, such as the cortical collecting duct and connecting tubules of the kidney, which absorb Na+ via ENaC under control of the mineralocorticoid aldosterone (35). In fact, the focus of a recent meeting on ENaC (Aldosterone and ENaC: from Genetics to Physiology, Banff, 2003) emphasized this class of tissues. However, some epithelia, such as the lung, have been found to respond to glucocorticoids (10); the SCCD clearly belongs to the latter group. The action of glucocorticoids in SCCD is via genomic regulation because the time course is on the order of hours; nongenomic action would be expected to occur on a much shorter time scale (47).
The lumen of the semicircular canals of the vestibular system is filled with endolymph, a fluid with a high concentration of K+ (149 mM) and a low concentration of Na+ (9 mM) (45). This uncommon composition for an extracellular fluid is required to support transduction of head acceleration into nerve impulses in the vestibular system. Disturbance to influx and/or efflux solute transport pathways is expected to lead to changes in fluid volume, as occur in Meniere's disease in which there is increased luminal fluid. A model of Meniere's disease surgically induced in guinea pigs has been studied extensively and there are conflicting reports on possible alterations to ion homeostasis. Two groups have reported an elevation of cochlear and vestibular endolymphatic Na+ concentration by two- to threefold (28, 38), whereas another reports no change in cochlear Na+ (22). Increased Na+ could bring additional water into the lumen, accounting for hydrops.
The Isc of SCCD is in the direction of cation absorption and anion secretion. In fact, both components of Isc have been demonstrated. It has recently been shown that SCCD epithelium contributes to the homeostasis of vestibular endolymph by secretion of chloride under adrenergic control (27). In the absence of adrenergic stimulation, there is still a minor Cl− secretion due to constitutive adenylyl cyclase activity that accounts for Isc (27). Inhibition of steroid-stimulated Isc by amiloride returns Isc to this baseline level (Table 2). Even after inhibition of Na+ absorption by amiloride, the cAMP pathway remains functional during stimulation by forskolin (Fig. 2, A and B).
Steroid-stimulated Isc across SCCD is mediated by apical ENaC because 1) amiloride and benzamil acted only on this side; 2) benzamil was ∼10 times more potent; and 3) EIPA, an inhibitor of Na+/H+ exchange, had no effect. The IC50 of amiloride was in the range previously found for mammalian ENaC in expression systems (26). These findings are in agreement with the pharmacological profile of highly selective ENaC (15, 21). Furthermore, transcripts for all three ENaC subunits were found in the presence and absence of steroid stimulation. The presence of transcripts in the latter case suggests that ENaC may be always present in the cells but not active at the membrane. There have been recent reports of ENaC in other parts of the inner ear determined by immunocytochemistry (9, 17) and electrophysiology (24), but none investigated the semicircular canals.
The concentration-response curves for the glucocorticosteroids were all in the ranges previously reported for GR (35); the physiological and therapeutic concentrations of the agonists fell on steep portions of the curves. By contrast, aldosterone only acted by its promiscuous action on GR rather than via MR. This scenario was supported by the effectiveness of a GR antagonist at a Kd (Fig. 6A) similar to reported values (20) and the ineffectiveness of an MR antagonist.
Our current understanding of glucocorticoid-mediated transepithelial Na+ transport by the SCCD epithelium is illustrated in Fig. 6B. The driving force for electrogenic Na+ absorption across the apical membrane is provided by the basolateral Na+-K+-ATPase in conjunction with the basolateral K+ channel. The pump decreases intracellular Na+ concentration to create a low chemical potential difference and increases intracellular K+ concentration to generate a negative intracellular potential across the K+ channel. The electrochemical potential difference for Na+ across the apical membrane would then be inwardly directed across the apical ENaC. Na+ entering the cell through the apical membrane is then removed at the basolateral side by the pump. K+ that enters the cell via Na+-K+-ATPase recirculates across the basolateral membrane via the Ba2+-sensitive K+ channel, resulting in net transepithelial movement of Na+. A Cl− secretory pathway is stimulated by β2-adrenergic agonists via cAMP (27). Both the Na+ absorption and Cl− secretion pathways are stimulated by chronic exposure to glucocorticosteroids.
The present study points to the SCCD as an important locus of Na+ absorption in the healthy mammalian vestibular labyrinth that maintains the unusually low concentration of Na+ in the endolymph. Our work also provides the basis for one possible mechanism of action of glucocorticoids in the treatment of Meniere's disease patients suffering from episodes of vertigo. Dexamethasone and prednisone probably restore endolymph volume and Na+ levels by inducing Na+ absorption through ENaC channels in SCCD epithelia.
The question arises whether the observed magnitude of Isc would be adequate to account for physiologically relevant luminal volume changes. The volume flux can be estimated from the canal diameter (∼250 μM), taking the amiloride-sensitive Isc to be totally Na+ flux, assuming Cl− to be the only osmotically active solute to follow Na+ and assuming the solutes to leave the lumen isosmotically. Every microampere per centimeter squared then corresponds to ∼4% luminal volume change per hour. Because the endolymph system is a closed compartment, the estimated volume changes can be taken to be clinically highly significant. The demonstration of GR in vestibular tissues provides a cellular means by which circulating glucocorticoids can directly affect vestibular physiology (32, 37).
This work was supported by National Institute for Deafness and Communication Disorders Grant R01-DC00212.
We thank Drs. L. Freeman, E. Minton, and B. Schultz for helpful discussions and suggestions.
Present address of P. G. Milhaud: Institut National de la Santé et de la Recherche Médicale U432 Vestibular Neurobiology, Université Montpellier II, 34095 Montpellier, France.
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 © 2004 the American Physiological Society