Enhancement of HCO₃ ⁻ Permeability across the Apical Membrane of Bovine Corneal Endothelium by Multiple Signaling Pathways

Abstract

purpose. In this study, the involvement of signaling pathways in the regulation of HCO₃ ⁻ permeability across the apical membrane of the corneal endothelium was examined.

methods. Cultured bovine corneal endothelial cells (CBCECs) were grown to confluence on permeable membranes. Apical and basolateral sides were perfused with a HCO₃ ⁻-rich Cl⁻-free Ringer’s solution (28.5 mM; pH 7.5). Relative changes in apical HCO₃ ⁻ permeability were assayed by pulsing the apical perfusion bath with a low-HCO₃ ⁻ Cl⁻-free Ringer’s solution (2.85 mM; pH 6.5), in the presence or absence of agonists or inhibitors, and comparing the rates of change in intracellular pH (pH_i), as measured with a pH-sensitive dye. Ca²⁺-activated signaling was measured with the Ca²⁺-sensitive dye Fura-2. Qualitative changes in membrane potential (E_m) were measured with a voltage-sensitive dye. RT-PCR using calcium–activated chloride channel (CLCA)–specific primers was used to examine the expression of CLCA in the corneal endothelium.

results. The adenoceptor agonist adenosine (20 μM) enhanced HCO₃ ⁻ permeability by a factor of 2. Forskolin (40 μM) exerted a 6.3-fold increase of HCO₃ ⁻ permeability, which was inhibited by the Cl⁻ channel blockers, glibenclamide (50 μM) and niflumic acid (100 μM). Adenosine triphosphate (ATP) and ATPγS, P₂ receptor agonists that increased intracellular Ca²⁺ in corneal endothelium, enhanced HCO₃ ⁻ permeability by 87% and 79%, respectively. ATPγS induced depolarization of the E_m, consistent with anion channel activation, rather than activation of Ca²⁺-dependent K⁺ channels, which could secondarily increase extrusion of anions by E_m hyperpolarization. Cyclopiazonic acid (CPA), an endoplasmic reticulum (ER) Ca²⁺-pump inhibitor that increased [Ca²⁺]_i, also enhanced HCO₃ ⁻ permeability by 95%. Both the calmodulin kinase II (CaMKII) inhibitor KN-62 and the PKC inhibitor bisindolylmaleimide I (BIMI), decreased HCO₃ ⁻ permeability induced by ATPγS. The PKC activator PMA also increased HCO₃ ⁻ permeability by a factor of 1.8. RT-PCR using CLCA-specific primers showed the expression of CLCA1 in both fresh and cultured BCECs.

conclusions. Activation of adenoceptors and purinoceptors enhances HCO₃ ⁻ permeability across the apical membrane of the cultured corneal endothelium. Multiple signaling pathways (PKA, PKC, and Ca²⁺/CaMKII) contribute to the HCO₃ ⁻ transport in cultured corneal endothelium. Both cAMP and Ca²⁺-activated Cl⁻ channels (possibly CLCA) may be involved in HCO₃ ⁻ transport.

The corneal endothelium is a thin monolayer of cells covering the posterior surface of the cornea. ^{1 2} Its primary function is to maintain corneal transparency through active transport of ions and fluid. The glycosaminoglycans of the corneal stroma exert a net swelling pressure that offers a constant potential fluid imbibition by the stroma. This fluid influx is counterbalanced by the endothelium with an ion-coupled fluid transport mechanism directed from stroma to aqueous humor. Numerous studies have shown that endothelial fluid transport is dependent on the presence of HCO₃ ⁻. ³ Studies in the past two decades have revealed at least four mechanisms that support HCO₃ ⁻ transport: (1) a potent Na⁺-dependent, 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid (DIDS)-sensitive, electrogenic Na⁺-nHCO₃ ⁻ cotransporter (NBC) on the basolateral membrane ^{3 4 5} ; (2) a Cl⁻-HCO₃ ⁻ exchanger ⁶ ; (3) anion/Cl⁻ channels on the apical membrane ⁷ ; and (4) cytosolic and membrane-bound carbonic anhydrases (CAs). ⁴ All these mechanisms have been demonstrated to exist in both fresh and cultured bovine corneal endothelial cells (BCECs), although the Cl⁻-HCO₃ ⁻ exchange activity is weak in cultured corneal endothelium. ⁶ It is well established that HCO₃ ⁻ enters the cell through the basolateral NBC. We have previously shown that basolateral HCO₃ ⁻ permeability is significantly higher than apical. ⁴ Thus, the rate-limiting step in transendothelial HCO₃ ⁻ transport is at the apical membrane. We speculate that, most probably, apical HCO₃ ⁻ permeability is controlled by Cl⁻ channels and/or CAs. ⁴  

We have recently shown that the cystic fibrosis transmembrane conductance regulator (CFTR) is expressed in fresh and cultured BCECs. ⁸ CFTR, which is activated by cAMP, has been shown to have substantial permeability to HCO₃ ⁻, ⁹ which may be involved in some of the pathogenic aspects of CF. Further, recent studies have identified some CF-causing mutants that demonstrate normal Cl⁻ channel activity but impaired HCO₃ ⁻ transport. ⁹ Thus, it is possible that CFTR has a significant role in HCO₃ ⁻ transport in corneal endothelium. It is well known that adenosine, which increases cAMP through the A₂ receptor can increase the ion and fluid transport across the corneal endothelium. ¹⁰ Our previous studies have also shown that Cl⁻ and HCO₃ ⁻ permeabilities are enhanced by cAMP in the cultured corneal endothelium. ⁷ Thus, we hypothesize that adenosine can also enhance HCO₃ ⁻ permeability across the apical membrane of BCECs by activating the cAMP-PKA signaling pathway. 

Agonists of P₂ purinergic receptors have been shown to mobilize Ca²⁺ in corneal endothelial cells. ¹¹ HCO₃ ⁻ has also been demonstrated to permeate a Ca²⁺-dependent anion channel in gallbladder. ¹² Therefore, we also investigated whether HCO₃ ⁻ permeability across the apical membrane can be activated by a Ca²⁺-signaling mechanism. One possible candidate for HCO₃ ⁻ permeability is the calcium-activated chloride channel (CLCA). ^{13 14 15} RT-PCR was used to determine the expression of CLCA at the mRNA level. Furthermore, activation of P₂ receptors can also generate diacylglycerol (DAG), which will activate PKC. Last, because constitutive phosphorylation by cellular PKC is a prerequisite for the activation of CFTR by PKA, ^{16 17} we tested to determine whether PKC is involved in HCO₃ ⁻ permeability across the apical membrane. 

Materials and Methods

Cell Culture

Bovine corneal endothelial cells (BCECs) were cultured to confluence on 13-mm permeable membranes (Whatman Anodisc; Fisher Scientific, Fairlawn, NJ) as previously described. ^{6 18} Briefly, primary cultures from fresh bovine eyes were established in T-25 flasks in 3 mL DMEM, 10% bovine calf serum with antibiotic-antimycotic agents (100 U/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin B), gassed with 5% CO₂-95% air at 37°C, and fed every 2 to 3 days. These cells were subcultured to three T-25 flasks and grown to confluence in 5 to 7 days. The resultant second-passage cultures were subcultured onto 13-mm permeable filters, reaching confluence within 7 days. Cells were transferred to 0.5% serum-DMEM for at least 24 hours before the experiments began. 

Solutions and Chemicals

The composition of the HCO₃ ⁻-rich (bicarbonate-rich; BR) Ringer’s solution was (in millimolar) 150 Na⁺, 4 K⁺, 0.6 Mg²⁺, 1.4 Ca²⁺, 118 Cl⁻, 1 HPO²⁻ ₄, 10 HEPES, 28.5 HCO₃ ⁻, 2 gluconate⁻, and 5 glucose. Ringer’s solutions were equilibrated with 5% CO₂, and pH was adjusted to 7.50 at 37°C. Low-HCO₃ ⁻ (low bicarbonate; LB) Ringer’s solution (2.85 mM; pH 6.5) was prepared by replacing 25.65 mM NaHCO₃ with sodium gluconate. Cl⁻-free Ringer’s was prepared by equimolar replacement of NaCl with sodium gluconate. Osmolarity was adjusted to 300 ± 5 mOsm with sucrose. 2′,7′-Bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM), bis-oxonol (DiBac4), and Fura-2 acetoxymethyl ester (AM) were obtained from Molecular Probes (Eugene, OR), cell culture supplies from Gibco BRL (Grand Island, NY), and all other chemicals from Sigma (St. Louis, MO). Stock solutions of BCECF-AM (10 mM in DMSO) and nigericin (10 mM in ethanol) were stored desiccated at −20°C. 

Perfusion

For independent perfusion of the apical and basolateral sides of cell-coated filters (Anodisc; Fisher Scientific), a double-sided perfusion chamber was used. ⁶ The assembled chamber was placed on a water-jacketed (37°C) brass collar held on the stage of an inverted microscope (Diaphot; Nikon, Melville, NY). The apical side was viewed with a long-working-distance objective (×40, 1.2-mm working distance, 0.75 NA; Carl Zeiss, Oberkochen, Germany). The apical and basolateral compartments were connected to separate sections of tubing (Phar-Med; Fisher Scientific), which, in turn, were connected to syringes containing Ringer’s in a Plexiglas warming box (37°C). Both HCO₃ ⁻-rich and low-HCO₃ ⁻ Ringer’s solutions were continually bubbled with 5% CO₂. The flow of the perfusate (∼0.5 mL/min) was achieved by gravity. Two independent eight-way valves were used to select the desired perfusate for the apical and basolateral chambers. 

Measurements of Intracellular pH

Intracellular pH (pH_i) was measured with the pH-sensitive fluorescent dye BCECF as previously described. ⁴ Fluorescence ratios (F₄₉₅-F₄₄₀) obtained at 1 second were calibrated against pH_i by the high-K⁺–nigericin technique. Initial rates of intracellular pH over time (dpH_i/dt; i.e., maximum slope) were measured for 20 seconds after initial responses or pH_i decreases. 

Measurements of Intracellular Ca²⁺

Intracellular Ca²⁺ ([Ca²⁺]_i) was measured with Fura-2, a calcium-sensitive fluorescent dye. The cells were loaded by incubation in Ringer’s solution containing 5 μM Fura-2-AM for 30 minutes at room temperature, followed by a wash for 45 minutes. Ca²⁺ measurements were performed at room temperature with a fluorescence imaging system (MetaFluor; Universal Imaging Co., West Chester, PA). Fluorescence emission at 505 nm was monitored with excitation at 340 and 380 nm. 

Measurements of Membrane Potential

Bis-oxonol, a voltage-sensitive fluorescent dye, was used to measure relative changes in membrane potential (E_m). Depolarization partitions the dye into the plasma membrane, making it more fluorescent. Hyperpolarization releases the dye from the membrane, resulting in a decrease of fluorescence. For continuous E_m measurements, bis-oxonol was included in the perfusing Ringer’s solutions at a concentration of 200 nM. The bis-oxonol fluorescence was excited at 495 ± 10 nm and the emission collected through a barrier filter centered at 520 ± 20 nm. 

Reverse Transcription–Polymerase Chain Reaction

Total RNA was extracted from cultured and fresh bovine corneal endothelium using extraction reagent (TRIzol; Gibco BRL), according to the manufacturer’s instructions. Total RNA extracted from fresh bovine corneal epithelium served as a positive control. To generate the first-strand cDNA, extracted total RNA (0.5–5 μg) was reverse-transcribed (total incubation mixture, 20 μL) at 42°C for 50 minutes in first-strand buffer (50 mM Tris, 75 mM KCl, and 3.0 mM MgCl₂ [pH 8.4]) that contained 10 mM dithiothreitol, 0.5 mM of each dNTP, oligo dT (3 μg/uL, 1.5 μL; Gibco BRL), and reverse transcriptase (40 U/μL, SuperScript II RT; Gibco BRL). First-strand cDNA was used in PCR amplification reactions (total incubation mixture, 25 μL) in a reaction buffer that contained 20 mM Tris (pH 8.4), 50 mM KCl, 2.0 mM MgCl₂, 2.5 units polymerase (Ex Taq; TaKaRa Shuzo, Kyoto, Japan), and 0.2 mM of each dNTP, with the specific primer set. Bovine CLCA1- and CLCA2-specific primers (TC-1 and -3 and Lu-1 and -3, respectively) were synthesized according to Elble et al. ¹⁴ The final concentration of primers was 0.1 μM. Ultrapure water (Nanopure Filtration Systems; Barnstead International, Dubuque, IA) water substituted for first-strand cDNA served as the negative control. 

PCR amplifications were performed in a thermocycler under the following conditions: denaturation at 94°C for 3 minutes for one cycle, 30 cycles of denaturation at 94°C for 1 minute each, annealing at 61°C for 1 minute, extension at 72°C for 2 minutes, and a final extension for one cycle at 72°C for 15 minutes. The PCR products were loaded onto 1% agarose gel, electrophoresed, and stained with 0.5 μg/mL ethidium bromide. 

Subcloning and Sequencing

The PCR product was purified with a gel extraction kit (Valencia, Chatsworth, CA). Freshly purified products were mixed for 5 minutes with a vector (pcDNA3.1/V5-His TOPO; Invitrogen, San Diego, CA). The cloning reaction was added into a vial of cells (One Shot; Invitrogen) for plasmid transformation. The transformed bacteria were plated on agar culture media that contained ampicillin (50 μg/mL) and incubated at 37°C overnight. The vectors with predicted inserts were isolated using a plasmid miniprep kit (Qiagen, Valencia, CA). Sequencing was performed using a dye terminator cycle-sequencing, ready-reaction mix (Prism BigDye kit; PE-Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. Sequencing electrophoresis was run on a DNA sequencer (Prism 377; PE-Applied Biosystems) at the Indiana University Molecular Biology Institute. Sequences were assembled and compared on computer (Vector NTI version 5.2 software; InforMax, Bethesda, MD). 

Statistics

Initial slopes of the descending pH_i responses for the first 20 seconds during LB pulses were calculated and served as a relative measure of HCO₃ ⁻ permeability across the apical membrane. Quantitation of the pH_i changes is expressed as the mean ± SE. Paired t-test was used for statistical analysis and P < 0.05 was considered significant. 

Results

Regulation of HCO₃ ⁻ Permeability across the Apical Membrane by the cAMP-PKA Pathway

In this study, we intended to characterize the signaling pathways involved in HCO₃ ⁻ transport across the apical side of cultured bovine corneal endothelial cells (CBCECs). We have shown that adenosine and forskolin enhance chloride and HCO₃ ⁻ permeability in BCECs cultured on coverslips. ⁷ Because chloride ions can compete with HCO₃ ⁻ to permeate anion channels, ^{4 7} we used chloride-free Ringer’s solution in all experiments to measure HCO₃ ⁻ permeability. Furthermore, the use of Cl⁻-free (gluconate-substituted) solutions eliminated any possible involvement of anion exchanger activity in our HCO₃ ⁻ permeability measurements. To measure relative changes in apical HCO₃ ⁻ permeability, we used a constant CO₂ protocol. ⁴ Constant [CO₂] on both sides removes any effect on pH_i due to rapidly diffusing CO₂. Permeable filter–raised cultures perfused in a double-sided chamber were exposed to a BR Cl⁻-free solution (5% CO₂, 28.5 mM HCO₃ ⁻ [pH 7.5]) on the basolateral and apical sides at 37°C. Apical perfusion was then quickly changed to an LB Cl⁻-free solution (5% CO₂, 2.85 mM HCO₃ ⁻ [pH 6.5]). 

Because the CO₂ concentration is the same on both sides, pH_i was affected by HCO₃ ⁻ efflux and H⁺ influx across the apical membrane. It has been shown that only 18% of the initial dpH_i/dt is due to H⁺ fluxes. ⁴ Considering that the amplitude of pH_i changes during LB pulses may be influenced by basolateral transporters such as NBC, we used only the initial rate of pH_i change instead of the amplitude as the criteria for comparison of HCO₃ ⁻ permeability across the apical membrane in the presence or absence of applied agents. When BR was replaced by LB on the apical side, pH_i decreased significantly (Fig. 1A) . After decreasing, pH_i stabilized or recovered slightly to a level that was still much lower than the baseline pH_i before readdition of BR. Readdition of BR caused a small decline in pH_i before it increased to baseline. These pH_i changes during LB pulses are consistent with our previous study. ⁴ Two sequential LB pulses in the absence of test drugs appeared to have the same initial dpH_i/dt (P = 0.335, n = 5, not shown). In this study, DMSO was often used to solubilize test drugs, and so we tested the effects of DMSO on HCO₃ ⁻ permeability. After a control pulse, the second pulse was performed in the presence of 0.5% (vol/vol) DMSO in perfusion solutions. The initial dpH_i/dt for the two LB pulses were not significantly different (data not shown, P = 0.614, n = 7), indicating that DMSO did not affect HCO₃ ⁻ permeability. 

When 20 μM adenosine was applied in the BR solution, baseline pH_i decreased 0.033 ± 0.008 pH units (n = 7; Fig. 1A ). Furthermore, in the presence of adenosine, the LB pulse caused an increased rate of pH_i change by a factor of 1.98 ± 0.20, compared with the control LB pulse (n = 7, P < 0.005, Fig. 1B ). These results suggest that apical HCO₃ ⁻ permeability can be enhanced by a cAMP-mediated process. 

We further examined whether forskolin, a potent activator of adenylate cyclase could increase apical HCO₃ ⁻ permeability. As shown in Figure 1C , after application of 40 μM forskolin, an immediate sharp decrease in pH_i (0.096 ± 0.023, n = 6) was observed, which gradually reached a new steady state level below baseline (0.097 ± 0.013, n = 6). After pH_i became stable, a second LB pulse was made in the continued presence of 40 μM forskolin. The rate of pH_i decline was 6.3 ± 2.2 times faster than in the control (P < 0.005, n = 6, Fig. 1D ), indicating that increasing cAMP leads to a significant increase in apical HCO₃ ⁻ permeability. 

To test whether HCO₃ ⁻ is transported through cAMP-activated chloride channels, we used the chloride channel blockers niflumic acid and glibenclamide in the presence of forskolin stimulation. After application of 100 μM niflumic acid or 50 μM glibenclamide, pH_i increased (Figs. 2A 2C ; 0.094 ± 0.008, n = 4 and 0.067 ± 0.016, n = 5, respectively). Niflumic acid and glibenclamide significantly inhibited forskolin-activated dpH_i/dt by 94.8% ± 1.4% and 51% ± 10% (P < 0.005, n = 4; P < 0.01, n = 5, respectively; Figs. 2B 2D ). After removing niflumic acid or glibenclamide, the effects of the inhibition were totally reversed by 10 to 15 minutes of washout (Figs. 2A 2C) . The inhibitory effects of chloride channel blockers are consistent with the presence of HCO₃ ⁻-permeable anion channels in BCECs. 

Regulation of HCO₃ ⁻ Permeability across the Apical Membrane by the Ca²⁺-Signaling Pathway

We attempted to examine whether Ca²⁺ signaling is involved in HCO₃ ⁻ permeability on the apical membrane in the corneal endothelium by stimulating purinergic receptors, because this has been shown to increase [Ca²⁺]_i in cultured CBCECs. ATP (100 μM) induced [Ca²⁺]_i responses (not shown) similar to those reported previously with BCECs cultured on coverslips. ¹¹ The response includes a peak and a plateau elevation, which are thought to be primarily composed of Ca²⁺ mobilization and capacitative calcium entry, respectively. ¹⁹ ATP (100 μM) also decreased baseline pH_i gradually (0.04 ± 0.019, n = 6; Fig. 3A ). The second LB pulse in the presence of ATP induced an 87% ± 27% increase in dpH_i/dt (Figs. 3A 3D) . ATP may be converted to adenosine by cell surface ecto-adenosine triphosphatases (ATPases; CD39). ²⁰ To exclude the possibility that the increased HCO₃ ⁻ permeability is due to adenosine, we used ATPγS, a nonhydrolyzable analogue of ATP, as the purinergic receptor agonist. ATPγS (100 μM) also induced a biphasic [Ca²⁺]_i elevation (not shown). Application of 100 μM ATPγS decreased pH_i immediately by 0.047 ± 0.009 (n = 7). The second LB pulse in the presence of ATPγS showed a 79% ± 12% increase in dpH_i/dt (Figs. 3B 3D) , indicating that elevated [Ca²⁺]_i can lead to increased apical HCO₃ ⁻ permeability. 

To further confirm the involvement of Ca²⁺ signaling in regulation of HCO₃ ⁻ permeability, we used an alternative strategy to increase [Ca²⁺]_i. Cyclopiazonic acid (CPA) is a sarcoendoplasmic reticulum Ca²⁺-ATPase (SERCA) inhibitor, which depletes the calcium store and causes further Ca²⁺ entry through store-operated calcium channels (SOCs) causing capacitative calcium entry (CCE). ¹⁹ CPA (20 μM) caused a two-phase Ca²⁺ increase, which is composed of a peak and an elevated sustained plateau (not shown). This response is also consistent with previous reports. ¹¹ Application of 20 μM CPA decreased pH_i, followed by a slow settling of pH_i back to the baseline. The second LB pulse in the presence of CPA showed an increase in dpH_i/dt 1.95 ± 0.24 times greater than in the control (Figs. 3C 3D) . 

Another possibility for the Ca²⁺-induced apparent increase in HCO₃ ⁻ permeability is an increased driving force for HCO₃ ⁻ efflux through E_m hyperpolarization caused by Ca²⁺-activated K⁺ channels. If this were the case, we would expect a hyperpolarization during the [Ca²⁺]_i elevation by ATPγS. A depolarization would be an indication of increased anion (HCO₃ ⁻) channel permeability. Therefore, we measured the E_m with the voltage-sensitive dye bis-oxonol. Under Cl⁻-rich conditions (Fig.4) , ATPγS depolarized the E_m for approximately 10 minutes, and then the E_m repolarized to baseline (n = 4). Under Cl⁻-free conditions, we observed similar, but smaller, changes in E_m (not shown). These results support a more direct role of Ca²⁺ in anion channel permeability, rather than as an increased driving force for HCO₃ ⁻ efflux through E_m hyperpolarization. 

Increased [Ca²⁺]_i has been shown to activate chloride channels through calmodulin kinases (CaMK). ¹⁵ In this study, we examined whether CaMKs are involved in the regulation of HCO₃ ⁻ permeability by using the CaMKII-specific inhibitor KN-62. ²¹ It has been reported that CaMKII is ubiquitously present in mammalian tissues. ²² Figure 5A shows a control response to LB and a response in the presence of both ATPγS and KN-62. The initial dpH_i/dt in the presence of both ATPγS (100 μM) and KN-62 (2 μM) was much slower than in the presence of ATPγS (100 μM) alone (Fig. 3B) . Figure 5B summarizes these results, showing that KN-62 in the presence of ATPγS inhibited the initial dpH_i/dt by 49% ± 9% compared with the control. We also found that LB-induced apical HCO₃ ⁻ efflux was inhibited 60% ± 6% by KN-62 alone (P < 0.01, n = 3, data not shown), which indicates that CaMK phosphorylation may be a potentiating factor, even in resting cells. 

The Expression of CLCA in Corneal Endothelial Cells

Recently, two CLCA genes have been described. ^{13 14} To examine the expression of CLCA genes in the corneal endothelial cells, we used bovine (b)CLCA1 and bCLCA2 (lung-endothelial cell adhesion molecule (Lu-ECAM)-1)-specific primers in RT-PCR for fresh and cultured BCECs. It has been reported that CLCA1 is expressed in the corneal epithelium, ²³ which we used as a positive control. Our RT-PCR results showed expression of bCLCA1 in both fresh and cultured BCECs (Fig. 6) , whereas bCLCA2 mRNA was not detected (data not shown). Sequencing of the PCR product (231 bp) showed 99% homology to published bCLCA1. The detection of mRNA expression of bCLCA1 in the fresh and cultured corneal endothelial cells is consistent with our data showing Ca²⁺-dependent HCO₃ ⁻ permeability. 

Possible Regulation of HCO₃ ⁻ Permeability across the Apical Membrane by PKC

Activation of phospholipase C can activate PKC through DAG or Ca²⁺. ²⁴ Furthermore, PKC can potentiate the activity of CFTR. ^{16 17} Because of the permeability of CFTR to HCO₃ ⁻, we hypothesized that PKC activation induced by purinoceptor agonists may be one of the signaling pathways leading to HCO₃ ⁻ permeability. To examine the involvement of PKC in HCO₃ ⁻ permeability regulation, we tested the effects of phorbol 12-myristate 13-acetate (PMA), a potent PKC activator, and its ineffective analogue, 4α-phorbol 12-myristate 13-acetate. LB pulses in the presence of PMA (1 μM) increased dpH_i/dt by 79% ± 29% (P < 0.01, n = 9; Fig. 7A 7B ), whereas the PMA analogue (1 μM) appeared to have no significant effect (P > 0.05, n = 6). We also applied ATPγS together with bisindolylmaleimide I (BIMI), a specific PKC inhibitor. ²⁵ Combined application of ATPγS and BIMI (1 μM) significantly inhibited dpH_i/dt by 71% ± 7% (n = 4; Fig 8A ) compared with control, whereas ATPγS alone increased dpH_i/dt by 79% ± 12% (Figs. 3B 3D 8B) . We also found that LB-induced apical HCO₃ ⁻ efflux was inhibited 50% by BIMI alone (P [lt] 0.05, n = 6, not shown). Moreover, in the presence of BIMI, forskolin did not enhance HCO₃ ⁻ permeability (not shown). These results are consistent with the prerequisite role of PKC in CFTR activation and indicate that PKC is one of the signaling pathways for HCO₃ ⁻ permeability regulation. 

Discussion

In this study, we showed for the first time that activation of purinergic receptors can enhance apical HCO₃ ⁻ permeability in corneal endothelial cells, possibly through Ca²⁺-CaMKII- and PKC-signaling pathways, and that CLCA1 is expressed in the corneal endothelium. Together with the stimulatory effect of cAMP-PKA on HCO₃ ⁻ permeability, multiple pathways including Ca²⁺-CaMKII and PKC can converge to cause the upregulation of HCO₃ ⁻ permeability. 

We used a two-LB pulse, constant CO₂ protocol to examine the effects of intervention drugs on HCO₃ ⁻ permeability across the apical membrane of the corneal endothelium. During the initial part of apical LB perfusion, HCO₃ ⁻ leaves the cell and H⁺ enters from the bath across the apical membrane. Because more than 80% of the initial dpH_i/dt is due to HCO₃ ⁻ flux, ⁴ we can use this rate as a relative measure of HCO₃ ⁻ flux. Further, because the experiments are paired and the driving force for HCO₃ ⁻ efflux is the same during control and test pulses, the differences in initial dpH_i/dt reflect differences in HCO₃ ⁻ permeability. 

Both the adenoceptor agonist adenosine and AC activator forskolin enhanced HCO₃ ⁻ permeability across the corneal endothelium. Adenosine has been used as a stimulator of ion and fluid transport across the corneal endothelium for 30 years. ^{10 26 27} Adenosine binds to A₂ receptors leading to activation of the cAMP pathway. We have shown that cAMP activates Cl⁻ permeability across cultured BCECs. ⁷ Our current results show that forskolin increased HCO₃ ⁻ permeability across the apical membrane (Fig. 2) . The upregulation of HCO₃ ⁻ permeability by the cAMP pathway is not surprising. We recently found that CFTR, a cAMP-activated chloride channel, is expressed in corneal endothelial cells. ⁸ Preliminary studies from our laboratory indicate that CFTR is localized only to the apical membrane of the corneal endothelium (Sun XC, written communication, August 2001). The apical localization of CFTR is consistent with our observation that adenosine and forskolin can enhance HCO₃ ⁻ permeability across the apical membrane. Thus, the immediate decrease of pH_i caused by either adenosine or forskolin may be due to activation of a cAMP-dependent anion channel, causing a quick increase in HCO₃ ⁻ efflux from the cell (Fig. 1C) . The reduced pH_i and E_m depolarization would both lead to increased basolateral HCO₃ ⁻ influx through the electrogenic NBC, which may explain the following transient pH_i fluctuations. The Cl⁻ channel blockers glibenclamide and niflumic acid inhibited the increased HCO₃ ⁻ permeability induced by forskolin, indicating the involvement of Cl⁻ channels in HCO₃ ⁻ transport (Fig. 2) . 

ATP, ATPγS, and CPA also enhanced HCO₃ ⁻ permeability. All these agents increased [Ca²⁺]_i. ATPγS depolarized E_m (Fig. 4) , consistent with increased anion conductance. After the depolarization in the continued presence of ATP, E_m repolarized. This may indicate that increased [Ca²⁺]_i also activates K⁺ channels, thereby maintaining the driving force for HCO₃ ⁻ efflux. Ca²⁺-activated chloride channels have been cloned in the bovine tracheal epithelium and lung endothelium. ^{13 14} HCO₃ ⁻ has also been demonstrated to permeate a Ca²⁺-dependent anion channel in gallbladder. ¹² Our RT-PCR results showed the expression of bCLCA1 in the fresh and cultured corneal endothelial cells (Fig. 6) . This is consistent with the enhancing effects of ATP, ATPγS, and CPA on HCO₃ ⁻ permeability. Further, studies on the expression and localization of CLCA are needed to determine the role of CLCA in corneal endothelial HCO₃ ⁻ transport. 

Constitutive phosphorylation of CFTR by PKC is thought to be required for subsequent CFTR activation by PKA. ^{16 17} Due to the permeability of CFTR to HCO₃ ⁻, it is not surprising that the PKC specific inhibitor BIMI inhibited HCO₃ ⁻ permeability dramatically, even in the presence of ATPγS (Fig. 8) . The enhanced HCO₃ ⁻ permeability caused by the PKC activator PMA further confirms the importance of PKC phosphorylation in the regulation of HCO₃ ⁻ permeability in BCECs (Fig. 7) . 

Because the corneal endothelium is essential for the dehydration and transparency of the cornea, the regulation of the transport functions of this monolayer would have significant effects on the hydration of the cornea. Involvement of multiple pathways and receptor systems has the advantage of using multiple factors to enhance the transport function of the corneal endothelium. In addition to the adenoceptor-cAMP-PKA mechanism, activation of purinergic receptors can enhance HCO₃ ⁻ permeability through both Ca²⁺ and PKC pathways. Purinoceptor agonists lead to a two-phase [Ca²⁺]_i elevation, including Ca²⁺ mobilization from the endoplasmic reticulum (ER) through IP₃ receptors and Ca²⁺ entry through store-operated calcium channels. Ca²⁺ may activate CLCA through CaMKII. However, we could not exclude the possibility that Ca²⁺ can directly activate CLCA. Furthermore, activation of purinoceptors can stimulate PKC through Ca²⁺ or DAG, which also increased HCO₃ ⁻ permeability, probably through activation of CFTR. 

²  Contributed equally to this work.

 Supported by Grant EY-08834 from the National Eye Institute.

 Submitted for publication August 3, 2001; revised December 3, 2001; accepted December 21, 2001.

 Commercial relationships policy: N.

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

 

Figure 1.

 

Adenosine and forskolin enhanced apical HCO₃ ⁻ permeability across CBCECs. CBCECs grown on permeable membranes were loaded with the pH-sensitive fluorescent dye BCECF. Apical and basolateral compartments were perfused with bicarbonate-rich, chloride-free Ringer’s. Black boxes: LB pulse applied to the apical side. (A) Effects of adenosine (20 μM) on HCO₃ ⁻ permeability. (B) Comparison of the initial rates of pH_i changes (dpH_i/dt) between the control and adenosine LB pulses. (C) Effects of forskolin (40 μM) on HCO₃ ⁻ permeability. (D) Comparison of the initial dpH_i/dt between the control and forskolin LB pulses.

Figure 2.

 

Glibenclamide and niflumic acid inhibited the FSK-induced enhancement of HCO₃ ⁻ permeability across the apical side of CBCECs. (A) Effect of niflumic acid (100 μM) on forskolin-stimulated HCO₃ ⁻ permeability. (B) Comparison of the initial dpH_i/dt between LB pulses in the presence of forskolin alone and forskolin plus niflumic acid (NA). Black boxes: LB pulses applied to the apical side. (C) The effects of glibenclamide (GB; 50 μM) on forskolin-enhanced HCO₃ ⁻ permeability. (D) Comparison of the initial dpH_i/dt between LB pulses in the presence of forskolin alone and forskolin plus glibenclamide.

Figure 3.

 

ATP, ATPγS, and CPA enhanced HCO₃ ⁻ permeability across the apical side of CBCECs. (A) The effects of ATP (100 μM; n = 6), (B) ATPγS (100 μM; n = 7), and (C) CPA (20 μM; n = 7) on apical HCO₃ ⁻ permeability. Black boxes: LB pulses applied to the apical side. (D) Summary of the effects of ATP, ATPγS, and CPA on initial dpH_i/dt during LB pulses.

Figure 4.

 

The effects of ATPγS (100 μM) on E_m. Cells were perfused in HCO₃ ⁻- and Cl⁻-rich solutions containing 200 nM bis-oxonol. Increasing fluorescence indicates depolarization of E_m, and decreasing fluorescence indicates hyperpolarization of E_m.

Figure 5.

 

The CaMKII inhibitor KN-62 inhibited ATPγS-enhanced apical HCO₃ ⁻permeability across CBCECs. (A) The effects of the CaMKII inhibitor KN-62 were studied by simultaneous application of ATPγS (100 μM) and KN-62 (2 μM). Black boxes: LB pulses applied to the apical side. (B) Comparison of the effects of ATPγS alone and ATPγS together with KN-62 on initial dpH_i/dt during LB pulses.

Figure 6.

 

RT-PCR analysis of the expression of bCLCA1 in corneal endothelial cells. The expected size is 231 bp. Lane 1: marker; lane 2: fresh epithelium; lane 3: fresh endothelium; lane 4: cultured endothelium; lane 5: negative control.

Figure 7.

 

Enhancement of HCO₃ ⁻ permeability across the apical side of CBCECs by PMA. (A) The effects of PMA (1 μM) on apical HCO₃ ⁻ permeability. (B) Summary of the effects of PMA and the PMA analogue (ANA) on initial dpH_i/dt during LB pulses. Black boxes: LB pulses applied to the apical side.

Figure 8.

 

The PKC inhibitor BIMI inhibits ATPγS-enhanced apical HCO₃ ⁻ permeability across CBCECs. (A) The effects of the PKC inhibitor BIMI were studied by simultaneous application of ATPγS (100 μM) and BIMI (1 μM). (B) Comparison of the effects of ATPγS alone and ATPγS together with BIMI on initial dpH_i/dt during LB pulses. Black boxes: LB pulses applied to apical side.

The authors thank Miao Cui and Kah Tan Allen for excellent technical support.