Role of NBC1 in apical and basolateral ${\mathrm{HCO}}_3^{-}$ permeabilities and transendothelial ${\mathrm{HCO}}_3^{-}$ fluxes in bovine corneal endothelium

Abstract

Corneal transparency and hydration control are dependent on ${\mathrm{HCO}}_3^{-}$ transport properties of the corneal endothelium. Recent work (13) suggested the presence of an apical $1{\mathrm{Na}}^{+}-3{\mathrm{HCO}}_3^{-}$ cotransporter (NBC1) in addition to a basolateral $1{\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$ cotransporter. We examined whether the NBC1 cotransporter contributes significantly to basolateral or apical ${\mathrm{HCO}}_3^{-}$ permeability and whether the cotransporter participates in transendothelial net ${\mathrm{HCO}}_3^{-}$ flux in cultured bovine corneal endothelium. NBC1 protein expression was reduced using small interfering RNA (siRNA). Immunoblot analysis showed that 5–15 nM siRNA decreased NBC1 expression by 80–95%, 4 days posttransfection. Apical and basolateral ${\mathrm{HCO}}_3^{-}$ permeabilities were determined by measuring the rate of pH_i change when ${\mathrm{HCO}}_3^{-}$ was removed from the bath under constant pH or constant CO₂ conditions. Using either protocol, we found that cultures treated with NBC1 siRNA had sixfold lower basolateral ${\mathrm{HCO}}_3^{-}$ permeability than untreated or siCONTROL siRNA-treated cells. Apical ${\mathrm{HCO}}_3^{-}$ permeability was unaffected by NBC1 siRNA treatment. Net non-steady-state ${\mathrm{HCO}}_3^{-}$ flux was 0.707 ± 0.009 mM·min⁻¹·cm² in the basolateral-to-apical direction and increased to 1.74 ± 0.15 when cells were stimulated with 2 μM forskolin. Treatment with 5 nM siRNA decreased basolateral-to-apical flux by 67%, whereas apical-to-basolateral flux was unaffected, significantly decreasing net ${\mathrm{HCO}}_3^{-}$ flux to 0.236 ± 0.002. NBC1 siRNA treatment or 100 μM ouabain also eliminated steady-state ${\mathrm{HCO}}_3^{-}$ flux, as measured by apical compartment alkalinization. Collectively, reduced basolateral ${\mathrm{HCO}}_3^{-}$ permeability, basolateral-to-apical fluxes, and net ${\mathrm{HCO}}_3^{-}$ flux as a result of reduced expression of NBC1 indicate that NBC1 plays a key role in transendothelial ${\mathrm{HCO}}_3^{-}$ flux and is functional only at the basolateral membrane.

THE ION AND FLUID TRANSPORT PROPERTIES of the corneal endothelium are responsible for maintaining the hydration and transparency of the cornea. Corneal endothelial fluid transport requires the presence of ${\mathrm{HCO}}_3^{-}$ (15, 18, 29) and Cl⁻ (44), is sensitive to carbonic anhydrase inhibitors (19, 22, 29), and is completely eliminated by ouabain, a Na⁺-K⁺-ATPase inhibitor. Net stroma to anterior chamber ${\mathrm{HCO}}_3^{-}$ flux is responsible for the measured short-circuit current and the small, negative transendothelial potential (19), suggesting that ${\mathrm{HCO}}_3^{-}$ is the primary secreted anion. Although significant progress has been made in identifying and locating plasma membrane transporters in the corneal endothelium, the contribution of the transporters to net ${\mathrm{HCO}}_3^{-}$ transport is largely unknown. In this study we have examined the role of the sodium bicarbonate cotransporter (NBC1) in transendothelial ${\mathrm{HCO}}_3^{-}$ transport.

Previous studies have shown that the uptake of ${\mathrm{HCO}}_3^{-}$ across the basolateral membrane of corneal endothelial cells occurs by a potent Na⁺-dependent, Cl⁻-independent, DIDS-sensitive, and electrogenic ${\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$ cotransporter (5, 8, 21, 35). The activity of this cotransporter has a significant effect on intracellular pH (pH_i), and it appears to be the major entry point for ${\mathrm{HCO}}_3^{-}$ flux across the endothelium (5, 8). Recent molecular cloning experiments have identified several Na⁺-dependent bicarbonate transporters (3, 10, 20, 23, 27, 31, 40). Two variants of NBC1 have been found: the kidney proximal tubule form of NBC (kNBC) (11, 30) has a 1:3 stoichiometry, and the pancreas form of NBC (pNBC) (1, 38) has a 1:2 stoichiometry. However, more recent studies have shown that the stoichiometry of either kNBC or pNBC can change depending on the cell type in which it is expressed (17).

Our previous studies have shown that human (35) and bovine corneal endothelial cells (36) express the pNBC isoform only. An earlier report (42), however, suggested that both pNBC and kNBC are expressed in human corneal endothelium. Immunohistochemistry studies in cultured and fresh bovine (36), rat (4), and human endothelium (35, 41) indicate that NBC1 exclusively locate to the basolateral membrane; however, a recent report (13) suggests apical expression as well. Whereas ${\mathrm{HCO}}_3^{-}$ uptake by a basolateral ${\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$ cotransporter is certain, the mechanism for apical efflux is not clear. Evidence has been provided suggesting that ${\mathrm{HCO}}_3^{-}$ can exit the endothelial cells through anion channels such as the cystic fibrosis transmembrane conductance regulator (CFTR) and Ca²⁺-activated Cl⁻ channels (CaCC) (13, 34, 45). In addition, CO₂ diffusion and conversion to ${\mathrm{HCO}}_3^{-}$ by an apical membrane-bound extracellular carbonic anhydrase (CAIV) could also provide for net apical efflux (5, 6). If an apical NBC1 exists, then a $1{\mathrm{Na}}^{+}:3{\mathrm{HCO}}_3^{-}$ stoichiometry could also potentially contribute to the apical efflux pathways (13).

In the present study we have investigated the role of NBC1 in ${\mathrm{HCO}}_3^{-}$ permeability and transendothelial ${\mathrm{HCO}}_3^{-}$ fluxes in cultured corneal endothelial cells by using a short interfering RNA (siRNA) knockdown approach. siRNA has significant advantages over pharmacological agents such as DIDS, which can block several anion transporters and channels. We found that siRNA transiently inhibited NBC1 expression, significantly reduced basolateral but not apical ${\mathrm{HCO}}_3^{-}$ permeability, and reduced non-steady-state basolateral-to-apical ${\mathrm{HCO}}_3^{-}$ flux and net transendothelial ${\mathrm{HCO}}_3^{-}$ flux, indicating that an apical NBC1, if present, does not significantly contribute to net ${\mathrm{HCO}}_3^{-}$ flux.

MATERIALS AND METHODS

Cell culture

Bovine corneal endothelial cells (BCEC) were cultured to confluence onto 25-mm round coverslips, 13-mm Anodisc filters, or T-25 flasks as previously described (7, 24). Briefly, primary cultures from fresh cow eyes were established in T-25 flasks with 3 ml of Dulbecco’s modified Eagle’s medium (DMEM), 10% bovine calf serum, and antibiotic (100 U/ml penicillin, 100 U/ml streptomycin, and 0.25 μg/ml Fungizone), gassed with 5% CO₂-95% air at 37°C, and fed every 2–3 days. Primary cultures were subcultured to three T-25 flasks and grown to confluence in 5–7 days. The resulting second-passage cultures were then further subcultured onto coverslips or Anodiscs and allowed to reach confluence within 5–7 days.

siRNA transfection

Four sense and antisense oligonucleotides corresponding to the following NBC1 cDNAs were designed and blasted using the Ambion siRNA-targeting design tool and were purchased from Invitrogen: AAGTTTGAAGAAAAAGTGGAA (397–417), AAAGAATATGTACTCAGGTGG (1209–1229), AATTGTGCCAAGTGAGTTCAA (2259–2279), and AAAAAGAAGGAGGATGAGAAG (3034–3054). Using these oligonucleotides, four siRNAs for NBC1 were synthesized using the Silencer siRNA construction kit from Ambion. A siCONTROL nontargeting siRNA (no known mammalian homology) was purchased from Dharmacom. Cells were transfected when 70–80% confluent by using Oligofectamine (Invitrogen) according to the manufacturer’s protocol in the presence of siRNA. Cell-coated coverslips or Anodiscs in six-well plates were incubated with 1 ml of Opti-MEM I (GIBCO) containing siRNA for 4 h, followed by addition of 2 ml of standard DMEM with serum. T-25 flasks were treated with 2 ml of Opti-MEM I containing siRNA, followed by addition of 4 ml of culture medium. Medium was then changed every 2 days.

Immunoblotting

BCEC were dissolved directly in 2% SDS sample buffer that contained protease inhibitors. The preparations were sonicated (Branson 250) briefly on ice and centrifuged at 10,000 g for 5–10 min. An aliquot of the supernatant was taken for protein concentration measurement using the Bradford assay (Bio-Rad). Samples (30 μg, not heated) were resolved on SDS-PAGE and transferred to polyvinylidene difluoride membrane (Bio-Rad). Blots were then probed with NBC1 polyclonal antibody (AB-3212, 1:2,000; Chemicon) or α-subunit of Na⁺-K⁺-ATPase antibody (1:1,000; Developmental Hybridoma Bank, Iowa University), and bound antibody was detected using enhanced chemiluminescence (ECL). The membrane was then stripped using Re-blot plus strong antibody stripping solution (Chemicon) to remove NBC1 antibody or Na⁺-K⁺-ATPase antibody, and blots were incubated with β-actin polyclonal antibody (1:10,000; Sigma) and developed using ECL. Films were scanned to produce digital images that were then assembled and labeled using Photoshop software.

Immunofluorescence

Cultured cells grown to confluence on coverslips were washed three to four times with warmed (37°C) PBS and fixed for 30 min in warmed PLP fixation solution (2% paraformaldehyde, 75 mM lysine, 10 mM sodium periodate, and 45 mM sodium phosphate, pH 7.4) on a rocker. After fixation, the cells were washed three to four times with PBS. Coverslips were then kept for 5 min in PBS that contained 1% SDS to unmask epitopes and washed three times in PBS. Cells were blocked for 1 h in PBS that contained 0.2% bovine serum albumin, 5% goat serum, 0.01% saponin, and 50 mM NH₄Cl. Rabbit polyclonal NBC1 antibody and rat monoclonal ZO-1 antibody (MAB1520; Chemicon) diluted 1:100 together in PBS-goat serum (1:1) were added onto coverslips and incubated for 1 h at room temperature or overnight at 4°C. Coverslips were washed three times for 15 min in PBS that contained 0.01% saponin. Secondary antibodies conjugated to Alexa 488 (NBC1) (1:1,000; Molecular Probes) and Alexa Fluor 594 (ZO-1) (1:1,000; Molecular Probes) were applied for 1 h at room temperature. Coverslips were washed and then stained with 1 μg/ml Hoechst nuclear dye for 5 min. Coverslips were washed with water and mounted with Prolong antifade medium (Molecular Probes) according to the manufacturer’s instructions. Fluorescence was observed with a standard epifluorescence microscope equipped with a cooled charge-coupled device camera.

Microscope perfusion

For measurement of pH_i and ${\mathrm{HCO}}_3^{-}$ transendothelial flux, cells were cultured to confluence on 13-mm diameter, 0.2-μm Anodisc membranes. Anodiscs were placed in a double-sided perfusion chamber designed for independent perfusion of the apical and basolateral sides (9). The assembled chamber was placed on a water-jacketed (37°C) brass collar held on the stage of an inverted microscope (Nikon Diaphot 200) and viewed with a long-working-distance (2 mm) water-immersion objective (×40; Nikon). Apical and basolateral compartments were connected to hanging syringes that contained Ringer solution in a Plexiglas warming box (37°C) by using Phar-Med tubing. The flow of the perfusate (~0.5 ml/min) was achieved by gravity. Two independent eight-way valves were employed to select the desired perfusate for the apical and basolateral chambers. The composition of the standard ${\mathrm{HCO}}_3^{-}$-rich Ringer solution used throughout the study was (in mM) 150 Na⁺, 4 K⁺, 0.6 Mg²⁺, 1.4 Ca²⁺, 118 Cl⁻, 1 ${\mathrm{HPO}}_4^{2-}$, 10 HEPES⁻, 28.5 ${\mathrm{HCO}}_3^{-}$, 2 gluconate⁻, and 5 glucose, equilibrated with 5% CO₂ and pH adjusted to 7.50 at 37°C. ${\mathrm{HCO}}_3^{-}$-free Ringer solution (pH 7.50) was prepared by equimolar substitution of NaHCO₃ with sodium gluconate. Low-${\mathrm{HCO}}_3^{-}$ Ringer solution (2.85 mM ${\mathrm{HCO}}_3^{-}$, pH 6.5) was prepared by replacing 25.65 mM NaHCO₃ with sodium gluconate.

Measurement of ${\mathit{HCO}}_3^{-}$ permeability

BCEC cultured onto permeable Anodisc filters were loaded with the pH-sensitive fluorescent dye 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF) by incubation in ${\mathrm{HCO}}_3^{-}$-free Ringer solution that contained 1 μM BCECF-AM at room temperature for 30–60 min. Dye-loaded cells were then kept in Ringer solution for at least 30 min before use. Fluorescence excitation (495 and 440 nM) and data collection were obtained using a DeltaRam ratio fluorescence system (Photon Technology International, Monmouth Junction, NJ) controlled by Felix software. Fluorescence ratios were obtained at 1 s⁻¹ and were calibrated against pH_i by using the high-potassium-nigericin technique (39). A calibration curve, which follows a pH titration equation, has been constructed for BCEC (7). ${\mathrm{HCO}}_3^{-}$ permeabilities of apical and basolateral membranes was determined using the constant-CO₂ or constant-pH protocols as described previously (5). Briefly, in the constant-pH protocol, the ${\mathrm{HCO}}_3^{-}$-rich Ringer on the apical or basolateral side is replaced with a CO₂- and ${\mathrm{HCO}}_3^{-}$-free Ringer of the same pH (HEPES buffered). Under this protocol, the initial pH_i change is due to rapid CO₂ efflux (increase in pH_i), followed by ${\mathrm{HCO}}_3^{-}$ efflux (decrease in pH_i). The maximum slope of the pH_i decrease is taken as an estimate of ${\mathrm{HCO}}_3^{-}$ permeability. In the constant-CO₂ protocol, the ${\mathrm{HCO}}_3^{-}$-rich Ringer (28.5 mM ${\mathrm{HCO}}_3^{-}$, 5% CO₂, pH 7.5) is replaced with a low-${\mathrm{HCO}}_3^{-}$ solution (2.85 mM ${\mathrm{HCO}}_3^{-}$, 5% CO₂, pH 6.5). Under this protocol, the initial pH_i change is predominantly due to ${\mathrm{HCO}}_3^{-}$ efflux, because there is no CO₂ gradient. However, there is a pH gradient that can contribute (~15%) to the pH_i decrease (5). Separate control experiments are performed in the absence of ${\mathrm{CO}}_2∕{\mathrm{HCO}}_3^{-}$ to estimate this contribution.

Measurement of transendothelial ${\mathit{HCO}}_3^{-}$ flux

BCEC cultured onto permeable Anodisc filters, perfused in a double-sided chamber, were exposed to the standard ${\mathrm{HCO}}_3^{-}$-rich solution (5% CO₂, 28.5 mM ${\mathrm{HCO}}_3^{-}$, pH 7.5) on the basolateral and apical sides at 37°C. A low-${\mathrm{HCO}}_3^{-}$ solution (5% CO₂/2.85 mM ${\mathrm{HCO}}_3^{-}$, pH 6.5) without HEPES buffer and containing 1 μM BCECF free acid was then quickly exchanged on one side of the chamber, and the exit tube on that side was clamped. The pH of the low-${\mathrm{HCO}}_3^{-}$ solution was estimated (1 Hz) at ~200 μm from the surface of the cells by measuring the fluorescence ratio of BCECF using the microscope fluorometer. The pH of the low-${\mathrm{HCO}}_3^{-}$ solution rose from 6.5, and the initial rate of change over the first 20 s after clamping was estimated. After the initial pH change was recorded, ${\mathrm{HCO}}_3^{-}$-rich solution was returned, the chamber was flipped over on the microscope stage, and the same measurement procedure was repeated so that unidirectional fluxes in the apical-to-basolateral and basolateral-to-apical directions were obtained for each Anodisc. Separate control experiments also were performed with the same pH gradient across the cells in 1 mM HEPES⁻ but in the complete absence of ${\mathrm{CO}}_2∕{\mathrm{HCO}}_3^{-}$. This pH change was used to estimate the initial OH⁻ flux (J_OH) across the Anodisc, which was subtracted from the total flux calculated in the presence of ${\mathrm{HCO}}_3^{-}$. Each unidirectional ${\mathrm{HCO}}_3^{-}$ flux can be calculated as

$$J_{{\mathrm{HCO}}_3}=\left[\frac{\mathrm{dpH}}{\mathrm{d}t}\times \frac{\beta }{A}\right]-J_{\mathrm{OH}}$$

where J is flux (mM·min⁻¹·cm²), A is the area of the exposed Anodisc (0.23 cm²), dpH/dt is the initial rate of pH change from 6.5, and β is the buffering capacity, which is 2.3 × [${\mathrm{HCO}}_3^{-}$], or 6.55 mM/pH. J is the OH⁻ OH flux calculated in the same manner, where β is 0.5 mM/pH. The difference between basolateral-to-apical and apical-to-basolateral unidirectional fluxes is the net ${\mathrm{HCO}}_3^{-}$ flux. This is called non-steady-state flux because it was measured in the presence of a large ${\mathrm{HCO}}_3^{-}$ gradient.

Measurement of steady-state ${\mathit{HCO}}_3^{-}$ flux

BCEC cultured to confluence on 0.2-μm Anopore membrane tissue culture inserts were washed with DMEM containing 2% bovine calf serum. Then, 200 μl of this culture medium containing 1 μM BCECF free acid were placed on the apical side and 300 μl on the basolateral side. After 6 h in a standard 5% CO₂ incubator at 37°C, cultures were placed in a large glove box equilibrated with 5% CO₂ at 37°C. Samples (50 μl) were taken from the apical and basolateral sides with separate glass capillary tubes, and both ends were sealed with wax. The tubes were then taken to the microscope fluorometer, and the fluorescence ratio of BCECF was measured. The pH of each sample was then determined using a standard curve constructed by using solutions of known pH that had been placed within capillary tubes. The difference in pH was calculated as ΔpH = apical pH − basolateral pH. A positive ΔpH indicates apical ${\mathrm{HCO}}_3^{-}$ accumulation.

RESULTS

NBC1 siRNA was used to knock down NBC1 expression in cultured BCEC. To confirm the reduction and select the best knockdown condition, we performed immunoblotting with rabbit anti-NBC1 antibody. BCEC were transfected with four siRNAs, and Western blotting results showed that the siRNA against the target AAAAAGAAGGAGGATGAGAAG was the most efficient (data not shown), so this was selected for use in the remainder of the study. Figure 1, A and B, shows that NBC1 expression in BCEC was slightly decreased 2 or 3 days after cells were treated with siRNA. At the highest concentration, 25 nM, NBC1 expression was reduced 46 and 49%, at 2 and 3 days posttransfection, respectively, compared with control. However, 4 or 5 days after being treated with siRNA, most of the NBC1 expression was suppressed. At 4 days posttransfection with 5 nM siRNA, at least 80% of protein expression was inhibited compared with control, with slightly better knockdown at higher starting siRNA concentration. Figure 1C shows that immunofluorescence for NBC1 was barely detectable at 4 days posttransfection with 15 nM siRNA, while cell and tight junction morphology were normal. NBC1 expression began to recover by the sixth day posttransfection. Although morphological appearance and cell confluence were normal, we did notice an increase in numbers of floating cells in cultures treated at the highest (25 nM) siRNA concentration. We chose to use lower concentrations, 15 nM or less, for the remainder of the study.

Fig. 1.

Effect of $1{\mathrm{Na}}^{+}-3{\mathrm{HCO}}_3^{-}$ cotransporter (NBC1) small interfering RNA (siRNA) on NBC1 expression. A: Western blot analysis of NBC1 in bovine corneal endothelial cells (BCEC). Total cellular protein was extracted from BCEC treated or untreated with different concentrations of NBC1 siRNA and examined at different time points. Protein (30 μg) was subjected to SDS-PAGE and transferred to polyvinylidene difluoride membranes overnight and stained with enhanced chemiluminescence (ECL) using rabbit polyclonal anti-NBC1. The major band position of NBC1 is at ~130 kDa. B: densitometric analyses of protein expression from experiments as shown in A. Data were quantified relative to control band intensity, respectively (n = 3). C: decrease in NBC1 protein 4 days after 15 nM siRNA treatment examined using immunofluorescence. Green, NBC1; red, ZO-1; blue, nuclei.

To determine that the reduction of NBC1 expression mediated by siRNA was not a nonspecific effect of siRNA transfection, we used a siCONTROL nontargeting siRNA to transfect BCEC and performed immunoblotting with anti-NBC1 antibody. As Fig. 2 shows, there was no significant change in NBC1 expression 4 days after cells were treated with 5 and 15 nM siCONTROL siRNA compared with untreated cells, whereas an obvious decrease in NBC1 expression occurred when cells were treated with 5 nM NBC1 siRNA. These results show that the reduction in NBC1 expression is not simply due to the introduction of foreign siRNA.

Fig. 2.

Western blot analysis of NBC1 and Na⁺-K⁺-ATPase. Total cellular proteins were extracted from BCEC treated or untreated (control, Con) with different concentrations of NBC1 siRNA or siCONTROL siRNA (si-Con), and Western blot analysis was performed using NBC1 antibody, α-subunit of Na⁺-K⁺-ATPase antibody, or β-actin antibody as detailed in MATERIALS AND METHODS.

More importantly, we must consider that any physiological effects of NBC1 siRNA treatment could be due to the nonspecific knockdown of expression of other transporters that could participate in “NBC1-like” transport. Given our current knowledge of transport mechanisms in corneal endothelium (6), candidates that might affect ${\mathrm{HCO}}_3^{-}$ flux include the Na⁺/H⁺ exchanger, ${\mathrm{Cl}}^{-}∕{\mathrm{HCO}}_3^{-}$ exchanger, and the Na⁺-K⁺-ATPase. In some systems, Na⁺/H⁺ exchange activity can indirectly contribute to ${\mathrm{HCO}}_3^{-}$ uptake. However, in corneal endothelium perfused in the presence of ${\mathrm{HCO}}_3^{-}$, inhibition of Na⁺/H⁺ exchange with amiloride could slow recovery of pH_i from acid loading but would have no effect on steady-state pH_i, indicating that there is no significant Na⁺/H⁺ activity in resting cells. The ${\mathrm{Cl}}^{-}∕{\mathrm{HCO}}_3^{-}$ exchanger AE2 is present on the basolateral membrane of fresh dissected endothelium (6), but it is not expressed in cultured endothelium and no ${\mathrm{Cl}}^{-}∕{\mathrm{HCO}}_3^{-}$ exchange activity can be detected in cultured cells (9). On the other hand, reduced expression of the Na⁺-K⁺-ATPase would collapse the Na⁺ gradients that drive ${\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$ cotransport, and we found in the present study (see below) that ouabain treatment significantly inhibited net ${\mathrm{HCO}}_3^{-}$ fluxes. Therefore, we tested whether NBC siRNA treatment might affect Na⁺-K⁺-ATPase expression. Figure 2 shows that Na⁺-K⁺-ATPase expression level did not change 4 days after BCEC cells were treated with NBC1 siRNA, indicating that physiological effects of NBC1 siRNA treatment are not due to nonspecific knockdown of the Na⁺-K⁺-ATPase.

Previous investigations have shown that ${\mathrm{HCO}}_3^{-}$ plays a key role in regulation of endothelial cell pH_i and that NBC1 has an exclusive role in basolateral ${\mathrm{HCO}}_3^{-}$ transport (5, 8, 21, 35). These conclusions are based on experiments showing Na⁺ dependent and DIDS-sensitive, electrogenic ${\mathrm{HCO}}_3^{-}$ transport at the basolateral membrane (36). More recently, data have been presented suggesting that NBC1 is also present on the apical membrane and could have a role in transendothelial ${\mathrm{HCO}}_3^{-}$ flux (13). We used a constant-pH and a constant-CO₂ protocol (5) to examine the apical and basolateral ${\mathrm{HCO}}_3^{-}$ permeabilities of cultured corneal endothelial cells. In the constant-pH protocol, the ${\mathrm{HCO}}_3^{-}$-rich Ringer is replaced with ${\mathrm{HCO}}_3^{-}$-free Ringer. The initial pH_i change is due to rapid CO₂ efflux. Although ${\mathrm{HCO}}_3^{-}$ efflux is occurring concomitantly, a decrease in pH_i is not observed until the [CO₂] has equilibrated (~15 s). The maximum rate of pH_i decrease (averaged over 20 s) is taken as an indirect measure of ${\mathrm{HCO}}_3^{-}$ efflux. In the constant-CO₂ protocol, the ${\mathrm{HCO}}_3^{-}$-rich Ringer (28.5 mM, pH 7.5) is replaced with low-${\mathrm{HCO}}_3^{-}$ (2.85 mM, pH 6.5) Ringer. The initial pH_i drop is predominantly from ${\mathrm{HCO}}_3^{-}$ efflux with a small (15%) contribution from H⁺ flux (5). Figure 3 shows that there was a sharp and rapid acidification due to ${\mathrm{HCO}}_3^{-}$ efflux (−0.004467 ± 0.003187 pH_i/s, n = 6) below the baseline after the initial alkalinization from CO₂ efflux during the constant-pH protocol on the basolateral side; however, there was a significantly slower acidification (−0.00073 ± 0.0000544 pH_i/s, n = 6, P < 0.001) in cells that had been treated with 5 nM siRNA, indicating a more than sixfold decrease in basolateral ${\mathrm{HCO}}_3^{-}$ permeability. When the constant-CO₂ protocol was used to remove the effect of CO₂ fluxes, similar results were obtained. There was a significantly slower drop in pH_i (−0.000812 ± 0.0000918 pH_i/s, n = 6, P < 0.001) in the siRNA-treated cells compared with the control cells (−0.00506 ± 0.000338 pH_i/s, n = 6). Conversely, there was no significant difference in the rate of pH_i decrease on the apical side between the control and siRNA-treated cells, under either the constant-pH or constant-CO₂ protocol. To test that the procedure of introducing foreign siRNA could have nonspecific effects on the measured ${\mathrm{HCO}}_3^{-}$ permeabilities, we transfected the siCONTROL siRNA into BCEC and measured pH_i using the constant-CO₂ protocol. Figure 4 shows that there was no significant difference in ${\mathrm{HCO}}_3^{-}$ permeabilities between control and siCONTROL siRNA-treated cells on either the basolateral or apical side. These results indicate that NBC1 plays a major role in the regulation of basolateral ${\mathrm{HCO}}_3^{-}$ permeability but not apical ${\mathrm{HCO}}_3^{-}$ permeability in corneal endothelial cells.

Fig. 3.

Effect of NBC1 siRNA treatment on basolateral and apical ${\mathrm{HCO}}_3^{-}$ fluxes. A: cells untreated or treated with 5 nM NBC siRNA were perfused in ${\mathrm{CO}}_2∕{\mathrm{HCO}}_3^{-}$-rich Ringer on both sides. Basolateral ${\mathrm{CO}}_2∕{\mathrm{HCO}}_3^{-}$ was totally removed under the constant-pH protocol (1: BF, bicarbonate free), and then, after reequilibration, basolateral ${\mathrm{HCO}}_3^{-}$ concentration was reduced from 28.5 to 2.85 mM under the constant-CO₂ protocol (2: LB, low bicarbonate). This sequence was then repeated on the apical side using the same protocols (3 and 4). 1′–4′ indicate the same procedures using siRNA-treated cells. B: maximum change in intracellular pH (pHi) over the initial 20 s (−dpHi/dt). Values are means ± SE (n = 6). **P ≤ 0.001 vs. control.

Fig. 4.

Effect of siCONTROL siRNA on basolateral and apical ${\mathrm{HCO}}_3^{-}$ fluxes. A: cells untreated or treated with 5 nM siCONTROL siRNA were perfused in ${\mathrm{CO}}_2∕{\mathrm{HCO}}_3^{-}$-rich Ringer on both sides. The basolateral and then apical ${\mathrm{HCO}}_3^{-}$ concentration was reduced from 28.5 to 2.85 mM under the constant-CO₂ protocol. B: maximum change in pH_i over the initial 20 s. Values are means ± SE.

We next focused on the effect of NBC1 on the transendothelial ${\mathrm{HCO}}_3^{-}$ flux in corneal endothelial cells. Figure 5, A and B, shows that basolateral-to-apical exceeded apical-to-basolateral unidirectional ${\mathrm{HCO}}_3^{-}$ fluxes by approximately sixfold in control cells, consistent with a net stroma-to-anterior chamber ${\mathrm{HCO}}_3^{-}$ flux. Increasing cell cAMP concentration with forskolin increases the permeability of the apical CFTR to ${\mathrm{HCO}}_3^{-}$ by twofold without having a significant effect on basolateral permeability (34). Figure 5 shows that 2 μM forskolin increased basolateral-to-apical but not apical-to-basolateral unidirectional ${\mathrm{HCO}}_3^{-}$ fluxes and increased net flux by more than twofold, consistent with the effect on apical permeability. In the NBC1 siRNA treated cells, basolateral-to-apical flux (0.010405 ± 0.000618 pH/s) was significantly slowed compared with the untreated cells (0.0294 ± 0.00319 pH/s), but there was no difference in apical-to-basolateral unidirectional fluxes. When cells were treated with 100 μM ouabain, basolateral-to-apical flux (0.00863 ± 0.000874 pH/s) also was significantly slow compared with the untreated cells. Figure 5C summarizes the results from six separate experiments and shows that the net initial ${\mathrm{HCO}}_3^{-}$ flux was 0.7066 ± 0.009144 mM·min⁻¹·cm², and it was increased to 1.74 ± 0.15 mM·min⁻¹·cm² when cells were stimulated with 2 μM forskolin (P < 0.001). In siRNA-treated cells, net initial ${\mathrm{HCO}}_3^{-}$ flux was reduced threefold to 0.236 ± 0.0024 mM·min⁻¹·cm² (P < 0.005). A similar result was observed in ouabain-treated cells, where net initial ${\mathrm{HCO}}_3^{-}$ flux was reduced to 0.134 ± 0.0336 mM·min⁻¹·cm² (P < 0.005).

Fig. 5.

Effect of NBC1 siRNA on non-steady-state transendothelial net ${\mathrm{HCO}}_3^{-}$ flux. A: cells treated or untreated with 5 nM NBC1 siRNA were initially perfused in ${\mathrm{CO}}_2∕{\mathrm{HCO}}_3^{-}$-rich Ringer on the basolateral and apical sides. LB solution without HEPES containing 1 μM 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF) free acid was then quickly exchanged on the apical side followed by clamping the exit tube on that side. After the initial rise in pH was recorded, the apical solution was replaced with ${\mathrm{CO}}_2∕{\mathrm{HCO}}_3^{-}$-rich ringer. The chamber was then flipped over, and the same experiment was done for the basolateral side. For forskolin or ouabain experiments, LB solution containing 2 μM forskolin or 100 μM ouabain was perfused across the cells for 2 min before clamping on apical or basolateral side as described above. B to A, basolateral to apical flux; A to B, apical to basolateral. B: maximum slope (ΔpH/s) over the first 20 s. Values are means ± SE (n = 6). *P ≤ 0.005 vs. control. **P ≤ 0.001 vs. control. C: transendothelial net ${\mathrm{HCO}}_3^{-}$ flux. Values are means ± SE (n = 6). *P ≤ 0.005 vs. control. **P ≤ 0.001 vs. control.

To obtain a measure of the contribution of NBC1 to steady-state transendothelial ${\mathrm{HCO}}_3^{-}$ flux, BCEC were cultured on 0.2-μm Anopore membrane tissue culture inserts and transfected with siRNA. Four days posttransfection, the cultures were washed with DMEM containing 2% bovine calf serum and 1 μM BCECF free acid; 200 μl of culture medium were placed on the apical side and 300 μl on the basolateral side. Six hours later, samples of apical and basolateral medium were collected and the BCECF fluorescence ratio was determined. Consistent with net basolateral-to-apical ${\mathrm{HCO}}_3^{-}$ flux, control untreated filters showed a net relative alkalinization (ΔpH = apical pH − basolateral pH) of the apical medium (+0.077 ± 0.018 pH unit). As a further control, cultures were treated with ouabain, which was expected to inhibit all Na⁺-dependent active transport processes and reduce the net pH change to zero. However, we found that 100 μM ouabain reversed the pH gradient to −0.073 ± 0.017 pH unit. NBC1 siRNA treatment caused a small alkalinization of the basolateral medium and a larger acidification of the apical medium, producing a net pH change of −0.098 ± 0.032, very similar to ouabain treatment. Figure 6 summarizes these data and indicates that NBC1 plays a major role in transendothelial ${\mathrm{HCO}}_3^{-}$ transport.

Fig. 6.

Effect of NBC1 siRNA on steady-state net ${\mathrm{HCO}}_3^{-}$ flux. Cells cultured on Anopore membrane tissue culture inserts were treated or untreated with 15 nM siRNA. Four days later when cells were confluent, the inserts were washed and 200 μl of DMEM containing 1 μM BCECF free acid were placed in the apical compartment, 300 μl were placed on the basolateral side, and cells were incubated for 6 h. Similarly, non-siRNA-treated inserts were prepared in the same manner but were incubated with 100 μM ouabain for 24 h. Samples from apical and basolateral compartments were collected and pH was measured as described in MATERIALS AND METHODS. ΔpH = apical pH − basolateral pH. Values are means ± SE (n = 6). **P ≤ 0.001 vs. control.

DISCUSSION

We have previously reported that a basolateral ${\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$ cotransporter (NBC1) provides for robust ${\mathrm{HCO}}_3^{-}$ influx in bovine (36) and human (35) corneal endothelial cells. This activity is Na⁺ dependent, DIDS sensitive, and electrogenic. To examine the role of this transporter in membrane ${\mathrm{HCO}}_3^{-}$ permeabilities and transendothelial ${\mathrm{HCO}}_3^{-}$ flux, we inhibited NBC1 expression by using a specific siRNA approach. Our findings are consistent with a major role of a basolateral NBC1 in transendothelial ${\mathrm{HCO}}_3^{-}$ flux but do not support the function of an apical NBC1. This is in contrast to immunofluorescence data recently reported that suggested the presence of an apical NBC1; however, no functional tests for apical ${\mathrm{Na}}^{+}-3{\mathrm{HCO}}_3^{-}$ exchange were performed (13).

The RNA interference (RNAi) technique has been extended to a wide range of commonly used mammalian cells. RNAi is a highly specific form of posttranscriptional gene silencing using 21- to 23-nt siRNA molecules and has been shown to be an effective and specific method for examining functional roles of specific proteins (2, 14, 26). RNAi is very attractive for reducing NBC1 activity, because the primary pharmacological inhibiting agent DIDS is nonspecific and can block other anion transporters and channels. Moreover, the information obtained from selective application of DIDS is difficult to interpret. For example, slowing ${\mathrm{HCO}}_3^{-}$ efflux through a DIDS-sensitive apical transporter or channel should increase pH_i. However, we have observed that apical application of DIDS produces a delayed reduction in pH_i of corneal endothelial cells. This result suggests that little, if any, apical DIDS-sensitive ${\mathrm{HCO}}_3^{-}$ flux is present and that the drug was diffusing to the lateral membrane and inhibiting the basolateral ${\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$ cotransporter. In the present study, we have shown that transfection of siRNA designed to target NBC1 markedly attenuated the expression of NBC1 in BCEC. In general, it is best to use the lowest effective concentration of siRNA, because high doses can have nonspecific effects (32). We found that high concentrations of siRNA (≥20 nM) showed some toxicity to the cells 4 days posttransfection, so we chose 5–15 nM siRNA as the optimal concentration throughout this study. Some studies have indicated that there is no significant nonspecific interference and degradation of endogenous mRNAs in mammalian cells subject to siRNA (12, 32); however, small mismatches with non-target genes can risk inhibiting expression of these genes (33). In addition to the posttranscriptional, sequence-specific effect mediated by the siRNA, the introduction of sequence nonspecific siRNA can sometimes have unpredictable effects on gene expression (14). To exclude these nonspecific effects, we used a siCONTROL nontargeting siRNA to determine whether the activation of the RNAi-induced silencing complex (RISC) could nonspecifically affect NBC1 expression or function. Figure 2 shows that siCONTROL siRNA did not have any effect on NBC1 expression, whereas NBC1 siRNA decreased NBC1 expression but did not affect Na⁺ K⁺-ATPase expression. We chose to examine Na⁺-K⁺-ATPase expression because inhibition of this protein could produce functional changes similar to NBC1 inhibition. Furthermore, transfection with siCONTROL had no functional effect on ${\mathrm{HCO}}_3^{-}$ permeabilities (Fig. 4). These results suggest that siRNA targeted to NBC1 is specific and effective in reducing expression in BCEC.

Basolateral, but not apical, ${\mathrm{HCO}}_3^{-}$ permeability was significantly affected by NBC1 siRNA treatment. This was determined by measuring the rate of pH_i change after removal of ${\mathrm{HCO}}_3^{-}$ from apical or basolateral perfusion chambers. Removal of ${\mathrm{HCO}}_3^{-}$ and CO₂ from the perfusing Ringer solution maintains solution pH (constant pH protocol); however, the pH_i changes induced are confounded by initial CO₂ efflux, which must be cleared before the rate of pH_i decrease can be measured. The constant-CO₂ protocol avoids confounding by CO₂ fluxes but adds instead the reduced pH of the perfusate. Previously, we (37) showed that the pH_i changes measured due to the bath pH changes contributed only ~15% to the initial rate of pH_i change. Nevertheless, both protocols showed similar sixfold drops in basolateral ${\mathrm{HCO}}_3^{-}$ permeability and no effect on apical permeability. The absence of an effect on apical ${\mathrm{HCO}}_3^{-}$ permeability is consistent with previous studies showing no change in apical ${\mathrm{HCO}}_3^{-}$ permeability under low-Na⁺ conditions (5) and an influx of Na⁺ upon removal of apical ${\mathrm{HCO}}_3^{-}$ (36). A siCONTROL siRNA was also transfected into endothelial cells, but this had no effect on ${\mathrm{HCO}}_3^{-}$ permeability in both basolateral and apical sides, indicating that NBC1 siRNA affects the basolateral permeability specifically.

The role of NBC1 in transendothelial ${\mathrm{HCO}}_3^{-}$ flux was measured under non-steady-state and steady-state conditions. Under non-steady-state conditions we introduced a large ${\mathrm{HCO}}_3^{-}$ gradient across the cell monolayer and measured initial unidirectional ${\mathrm{HCO}}_3^{-}$ fluxes, which were adjusted for OH⁻ fluxes. In control cultures, we found that basolateral-to-apical flux was significantly larger than apical-to-basolateral flux. Forskolin, which produces a rapid increase in cell cAMP concentration, has been shown to stimulate endothelial fluid transport (28) and activate an apical CFTR channel, leading to increased ${\mathrm{HCO}}_3^{-}$ permeability across the apical, but not basolateral, membrane (34, 46). We used forskolin as a positive control for the non-steady-state flux experiments. We found (Fig. 5) that 2 μM forskolin increased basolateral-to-apical but not apical-to-basolateral fluxes. The absence of a change in apical-to-basolateral fluxes in the presence of forskolin may be explained by a lack of significant change in driving force for ${\mathrm{HCO}}_3^{-}$ from the outside to the inside of cells across an apical membrane channel. When ${\mathrm{HCO}}_3^{-}$ is totally removed, the membrane potential depolarizes by only 5 mV (21). In apical-to-basolateral fluxes, basolateral ${\mathrm{HCO}}_3^{-}$ concentration is reduced from 28 to 2.85 mM. The initial intracellular ${\mathrm{HCO}}_3^{-}$ concentration is ~20 mM (8), so there is a significant electrochemical gradient for ${\mathrm{HCO}}_3^{-}$ exit across the basolateral membrane, but initially there is no change in chemical gradient across the apical membrane and only a small change in electrical gradient that would still limit ${\mathrm{HCO}}_3^{-}$ influx across an apical channel. On the other hand, NBC1 siRNA inhibited basolateral-to-apical but not apical-to-basolateral fluxes, consistent with NBC1 contributing to a net ${\mathrm{HCO}}_3^{-}$ flux from the stroma to the anterior chamber (19, 25, 43).

Steady-state ${\mathrm{HCO}}_3^{-}$ flux was determined by measuring the relative pH change in apical and basolateral compartments in cultures bathed with identical ${\mathrm{HCO}}_3^{-}$-rich starting culture media. After 6 h of incubation, the apical side was relatively alkaline in control cultures. Similar types of measurements have been made in cultured endothelium using a pH-stat technique (13). Ouabain, which is expected to inhibit the accumulation of ${\mathrm{HCO}}_3^{-}$, actually produced a relative apical acidification. The reason for the apical acidification is not clear. One possibility is acidification by an apical lactate-H⁺ cotransporter (16), which would be unaffected by ouabain’s effect on Na⁺ gradients. siRNA treatment, which also was expected to inhibit apical compartment ${\mathrm{HCO}}_3^{-}$ accumulation, also produced a relative apical acidification consistent with the ouabain treatment and inhibition of net basolateral-to-apical ${\mathrm{HCO}}_3^{-}$ flux.

In summary, our results show conclusively that basolateral NBC1 ($1{\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$) has a significant role in transendothelial ${\mathrm{HCO}}_3^{-}$ flux. The basolateral ${\mathrm{Na}}^{+}-2{\mathrm{HCO}}_3^{-}$ cotransporter loads endothelial cells with ${\mathrm{HCO}}_3^{-}$. ${\mathrm{HCO}}_3^{-}$ flux from the cell across the apical membrane can be significantly stimulated by increasing cAMP concentration, consistent with an increase in apical ${\mathrm{HCO}}_3^{-}$ permeability through CFTR. However, we found no evidence for NBC1 on the apical membrane.