Studies on Bicarbonate Transporters and Carbonic Anhydrase in Porcine Nonpigmented Ciliary Epithelium

Abstract

purpose. Bicarbonate transport plays a role in aqueous humor (AH) secretion. The authors examined bicarbonate transport mechanisms and carbonic anhydrase (CA) in porcine nonpigmented ciliary epithelium (NPE).

methods. Cytoplasmic pH (pH_i) was measured in cultured porcine NPE loaded with BCECF. Anion exchanger (AE), sodium bicarbonate cotransporter (NBC), and CA were examined by RT-PCR and immunolocalization. AH secretion was measured in the intact porcine eye using a fluorescein dilution technique.

results. Anion exchanger AE2, CAII, and CAIV were abundant in the NPE layer. In cultured NPE superfused with a CO₂/HCO₃ ⁻-free HEPES buffer, exposure to a CO₂/HCO₃ ⁻-containing buffer caused rapid acidification followed by a gradual increase in pH_i. Subsequent removal of CO₂/HCO₃ ⁻ with HEPES buffer caused rapid alkalinization followed by a gradual decrease in pH_i. The rate of gradual alkalinization after the addition of HCO₃ ⁻/CO₂ was inhibited by sodium-free conditions, DIDS, and the CA inhibitors acetazolamide and methazolamide but not by the Na-H exchange inhibitor dimethylamiloride or low-chloride buffer. The phase of gradual acidification after removal of HCO₃ ⁻/CO₂ was inhibited by DIDS, acetazolamide, methazolamide, and low-chloride buffer. DIDS reduced baseline pH_i. In the intact eye, DIDS and acetazolamide reduced AH secretion by 25% and 44%, respectively.

conclusions. The results suggest the NPE uses a Na⁺-HCO₃ ⁻ cotransporter to import bicarbonate and a Cl⁻/HCO₃ ⁻ exchanger to export bicarbonate. CA influences the rate of bicarbonate transport. AE2, CAII, and CAIV are enriched in the NPE layer of the ciliary body, and their coordinated function may contribute to AH secretion by effecting bicarbonate transport into the eye.

Aqueous humor (AH) is formed by the active transport of solutes followed by the osmotic flux of water across the double-layer ciliary epithelium. ^{1 2} This process requires the coordinated action of several different ion transport mechanisms and ion channels. ^{3 4} The ciliary epithelium bilayer consists of two types of cells joined by gap junctions in an apex-to-apex orientation. ⁵ The inner layer of pigmented ciliary epithelium (PE) faces the stromal blood supply, and the outer nonpigmented ciliary epithelium (NPE) faces the AH in the posterior chamber of the eye. Solute transport across the bilayer gives rise to osmotic water movement, resulting in AH formation. The ability of carbonic anhydrase (CA) inhibitors to reduce the rate of AH formation points to an important role for CA. ^{6 7 8 9}  

Mammalian CA is a family of enzymes consisting of at least 10 members, some localized in the cytoplasm and some associated with the cytoplasmic membrane. ^{10 11} All members catalyze the reversible interconversion of CO₂ and HCO₃ ⁻. ^{12 13} In human erythrocytes it has been shown that the speed of the reaction is increased 23,000-fold by CA. ¹⁴ By increasing the availability of HCO₃ ⁻, CA can have a significant effect on bicarbonate and proton transporter mechanisms. Thus, sodium-hydrogen exchanger Na-H exchanger (NHE) 1 function is influenced by CA inhibitors, ^{15 16} as is anion exchanger (AE)-mediated bicarbonate and chloride transport. ^{17 18 19} Here we show the enrichment of AE2, CAII, and CAIV in the porcine ciliary NPE layer. For consistency we also used the porcine eye to examine aqueous secretion and as a source of NPE cells for primary culture. In cultured porcine NPE, studies were conducted to determine the effects of DIDS and acetazolamide on cytoplasmic pH responses to the addition or removal of HCO₃ ⁻ _. We also provide evidence for a reduced AH formation rate in arterially perfused intact eyes treated with DIDS and acetazolamide, inhibitors of bicarbonate transporters and CA, respectively. 

Materials and Methods

Perfused Eye Preparation

Fresh pig eyes were obtained from a local abattoir and perfused based on a similar method described earlier. ^{20 21} The eye was placed in a circulating warming jacket, maintained at 37°C, and covered with an insulated plastic cup. The ophthalmic artery was cannulated, and the eye was perfused with Krebs solution at 37°C (pH 7.4) containing 118 mM NaCl, 4.0 mM KCl, 1.2 mM MgSO₄, 2.0 mM CaCl₂, 25 mM NaHCO₃, 1.2 mM KH₂PO₄, 10 mM glucose, 1.0 mM reduced glutathione, 0.05 mM ascorbate, and 1.8 mM allopurinol. The solution was bubbled with O₂ containing 5% CO₂. Allopurinol, a xanthine oxidase inhibitor, was added to the perfusate to suppress oxidative damage and reperfusion injury. With the use of a peristaltic pump (505S; Watson Marlow, Wilmington, MA), perfusion flow was increased stepwise to 1.5 mL/min. 

The anterior chamber was cannulated with two 23-gauge needles connected to silicon rubber tubing filled with 1.04 mL artificial AH containing 110 mM NaCl, 3 mM KCl, 1.4 mM CaCl₂, 0.5 mM MgCl₂, 0.9 mM KH₂PO₄, 30 mM NaHCO₃, 6 mM glucose, 3 mM ascorbic acid, and 0.0186 mM sodium and fluorescein at pH 7.6 and equilibrated with 95% O₂/5% CO₂. The artificial AH formed a loop that circulated through the pump at 0.2 mL/min from the anterior chamber to a spectrophotometer cuvette (Spectronic 2000; Pharmacia Biotech, Uppsala, Sweden) and back to the anterior chamber. Absorbance was recorded every 5 minutes at 490 nm. The two needles were kept wide apart to optimize fluid mixing. A third 23-gauge needle in the anterior chamber was connected to a water manometer to measure intraocular pressure. 

The AH formation rate was estimated from the rate of fluorescein dilution (decrease in absorbance) that occurred because of the continuous secretion of AH into the eye. After a settling-in period, the plot of log_e (absorbance) versus time (minute) was a straight line whose slope was the rate constant, K_out (minute) of aqueous flow. Test agents were added to the arterial perfusate (i.e., to the stromal side) at a fixed concentration (100 μM for DIDS, 500 μM for acetazolamide). K_out determined 30 minutes before the addition of DIDS or acetazolamide to the arterial perfusate was considered the control value for comparison with K_out, measured in the presence of the test compound. 

Isolation and Culture of Porcine NPE

NPE was established in primary culture using a modification of our previous technique. ²² Porcine eyes were dissected to obtain the entire ring of NPE that remained attached to the vitreous, leaving the pigmented cell (PE) layer attached to the ciliary body. The NPE ring was separated from the vitreous using fine scissors and cut into 1- to 2-mm pieces. NPE from 5 to 7 eyes was pooled and transferred to a 90-mm petri dish containing 15 mL of 0.015% collagenase A and 500 U/mL hyaluronidase (Sigma, St. Louis, MO) in a collagenase buffer containing 66.7 mM NaCl, 13.4 mM KCl, 3.8 mM HEPES, 4.8 mM CaCl₂, pH 7.4. The petri dish was placed for 5 to 7 minutes on a rotary shaker in a 37°C incubator and then removed from the shaker; collagenase and hyaluronidase were neutralized by adding the excess (7 mL) of a 1:1 mixture of newborn calf serum (NCS) and fetal bovine serum (FBS). NPE cells were dispersed by gentle trituration using a round-tipped Pasteur pipette and pelleted by centrifugation at 2000 rpm (670g) for 10 minutes at 4°C. The pellet was dispersed, and the cells were incubated without changing the medium for 3 to 4 days in a small volume of Dulbecco’s modified Eagle’s medium (DMEM; Sigma) containing 10% FBS and 100 IU/mL gentamicin at 37°C in 5% CO₂/95% air. Thereafter, the medium was changed on alternate days. The cells grew to confluence in 7 to 10 days. At confluence, the cells were trypsinized and seeded at a density of 2 × 10⁴ to 5 × 10⁴ cells/cm² for subsequent passages. In the studies reported here, only fourth- passage cells before confluence were used. 

Immunolocalization

Immunolocalization studies were conducted with fresh pig eyes. The cornea was removed, and 4 to 6 mm of the anterior sclera was carefully peeled from the choroid around the globe with a pair of curved scissors. The corneal remnant at the limbus was used as the handle to facilitate this dissection. The whole iris-ciliary body along with the lens and anterior vitreous was then removed from the posterior part of the globe. Tissue was placed in a petri dish (corneal face down) containing ice-cold Ringer solution consisting of 113 mM NaCl, 4.6 mM KCl, 21.0 mM NaHCO₃, 0.6 mM MgSO₄, 7.5 mM d-glucose, 1.0 mM glutathione (reduced form), 1.0 mM Na₂HPO₄, 10.0 mM HEPES, and 1.4 mM CaCl₂, pH adjusted to 7.4, and the lens was removed by cutting the zonules. Vitreous was carefully trimmed, and the iris was dissected away to leave the ciliary body, which was fixed in formalin and used to prepare 5- to 7-μm paraffin sections. Cultured NPE cells also were used for immunolocalization. The cells, which were grown on specially designed chamber slides (Laboratory-Tek II Chamber Slides; Nalge Nunc, Rochester, NY), were washed with PBS containing 1.0 mM MgCl₂ and 0.1 mM CaCl₂ and were fixed in acetone for 2 minutes at room temperature. To probe for the cytoplasmic protein CAII, the cells were permeabilized before fixation using 0.1%Triton X-100 in PBS. 

Tissue sections or cultured NPE cells before confluence were incubated at room temperature for 90 minutes in 10% goat serum in PBS (blocking buffer). Primary antibodies directed against either AE2 (10 μg/mL), CAII (10 μg/mL), or CAIV (20 μg/mL) were added for 24 hours at 4°C. Control specimens received only the blocking buffer. The specimens were washed with PBS and incubated for 24 hours at 4°C with fluorescent secondary antibody (Alexa Fluor 488 or 546 conjugated to either goat anti–rabbit or anti–mouse IgG; 1:200 dilution). Specimens were examined under a microscope (Axiovert 200 M; Carl Zeiss, Thornwood, NY) and photographed with a digital camera (C4742–95; Hamamatsu, Bridgewater, NJ). 

Some samples were prepared for confocal microscopic study. In these cases, nuclear counterstaining with TO-PRO-3 iodide (Invitrogen, Carlsbad, CA) was used to visualize the nucleus, and the images were captured with a confocal microscope (LSM 510 meta-NLO; Carl Zeiss). Fluorescence excitation was achieved with 488-, 543-, and 633-nm laser excitation wavelengths for FITC, rhodamine, and TO-PRO-3 iodide, respectively. 

Reverse Transcription-Polymerase Chain Reaction

Design of Oligonucleotide Probes for Porcine AE1, AE2, AE3, and kNBC1.

No sequence has been published for the porcine ortholog of AE1. The bovine AE1 sequence (GenBank accession no. NM181036) was used to probe for a related sequence in the partially sequenced porcine genome (NCBI). This query resulted in several highly conserved hits located in proximity on chromosome 12 (genomic locales: 98698–98874 bp, 101907–102086 bp, 102522–102711 bp, 105983–106259 bp). Nucleotide sequences in these locales were 83% to 87% identical with the open-reading frame of human AE1 but less than 80% (60%–80%) identical with human AE2 and AE3, suggesting that the mined sequence is that of porcine AE1 (pAE1). Oligonucleotide probes used were sense 5′-GTGACATCACAGACGCCTTGA-3′ and antisense 5′-CTCTGGTTTGCTGACGATCA-3′. The porcine ortholog of AE2 has been fully sequenced (GenBank accession no. AF120099), and the oligonucleotide probes used were sense 5′-AGGAGATCTTCGCCTTCCTC-3′ and antisense 5′-AGCATCCAGGCATTTCATCT-3′. No sequence has been published for pAE3, and a sequence similar to hAE3 could not be found in the porcine genome. Nucleotide sequences conserved among the human, mouse, rabbit, rat, and monkey AE3 cDNA sequences were used to design oligonucleotide probes for pAE3 (sense, 5′-AAGACCTTGGCTGTGAGCAG-3′; antisense, 5′-GCTGCTCCAAGAAAGGCAC-3′). The porcine ortholog of kNBC1 has been cloned (NM001030533), and the oligonucleotide probes used were sense 5′-TCTTTTGCCTCTTTGCTGGT-3′ and antisense 5′-GCTTGAACTCACTTGGCACA-3′. 

RNA Isolation and RT-PCR.

Total RNA from porcine renal cortex, cardiac muscle, native NPE, and primary cultures of NPE at passage 4 was isolated (RNA-Bee; Tel-Test, Inc., Friendswood, TX). Porcine kidney was used as a positive control for AE1 and AE2, whereas porcine myocardial tissue served as a control for AE3. One microgram of total RNA was reverse transcribed with a reverse transcription kit according to the manufacturer’s protocol (QuantiTect; Qiagen, Valencia, CA). This protocol includes a step to remove genomic DNA. Four microliters of cDNA was used in the PCR reaction. PCR components were assembled, and a single denaturing step was performed for 2 minutes at 94°C. This was followed by 35 cycles of 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 2 minutes. A final elongation step of 7 minutes (72°C) was included after the last cycle. PCR products were separated on 1% agarose gels and visualized with ethidium bromide. Amplified products were purified with a gel extraction kit (QIAquick; Qiagen), and sequences were confirmed with the use of a DNA analyzer (3730xl; Applied Biosystems, Foster City, CA) at the University of Arizona sequencing facility. The partially cloned pAE1 and pAE3 sequences were 87% and 92% identical with the hAE1 and hAE3 sequences, respectively. The cloned pAE2 sequence was identical with the previously published pAE2 sequence. 

Measurement of Cytoplasmic pH

NPE cells grown to preconfluence on 35-mm plastic dishes (Corning, Corning, NY) were loaded for 10 minutes with the pH-sensitive dye BCECF-AM (5.0 μM), as described, ²³ placed in a temperature-controlled perfusion microincubator (PDMI-2; Harvard Biosciences, Holliston, MA) on the stage of an upright epifluorescence microscope (Eclipse 80i; Nikon, Tokyo, Japan), and superfused with a HEPES buffer containing 137 mM NaCl, 4.5 mM KCl, 6.0 mM d-glucose, 1.0 mM MgCl₂, 1.5 mM CaCl₂, and 10.0 mM HEPES, adjusted with NaOH to pH 7.35. The flow rate was 3.0 mL/min. Cytoplasmic pH (pH_i) was recorded with the use of an imaging system (Incyt; Intracellular Imaging Inc., Cincinnati, OH) with an emission wavelength of 535 nm and alternating excitation wavelengths of 488 nm and 460 nm. pH_i was calculated from the fluorescence intensity ratio I₄₈₈/I₄₆₀. 

Cells were first superfused with the HEPES buffer for 5 minutes to obtain a stable pH_i baseline, after which the superfusate was switched to bicarbonate/CO₂ buffer containing 117 mM NaCl, 4.5 mM KCl, 20 mM NaHCO₃, 6.0 mM d-glucose, 1.0 mM MgCl₂, 1.5 mM CaCl₂, and 10.0 mM HEPES, adjusted to pH 7.35 and equilibrated by gassing with 5% CO₂ and 95% air. In some experiments either a sodium-free or a low-chloride buffer was used. The sodium-free bicarbonate/CO₂ solution contained 117 nM choline chloride, 20 mM choline bicarbonate, 4.5 mM KCl, 6.0 mM d-glucose, 1.0 mM MgCl₂, 1.5 mM CaCl₂, and 10 mM HEPES. The low-chloride buffer contained 117 mM sodium gluconate, 4.5 mM potassium gluconate, 20 mM NaHCO₃, 6.0 mM glucose, 1.0 mM MgCl₂, 2.5 mM CaCl₂, and 10 mM HEPES. To prepare sodium-free/bicarbonate-free or low-chloride/bicarbonate-free buffers, the bicarbonate was omitted and replaced with an equimolar amount of sodium gluconate or choline chloride, respectively. 

Reagents

Mouse anti–human CAIV monoclonal antibody was obtained from R&D Systems (Minneapolis, MN). Rabbit anti–CAII polyclonal antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti–AE2 polyclonal antibody was obtained from Alpha Diagnostics (San Antonio, TX). Secondary antibodies used to probe the bound primary antibodies were Alexa Fluor 488 goat anti–rabbit IgG and Alexa Fluor 546 goat anti–mouse IgG (Invitrogen). Acetazolamide, methazolamide, 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS), and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich. BCECF (2′,7′-bis(2-carboxyl)-5 ⁶ -carboxyfluorescein-acetoxyethyl ester) was purchased from Invitrogen. All other chemicals were purchased from Sigma-Aldrich. Stock solutions of test compounds were prepared in DMSO before addition to the cell superfusate or to the whole eye perfusate. Control solutions received DMSO only. 

Statistical Analysis

A two-sample t-test was used to analyze unpaired data, and a paired t-test was used to compare paired samples. One-way analysis of variance (ANOVA) followed by Bonferroni post hoc multiple comparison tests was used to compare differences for more than two groups of data. P < 0.05 was considered significant. 

Results

Influence of DIDS and Acetazolamide on Aqueous Humor Formation

The rate of AH formation was measured in the intact, arterially perfused eye. Under control conditions, the rate constant for AH formation (K_out/min × 10⁻⁴) was 28.0 ± 2.0 (n = 5). To obtain the rate constant for AH formation under drug-treated conditions, the anion transport inhibitor DIDS or the CA inhibitor acetazolamide was added to the arterial perfusate, representing stromal application. DIDS was added at a final concentration of either 10 or 100 μM, and acetazolamide was added at a final concentration of 500 μM. At a concentration of 100 μM, DIDS significantly reduced the rate constant for AH to 20.0 ± 2.0 (n = 5; Fig. 1A ). The CA inhibitor acetazolamide (500 μM) produced more pronounced reduction of the rate constant for AH formation to 15.0 ± 2.0 (n = 6; Fig. 1B ). 

Expression and Localization of Chloride-Bicarbonate Exchanger (AE2) and CA in Porcine NPE

Given that DIDS and acetazolamide reduce the rate of AH formation, experiments were conducted to examine chloride-bicarbonate exchanger and CA expression in ciliary epithelium. Immunolocalization and RT-PCR studies were carried out to examine chloride-bicarbonate exchangers (AEs) in the ciliary epithelium bilayer and cultured NPE. AE2 was detected in the NPE cell layer, where it was enriched at the basolateral membrane (Fig. 2A) . AE2 immunoreactivity was also present in cultured NPE (Fig. 2B) . AE1 and AE3 were not detected (result not shown). RT-PCR studies showed the presence of AE2 mRNA in native and cultured porcine NPE. AE1 or AE3 mRNA was not found (Fig. 3) . CAIV was localized to the surface of the NPE (Fig. 4)but was not detectable in the pigmented cell layer. However, the presence of heavy pigmentation in the pigmented cells might have masked the detection of fluorescence in these cells. In contrast, CAII distribution in the ciliary body was widespread, as revealed by laser confocal microscopy (Fig. 5A) . CAII was abundant in the cytoplasm of the NPE. CAIV and CAII immunoreactivity also was detected in the cultured NPE by laser confocal (Fig. 4B)and epifluorescence microscopy (Fig. 5B) , respectively. 

Cytoplasmic pH Response of Cultured NPE to HCO₃ ⁻/CO₂ Addition

Figure 6shows a typical cytoplasmic pH (pH_i) response of cultured NPE when the bathing medium was switched for 5 minutes from bicarbonate-free HEPES-buffered solution to a HCO₃ ⁻/CO₂-buffered solution. Baseline pH_i in HCO₃ ⁻/CO₂-free HEPES buffer was 7.25 ± 0.13. When the superfusate was switched to HCO₃ ⁻/CO₂ buffer, there was a rapid decrease in pH_i (0.4 ± 0.02 U; n = 10) followed by gradual recovery toward baseline. Subsequent replacement of HCO₃ ⁻/CO₂ medium with HEPES-buffered solution caused pH_i to increase sharply (0.4 ± 0.04 U; n = 10) and then gradually recover toward baseline. The pH_i response was examined under several different conditions: in sodium-free buffer, in low-chloride buffer, in the presence of DMA, in the presence of DIDS, and in the presence of acetazolamide or methazolamide. 

The addition of HCO₃ ⁻/CO₂ acidified the cells because of rapid diffusion of CO₂ into the cells, its CA-mediated conversion to H₂CO₃, and subsequent dissociation of H₂CO₃ to HCO₃ ⁻ and H⁺. After this, we suggest the gradual alkalinization toward baseline occurred as the result of bicarbonate entry. DIDS at a concentration of 100 μM prevented the gradual alkalinization (Figs. 7 8) . Sodium-free buffer also abolished the gradual alkalinization, and, on replacement of external sodium, the rate of pH_i recovery returned to normal (Fig. 8) . However, the rate of gradual alkalinization was not significantly different in low-chloride buffer or in the presence of 100 μM DMA (Fig. 8) . These findings are consistent with the notion that the gradual increase of pH_i results from bicarbonate entry through a sodium-dependent anion transporter. Acetazolamide (500 μM) and methazolamide (100 and 500 μM) significantly reduced the rate of gradual alkalinization (Fig. 8) . 

The removal of external HCO₃ ⁻/CO₂ caused pH_i to increase sharply because of the rapid exit of CO₂. After this, we suggest the gradual acidification toward baseline resulted from chloride-sensitive HCO₃ ⁻ efflux. The evidence is as follows: DIDS significantly inhibited the gradual acidification (Fig. 9) . The rate of gradual acidification also was reduced in low-chloride buffer (Fig. 9) . The rate of gradual acidification measured in the presence of the sodium-hydrogen exchange inhibitor, DMA (100 μM), was not significantly different from the control rate (Fig. 9) . These findings are consistent with the gradual reduction of pH_i because of bicarbonate exit through a sodium-independent anion exchanger. CA inhibitors acetazolamide (500 μM) and methazolamide (100 and 500 μM) completely inhibited gradual acidification (Fig. 9) . 

Effect of DIDS and Acetazolamide on Baseline Cytoplasmic pH

Given that DIDS blocked and acetazolamide reduced the rate of cytoplasmic pH recovery in bicarbonate-containing buffer and that both drugs reduced AH formation in isolated intact eye preparations, we studied the effect of these drugs on the baseline cytoplasmic pH of cultured NPE. DIDS (100 μM) caused a significant progressive reduction of baseline cytoplasmic pH (Fig. 10) . After 20 minutes, when AH formation began in the experiments on intact eye, DIDS lowered the pH by approximately 0.6 U. In cells exposed to acetazolamide (500 μM), pH_i was consistently slightly lower than control pH_i, but at any one time point the difference was not significant (Fig. 10) . A slight and gradual drift in cytoplasmic pH occurred in control cells, possibly because of dye bleaching. 

Discussion

Three lines of evidence point to expression of the AE2 chloride-bicarbonate exchanger in porcine NPE: RT-PCR detection of mRNA, protein immunolocalization, and chloride-sensitive pH responses. Consistent with previous findings on the human ciliary body, ²⁴ RT-PCR detected neither AE1 nor AE3. The ability of the ciliary epithelium bilayer to form aqueous humor is determined by the location of transport proteins, and immunolocalization studies revealed the expression of AE2 in the NPE layer. AE2 appeared most dense at the NPE basolateral surface, but it was not strictly limited to the cell border. The apparent cytoplasmic signal could stem from nonspecific antibody binding or the intracellular presence of AE2-trafficking vesicles. AE2 expression is consistent with earlier functional evidence for an electroneutral sodium-independent and DIDS-sensitive Cl⁻/HCO₃ ⁻ exchange mechanism in native rabbit NPE. ²⁵ Together with AE2, the porcine NPE layer also displayed abundant CAII, which appeared in the cytoplasm, and CAIV, which was localized to the membrane. To our knowledge, this is the first report on the localization of AE2 in native porcine ciliary epithelium, but the presence of CA has been demonstrated earlier in rabbit, monkey, and human ciliary body with the use of histochemical methods. ^{26 27 28} In human NPE, CA was reported at the basal and lateral membranes and in the cytoplasm. ²⁸ In addition to CAII, other CAs reported in human ciliary body include the membrane-bound isoforms CAIX and CAXII. ²⁹ Western blot analysis and functional studies using a membrane-impermeable CA inhibitor pointed to the presence of the membrane-bound CAIV in rabbit NPE. ³⁰ Matsui et al. ⁹ suggested that CAIV could be linked to chloride/bicarbonate exchanger function in the NPE. 

With the use of cultured porcine NPE, we were able to examine anion transport indirectly by measuring pH_i responses. Although significant differences between native and cultured cells are likely, porcine NPE in primary culture maintained the expression of AE2, CAII, CAIV, and Na⁺-HCO₃ ⁻ cotransporter (NBC), as judged by RT-PCR and immunolocalization. Earlier we showed similar patterns of nitric oxide synthase (NOS1, NOS2, and NOS3 isoforms), NaK-ATPase (alpha 1, alpha 2, and alpha 3 isoforms), ²² NHE (NHE1, NHE3, and NHE4 isoforms), aquaporin (AQP-1 and AQP-4), and connexins (Cx50 and Cx43; Shahidullah M, Delamere NA, unpublished observation, 2007) in cultured and native porcine NPE. 

In cells superfused with a CO₂/HCO₃ ⁻-free HEPES buffer, exposure to a CO₂/HCO₃ ⁻-containing buffer caused rapid acidification followed by a gradual increase in pH_i. Subsequent replacement of CO₂/HCO₃ ⁻ with HEPES buffer caused rapid alkalinization followed by a gradual decrease in pH_i. A similar response to CO₂/HCO₃ ⁻ addition and removal was reported by Wolosin et al. ^{25 31} in native rabbit NPE cells, suggesting the initial rapid pH_i changes likely resulted from rapid entry or exit of CO₂. In the present study, we focused on the subsequent slower pH_i changes that are dependent, at least in part, on bicarbonate transport. 

A gradual pH_i increase toward baseline was observed after rapid acidification caused by the addition of CO₂/HCO₃ ⁻. Alkalinizing systems reported in the ciliary epithelium include NBC, ³² NHE, ³³ and vacuolar H-ATPase (V-ATPase). ^{34 35} Here, the rate of alkalization was completely inhibited by Na-free solutions and DIDS but was not significantly altered in cells exposed to low-chloride solution or to the NHE inhibitor DMA. Our results are consistent with an Na⁺-HCO₃ ⁻ cotransporter that enables external bicarbonate to enter the cell. This notion fits with an earlier proposal that an Na⁺-HCO₃ ⁻ cotransporter and the Cl⁻/HCO₃ ⁻ exchanger are the two principal determinants of NPE cytoplasmic pH. ²⁵ The lack of sensitivity to DMA suggests NHE-mediated proton export does not contribute to the observed alkalinization response. This apparently contrasts with our previous finding that porcine native and cultured NPE express abundant quantities of NHE1, NHE3, and NHE4 and that DMA significantly inhibits pH_i recovery in this cell after acidification by a 20-mM ammonium chloride pulse (Shahidullah M, et al. IOVS 2007;48:ARVO E-Abstract 3997). The explanation may lie in the fact that the smaller degree of acidification caused by exposure to CO₂/HCO₃ ⁻ (approximately 0.4 pH U below baseline) compared with a 20-mM ammonium chloride pulse (>1 pH U below baseline) is insufficient to cause the activation of NHE. It is well known that NHE activity is regulated primarily by pH_i and increases markedly in response to intracellular acidosis. ³⁶ Such NHE activation is thought to occur through the interaction of H⁺ with an allosteric modifier site within the transport domain. ^{37 38} The ability of DIDS to completely inhibit the NPE alkalinization response after acidification by the addition of CO₂/HCO₃ ⁻ argues against the contribution of NHE or H-ATPase to the pH_i increase observed in the present studies. 

Gradual acidification toward baseline was observed after the rapid increase of pH_i on the removal of extracellular CO₂/HCO₃ ⁻ and the return to HEPES buffer. The rate of gradual acidification was inhibited by DIDS and low-chloride solution. These results are consistent with a chloride-bicarbonate exchanger, such as AE2, that enables bicarbonate to exit the cell in exchange for external chloride entry. Removal of extracellular chloride results in the reversal of the exchanger because of a favorable gradient for AE2-mediated chloride efflux in the presence of ample extracellular bicarbonate. Gradual acidification could be abolished by a CA inhibitor, either acetazolamide or methazolamide. The sodium-hydrogen exchange inhibitor DMA (100 μM) had no effect on gradual acidification. Our results are consistent with findings in other tissues in which AE2-mediated bicarbonate export effects recovery from alkaline loading. ^{39 40} The sensitivity of the gradual acidification to acetazolamide or methazolamide can be explained if bicarbonate transport is rate limited by the availability of cytoplasmic HCO₃ ⁻ and can be inhibited by the build-up of extracellular HCO₃ ⁻. On this basis, cytosolic CAII might provide the “push” for bicarbonate transport by making HCO₃ ⁻ available to the cytoplasmic face of AE2, and CAIV-catalyzed conversion of HCO₃ ⁻ to CO₂ in the extracellular unstirred layer might provide the “pull” by diminishing the concentration of HCO₃ ⁻ at the extracellular AE2 face. Such a push-pull mechanism, resulting from CA activity on both sides of the NPE basolateral membrane, could accelerate AE2-mediated bicarbonate transport into the eye. The pH alkalinization response described, which we propose is mediated by bicarbonate uptake through the Na⁺-HCO₃ ⁻ cotransport, was inhibited only partially by the CA inhibitors acetazolamide and methazolamide, possibly reflecting the greater availability of HCO₃ ⁻ in the extracellular solution compared with cytoplasm. 

The fact that gradual alkalinization after cytoplasmic acidification was abolished by DIDS and sodium-free solution but was insensitive to extracellular chloride removal indicates that alkalinization was caused by NBC-mediated bicarbonate entry into the cells. Because AE2 localized mostly on the basolateral surfaces of the NPE cells and because AE2 enabled bicarbonate to exit the cell in exchange for external chloride entry, it is tempting to speculate that the NBC might be localized on the apical membrane of the NPE cells. Apical localization of NBC has been reported in other secretory tissues, such as the human parotid and sublingual duct ⁴¹ and the rat pancreatic duct ⁴² epithelium. 

The ability of acetazolamide and DIDS to abolish HCO₃ ⁻ export by cultured porcine NPE is interesting in light of the ability DIDS and acetazolamide to reduce AH secretion in the intact porcine eye. These findings are consistent with a contribution of bicarbonate transport to AH formation by the porcine eye. Acetazolamide reduced the rate of aqueous humor secretion by 40%, and DIDS inhibited AH secretion by 25%. DIDS previously had been reported to reduce aqueous formation in the bovine eye. ²⁰ DIDS significantly reduced basal cytoplasmic pH in cultured NPE by lowering pH_i by approximately 0.6 pH U after 20 minutes. In the experiments on the perfused intact eye, a 20-minute interval was allowed to establish the drug effect before measuring the effect of DIDS on the AH formation rate. DIDS is not a selective inhibitor of Cl⁻-HCO₃ ⁻ exchangers such as AE2. It also inhibited Na-HCO₃ ⁻ cotransport ^{24 43} and chloride channels. ⁴⁴ The CA inhibitor acetazolamide had little effect on basal cytoplasmic pH in the cultured NPE, even though it had a robust effect on reducing the AH formation rate in the perfused intact eye. One explanation for this apparent discrepancy might be that in vitro and with continuous CO₂ bubbling of the bathing medium, sufficient hydration of CO₂ might occur in the absence of the catalysis by the CA. Given that CO₂ is membrane permeable, such noncatalytic hydration would be equally effective in the bathing medium and in the cytoplasm of cultured NPE. Perhaps this could help maintain the normal HCO₃ ⁻ transport and acid-base balance in the cultured cells. In the intact eye preparation or in vivo, aqueous humor CO₂ equilibration might be less efficient; hence, sufficient hydration of CO₂ would be more dependent on CA. Consequently, the inhibition of CA would cause insufficient hydration of CO₂ and, therefore, insufficient production of HCO₃ ⁻ inside the cells for transport into the aqueous by basolateral AE2. On this basis, the inhibition of AE2 by DIDS would inhibit the transport of bicarbonate into the aqueous in exchange for Cl⁻. The inhibition of CA would cause the depletion of HCO₃ ⁻ required for transport by the AE2 and would have a similar effect—less bicarbonate transport to the aqueous and lower rate of AH formation. The reduction of pH_i, observed in response to DIDS in the present experiments, may also slow down the overall metabolic rate in the NPE in a way that leads to the inhibition of AH formation. 

In contrast to the minimal effect of acetazolamide on pHi in porcine NPE in the present study, a reduction in pH_i by 0.2 U was detected in rabbit-transformed NPE, ³⁰ possibly because of a difference in the contribution of bicarbonate transport to the formation of AH in different species. For example, it had been shown previously that bicarbonate depletion in the bathing medium of rabbit ciliary body preparation completely reversed electrical polarity, ⁴⁵ whereas the same maneuver only inhibited the short circuit current (I _sc) by approximately 30% in the ox ⁴⁶ and by approximately 90% in the pig. ⁴⁷ According to these results, the importance of HCO₃ ⁻ transport in the porcine ciliary body is between that of rabbit and ox. 

In summary, our results suggest that porcine NPE uses an Na⁺-HCO₃ ⁻ cotransporter to import bicarbonate and a Cl⁻/HCO₃ ⁻ exchanger to export bicarbonate. CA influences the rate of bicarbonate transport. The Cl⁻/HCO₃ ⁻ exchanger AE2 and the carbonic anhydrases CAII and CAIV are enriched in the NPE layer of the ciliary body, and their coordinated function may contribute to aqueous humor secretion by effecting bicarbonate transport into the eye. 

 Supported by National Institutes of Health Grant EY006915.

 Submitted for publication June 24, 2008; revised September 12 and October 23, 2008; accepted February 27, 2009.

 Disclosure: M. Shahidullah, None; C.-H. To, None; R.M. Pelis, None; N.A Delamere, None

 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.

 

Effects of DIDS (10 μM, 100 μM; A) and acetazolamide (500 μM); B) on AH formation rate measured in the porcine isolated perfused whole eye preparation. AH formation was measured by a fluorescein dilution technique. Results are expressed as a rate constant (K_out/min × 10⁻⁴) and are shown as mean ± SEM of five independent experiments for each condition. The rate measured during the first 30 minutes before the addition of drug was taken as the control value. After the addition of drug, a 20-minute period was allowed to establish the drug effect, and the rate was measured over the subsequent 60 minutes. Significant differences from control are indicated by **P < 0.01 (A) and ***P < 0.001 (B).

Figure 2.

 

Immunolocalization of Cl⁻/HCO₃ ⁻ exchanger AE2 in the porcine ciliary body, where it appears on the NPE basolateral membrane (A). AE2 also was detected in cultured NPE (B; fourth-passage cells are shown). Negative controls, in which the primary antibody was replaced by PBS, show no staining. (A, B) Original magnification, 200×.

Figure 3.

 

RT-PCR for AE1, AE2, AE3, and kNBC1 in native and primary cultures of porcine NPE. Porcine renal cortex (kidney) served as positive control for AE1 and AE2, and porcine cardiac muscle (heart) served as positive control for kNBC1 and AE3. Amplified products were separated on 1% agarose gels and visualized with ethidium bromide. The resultant amplified cDNA products were gel purified, and their sequences were confirmed with DNA sequencing.

Figure 4.

 

Immunolocalization of carbonic anhydrase IV in the porcine ciliary body, where it appears on the NPE membrane (A). Original magnification, 200×. CAIV also was detected in cultured NPE by laser confocal microscopy (B; fourth-passage cells are shown). Negative controls, in which the primary antibody was replaced by PBS, show no staining. Original magnification, 400×.

Figure 5.

 

Immunolocalization of CAII in the porcine ciliary body by laser confocal microscopy, where it appears within the cytoplasm of the NPE (A). CAII also was detected in cultured NPE (B; fourth-passage cells are shown). Negative controls, in which the primary antibody was replaced by PBS, show no staining. (A, B) Original magnification, 200×.

Figure 6.

 

Typical pH_i response of BCECF-loaded porcine NPE cells to the addition and subsequent removal of HCO₃ ⁻/CO₂ from the bathing solution. The cells first were superfused with bicarbonate-free HEPES buffered Krebs solution for 5 minutes to obtain a stable baseline. The superfusate was then switched to HCO₃ ⁻/CO₂ buffer. This caused a rapid decrease in pH_i, which gradually recovered toward baseline. Subsequent removal of HCO₃ ⁻/CO₂ and replacement with HEPES buffer caused a rapid pH_i increase followed by a gradual recovery toward baseline.

Figure 7.

 

Typical pH_i response of BCECF-loaded NPE to the addition and subsequent removal of HCO₃ ⁻/CO₂ from the bathing solution in the presence of DIDS (100 μM). The cells first were superfused with bicarbonate-free HEPES buffered Krebs solution for 5 minutes to obtain a stable baseline, and then the superfusate was switched to HCO₃ ⁻/CO₂ buffer for 5 minutes.

Figure 8.

 

Effect of DIDS (100 μM), sodium-free buffer, low-chloride buffer, CA inhibitors acetazolamide (500 μM) and methazolamide (100 μM, 500 μM), and sodium-hydrogen exchange inhibitor DMA (100 μM) on the rate of gradual alkalinization toward baseline after rapid pH_i decrease caused by the addition of HCO₃ ⁻/CO₂ to the bathing medium. Results are the mean ± SEM of data from 5 or 10 independent experiments. Significant differences from control are indicated by *P < 0.05, **P < 0.01, and ***P < 0.001.

Figure 9.

 

Effect of DIDS (100 μM), low-chloride buffer, CA inhibitors acetazolamide (500 μM) and methazolamide (100 μM, 500 μM), and sodium-hydrogen exchange inhibitor DMA (100 μM) on the rate of gradual acidification toward baseline after the rapid pH_i increase caused by the removal of HCO₃ ⁻/CO₂ and replacement with HEPES buffer. Results are mean ± SEM of data from 6 or 10 independent experiments. Significant difference from control is indicated by ***P < 0.0001.

Figure 10.

 

Effect of DIDS and acetazolamide on baseline cytoplasmic pH of porcine cultured NPE. BCECF-loaded cells were first superfused with bicarbonate-containing buffer for 3 minutes to establish baseline cytoplasmic pH. At this point, 100 μm DIDS, 500 μM acetazolamide, or vehicle 0.1% dimethyl sulfoxide was introduced (arrow), and data collection was continued for another 20 minutes. Results are shown as mean ± SEM of 7 to 10 independent experiments. At the final time point, pH_i in DIDS-treated cells showed a significant difference from control, indicated by ***P < 0.001.