CFTR drives Na+- Formula cotransport in pancreatic duct cells: a basis for defective Formula secretion in CF

Holli Shumaker, Hassane Amlal, Raymond Frizzell, Charles D. Ulrich II, Manoocher Soleimani

Abstract

Pancreatic dysfunction in patients with cystic fibrosis (CF) is felt to result primarily from impairment of ductal HCO3 secretion. We provide molecular evidence for the expression of NBC-1, an electrogenic Na+- HCO3 cotransporter (NBC) in cultured human pancreatic duct cells exhibiting physiological features prototypical of CF duct fragments (CFPAC-1 cells) or normal duct fragments [CAPAN-1 cells and CFPAC-1 cells transfected with wild-type CF transmembrane conductance regulator (CFTR)]. We further demonstrate that1) HCO3 uptake across the basolateral membranes of pancreatic duct cells is mediated via NBC and2) cAMP potentiates NBC activity through activation of CFTR-mediated Cl secretion. We propose that the defect in agonist-stimulated ductal HCO3 secretion in patients with CF is predominantly due to decreased NBC-driven HCO3 entry at the basolateral membrane, secondary to the lack of sufficient electrogenic driving force in the absence of functional CFTR.

  • cystic fibrosis
  • cystic fibrosis transmembrane conductance regulator
  • electrogenic sodium-bicarbonate cotransporter
  • pancreatic dysfunction

cystic fibrosis (CF) remains a major healthcare problem worldwide (29). An autosomal recessive disease, CF results from mutational inactivation of a cAMP-sensitive Cl channel [CF transmembrane conductance regulator (CFTR)] with resultant impairments in the respiratory, pancreatic, hepatobiliary, and genitourinary systems (29). With regard to the pancreas, pancreatic dysfunction is felt to result primarily from impairment of secretin-stimulated ductal Cl and HCO3 secretion (17, 22,23). Experimental and histopathological evidence suggests that the reduction in secretin-stimulated HCO3 secretion from pancreatic duct epithelial cells alters intraductal pH sufficiently to precipitate proteins secreted from acinar cells, resulting in protein plugs, and to disrupt vesicular trafficking in the apical domain of the acinar cell (16, 31). Together, these alterations lead to pancreatic fibrosis and insufficiency in 80% of cases (31).

Although gene therapy has provided the scientific and medical communities with an appealing potential therapeutic option in the treatment of the CF defect, important obstacles have yet to be overcome (2, 21, 42). An alternative hope is that a better understanding of the physiological role(s) of CFTR in various organs will lead to novel strategies capable of correcting the functional defect in the absence of functional CFTR. The currently accepted model of pancreatic ductal HCO3 secretion suggests that1) intracellular HCO3 accumulates due to basolateral diffusion of CO2 and subsequent action of carbonic anhydrase, 2) G protein-coupled receptors (e.g., secretin and vasoactive intestinal peptide) activate cAMP-sensitive CFTR, and3) resultant increases in luminal Cl drive a Cl/ HCO3 exchanger (reviewed in Refs. 11, 18, 34). On the basis of the belief that HCO3 accumulates intracellularly in a passive, essentially unregulated fashion, recent physiological studies have focused on alternate channels capable of conducting Cl into the pancreatic duct lumen (43, 44).

However, a recent report indicated that HCO3 uptake at the basolateral membrane of pancreatic duct cells in guinea pig is Na+ dependent (19). We have recently reported the molecular cloning, cellular distribution, and functional expression of NBC-1, an electrogenically driven Na+- HCO3 cotransporter (NBC) found in human renal epithelial cells (9). Northern blot hybridization utilizing an NBC-1 cDNA probe detected a similarly sized transcript in normal human kidney and pancreas (9). Taken together, these findings raise the possibility that either NBC-1 or a highly homologous electrogenically driven Na+- nHCO3 cotransporter may mediate HCO3 uptake in human pancreatic duct cells (wheren represents an unknown number ≥2). If this is in fact the case, an important testable hypothesis is that cAMP-mediated activation of CFTR creates an electrogenic potential that drives HCO3 entry into pancreatic duct epithelium via a Na+- nHCO3 cotransporter.

In an attempt to identify a model system(s) that would allow us to gain initial insights into the functional role of Na+- HCO3 cotransport in both normal and CF pancreatic duct epithelium, duct cells prototypical of normal or CF duct fragments (CAPAN-1 or CFPAC-1 cells, respectively) were utilized. Interestingly, CAPAN-1 cells, derived from a well-differentiated human ductal pancreatic adenocarcinoma with ductal morphology, represent the only human tumor-derived cell line known to express the cAMP-sensitive CFTR in a confluence-dependent manner and exhibit other functional characteristics prototypical of normal pancreatic ducts (7, 12, 24). CFPAC-1 cells were derived from a well-differentiated human ductal pancreatic cancer resected from a patient with CF due to the F508 mutation in CFTR (32). These cells have been extensively studied and documented to have normal pancreatic ductal physiology, excepting the lack of cAMP-sensitive Clsecretion (15, 32). Furthermore, a recent report from our group indicated the development of stably transfected CFPAC-1 cell lines bearing functional wild-type CFTR (in this paper referred to as CFPAC-WT), and confirmed the recovery of cAMP-sensitive Cl secretion in those cells (15).

In this work, we report the results of studies assessing1) the molecular identity and functional expression of this pancreatic Na+- nHCO3 cotransporter in CAPAN-1, CFPAC-1, and CFPAC-WT cells,2) the effect of membrane depolarization on Na+-dependent, HCO3 cotransport following acid loading of these cells, and 3) the ability of cAMP to influence intracellular pH (pHi) recovery in the same cells either bearing or lacking functional CFTR. Our findings suggest that HCO3 uptake in these prototypical pancreatic duct cell models is electrogenically regulated by cAMP-sensitive CFTR. These findings have important ramifications regarding our understanding of the physiological role of CFTR in these cells and provide important novel insights into the pancreatic pathophysiology of CF.

MATERIALS AND METHODS

Cell lines.

CAPAN-1 and CFPAC-1 cells were obtained from the American Type Culture Collection and cultured as previously described by our group (15, 32). Stably transfected CFPAC-1 cells bearing functional CFTR (CFPAC-WT) were cultured in a similar fashion, excepting the addition of G418 (1 mg/ml) to the medium (15).

RNA isolation and Northern blot hybridization.

Total cellular RNA was extracted from CAPAN-1 and CFPAC-1 cells according to the established methods (13), quantitated spectrophotometrically, and stored at −80°C. Total RNA samples (30 μg/lane) were fractionated on a 1.2% agarose-formaldehyde gel, transferred to Magna NT nylon membranes, cross-linked by ultraviolet light, and baked. Hybridization was performed according to Church and Gilbert (14). The membranes were washed, blotted dry, and exposed to a PhosphorImager screen (Molecular Dynamics, Sunnyvale, CA). A32P-labeled cDNA fragment corresponding to nucleotides 509–3237 of the mRNA encoding human NBC-1 was used as a specific probe.

Functional localization of Na+- HCO3 cotransport.

CFPACWT cells were grown in collagen-coated 30-mm Millicell HA culture dish inserts (porosity 0.45 μm) for 10 days and assayed for the presence of HCO3 -dependent22Na+uptake from the luminal or basolateral surface. The integrity of confluent cell monolayers was assessed by a lack of significant transport of22Na+from the upper compartment to the lower compartment under control conditions (>1% of the22Na+appeared in the lower compartment after 5 min, which is a good indicator of the integrity of the monolayer grown on a filter). Acid loading was accomplished by NH4Cl pulse at each side of the filter (38). The NH4Cl-containing solution consisted of (in mM) 75N-methyl-d-glucamine (NMDG) chloride, 25 KHCO3, and 40 NH4Cl. The uptake medium consisted of (in mM) 10 22NaCl, 105 NMDG chloride, and 25 KHCO3; 0.5 mM ouabain was added to the basolateral uptake medium (to inhibit Na+-K+-ATPase), and 2 mM amiloride was added to the apical or basolateral uptake medium (to inhibit Na+/H+exchange). A Na+-free solution (Na+ was replaced by equimolar NMDG as chloride salt) that in addition contained 2 mM amiloride was used to bathe the side opposite the uptake medium. All solutions contained (in mM) 4 KCl, 1 MgCl2, 1 CaCl2, and 10 HEPES (pH 7.4).

Cell pH measurement.

Changes in pHi were monitored using the pH-sensitive fluorescent dye 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF)-AM as described (3, 38, 39). Cells were grown to confluence on glass coverslips and incubated in the presence of 5 μM BCECF in a solution consisting of (in mM) 115 tetramethylammonium (TMA) chloride, 25 KHCO3, and 10 HEPES (pH 7.4) and gassed with 5% CO2-95% O2.1The monolayer was then perfused with the appropriate solutions in a thermostatically controlled holding chamber (37°C) in a Delta Scan dual-excitation spectrofluorometer (PTI, South Brunswick, NJ). The fluorescence ratio at excitation wavelengths of 500 and 450 nm was utilized to determine pHi values. Calibration curves were established daily by the KCl-nigericin technique. HCO3 -free or HCO3 -containing solutions were used to determine the HCO3 dependence of the transporter. The Na+ and HCO3 dependence of the transporter was examined by monitoring pHirecovery (dpHi/dt, in pH unit/min) from an acid load following NH3/ NH4+ withdrawal, as employed for Refs. 4 and 9. The NH4+ -containing solution consisted of (in mM) 75 TMA (or NMDG) chloride, 40 NH4Cl, and 25 KHCO3. The Na+-containing solution consisted of 115 mM NaCl and 25 mM KHCO3. For voltage-clamp studies of cAMP effect, the Na+-containing solution consisted of (in mM) 45 NaCl, 25 NaHCO3, and either 70 TMA chloride or 70 KCl. All solutions contained (in mM) 0.8 K2HPO4, 0.2 KH2PO4, 1 CaCl2, 1 MgCl2, and 10 HEPES (pH 7.4).

Statistical analyses.

Values are expressed as means ± SE. The significance of differences between mean values was examined using ANOVA.P < 0.05 was considered statistically significant.

RESULTS

Molecular expression of NBC-1 in pancreatic duct cells.

On the basis of the high level of expression of NBC-1 mRNA in whole normal human pancreas (9), we first examined whether this transporter is expressed in cultured human duct cells, CAPAN-1 and CFPAC-1 cells. Northern hybridizations utilizing a32P-labeled probe corresponding to nucleotides 509–3237 of human NBC-1 identified an 8.1-kb transcript in both CAPAN-1 cells and CFPAC-1 cells (Fig.1). The expression of this pancreatic NBC was compared with rat kidney cotransporter (Fig. 1). NBC-1 mRNA levels in the duct cells (Fig. 1) were lower than in the whole pancreas (9).

Fig. 1.

Northern hybridization of Na+- Formula cotransporter-1 (NBC-1) in cultured CAPAN-1 and CFPAC-1 cells.

Functional expression of NBC in pancreatic duct cells.

Cells were grown to confluence on glass coverslips and incubated with BCECF. In the presence of 1 mM amiloride (an inhibitor of Na+/H+exchange), switching to a Na+-containing solution resulted in rapid pHi recovery from acidosis in the presence of HCO3 (Fig.2, A andB). The recovery from cell acidosis was inhibited in both cell lines by 200 μM DIDS (Fig. 2). The results of four separate experiments showed DIDS-sensitive, Na+-dependent HCO3 -mediated pHi recovery rates of 0.11 ± 0.03 and 0.16 ± 0.03 pH unit/min in CAPAN-1 and CFPAC-1 cells, respectively (Fig. 2, C andD). NBC activity in CFPAC-WT cells showed rates very similar to CFPAC-1 cells (0.15 ± 0.03 pH unit/min, n = 4). In the absence of HCO3 but in the presence of Na+ and amiloride, no pHi recovery from acidosis was detected, confirming the HCO3 dependence of pHi recovery in Fig.2. In the absence of Na+ and HCO3 , there was no pHi recovery from acidosis, indicating functional absence of H+-ATPase in both cell lines (Fig.2, E andF).

Fig. 2.

A andB: representative tracings indicating functional expression of NBC in CAPAN-1 (A) and CFPAC-1 (B) cells. TMA, tetramethylammonium.C andD: summary of 4 separate experiments on NBC activity in CAPAN-1 (C) and CFPAC-1 (D) cells.E andF: representative tracings indicating functional absence of H+-ATPase in CAPAN-1 (E) and CFPAC-1 (F) cells. pHi, intracellular pH.

Basolateral localization of Na+- HCO3 cotransport in pancreatic duct cells.

To examine the localization of the NBC in pancreatic duct cells, CFPAC-WT were grown to confluence in 30-mm Millicell HA culture dish inserts (porosity 0.45 μm) for 8–10 days and assayed for the presence of acid-stimulated, HCO3 -dependent22Na+uptake from the upper compartment (luminal surface) or lower compartment (basolateral surface) (see Functional localization of Na+- HCO3 cotransport). As illustrated in Fig.3 A, HCO3 -dependent, acid-stimulated22Na+influx was observed predominantly at the basolateral surface (there is a minimal degree of22Na+influx at the luminal surface that could be due to overgrowth of cells on the filters, which can disturb complete polarization of the cells). Taken together, these experiments indicate that the electrogenic NBC mediates HCO3 influx into the duct cells across the basolateral membrane.

Fig. 3.

A: functional localization of NBC in CFPAC-WT cells. NBC is localized on basolateral membranes of cells, as shown by Formula -dependent22Na+influx studies. B: NBC in duct cells is electrogenic. Representative tracings indicating that activity of NBC is reduced at low external K+concentration ([K+]o) compared with high [K+]o. K+ concentrations of 1.8 and 72 mM were used for low- and high-K+experiments, respectively. Na+-containing solution was 70 mM Na+. TMA orN-methyl-d-glucamine (NMDG), at 70 mM, was used to adjust osmolality of low-K+ solution.

Electrogenicity of Na+- HCO3 cotransport in pancreatic duct cells.

The kidney NBC is highly electrogenic and transports three HCO3 for each Na+ (35, 37, 45), indicating that it carries a net negative charge. As such, this transporter is very sensitive to alterations in the membrane potential (6, 8, 28, 36). Indeed, in the kidney, the Na+- 3HCO3 cotransporter that is located on the basolateral membrane of the proximal tubule and carries HCO3 from the cell to the blood is only driven by inside-negative membrane potential (6, 8, 28, 36). To determine whether alterations in the membrane potential can affect the rate of Na+- HCO3 cotransport in pancreatic duct cells, CFPAC-1 cells were acid loaded by the NH4+ pulse method and switched to a Na+-containing solution (70 mM NaCl) that had either low (1.8 mM) or high (72 mM) K+. As indicated in Fig.3 B, in the presence of valinomycin (which increases permeability of the membrane to K+ and causes the flow of K+ down its gradient), the rate of Na+-dependent HCO3 influx decreased significantly in low-K+ solution (summary of 4 separate experiments showed that NBC activity decreased by 54% in low-K+ solution vs. high-K+ solution,P < 0.02,n = 4). Taken together, these results indicate that changing the membrane potential regulates Na+- HCO3 cotransport activity.

Effects of cAMP on Na+- HCO3 cotransport in pancreatic duct cells.

It has been shown that stimulation of CFTR by secretin, which works via cAMP, increases HCO3 secretion (5, 11,34). In view of the fact that Na+- HCO3 cotransport is highly electrogenic and is regulated by membrane potential (Fig. 3 B), we entertained the possibility that stimulation of ductal HCO3 secretion by CFTR activation could be due to enhanced HCO3 entry at the basolateral membrane via stimulation of the NBC (resulting from depolarization of the membrane potential secondary to Cl secretion by CFTR). We therefore examined the effect of 8-bromoadenosine 3′5′-cyclic monophosphate (8-BrcAMP) on Na+- HCO3 cotransport in cells with functional or mutant CFTR. Cells were exposed to cAMP during the NH4+ withdrawal step. As indicated, addition of 500 μM 8-BrcAMP significantly enhanced the NBC activity in CAPAN-1 cells, increasing it by ∼2.8-fold (Fig.4, A andC). The rate of pHi recovery increased from 0.10 ± 0.02 pH unit/min in the absence of cAMP to 0.28 ± 0.04 pH unit/min in the presence of cAMP (P< 0.001, n = 4). These experiments were performed in the presence of an outward Cl gradient (extracellular Cl was replaced with gluconate). However, potentiation of Na+- HCO3 cotransport activity by cAMP was significantly attenuated when external Cl was present (apparently due to the lack of a favorable chemical gradient for the outward movement of Cl), strongly suggesting that the stimulatory effect of cAMP on the NBC is indirect and is mediated via activation of CFTR (data not shown).

Fig. 4.

A: representative tracings demonstrating effect of cAMP on Na+- Formula cotransport in CAPAN-1 cells. B: representative tracings demonstrating lack of potentiating effect of cAMP on Na+- Formula cotransport in CFPAC-1 cells. C: summary of 4 experiments demonstrating stimulatory effect of cAMP on Na+- Formula cotransport in CAPAN-1 cells. Na+-dependent Formula transport was assayed in presence of an outward Cl gradient. For this purpose, sodium gluconate (rather than NaCl) was used during recovery from acidosis. Na+-free gluconate solutions (i.e., NMDG gluconate) had no effect on pHi recovery (data not shown). Na+-containing solution consisted of 70 mM Na+ (45 mM sodium gluconate and 25 mM NaHCO3) and 70 mM NMDG gluconate. Stimulatory effect of cAMP on NBC was blocked in presence of Cl in medium (or absence of an outward Cl gradient).D: summary of experiments demonstrating lack of stimulatory effect of cAMP on Na+- Formula cotransport in CFPAC-1 cells. Experiments were performed in a manner similar to those on CAPAN-1 cells inA.

Potentiation of Na+- HCO3 cotransport by cAMP is only observed in duct cells expressing functional CFTR.

To determine whether the presence of a functional CFTR is indeed necessary for enhancement of the NBC by cAMP (Fig. 4), CFPAC-1 cells that express the defective CFTR were examined. As shown in Fig. 4,B andD, in the presence of an outward Cl gradient, addition of cAMP had no significant effect on the NBC activity. To determine whether stimulation of Na+- HCO3 cotransport by cAMP is indeed due to functional CFTR and is not a cell-specific property, CFPAC-1 cells stably transfected with wild-type CFTR (CFPAC-WT) were examined. Addition of cAMP resulted in an approximately twofold increase in the rate of Na+- HCO3 cotransport activity in CFPAC-WT cells (Fig.5, Aand C). The rate of pHi recovery increased from 0.15 ± 0.02 pH unit/min in the absence of cAMP to 0.29 ± 0.04 pH unit/min in the presence of cAMP (P< 0.005, n = 4). The stimulatory effect of cAMP on Na+- HCO3 cotransport was prevented when extracellular Cl was present (data not shown).

Fig. 5.

A: representative tracings demonstrating stimulatory effect of cAMP on Na+- Formula cotransport in CFPAC-WT cells. cAMP was added during Formula withdrawal and before switch to Na+-containing solution.B: representative tracings demonstrating effect of voltage clamping on cAMP stimulation of NBC in CFPAC-WT cells.C: summary of experiments demonstrating stimulatory effect of cAMP on Na+- Formula cotransport in CFPAC-WT cells. Na+-dependent Formula transport was assayed in presence of an outward Cl gradient in a manner similar to CAPAN-1 cells (see Fig. 4). For this purpose, sodium gluconate (and not NaCl) was used during recovery from acidosis. Stimulatory effect of cAMP on NBC was blocked in presence of external Cl.D: summary of experiments demonstrating that stimulatory effect of cAMP on Na+- Formula cotransport in CFPAC-WT cells is blocked under voltage-clamped conditions (high [K+]oand valinomycin). Na+-containing solution consisted of 70 mM Na+(45 mM sodium gluconate and 25 mM NaHCO3) and 70 mM potassium gluconate. E: representative tracing demonstrating that addition of cAMP [and subsequent stimulation of cystic fibrosis transmembrane conductance regulator (CFTR)] has no effect on pHi in CFPAC-WT. Experiments were performed in Na+-free solution to keep NBC inactive. Experiments were performed in presence of an outward Cl gradient ([Cl]o = 0 mM). Results of 4 separate experiments were identical.

Potentiation of Na+- HCO3 cotransport by cAMP in pancreatic duct cells is via membrane depolarization.

The above results indicate that cAMP stimulates the NBC in cells expressing only functional CFTR. They further indicate that stimulation of the NBC by cAMP is due to activation of CFTR. Activation of CFTR by cAMP in vitro or in vivo leads to Cl secretion and subsequent cell membrane depolarization (reviewed in Refs. 11, 18, 34). To determine whether alteration in cell membrane potential was responsible for stimulation of Na+- HCO3 cotransport, Na+-dependent HCO3 transport in CFPAC-WT cells was examined in response to cAMP under voltage-clamped conditions. Cells were acidified and exposed to cAMP in the presence of the K+ ionophore valinomycin and high external K+ to clamp the membrane potential (35, 37). As shown in Fig. 5,B andD, the stimulatory effect of cAMP on the NBC was blocked by voltage clamping in high-K+ solution.

CFTR does not carry HCO3 in pancreatic duct cells.

It has been shown that stimulation of CFTR by secretin, which works via cAMP, increases HCO3 secretion (5, 11). One possibility with respect to the underlying stimulatory mechanism of ductal HCO3 secretion by cAMP was that CFTR had the capability to carry HCO3 , analogous to findings in tracheal epithelial cells (27, 33). To determine whether CFTR mediates HCO3 efflux, duct cells were loaded with HCO3 and monitored for pHi in the presence of HCO3 but the absence of Na+ in the media (to keep the NBC inactive). As indicated in Fig. 5 E, addition of cAMP had no significant effect on pHi in the presence of an outward Cl gradient (0 mM external Cl). Similarly, cAMP had no effect on pHi in the absence of an outward Cl gradient. These results indicate that CFTR has minimal affinity for HCO3 in the pancreatic duct cells.

RT-PCR analysis of pancreatic NBC-1 cDNA.

To determine whether the pancreas NBC-1 expresses the protein kinase A (PKA) binding site, a region around the PKA phosphorylation consensus site of the kidney NBC-1 was amplified using RT-PCR, cloned, and sequenced. Accordingly, total RNA from human pancreas was subjected to reverse transcription using an oligo(dT) primer. After the reverse transcription and digestion with RNase H, the pancreatic cDNA was purified on a Microcon-30 ultrafilter (Amicon) to remove the oligo(dT) primers and digested RNA. PCR was then performed using the primers 5′-CAA AGA GTC ACT GGA ACC CTT G and 5′-AAG AGA GAA GCA GAG AGA GCG. Analysis by gel electrophoresis revealed a single band, indicating a single amplification product. The band was excised from the agarose gel, purified, and ligated into a pGEM-T plasmid (Promega). After transformation by electroporation and antibiotic selection, two independent clones were selected. The two clones were grown up, and plasmids were isolated from each for sequencing. Because of the central location of the PKA consensus site within each of the cDNA inserts, it was possible to get high-quality sequence of this region by sequencing from each end of the two plasmids. Thus the sequence of the PKA consensus phosphorylation site was determined by sequencing in both directions from two independent clones. Because bidirectional sequencing was not performed on the flanking regions, some sequence ambiguities remain in those areas. However, all four overlapping sequences showed an intact PKA consensus sequence.

DISCUSSION

Cultured pancreatic duct cells (CAPAN-1 and CFPAC-1) express the NBC-1 mRNA (Fig. 1) and NBC activity (Fig. 2). The NBC is located on the basolateral membrane domain and dictates HCO3 uptakes in duct cells (Fig.3 A). This transporter is highly electrogenic and is regulated by alterations in membrane potential (Fig. 3 B). cAMP potentiates HCO3 uptake via NBC only in duct cells expressing functional CFTR (Figs. 4 and5 A), by depolarizing the cell membrane (Fig. 5 B).

According to the currently accepted model of HCO3 secretion in the pancreatic duct cells, basolateral HCO3 transport is via diffusion of CO2 and subsequent action of cytosolic carbonic anhydrase (5, 11, 34). The resultant carbonic acid dissociates into H+ and HCO3 , with H+ transported back to the blood via a basolateral H+-ATPase and/or a Na+/H+exchanger (5, 11, 34). Our data indicate that intracellular accumulation of HCO3 occurs as a result of direct transport of HCO3 via a basolateral electrogenic NBC rather than by diffusion of CO2. This is consistent with the expression of NBC-1 in the pancreatic duct cells (Fig. 1) and is in agreement with recent reports indicating functional expression of NBC in basolateral membranes of pancreatic duct cells in guinea pig and rat (19).

NBC-1 is expressed in both renal proximal tubule cells (10) and pancreatic duct cells (Figs. 1-4). The pancreatic NBC-1 is apparently a spliced variant of kidney NBC-1 (1). In the kidney proximal tubule cells, the Na+- 3HCO3 cotransporter mediates the transport of Na+ and HCO3 from cell to blood against chemical gradients for both ions, as cellular concentrations of Na+ and HCO3 are less than those of blood (6,8, 28, 36). As such, the only driving force for outward movement of Na+- 3HCO3 cotransport is the favorable membrane potential (inside-negative potential). Indeed, with basolateral membrane potentials at −65 to −70 mV in the kidney proximal tubule cells (8, 28, 36, 45), the stoichiometry of three HCO3 per Na+ allows for the efflux of the Na+- 3HCO3 cotransporter from cell to blood against chemical gradients for both Na+ and HCO3 . Membrane potentials of −60 mV or lower (i.e., −50 or −40 mV) will not be able to overcome the uphill chemical gradients for Na+ and HCO3 (45). In the pancreatic duct cells, basolateral membrane potentials of −35 mV or lower (which has been shown to be the case in the stimulated state) (26) guarantee that the direction of the NBC is from the blood to the cell. In other words, this transporter mediates the influx of HCO3 to duct cells, a phenomenon completely opposite to that in the kidney proximal tubule cells. This change in the direction of the transport results from the activity of the apical CFTR that is absent in the kidney proximal tubule cells.

We have examined the effect of cAMP on NBC-1 in kidney HEK-293 cells (transiently transfected with the NBC-1 cDNA) and found that its activity is decreased in response to cAMP (dpHi/dt decreased from 0.24 to 0.16 pH unit/min in response to cAMP,P < 0.04,n = 4 for each group). Inhibition of Na+- 3HCO3 cotransport in the kidney by cAMP is consistent with published reports on regulation of this transporter in kidney proximal tubule cells (30). Kidney NBC-1 has only a single PKA phosphorylation site (at the 3′ end of the coding region at amino acid 979), which likely mediates the inhibitory effect of cAMP (9). One possibility with respect to the opposite effects of cAMP on Na+- 3HCO3 cotransport activity in the kidney and pancreas (inhibition vs. stimulation, respectively) was that pancreatic NBC lacked the PKA binding site (resulting from either deletion or alternate splicing), leaving the stimulatory effect of membrane depolarization unopposed. Sequence analysis of the 3′ end of the pancreatic NBC-1 cDNA, however, showed an intact PKA consensus sequence in pancreatic NBC-1 (Fig.6).2The stimulatory effect of cAMP on the pancreas cotransporter strongly suggests that the inhibitory phosphorylation effect would have to be superseded by the depolarizing effect of CFTR activation. Alternatively, it is plausible that the inhibitory effect of PKA in the kidney proximal tubule cells is mediated via an intermediary protein (41), and that protein could be absent in the pancreatic duct cells, leaving the stimulatory effect of membrane depolarization unopposed.

Fig. 6.

Sequence analysis of 3′ end of coding region of pancreatic duct NBC-1. HPanc, human pancreas. Protein kinase A binding site at amino acid 979 (shaded box) is present in both cDNAs.

Stimulation of NBC by cAMP (which is also observed at baseline pHi, data not shown) in the pancreatic duct cells expressing functional CFTR is intriguing and highlights the coordinated regulation of apical and basolateral HCO3 transporters in the duct cells. In the currently accepted model of HCO3 secretion, the defect in HCO3 secretion presumably lies at the apical membrane of the duct cells, where decreased secretion of Cl via CFTR inhibits the apical Cl/ HCO3 exchange. It is worth mentioning that, although an apical Cl/ HCO3 exchanger is expressed in the duct cells (20, 46), its role in mediating secretin-stimulated HCO3 secretion remains controversial (20). Indeed, a recent report demonstrated that removal of luminal Cl or addition of DIDS only partially inhibited secretin-stimulated HCO3 secretion (<25%) in the guinea pig pancreatic duct cells (20), strongly suggesting that the apical Cl/ HCO3 exchanger does not play a major role in secretin-stimulated HCO3 secretion. In addition, CFTR also did not show any significant affinity for HCO3 in the duct cells (Fig.7), which would have otherwise explained the ductal HCO3 secretion defect in CF. Taken together, these results suggest that the ductal HCO3 secretion defect in CF has an alternative molecular basis than the currently accepted one.

Fig. 7.

Schematic diagram demonstrating Formula transport in human pancreatic duct cells; + inside cell shows membrane depolarization due to loss of anions secondary to CFTR stimulation. AE, Cl/ Formula exchanger (anion exchanger); n, no. of Formula ≥2.

Our results indicating lack of stimulation of NBC in cells expressing the mutant CFTR are consistent with a decreased basolateral HCO3 transport into the duct cells in patients with CF. Although we acknowledge the need for confirmatory studies in primary cultures of both normal and CF human pancreatic duct cells, our findings in these prototypical cells suggest that the limiting step in ductal HCO3 secretion in CF patients lies at the basolateral membrane of these cells, where decreased electrogenic driving force (resulting from lack of depolarization of the membrane) prevents the entry of HCO3 via the NBC. Accordingly, we propose the following HCO3 transport model for the pancreatic duct cells (Fig. 7). According to this model, secretin increases intracellular cAMP, which then results in the activation of CFTR and secretion of Cl, leading to depolarization of both luminal and basolateral membranes. The depolarization of the basolateral membrane increases the driving force for NBC and, as a result, enhances entry into the duct cells of HCO3 , which is then secreted at the apical membrane (mostly via a transporter that is poorly defined at the present and partially via an apical Cl/ HCO3 exchanger).

Because the identity of the apical transporter mediating the majority of the ductal HCO3 secretion remains poorly understood (20), and because all acid-base transporters seem to be expressed in both normal and CF cells, it is logical to conclude that an alternative strategy to gene therapy for correcting the HCO3 secretion defect in CF patients could be via activation of NBC (either directly or indirectly). According to this hypothesis, stimulation of NBC either directly (i.e., via protein kinase C analogs or angiotensin II) or indirectly (i.e., via cell membrane depolarization) should increase entry into the duct cells of HCO3 , which can be secreted into the lumen via the apical HCO3 transporter(s). Depolarizing agents such as basolateral K+ channel blockers are excellent candidates to indirectly stimulate NBC.

In summary, a basolateral electrogenic NBC dictates HCO3 uptake in the pancreatic duct cells. This cotransporter is stimulated by cAMP only in cells expressing functional CFTR. The stimulation of the NBC by cAMP is mediated via membrane depolarization. We propose that the defect in ductal HCO3 secretion in patients with CF is predominantly due to decreased HCO3 entry at the basolateral membrane as a result of decreased electrogenic driving force and subsequent inhibition of NBC. Future studies should test the possibility that stimulating the NBC (either directly or indirectly) can increase HCO3 entry into the duct cells and therefore alleviate pancreatic HCO3 secretion defect in patients with CF.

Acknowledgments

We acknowledge the excellent contributions of Elizabeth Kopras, Charles Burnham, and Zhaohui Wang.

Footnotes

  • Address for reprint requests: M. Soleimani, Division of Nephrology and Hypertension, University of Cincinnati Medical Center, 231 Bethesda Ave., MSB 5502, Cincinnati, Ohio 45267-0585.

  • These studies were supported by National Institutes of Health Grants DK-46789, DK-52821, and DK-54430 (to M. Soleimani) and CA-74456 (to C. D. Ulrich) and a grant from Dialysis Clinic Inc. (to M. Soleimani).

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

  • 1 To use HCO3 in Na+-free solutions for NH4+ loading and withdrawal steps, we initially attempted to avoid choline bicarbonate, as choline has been shown to have nonspecific effects on pHi. We therefore used KHCO3. However, subsequent studies indicated similar experimental results when either 25 mM choline bicarbonate (and more physiological concentrations of K+) or 25 mM KHCO3 was used.

  • 2 Pancreatic NBC-1 has two consensus PKA binding sites (1). However, the additional phosphorylation site does not play an important role in mediating the stimulatory effect of cAMP on NBC-1, as addition of cAMP had no effect on Na+-dependent HCO3 cotransport in cells expressing the mutant CFTR.

REFERENCES

View Abstract