CFTR and Bicarbonate Secretion to Epithelial Cells
Abstract
Defective HCO3– and fluid secretion are hallmarks of thepathophysiology of the pancreas of cystic fibrosis patients.Recently, impaired HCO3– secretion has been shown in mosttissues known to express the cystic fibrosis transmembrane conductanceregulator (CFTR). New results suggest that CFTR plays an importantrole in the transcellular secretion of HCO3–.
Introduction
The cystic fibrosis transmembrane conductance regulator (CFTR)plays a crucial role in maintaining fluid secretion of epithelialcells of the airways and the intestine. Defective CFTR leadsto an imbalance between fluid absorption and secretion in thelungs of cystic fibrosis (CF) patients, resulting in a relativelydehydrated mucus layer on the airways. However, the onset ofclear symptoms of impaired lung function remains highly variable.A striking contrast can be found when one examines the exocrinepancreas. Among all CF patients, 70–90% are born withpancreatic insufficiency, which means that >98% of the pancreaticcapacity is already lost (17). Even in the seemingly pancreatic-sufficientpatients, the ratio between alkaline fluid and secreted digestiveenzymes is significantly decreased (8). Clinicians have beenusing the amount of residual pancreatic function to classifyCF patients into severe and mild cases. Under physiologicalconditions, the secreted HCO3–-rich fluid and electrolytesserve to flush the digestive enzymes from the acini and ductsof the pancreas. Thus impaired HCO3– secretion resultsin poor clearance of the digestive enzymes, and their prematureactivation eventually causes the destruction of the pancreasin CF. During the past few years, it has been shown that a similardefect in HCO3– secretion can also be found in the smalland large intestine (16) and, surprisingly, also in the airwaymucosa (18). We would assert that a similar sequela as in thepancreas follows from impaired HCO3– secretion in thesubmucosal glands and airways of CF patients. Analogous to thepancreas, the submucosal glands secrete mucins, protease inhibitors,antibiotic peptides, and enzymes that must be flushed from theglands onto the airway surface epithelium. Moreover, the physicalproperties of mucus are intrinsically dependent on the electrolytecomposition of the fluid. Thus the results of recent studieson submucosal gland serous cells should provide important newinsights into the mechanism of HCO3– secretion in airwayepithelial cells and have significant implications in the futuretreatment of CF.
HCO3– secretion in the exocrine pancreas
The human pancreas is the most severely affected organ in theonset of CF. Patients with CF exhibit a variety of symptomsthat are related to pancreatic insufficiency or even the lackof secretion of pancreatic juice. The enzymes secreted by theacinar cells of the pancreas and targeted for the small intestineremain stuck in the ducts, leading to subsequent destructionof the pancreatic tissue. In most of the pancreatic-insufficientpatients, the tissue damage has already taken place in utero;however, in some cases the process may develop over a periodof many years. Pancreatic insufficiency leads to maldigestionand severe steatorrhea, with concomitant loss of lipid-solublevitamins and essential fatty acids. The malnutrition rendersthe patients more susceptible to infections, thus also aggravatingthe lung symptoms of the patients. Fortunately, the deficienciesof pancreatic insufficiency can be surmounted in large measureby dietary supplementation.
The exocrine pancreas consists of two morphologically distinctstructures: 1) acinar cells that secrete enzymes, mucins, andNaCl and 2) duct cells that mainly secrete a HCO3–-richfluid. The cause for the pancreatic insufficiency in CF hasbeen attributed to a lack of ductal function, whereas the acinarcells show only small or no abnormalities at all. Thus it wasno surprise that immunocytochemical studies localized CFTR almostexclusively to the pancreatic ductal cells, although a few reportsdemonstrated some scanty expression in the acinar cells. Thegeneral concept of ductal fluid secretion has been developedover the past 60 years. Bro-Rasmussen and colleagues (1) wereamong the first to correlate the HCO3– concentration withthe secretory output of the pancreas after stimulation withthe hormone secretin, a cAMP-mediated agonist (1). The observedinverse correlation between the HCO3– and the Cl–content of the pancreatic juice can now be found in most textbooksof physiology. In a series of elegant experiments on perfusedpancreatic ducts (13), Ivana Novak and Rainer Greger demonstratedthat the ductal epithelial cell uses a unique mechanism forits secretory functions. The simplified model in Fig. 1 depictstheir basic hypothesis for ductal HCO3– secretion.
FIGURE 1. HCO3– secretion by the rat pancreatic duct cell. The lipid-permeable CO2 enters the cell through the basolateral membrane and serves as a pool for the generation of H2HCO3 and, subsequently, HCO3–. HCO3– leaves the cell via a luminal anion exchanger. The accumulated Cl– recycles vial luminal Cl– channels.
HCO3– ions are generated from CO2 that enters the cellfrom the basolateral side by passive diffusion. The activityof carbonic anhydrase in the duct cell catalyzes the formationof carbonic acid from CO2 and H2O and the subsequent dissipationinto HCO3– and protons. The latter are extruded throughthe basolateral membrane via a secondary active Na+/H+ exchanger.The driving force for the antiporter is provided by the Na+pump, establishing the concentration gradient for Na+. BasolateralK+ channels maintain a hyperpolarized basolateral membrane.For a number of species, an additional Na+-dependent HCO3–uptake mechanism on the basolateral membrane of pancreatic ductcells has been demonstrated. The HCO3– ions that are accumulatedby these mechanisms leave the cell on the apical membrane viaa disulfonic stilbene-sensitive pathway in exchange for Cl–.To this end, the molecular identity of this anion exchanger(AE) in the pancreatic duct has not yet been identified. Theclassic AE1 that was first identified in the red blood celland the other members of this family (AE2 and AE3) are not expressedon the apical membrane of HCO3–-secreting epithelial cells.Other proteins like downregulated in adenoma (DRA; SLC26A3),(19) pendrin (SLC26A4), and a member of the putative anion transporterfamily (PAT1; SLC26A6) have been implicated as likely candidatesfor this mechanism (12). The Cl– ions required for theexchange process are provided by the Cl–-rich acinar fluidand are recycled via luminal Cl– channels. The concertedactions of apical Cl– channels and basolateral K+ channelscreate a lumen-negative transepithelial voltage that draws Na+and H2O across the epithelium into the lumen. The proposed mechanismmight explain the earlier finding that with increasing HCO3–secretion the concentration of Cl–decreases. It alsohighlights the crucial role of CFTR in this mechanism. The seemingvoid of any other Cl– export mechanisms ties the functionof the AE inseparably to the only Cl– channel detectedin the apical membrane of this epithelium, which is consideredto be CFTR. A defect in the luminal Cl– conductance wouldeventually abolish both HCO3– and fluid secretion.
Does CFTR directly regulate anion exchange?
The coupling between HCO3– transport and CFTR has recentlybeen reinvestigated (10), and it was postulated that the activityof the putative AE per se is regulated by CFTR and is not dependenton Cl– movement through the channel. On the basis of theobservation that some mutations in the CFTR gene render CF patientspancreatic sufficient but most others do not, Choi and coworkers(2) expressed selected CFTR mutations in fibroblasts. In thesubsequent functional studies, the authors recognized that someof the mutations showed an apparent lack in their HCO3–transport, as assessed by the rate of alkalinization in Cl–-freemedium, whereas the Cl– transport (detected by a Cl–-sensitivefluorophore) seemed to be unaffected and vice versa. The comparisonof the results in the expression system with clinical data yieldedan interesting correlation. Mutations that led to impaired HCO3–transport in the expression system were only found in the pancreatic-insufficientpatients, and those mutations that solely produced a decreasedCl– conductance were found in pancreatic-sufficient patients.The authors proposed that the impaired HCO3– transportresulted from a disruption between CFTR and the anion exchangemechanism and therefore could be held responsible for the fatalpathogenesis in CF in all tissues expressing both proteins.This work has drawn quite a bit of attention toward the fieldof HCO3– transport. Nevertheless, the results of thisstudy should be regarded with caution. Some carriers of mutationsthat were seemingly associated with an intact Cl– conductancedid have abnormal sweat Cl– concentrations, indicativeof an impaired Cl– permeability in the sweat duct (20).Moreover, whereas the coupled AE/Cl– conductance modelis applicable for the rat pancreas where the maximal HCO3–concentration is ~70 mmol/l, it is not sufficient to explainwhy the human pancreas is able to secrete almost isotonic (140mmol/l) NaHCO3. Recent experiments performed on pancreatic ductsof guinea pigs, a species in which the HCO3– concentrationof the pancreatic juice is similar to that of humans, revealedthat the model depicted in Fig. 1 might need to be revised.In guinea pigs the basal HCO3–secretion is dependent onthe luminal Cl– concentration. However, the cAMP-stimulatedsecretion is unaffected by the removal of luminal Cl–.This observation indicated the presence of a Cl–-independentconductive pathway for HCO3– in the luminal membrane ofpancreatic duct cells. On the basis of these data, a modifiedhypothesis on ductal HCO3– and fluid secretion in thedistal ducts of the exocrine pancreas was proposed (Fig. 2)(7,14).
FIGURE 2. HCO3– secretion by distal pancreatic ducts. The decrease in the luminal Cl– concentration leads to a subsequent loss of driving force for the anion exchange mechanism. To sustain further Cl–-independent HCO3– secretion, an alternative HCO3– exit via an electrogenic pathway was postulated. The n (in nHCO3–) refers to the number of ions transported with a single Na+ ion and varies between 2 and 3.
Does CFTR conduct HCO3–?
The exact nature of the luminal HCO3-exit pathway in pancreaticducts is still under debate. A body of evidence initially pointedin the direction of CFTR. Several studies to date have addressedthe question of whether CFTR per se conducts HCO3-. Table 1summarizes the findings of these studies. Together these datasuggest that CFTR does have a finite permeability for HCO3–ions, but this permeability is at best 26% of that for Cl–ions. At this point, it seems justified to question whetherthis seemingly low permeability might in fact be sufficientto permit HCO3– transport through CFTR. Possibly the answerto this question can be found in a tissue that was generallynot considered to secrete HCO3– ions: the airways.
TABLE 1. HCO3– vs. Cl– permeability in cells endogenously expressing CFTR or wild-type CFTR expressed in heterologous systems
HCO3– secretion in the airways
In the past decade, a few studies on primary cultures of humanbronchial epithelial cells derived from non-CF and CF patientsundergoing lung transplant have demonstrated that non-CF cellssecrete HCO3– and that HCO3– secretion is impairedin CF cells (3,18). In non-CF cells, amiloride causes a 50–70%decrease in the short-circuit current (Isc). The residual Iscrequires HCO3– and not Cl– in the bathing solutionand is partially inhibited by serosal DNDS, a disulfonic stilbenethat blocks HCO3– transporters. cAMP causes a furtherincrease in the Isc, and this increase requires HCO3–in the bathing solution and is inhibited by serosal DNDS. Thusnon-CF cells display a basal level of HCO3– secretion,and this can be stimulated by cAMP. In CF cells, amiloride inhibitsnearly all of the Iscand cAMP fails to cause an increase inthe Isc. Studies of this nature led Smith and Welsh (18) toconclude that HCO3– secretion is impaired in CF airwayepithelia and that HCO3– exit at the apical membrane isthrough the anion channel that is defectively regulated in CFepithelia (18).
Secretion of airway serous cells: lessons from Calu-3 cell studies
Airway epithelia can be divided in two different functionalentities: primarily absorptive cells and secretory cells. Theabsorptive surface epithelia of the airways express high levelsof the epithelial Na+channel (ENaC), whereas the CFTR expressionis rather scanty. In contrast, the secretory serous cells ofthe submucosal glands lack ENaC expression and have been demonstratedto be the predominant site of CFTR expression in the airways,expressing manyfold higher levels of CFTR compared with thesurface airway epithelium. Recent studies on an airway serouscell line, Calu-3, have provided further support that airwaycells secrete HCO3– in response to cAMP (4,9). Calu-3cells resemble the characteristics of airway serous cells andcan be grown as polarized monolayers for transport studies.Thus the Calu-3 cells have served as a model for airway serouscells, and studies with the Calu-3 cells have provided importantinsight into the underlying mechanisms of HCO3–secretion.
Calu-3 cells display a low basal Isc that increases upon stimulationwith forskolin. Isotope flux studies revealed that the forskolin-stimulatedIsc was not the result of net Cl–, Na+, or K+ transport,leaving HCO3– secretion as the likely basis of the Isc.Ion substitution studies showed that the forskolin-stimulatedIsc did not require Cl– in the mucosal or serosal bathingsolutions but did require serosal HCO3– and serosal Na+.Bumetanide, an inhibitor of the Na+-K+-2Cl– cotransporter,also failed to block the forskolin-stimulated Isc. In contrast,serosal DNDS, but not mucosal DNDS, partially inhibited theforskolin-stimulated Isc. These results led us to conclude thatCalu-3 cells secrete HCO3– by an electrogenic mechanismin response to forskolin stimulation. We proposed that HCO3–influx across the basolateral membrane was mediated by a DNDS-sensitiveNa+-HCO3–cotransporter (NBC). HCO3– exit acrossthe apical membrane did not require luminal Cl–, nor wasit inhibited by mucosal DNDS. Thus we proposed that HCO3–exit was mediated by CFTR.
The switch from HCO3– to Cl– secretion: driving force matters
An important feature of anion secretion by the Calu-3 cellswas revealed by the use of 1-ethyl benzimidazolone (1-EBIO).1-EBIO activates Ca2+-activated K+ channels such as hIK-1, butunlike with agonists like acetylcholine the activation of theK+ channels is prolonged. The addition of 1-EBIO to forskolin-stimulatedCalu-3 cells caused a further increase in the Isc, as expectedfor the activation of basolateral membrane K+ channels and thehyperpolarization of the membrane potential. However, in contrastto stimulation with forskolin alone, stimulation with forskolinplus 1-EBIO caused the net secretion of Cl–, as revealedby isotope flux studies and the inhibition of the Isc by bumetanide.Indeed, the Isc of forskolin plus 1-EBIO-stimulated cells wasfully accounted for by the net secretion of Cl–. Figure3 illustrates these findings.
FIGURE 3. Effects of forskolin and 1-EBIO on Calu-3 cell short-circuit current (Isc) and 36Cl– fluxes. A: representative Isc recording in response to forskolin (2 µmol/l) and 1-ethyl benzimadazolone (1-EBIO; 1 mmol/l). B: summary of mucosal-to-serosal (Jms), serosal-to-mucosal (Jsm), and net 36Cl– (Jnet) fluxes under forskolin- and forskolin plus 1-EBIO-stimulated conditions. Reproduced from the Journal of General Physiology, 1999, vol. 113, pp. 743–760 by copyright permission of The Rockefeller University Press.
The activation of basolateral K+ channels caused Calu-3 cellanion secretion to switch from HCO3– secretion to Cl–secretion. These results led us to propose that the basolateralmembrane NBC was an electrogenic transporter carrying a greaternumber of HCO3– anions than Na+ cations. Hyperpolarizationof the basolateral membrane potential as a result of the activationof K+ channels by 1-EBIO would tend to inhibit HCO3– influxacross the basolateral membrane. Indeed, if the basolateralmembrane potential were to exceed the reversal potential ofthe NBC, HCO3– might exit rather than enter the cell acrossthe basolateral membrane. Concomitant with the inhibition ofan electrogenic NBC, the Na+-K+-2Cl– cotransporter isactivated in forskolin plus 1-EBIO-stimulated cells.
Consistent with the very high levels of CFTR expression, forskolincauses the apical membrane resistance to fall to a remarkablylow value. At the same time we observe a depolarization of theapical membrane to the equilibrium potential for Cl–.This effect is so dominant that it also depolarizes the basolateralmembrane due to the low shunt resistance of the paracellularpathway. The activation of basolateral membrane K+ channelsby forskolin, as is evident from the decrease in the basolateralmembrane resistance upon forskolin stimulation, is insufficientto maintain the basolateral membrane potential. Instead thebasolateral membrane depolarizes, and this provides a favorablemembrane potential for HCO3– entry on an electrogenicNBC. Whether forskolin also activates the NBC via PKA-mediatedphosphorylation is not known at this time. Thus the very highapical membrane anion conductance stimulated by forskolin 1)serves to mediate the conductive exit of HCO3–, 2) setsthe driving force for HCO3– exit across the apical membrane,and 3) sets the driving force for the entry of HCO3– acrossthe basolateral membrane on the NBC.
As expected for the activation of basolateral membrane K+ channelsby 1-EBIO, the basolateral membrane and apical membrane potentialsare seen to hyperpolarize from their forskolin-stimulated values.The hyperpolarization of the apical membrane would be expectedto increase the driving force for both HCO3– and Cl–exit and thus cannot explain the inhibition of HCO3– secretion.However, the hyperpolarization of the basolateral membrane potentialis expected to inhibit the influx of HCO3– mediated byan electrogenic NBC that moves a net anionic charge. The removalof Na+ or HCO3– from the serosal solution or the additionof DNDS both result in the depolarization of the basolateralmembrane potential, as expected for an electrogenic NBC (Tamada,Hug, and Bridges, unpublished observations). Indeed, one maydeduce from the basolateral membrane potential measurementsin forskolin and forskolin plus 1-EBIO-stimulated cells thatthe Na+-HCO3– stoichiometry of the NBC in Calu-3 cellsis 1:3. On the basis of these observations, we proposed a modelfor anion secretion by airway serous cells that is depictedin Fig. 4.
FIGURE 4. Proposed model for anion secretion in Calu-3 cells. A: forskolin-stimulated cells secrete HCO3–. B: forskolin plus 1-EBIO-stimulated cells secrete Cl–.
Summary
The studies with Calu-3 cells establish an electrochemical profileagainst which results from submucosal gland serous cells canbe compared to determine whether native serous cells secreteanions in a similar manner. If our results with Calu-3 cellsare representative of airway serous cells, then HCO3–secretion in the airways may be more important than has previouslybeen appreciated. In addition, these studies and our proposedmodel for HCO3– and Cl– secretion by the same cellmay help explain the pathophysiology of anion secretion in thepancreas and small intestine of CF patients. If our model iscorrect, CFTR serves as the conductive pathway for HCO3–exit across the apical membrane in HCO3–-secreting cells.Mutations in CFTR that impair the conductance of the channelfor HCO3– are expected to increase the severity of thedisease in those epithelia where HCO3– secretion is essentialfor the normal physiology of the organ. Impaired HCO3–secretion in the pancreas and small intestine in CF patientshas been known for many years. The results with primary culturesof human bronchial epithelial cells and Calu-3 cells suggestthat HCO3– secretion may also be important in the airways.It seems prudent to speculate that a similar mechanism to thatfound in Calu-3 cells might be attributable to other epitheliathat secrete HCO3–.
Acknowledgments
Our work is supported by Innovative Medizinische Forschung HU11 01 03 and National Institute of Diabetes and Digestive andKidney Diseases Grants RO1-DK-58782 and 1-P50-DK-56490.
References
Bro-Rasmussen F, Killmann SA, and Thaysen JH. The composition of pancreatic juice as compared to sweat, parotid saliva and tears. Acta Physiol Scand 37: 97–113, 1956.[Web of Science][Medline]
Choi JY, Muallem D, Kiselyov K, Lee MG, Thomas PJ, and Muallem S. Aberrant CFTR-dependent HCO3– transport in mutations associated with cystic fibrosis. Nature 410: 94–97, 2001.[Medline]
Devor DC, Bridges RJ, and Pilewski JM. Pharmacological modulation of ion transport across wild-type and F508 CFTR-expressing human bronchial epithelia. Am J Physiol Cell Physiol279: C461–C479, 2000.[Abstract/Free Full Text]
Devor DC, Singh AK, Lambert LC, DeLuca A, Frizzell RA, and Bridges RJ. Bicarbonate and chloride secretion in Calu-3 human airway epithelial cells. J Gen Physiol 113: 743–760, 1999.[Abstract/Free Full Text]
Gray MA, Pollard CE, Harris A, Coleman L, Greenwell JR, and Argent BE. Anion selectivity and block of the small-conductance chloride channel on pancreatic duct cells. Am J Physiol Cell Physiol 259: C752–C761, 1990.[Abstract/Free Full Text]
Illek B, Tam AW, Fischer H, and Machen TE. Anion selectivity of apical membrane conductance of Calu 3 human airway epithelium. Pflügers Arch 437: 812–822, 1999.[Web of Science][Medline]
Ishiguro H, Steward MC, Wilson RW, and Case RM. Bicarbonate secretion in interlobular ducts from guinea-pig pancreas. J Physiol 495: 179–191, 1996.[Abstract/Free Full Text]
Kopelman H, Corey M, Gaskin K, Durie P, Weizman Z, and Forstner G. Impaired chloride secretion, as well as bicarbonate secretion, underlies the fluid secretory defect in the cystic fibrosis pancreas. Gastroenterology 95: 349–355, 1988.[Web of Science][Medline]
Lee MC, Penland CM, Widdicombe JH, and Wine JJ. Evidence that Calu-3 human airway cells secrete bicarbonate. Am J Physiol Lung Cell Mol Physiol 274: L450–L453, 1998.[Abstract/Free Full Text]
Lee MG, Wigley WC, Zeng W, Noel LE, Marino CR, Thomas PJ, and Muallem S. Regulation of Cl-/HCO3– exchange. J Biol Chem 274: 3414–3421, 1999.[Abstract/Free Full Text]
Linsdell P, Tabcharani JA, Rommens JM, Hou YX, Chang XB, Tsui LC, Riordan JR, and Hanrahan JW. Permeability of wild-type and mutant cystic fibrosis transmembrane conductance regulator chloride channels to polyatomic anions. J Gen Physiol 110: 355–364, 1997.[Abstract/Free Full Text]
Lohi H, Kujala M, Kerkela E, Saarialho-Kere U, Kestila M, and Kere J. Mapping of five new putative anion transporter genes in human and characterization of SLC26A6, a candidate gene for pancreatic anion exchanger.Genomics 70: 102–112, 2000.[Web of Science][Medline]
Novak I and Greger R. Properties of the luminal membrane of isolated perfused rat pancreatic ducts. Effect of cyclic AMP and blockers of chloride transport. Pflügers Arch 411: 546–553, 1988.[Web of Science][Medline]
O’Reilly CM, Winpenny JP, Argent BE, and Gray MA. Cystic fibrosis transmembrane conductance regulator currents in guinea pig pancreatic duct cells: inhibition by bicarbonate ions. Gastroenterology 118: 1187–1196, 2000.[Web of Science][Medline]
Poulsen JH, Fischer H, Illek B, and Machen TE. Bicarbonate conductance and pH regulatory capability of cystic fibrosis transmembrane conductance regulator. Proc Natl Acad Sci USA 91: 5340–5344, 1994.[Abstract/Free Full Text]
Pratha VS, Hogan DL, Martensson BA, Bernard J, Zhou R, and Isenberg JI. Identification of transport abnormalities in duodenal mucosa and duodenal enterocytes from patients with cystic fibrosis. Gastroenterology 118: 1051–1060, 2000.[Web of Science][Medline]
Rosenstein BJ and Cutting GR. The diagnosis of cystic fibrosis: a consensus statement. Cystic Fibrosis Foundation Consensus Panel. J Pediatr 132: 589–595, 1998.[Web of Science][Medline]
Smith JJ and Welsh MJ. cAMP stimulates bicarbonate secretion across normal, but not cystic fibrosis airway epithelia. J Clin Invest 89: 1148–1153, 1992.[Web of Science][Medline]
Wheat VJ, Shumaker H, Burnham C, Shull GE, Yankaskas JR, and Soleimani M. CFTR induces the expression of DRA along with Cl–/HCO3– exchange activity in tracheal epithelial cells. Am J Physiol Cell Physiol 279: C62–C71, 2000.[Abstract/Free Full Text]
Wine JJ. Cystic fibrosis: the ‘bicarbonate before chloride’ hypothesis. Curr Biol 11: R463–R466, 2001.[Web of Science][Medline]