Journal of the Pancreas Open Access

Keywords

Bicarbonates; Chlorides; Secretions

Abbreviations

Cap: apical membrane capacitance; Cbl: basolateral membrane capacitance; CFTR: cystic fibrosis transmembrane conductance regulator; CTX: charybdotoxin; DNDS: 4,4’-dinitrostilben-2,2’- disulfonic acid; ECl: equilibrium potential for Cl–; EHCO3: equilibrium potential for HCO3–; ERevNBC: equilibrium reversal potential of the DNDS-sensitive NBC; 1-EBIO: 1-ethyl-2- benzimidazolinone; FRap: apical membrane fractional resistance; Gap: apical membrane conductance; GT : transepithelial conductance; KCa: Ca+-activated, charybdotoxin-sensitive K+ channels; NBC: sodium bicarbonate cotransporter; PKA: protein kinase A; Rap: apical membrane resistance; Rbl: basolateral membrane resistance; RP: resistance of the paracellular pathway; RT : transepithelial resistance; Vap: apical membrane potential; Vbl: basolateral membrane potential; VT : transepithelial potential; Zi: imaginary impedance; ZR: real impedance

Our model for anion secretion in Calu-3 cells is illustrated in Figure 1. We have demonstrated forskolin stimulated Calu-3 cells secrete HCO3– by an electrogenic, Cl–- independent, serosal Na+-dependent, serosal bumetanide-insensitive and serosal disulfonic stilbene (DNDS)- sensitive mechanism as judged by transepithelial currents, isotope fluxes, and the results of ion substitution, pharmacology and pH studies [1]. However, Calu-3 cells are not limited to the secretion of HCO3–. Instead, when stimulated by 1-EBIO, an activator of basolateral membrane, Ca+-activated, charybdotoxin-sensitive K+ channels (KCa), the Calu-3 cells secrete Cl– by an electrogenic bumetanide-sensitive mechanism and HCO3– secretion is diminished. Moreover, when stimulated by both forskolin and 1-EBIO, the secretion of HCO3– is diminished and Clsecretion dominates. A similar switch from HCO3– secretion to Cl– secretion in Calu-3 cells was reported by Lee et al. [2] using thapsigargin to activate KCa channels. To account for these results, we proposed a model of anion secretion whereby cystic fibrosis transmembrane conductance regulator (CFTR) serves as the cAMP/protein kinase A (cAMP/PKA) activated anion channel for both Cl– and HCO3– exit across the apical membrane. The driving force for HCO3– or Cl– exit across the apical membrane is equal to the apical membrane potential (Vap) minus the equilibrium potential for HCO3– or Cl–, EHCO3 or ECl, respectively. Activation of CFTR alone will tend to bring Vap to ECl, a value that is predicted to be greater than EHCO3 and thus provides the driving force for HCO3– exit. Stimulation by cAMP (forskolin) alone leaves the basolateral membrane potential (Vbl) less hyperpolarized than the reversal potential of the 4,4’-dinitrostilben-2,2’-disulfonic acid- (DNDS)-sensitive NBC (ERevNBC) and HCO3– is secreted. Subsequent activation of KCa by 1- EBIO or cholinergic agonists hyperpolarizes Vbl so that Vbl greater than ERevNBC, and this inhibits HCO3– uptake by the NBC but it provides the driving force for Cl– secretion because Vap becomes greater than ECl. Whether cAMP/PKA activates the NBC directly is unknown. The Na+:HCO3– stoichiometry of the Calu-3 cell NBC is also unknown. At a Na+:HCO3– stoichiometry of 1:3, ERevNBC is –49 mV and at a stoichiometry of 1:2 ERevNBC is –84 mV. Thus, the model in Figure 1 leads to several readily testable predictions. To begin to test these hypotheses, we have performed microelectrode and impedance studies using the Calu-3 cells as an experimental model of airway serous cells.

Calu-3 cells were grown on Snapwell filters as previously described [1] and studied in a horizontal chamber that allowed for the impalement of the cells with a microelectrode from the apical side. The transepithelial potential (VT) and Vap were recorded from cells maintained under open circuit conditions. A 50 μA transepithelial bipolar pulse was passed every 20 seconds and the transepithelial resistance (RT) and the apical fractional resistance (FRap) were calculated from the DVT and DVap/DVT ratio respectively as previously described [3]. Both the apical and basolateral surfaces were continuously perfused with a warm (37 °C) gassed (95%/5%, O2/CO2) solution. Results from a typical microelectrode experiment are shown in Figure 2. We were able to maintain the microelectrode impalement for 10 to 30 minutes on a routine basis. This allowed us to monitor the same cell under control, forskolin (2 μM) and forskolin plus 1- EBIO (1 mM) stimulated conditions. In the experiment shown in Figure 2 VT was approximately –5.5 mV, mucosal side negative, under control conditions and hyperpolarized to –22 mV with forskolin stimulation and further hyperpolarized to –35 mV with the subsequent addition of forskolin plus 1-EBIO. The RT decreased from a control value of 530 W cm2 to 320 W cm2 with forskolin and to 235 W cm2 with forskolin plus 1-EBIO. Using these VT and RT values one obtains an equivalent short circuit current of 10 μA/cm2, 68 μA/cm2 and 148 μA/cm2 for the control, forskolin and forskolin plus 1-EBIO conditions, respectively. These results are in excellent agreement with the results obtained under short circuit current conditions where it was demonstrated that Calu-3 cells secrete HCO3– when stimulated with forskolin and Cl– when stimulated with forskolin plus 1-EBIO [1].

Panel B of Figure 2 is the voltage measured by the microelectrode which upon impalement has a value of approximately –53 mV that improved to a value of –59 mV after a few minutes. Upon the addition of forskolin, Vap depolarized to a value of –22 mV and then repolarized to a value of –28 mV upon the subsequent addition of forskolin plus 1-EBIO. Panel C of Figure 2 is a plot of the FRap from the same experiment and shows that FRap fell from a control value of approximately 0.6 to approximately 0.1 upon stimulation with forskolin and was unchanged when forskolin plus 1-EBIO was added. Shortly after the addition of 1-EBIO the impalement was lost. This was frequently observed whenever 1- EBIO was added and is an effect we speculate may be due to cell shrinkage since advancing the electrode often reestablished the impalement.

Figure 3 summarizes the results of 25 similar experiments. As already noted the hyperpolarization of VT is consistent with the secretion of HCO3– in forskolin stimulated cells and Cl– in forskolin plus 1-EBIO stimulated cells. These changes in VT reflect the activation of an apical membrane conductance so that Vap depolarized and FRap decreased upon the addition of forskolin. Vbl also depolarized from a control value of –60±1.7 mV to –44±1.3 mV in forskolin stimulated cells. The depolarization of Vbl with forskolin can be explained by an activation of an apical membrane anion conductance that dominates the total cellular conductance. Support for this notion is reflected in the remarkably low FRap observed in the forskolin stimulated cells and this will be further supported by the impedance results given below. Our model predicts Vap should be near ECl in forskolin stimulated cells. Ideally, one should measure Vap and the intracellular Clactivity using a double barreled microelectrode and these experiments are in progress. Based on the observation that Cl– is not secreted by forskolin stimulated cells, the measured Vap allows one to predict an intracellular Clactivity of approximately 40 mM in the forskolin stimulated Calu-3 cells. This estimate of the intracellular Cl– activity agrees rather well with the measured values in other airway epithelial cells [4]. Perhaps more importantly the measured value of –21.5±1.3 mV for Vap in forskolin stimulated cells exceeds EHCO3 (EHCO3 = –13.5 mV) such that there is a net driving force for the conductive exit of HCO3– across the apical membrane of 7.8 mV. In addition the observed value for Vbl of –44.2±1.3 mV in forskolin stimulated cells is less than the ERevNBC of –49 mV and –84 mV for NBCs with 1:3 or 1:2 Na+:HCO3– stoichiometries and thus would allow HCO3– to enter the cell on a basolateral membrane NBC. Therefore, the observed effects of forskolin on VT, Vap, Vbl and FRap are consistent with the net secretion of HCO3– by the mechanism proposed in our model.

When 1-EBIO is added to the forskolin stimulated cells, HCO3– secretion is inhibited and Cl– secretion is stimulated [1]. The addition 1-EBIO was observed to cause Vap and Vbl to repolarize and these changes are the expected changes for the activation of basolateral membrane K+ channels. The repolarization of Vap would increase the driving force for conductive anion exit across the apical. This is true for both HCO3– and Cl– exit and therefore the change in Vap can not explain the switch from HCO3– secretion to Cl– secretion. However, 1-EBIO was seen to cause Vbl to repolarize from –44±1.3 mV to –60±2.3 mV. At a Vbl of –60 mV an NBC with a Na+:HCO3– stoichiometry of 1:3 (ERevNBC = –49 mV) would be inhibited and HCO3– would actually be expected to leave the cell rather then enter across the basolateral membrane. In contrast, if the Na+:HCO3– stoichiometry were 1:2 (ERevNBC = –84 mV) HCO3– would still enter the cell and secretion of HCO3– should continue. Thus, based on a measured Vbl of –60 mV and the observation that HCO3– secretion is inhibited when Calu-3 cells are stimulated by forskolin plus 1-EBIO, our results suggest the NBC responsible for HCO3– entry has a Na+:HCO3– stoichiometry of 1:3. This is a surprising outcome since the measured stoichiometries of the two NBC isovariants we have detected in Calu-3 cells, the kidney and pancreatic NBCs [5] both have Na+:HCO3– stoichiometries of 1:2 when expressed in Xenopus oocytes [5, 6, 7]. Therefore, there must exist an alternative NBC with a 1:3 Na+:HCO3– stoichiometry that is expressed in the Calu-3 cells. Alternatively, the kidney or pancreatic NBC isovariants may be subjected to some form of regulation that alters the Na+:HCO3– stoichiometry from 1:2 to 1:3 when expressed in Calu-3 cells. Additional studies are required to clarify this point.

The remarkably low FRap value of 0.1 we observed in the forskolin stimulated Calu-3 cells prompted us to obtain an independent estimate of the apical and basolateral membrane resistances in the Calu-3 cells. To obtain these values we elected to perform impedance analysis studies. The impedance measurements were made on short-circuited Calu-3 cells grown as described above and given in detail elsewhere [1]. The transepithelial impedance was measured in response to a series of 100 sine waves over a frequency range of 1 Hz to 22 kHz as previously described [8, 9]. The impedance values are presented as Nyquist plots and were fit to the equations describing the equivalent electric circuit shown in Figure 4 to obtain estimates of the apical and basolateral membrane resistances (Rap, Rbl) and capacitances (Cap, Cbl). Only the resistance values will be considered here. The resistance of the paracellular pathway (RP) was estimated from the y-intercept of a GT (transepithelial conductance) versus ISC plot [10] and was assumed to be constant under the different experimental conditions.

Shown in Figure 5A is an ISC trace of a Calu-3 cell monolayer under control, forskolin (2 μM), 1-EBIO (1 mM) and charybdotoxin (CTX) (50 nM) conditions. As previously reported [1], the Calu-3 cells display a control ISC of approximately 8 μA/cm2. Addition of forskolin caused an increase in ISC to 76 μA/cm2 and 1- EBIO further increased the ISC to 130 μA/cm2. CTX, an inhibitor of the 1-EBIO activated K+ channels, decreased the ISC to the pre-1-EBIO level. Panels B-E of Figure 5 show the Nyquist plots corresponding to each of the experimental conditions in Panel A. In this monolayer, the control impedance spectrum could be fit to the two membrane model shown in Figure 4 to yield estimates of Rap and Rbl given in Panel 5B. Using these values one obtains an FRap of 0.53 in good agreement with the FRap of 0.52±0.26 obtained in the microelectrode studies. Stimulation with forskolin reduced the total impedance and two semicircles were clearly resolved in the Nyquist plot (Figure 5C). Based on morphological considerations and pharmacological studies the smaller semicircle closest to the origin can be identified as the apical membrane and the larger semicircle to the right as the basolateral membrane. Stimulation with forskolin reduced Rap to 20 W cm2 and Rbl to 160 W cm2 to yield an FRap of 0.1, a value that is also in excellent agreement with the FRap of 0.08±0.005 obtained in the microelectrode studies. These results demonstrate that forskolin activates both apical membrane and basolateral membrane conductances. However, the change in the apical conductance far exceeds the change in the basolateral conductance and this likely explains why Vbl depolarizes in forskolin stimulated monolayers. The subsequent addition of 1-EBIO further reduced the total impedance and caused a decrease in Rbl to 77 W cm2 consistent with the activation of basolateral membrane K+ channels, and the repolarization of Vap and Vbl observed in the microelectrode studies. One would also anticipate from the impedance results that FRap should increase to 0.19 but this effect was not observed in the microelectrode studies. As expected for the blockade of the basolateral membrane 1-EBIO activated K+ channels, CTX caused the total impedance to increase and increased Rbl to 140 W cm2, a value approaching the pre 1-EBIO Rbl of 167 W cm2.

The above impedance results demonstrate forskolin and 1-EBIO have their expected effects on the apical and basolateral membranes. The results shown in Figure 5 were selected from a large number of impedance experiments to illustrate the presence of both the apical and basolateral membranes components in the Nyquist plots. However, in most experiments the apical membrane could not be discerned in the Nyquist plot once the cells were stimulated with maximal stimulatory concentrations of forskolin. Time course and dose response studies during the forskolin stimulated increase in ISC revealed the presence of the apical membrane component (Figure 6). The decrease in the apical membrane impedance continued to a point where it essentially vanished from the Nyquist plot as ISC continued to increase. If one uses data from the spectrum just prior to when the apical membrane vanishes, estimates of the apical membrane resistance in these monolayers falls below 5 W cm2. This is an astonishingly low value but nonetheless a value that is consistent with the very high levels of CFTR expressed by Calu-3 cells [11]. Thus, in most monolayers it was not possible by impedance analysis to obtain a true estimate of Rap in the forskolin or forskolin plus 1-EBIO stimulated monolayers. Rather, Rap would appear to be less than 5 W cm2 in the stimulated monolayers. This suggests that the apical membrane conductance (Gap = 1/Rap) is greater than 200 mS/cm2 in forskolin stimulated Calu-3 cells. If one uses the microelectrode derived estimate of FRap of 0.08±0.005 and the impedance derived estimate of Rbl of approximately 170 W cm2 an Rap of approximately 14 W cm2 is calculated for forskolin stimulated Calu-3 cells. Therefore, a conservative approximation of Gap in forskolin stimulated Calu-3 cells would be in the range of 50 to 200 mS/cm2.

The astonishingly high Gap of forskolin stimulated Calu-3 cells has several important implications for anion secretion by these cells. We propose that this Gap is used to clamp the apical membrane at ECl and thus insures a driving force for HCO3– secretion equal to Vap – EHCO3 where Vap = ECl in forskolin stimulated cells. Based on our measured values for Vap and assuming an intracellular HCO3– concentration of 15 mM (pHi = 7.15), there would be a driving force of 7.8 mV for the conductive exit of HCO3– across the apical membrane. Patch clamp estimates of the HCO3– to Cl– selectively of CFTR have yielded values in the 0.1 to 0.25 range [12, 13]. Machen and coworkers [14, 15] have obtained similar anion selectivity ratios in alpha toxin basolateral membrane permeabilized Calu-3 monolayers. Thus, if one does a calculation using the “worst case” estimates of a selectivity of 0.1 for HCO3– to Cland Gap of 50 mS/cm2 so that GapHCO3 = 5 mS/cm2 (GapHCO3 = 0.1 x 50 mS/cm2 = 5 mS/cm2) with a driving force of 7.8 mV, a HCO3– current of approximately 40 μA/cm2 is possible. On the hand if the selectivity were 0.25 and the Gap 200 mS/cm2, (GapHCO3 = 0.25 x 200 mS/cm2 = 50 mS/cm2) a current of 390 μA/cm2 could be observed. The actual measured rate of HCO3– secretion in forskolin stimulated Calu-3 cells is approximately 60 μA/cm2 suggesting the actual GapHCO3 will be closer to the lower value. Therefore, even though the conductance of CFTR is lower for HCO3– compared to Cl– there is an adequate driving force to account for the secretion of HCO3–. Moreover, these results suggest that in cAMP stimulated Calu-3 cells CFTR serves as a Cl– conductance to set the driving force for another anion, HCO3–, and not for the secretion of Cl– as is commonly held in other models of HCO3– secretion.

Why is Cl– not secreted by the forskolin stimulated Calu-3 cells? The reasons appear to be several fold. First, it would appear that the very high Gap dominates the total cellular conductance even though forskolin was observed to activate a basolateral membrane conductance. Thus, unlike other Cl– secretory cells the activation of this basolateral membrane conductance by cAMP is relatively small compared to the activation of Gap and does not adequately repolarize Vap to provide a driving force for Cl– exit across the apical membrane. Secondly, the Na+:K+:2Clcotransporter would appear to be inactive in forskolin stimulated Calu-3 cells [1]. Isotope flux studies failed to detect any bumetanide sensitive Cl– flux in forskolin treated Calu-3 cells. This is not because Calu-3 cells lack the Na+:K+:2Cl– cotransport because once stimulated by 1-EBIO the ISC becomes bumetanide sensitive and there is an increase in the serosal-to-mucosal flux of Cl– that is inhibited by bumetanide. Thus, only after 1- EBIO repolarizes Vap and Vbl does the Na+:K+:2Cl– cotransporter become active. Based on studies in other systems, the signals that activate the Na+:K+:2Cl– cotransporter are a decrease in cell volume and a fall in the intracellular Cl– concentration [16]. The activation of basolateral membrane K+ channels by 1-EBIO is expected to do both and studies are in progress to measure these changes. Based on the measured rate of Cl– secretion of approximately 130 μA/cm2 and a Gap of 50 to 200 mS/cm2 the necessary driving force to sustain this level of Cl– secretion is only 0.7 mV to 2.6 mV. If the intracellular Cl– activity were to remain unchanged when 1-EBIO was added, the repolarization of Vap by 5.9 mV would yield a current of 295 μA/cm2 (Gap = 50 mS/cm2) to 1180 μA/cm2 (Gap = 200 mS/cm2). Since the measured ISC is only 130 μA/cm2, we predict that the intracellular Cl– activity will fall by approximately 8 mM thereby causing a shift in ECl in keeping with the observed ISC and high Gap.

In summary, the microelectrode and impedance results reported here lend additional support to our proposed model of anion secretion in Calu- 3 cells (Figure 1). The studies with Calu-3 cells establish an electrochemical profile against which results from submucosal gland serous cells can be compared to determine whether native serous cells secrete anions in a similar manner. If our results with Calu-3 cells are representative of airway serous cells, then HCO3– secretion in the airways may be more important than has previously been appreciated. In addition, these studies and our proposed model for HCO3– and Cl– secretion by the same cell may help explain the pathophysiology of anion secretion in the pancreas and small intestine of cystic fibrosis patients. If our model is correct, 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 channel for HCO3– are expected to increase the severity of the disease in those epithelia where HCO3– secretion is essential for the normal physiology of the organ. Impaired HCO3– secretion in the pancreas and small intestine in cystic fibrosis patients has been known for many years. The results with Calu-3 cells suggest HCO3– secretion may also be important in the airways.

Acknowledgements

We wish to thank Ms. Maitrayee Sahu and Mr. Matthew Green for their excellent technical assistance and Ms. Michele Dobransky for her secretarial assistance. This work was supported by National Institutes of Health Grant Number 1RO1DK58782-01 and the Cystic Fibrosis Foundation Grant Number BRIDGE00G1 to Robert J. Bridges.

Figures at a glance

Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6

References

1. Devor DC, Singh AK, Lambert LC, DeLuca A, Frizzell RA, Bridges RJ. Bicarbonate and chloride secretion in Calu-3 human airway epithelial cells. J Gen Physiol 1999; 113:743-60. [99246328]
2. Lee MC, Penland CM, Widdicombe JH, Wine JJ. Evidence that Calu-3 human airway cells secrete bicarbonate. Am J Physiol 1998; 274:L450-3. [98191213]
3. Welsh MJ, Smith PL, Frizzell RA. Chloride secretion by canine tracheal epithelium: II. The cellular electrical potential profile. J Membr Biol 1982; 70:227- 38. [84036133]
4. Willumsen NJ, Davis CW, Boucher RC. Intracellular Cl^– activity and cellular Cl^– pathways in cultured human airway epithelium. Am J Physiol 1989; 256:C1033-44. [89244925]
5. Peters KW, Gangopadhyay NN, Devor DC, Watkins SC, Frizzell RA, Bridges RJ. Sodium bicarbonate cotransporter expression in airway epithelial cells. Pediatr Pulmonol Suppl 1999; 19:193.
6. Sciortino CM, Romero MF. Cation and voltage dependence of rat kidney electrogenic Na+-HCO₃^– cotransporter, rkNBC, expressed in oocytes. Am J Physiol 1999; 277:F611-23. [99447316]
7. M. Heyer, S. Muller-Berger, M.F. Romero, W.F. Boron, & E. Fromter. Stoichiometry of the rat kidney Na+-HCO₃^– cotransporter expressed in Xenopus laevis oocytes. Pflugers Arch 1999; 438:322-9. [99329282]
8. Margineau DG, Van Driessche W. Effects of millimolar concentrations of glutaraldehyde on the electrical properties of frog skin. J Physiol 1990; 427:567-81. [91012230]
9. Van Driessche W, De Vos R, Jans D, Simaels J, De Smet P, Raskin G. Transepithelial capacitance decrease reveals closure of lateral interspace in A6 epithelia. Pflugers Arch 1999; 437:680-90. [99189322]
10. Wills NK, Lewis SA, Eaton DC. Active and passive properties of rabbit descending colon: a microelectrode and nystatin study. J Membr Biol 1979; 28:81-108. [79196670]
11. Finkbeiner WE, Carrier SD, Teresi CE. Reverse transcription-polymerase chain reaction (RT-PCR) phenotypic analysis of cell cultures of human tracheal epithelium, tracheobronchial glands, and lung carcinomas. Am J Respir Cell Mol Biol 1993; 9:547-56. [94031082]
12. Gray MA, Pollard CE, Harris A, Coleman L, Greenwell JR, Argent BE. Anion selectivity and block of the small-conductance chloride channel on pancreatic duct cells. Am J Physiol 1990; 259:C752-61. [91051856]
13. Linsdell P, Tabcharani JA, Hanrahan JW. Multi-Ion mechanism for ion permeation and block in the cystic fibrosis transmembrane conductance regulator chloride channel. J Gen Physiol 1997; 110:365-77. [98021950]
14. Illek B, Tam AW, Fischer H, Machen TE. Anion selectivity of apical membrane conductance of Calu-3 human airway epithelium. Pflugers Arch 1999; 437:812- 22. [99299471]
15. Illek B, Yankaskas JR, Machen TE. cAMP and genistein stimulate HCO₃^–conductance through CFTR in human airway epithelia. Am J Physiol 1997; 272:L752- 61. [97287823]
16. Haas M, Forbush B 3rd. The Na-K-Cl cotransporter of secretory epithelia. Annu Rev Physiol 2000; 62:515- 34. [20303520]