Polarized distribution of HCO₃⁻ transport in human normal and cystic fibrosis nasal epithelia

Abstract

The polarized distribution of HCO₃⁻ transport was investigated in human nasal epithelial cells from normal and cystic fibrosis (CF) tissues. To test for HCO₃⁻ transport via conductive versus electroneutral Cl⁻/HCO₃⁻ exchange (anion exchange, AE) pathways, nasal cells were loaded with the pH probe 2′,7′-bis(carboxyethyl)-5(6)-carboxyfluorescein and mounted in a bilateral perfusion chamber. In normal, but not CF, epithelia, replacing mucosal Cl⁻ with gluconate caused intracellular pH (pH_i) to increase, and the initial rates (ΔpH min⁻¹) of this increase were modestly augmented (∼26 %) when normal cells were pretreated with forskolin (10 μm). Recovery from this alkaline shift was dependent on mucosal Cl⁻, was insensitive to the AE inhibitor 4,4′-diisothiocyanatodihydrostilbene-2,2′-disulfonic acid (H₂DIDS; 1.5 mm), but was sensitive to the cystic fibrosis transmembrane conductance regulator (CFTR) channel inhibitor diphenylamine-2-carboxylate (DPC; 100 μm). In contrast, removal of serosal Cl⁻ caused pH_i to alkalinize in both normal and CF epithelia. Recovery from this alkaline challenge was dependent on serosal Cl⁻ and blocked by H₂DIDS. Additional studies showed that serosally applied Ba²⁺ (5.0 mm) in normal, but not CF, cells induced influx of HCO₃⁻ across the apical membrane that was reversibly blocked by mucosal DPC. In a final series of studies, normal and CF cells acutely alkaline loaded by replacing bilateral Krebs bicarbonate Ringer (KBR) with Hepes-buffered Ringer solution exhibited basolateral, but not apical, recovery from an alkaline challenge that was dependent on Cl⁻, independent of Na⁺ and blocked by H₂DIDS. We conclude that: (1) normal, but not CF, nasal epithelia have a constitutively active DPC-sensitive HCO₃⁻ influx/efflux pathway across the apical membrane of cells, consistent with the movement of HCO₃⁻ via CFTR; and (2) both normal and CF nasal epithelia have Na⁺-independent, H₂DIDS-sensitive AE at their basolateral domain.

Studies of airway epithelial salt and water physiology have traditionally centred around measurements of Na⁺ and Cl⁻ transport, and these studies have provided a large body of information about cellular signalling mechanisms important for regulating these ion transport processes and their protective functions in the respiratory tract. In contrast, fewer studies have focused on identifying the membrane elements important for the transport of H⁺ and HCO₃⁻ (Nord et al. 1988; Willumsen & Boucher, 1992; Lubman et al. 1995; Paradiso, 1997), which are necessary for the maintenance of intracellular pH (pH_i) homeostasis. It is predicted that polarized airway epithelia distribute these mechanisms (apical/basolateral membrane) to establish salt and H⁺ gradients between the cytosol and secreted liquids, important for the regulation of airway surface liquid (ASL) ionic composition and pH.

Since the recognition that the cystic fibrosis transmembrane conductance regulator (CFTR) functions as a Cl⁻ channel (Quinton, 1983), Cl⁻ transport activity by CFTR-expressing tissues has been well studied (Quinton, 1983, 1990; Willumsen et al. 1989; Widdicombe & Wine, 1991; Anderson et al. 1991; Linsdell et al. 1999). However, the mechanisms underlying HCO₃⁻ secretion in CFTR-expressing epithelia remain poorly understood, even though HCO₃⁻ transport activity is impaired in cystic fibrosis (CF; Kaiser & Drack, 1974; Smith & Welsh, 1992). Of relevance to ASL acid-base homeostasis, a role for CFTR as a HCO₃⁻ transporter has been suggested in CF airway models, based on a variety of technical approaches comparing wild-type CFTR-expressing cells with ΔF508 CFTR-expressing CF cells (Smith & Welsh, 1992; Illek et al. 1997) or comparing wild-type mice with CFTR knockout (−/−) mice (Grubb & Gabriel, 1997). These studies suggest the possibility that a defect in HCO₃⁻ secretion through CFTR may lead to abnormally acidic ASL pH and thus have the potential to contribute to the pathophysiology of CF airway disease.

In addition to conductive CFTR-mediated HCO₃⁻ secretion, it is well known that a major mechanism for HCO₃⁻ transport across plasma membranes is Na⁺-independent, electroneutral Cl⁻/HCO₃⁻ exchange (anion exchange; AE). Both Cl⁻ and HCO₃⁻ can be transported via AE and the direction of transport is determined by the gradients of both anions across the plasma membrane. Under normal physiological conditions, the ratio of extracellular-to-intracellular Cl⁻ is greater than the inward-directed HCO₃⁻ gradient, and AE exchanges extracellular Cl⁻ for intracellular HCO₃⁻. When pH_i is increased, i.e. when cell HCO₃⁻ increases, the inward-directed HCO₃⁻ gradient is reduced, thus further favouring efflux of HCO₃⁻ via electroneutral AE from the cell. Thus, AE functions as an ‘acid-loader’ by removing cell HCO₃⁻ during acute cellular alkalosis.

Recent studies have raised the possibility that CFTR and AE may exhibit a complex regulatory interaction when both proteins are co-localized in the same membrane. In this regard, Lee and co-workers (1999b) reported regulation of AE by CFTR expressed in NIH 3T3 and HEK 293 cells, as well as regulation of luminal AE by CFTR in mouse submandibular and pancreatic ducts (Lee et al. 1999a). In contrast to these studies, Mastrocola et al. (1998) reported that the transfection of wt CFTR, ΔF508 CFTR or vector did not influence AE activity in C129 cell lines. Whether such a relationship exists in airway epithelia will, in part, depend on determining whether CFTR and AE indeed co-localize at the apical barrier of airway epithelia. Utilizing RT-PCR, Dudeja and co-workers (1999) reported that the message for the AE2 and brain AE3 isoforms, but not AE1 or cardiac AE3 isoforms, was present in proximal human airways. Currently, however, no definitive data are available as to the apical/basolateral membrane distribution of any AE, or on the relative contributions of AE and CFTR on overall HCO₃⁻ fluxes in human normal and CF nasal epithelia.

Consequently, the aim of the present study was to investigate the polarized distribution (apical/basolateral membrane) of HCO₃⁻ transport in nasal epithelial cells in the presence and absence of functional CFTR. To this end, polarized normal and CF preparations were loaded with the pH-sensitive probe 2′,7′-bis(carboxyethyl)-5(6)-carboxyfluorescein (BCECF) to measure pH_i, mounted in a bilateral perfusion chamber, and the response of pH_i to ion substitution and inhibitor protocols employed to determine the relative roles of conductive, i.e. CFTR-mediated, versus AE-mediated HCO₃⁻ transport.

Methods

Subjects

Primary human nasal epithelial tissues were obtained, with written consent, from 12 normal subjects (32 ± 5 years old; seven males, five females) undergoing elective surgery for standard medical indications (e.g. sleep apnoea secondary to nasal obstruction) and nine cystic fibrosis patients (17 ± 5 years old; five males, four females) undergoing polypectomy. All procedures were approved by the University of North Carolina Committee for the Rights of Human Subjects and conformed to the Declaration of Helsinki.

Chemicals and solutions

The acetoxymethyl ester of BCECF (BCECF/AM) and 4,4′-diisothiocyanatodihydrostilbene-2,2′-disulfonic acid (H₂DIDS) were purchased from Molecular Probes, Inc. (Eugene, OR, USA). Diphenylamine-2-carboxylate (DPC) was obtained from Research Biochemicals International (Natick, MA, USA). All other chemicals were obtained from Sigma Chemical Co. (St Louis, MO, USA).

The standard Krebs bicarbonate Ringer (KBR) solution contained (mm): 125 NaCl, 2.5 K₂HPO₄, 1.3 CaCl₂, 1.3 MgCl₂, 25 NaHCO₃ and 5 d-glucose (5 % CO₂/95 % O₂; pH 7.4). For Cl⁻-free KBR, NaCl and MgCl₂ were replaced mole-for-mole by sodium gluconate and MgSO₄, respectively, and CaCl₂ was replaced with 2 mm CaSO₄ to compensate for Ca²⁺ chelation by gluconate, as previously reported (Clarke & Boucher, 1992). The Hepes-buffered Na⁺- and Cl⁻-free Ringer solution contained (mm): 150 N-methyl-d-glucamine (NMG) gluconate, 2.5 K₂HPO₄, 2.0 CaSO₄, 1.3 MgSO₄, 5 d-glucose and 10 Hepes (pH 7.4). For the Hepes-buffered Na⁺-free Ringer solution containing Cl⁻, the chemicals were identical to Na⁺- and Cl⁻-free Ringer solution except that NMG gluconate was replaced mole-for-mole by NMGCl (pH 7.4).

Cell culture and perfusion chamber

Human nasal epithelial (HNE) cells were harvested from polyps by enzymatic digestion (Protease XIV (Sigma) for 24–48 h at 4 °C) as previously described (Wu et al. 1985). HNE cells were plated on porous Transwell Col filters (pore diameter = 0.40 μm; Costar, Corning, Inc., NY, USA) affixed to O-rings and maintained in serum-free Ham's F-12 medium supplemented with insulin (10 μg ml⁻¹), transferrin (5 μg ml⁻¹), triiodothyronine (3 × 10⁻⁸m), endothelial cell growth supplement (3.75 μg ml⁻¹), hydrocortisone (5 × 10⁻⁹m) and CaCl₂ (10⁻³m). Polarized monolayers were studied 10–12 days after seeding, as previously described (Paradiso et al. 2001).

After achieving confluence, polarized monolayers of HNE were loaded with BCECF (5 μm at 37 °C for 25 min) and mounted in a miniature chamber over an objective (Zeiss LD Achroplan ×40, NA 0.6; working distance 1.8 mm) of a Zeiss Axiovert 35 microscope. The cell chamber and the method for independently perfusing the apical and basolateral membranes of cells have been previously described in detail (Paradiso et al. 2001).

Fluorimeter and measurements of pH_i

Measurements of pH_i in polarized HNE cells were obtained using a RatioMaster fluorimeter (Photon Technology International, Brunswick, NJ, USA) attached via fibre optics to the microscope. BCECF fluorescence from 30–40 cells (spot diameter ∼65 μm) was acquired alternately at 440 and 490 nm (emission ≥ 520 nm). At a given excitation wavelength (440 or 490 nm), background light levels were measured in non-loaded cells and subtracted from the corresponding signal measured in BCECF-loaded cells prior to taking the ratio (490/440). The corrected ratio was converted to pH_i as previously described (Paradiso, 1997).

Data analysis

Means ±s.e.m. were calculated from the total number of measurements for a given experimental condition. Statistical significance was determined using the Student's paired t test, with P < 0.05 being considered significant.

Results

Basal pH_i in normal and CF HNE cells bilaterally perfused with nominally CO₂/HCO₃⁻-free Ringer solution or KBR

Employing H⁺-sensitive microelectrodes, an earlier study by Willumsen & Boucher (1992) reported that basal pH_i was 7.1–7.15 in normal and CF HNE cells bilaterally perfused with KBR (5 % CO₂/25 mm HCO₃⁻, pH 7.4). They also reported that basal steady-state pH_i was not affected by the removal of HCO₃⁻ from the bathing medium, suggesting that HCO₃⁻ may not be important for the maintenance of steady-state basal pH_i.

To reassess the role of a CO₂/HCO₃⁻ buffer system on basal pH_i, we compared the effects of Hepes-buffered NaCl Ringer solution and KBR on basal pH_i in normal and CF HNE cells. As shown in Fig. 1, when normal (Fig. 1A, n = 10, four individuals) and CF (Fig. 1B, n = 10, four individuals) HNE cells were bilaterally exposed to Hepes-buffered NaCl Ringer (pH 7.4) solution, basal pH_i was ∼7.1 in both cell preparations. When normal and CF cells were subsequently exposed to symmetrical KBR (Fig. 1A and B), pH_i first rapidly decreased (due to the hydration of CO₂) and then regulated back to a new steady-state basal pH_i, which was lower than the values measured in a Hepes-buffered Ringer solution, in both cell preparations. On average, in Hepes-buffered NaCl Ringer solution, basal pH_i was 7.16 ± 0.03 (n = 17; four individuals) and 7.13 ± 0.02 (n = 17; four individuals) in normal and CF HNE cells, respectively, consistent with our previous measurements of basal pH_i in Hepes-buffered Ringer solution (also refer to Table 1 of Paradiso, 1997). In contrast, in the presence of bilateral KBR, basal pH_i was 6.95 ± 0.02 (n = 138; twelve individuals) and 6.94 ± 0.01 (n = 102; nine individuals) in normal and CF HNE cells, respectively.

Figure 1. Basal intracellular pH (pH_i) in polarized monolayers of normal and cystic fibrosis (CF) human nasal epithelial (HNE) cells bilaterally perfused with nominally CO₂/HCO₃⁻-free Ringer solution or Krebs bicarbonate Ringer (KBR).

For both normal (A) and CF (B) cell preparations, HNE cells were initially bilaterally (BL) perfused with Hepes-buffered NaCl Ringer solution (pH 7.4) before changing to bilateral KBR (pH 7.4) at the times indicated in the tracings. Each trace is representative of ten separate experiments (four different individuals).

Our results demonstrate that steady-state basal pH_i in normal and CF HNE cells is ∼0.15 pH units more acidic in KBR compared to Hepes-buffered NaCl Ringer solution, clearly suggesting a role for a CO₂/HCO₃⁻ buffer system for the maintenance of cell pH. This issue is discussed further below.

Effects of amiloride on pH_i in response to an CO₂/HCO₃⁻ challenge in normal and CF HNE cells

In nominally CO₂/HCO₃⁻-free NaCl Ringer solution, we have previously identified an amiloride-sensitive Na⁺/H⁺ exchanger that was restricted to the basolateral membrane in both normal and CF HNE cells (Paradiso, 1997). To extend these observations, we examined whether the Na⁺/H⁺ exchanger was the major transporter that could account for the recovery of pH_i during an acid challenge in a CO₂/HCO₃⁻ buffer system (see Fig. 1). As shown in Fig. 2, when normal (Fig. 2A, n = 7, three individuals) and CF (Fig. 2B, n = 7, three individuals) HNE cells were symmetrically perfused with Hepes-buffered NaCl Ringer solutions, the addition of amiloride (500 μm) to the serosal compartment elicited no change in basal pH_i. However, in the presence of serosal amiloride, when the perfusate was subsequently changed to bilateral KBR, pH_i rapidly decreased to ∼6.6 and remained acidic in both normal and CF cell preparations; pH_i recovered towards a new steady-state basal level in both normal and CF HNE cells only when amiloride was removed from the serosal perfusate (Fig. 2).

Figure 2. Effects of amiloride on pH_i in response to a CO₂/HCO₃⁻ challenge in polarized monolayers of normal and CF HNE cells.

For both normal (A) and CF (B) cell preparations, HNE cells were initially bilaterally perfused with Hepes-buffered NaCl Ringer solution (pH 7.4) before adding amiloride (Amil., 500 μm) to the serosal (S) perfusate. In the presence of amiloride, the perfusate was subsequently changed to bilateral KBR (pH 7.4) and amiloride removed from the serosal compartment at the times indicated in the tracings. Each trace is representative of seven separate experiments (three different individuals).

The data thus far strongly suggest that the major membrane mechanism for the recovery of pH_i in response to an acid challenge in the presence (Fig. 2) and absence (Paradiso, 1997) of a CO₂/HCO₃⁻-buffered solution is a basolateral Na⁺/H⁺ exchanger in normal and CF HNE cells. Because amiloride completely and reversibly blocked the recovery phase of the acid-load, a role for a Na⁺-HCO₃⁻ cotransporter, i.e. the influx of Na⁺ and HCO₃⁻ across the apical/basolateral membrane, is unlikely. Finally, the observation that amiloride added to the serosal perfusate induced no change of pH_i in nominally CO₂/HCO₃⁻-free (Hepes-buffered) Ringer solution (Fig. 2) indicates that the basolateral Na⁺/H⁺ exchanger is inactive, i.e. there is no net efflux of H⁺ coupled to influx of Na⁺ via the basolateral Na⁺/H⁺ exchanger at a basal pH_i ≥ 7.1, as previously reported in normal and CF HNE cells (Paradiso, 1997).

Acute effects of H₂DIDS and amiloride on basal pH_i

To identify the membrane transporters that acutely modulate pH_i under basal conditions in a CO₂/HCO₃⁻ buffer system, we tested the effects of mucosal and serosal H₂DIDS (to block a putative AE) and amiloride (to inhibit the Na⁺/H⁺ exchanger) in normal and CF HNE cells bilaterally perfused with KBR.

As depicted in Fig. 3, exposing normal (Fig. 3A) or CF (Fig. 3B) cells to H₂DIDS (1.5 mm) added to the mucosal perfusate caused no changes in basal pH_i. In contrast, serosally applied H₂DIDS elicited an alkalinization in both normal (Fig. 3A) and CF (Fig. 3B) HNE cells. On average, the magnitude of the increase (peak-basal value) of pH_i in response to serosal H₂DIDS was 0.10 ± 0.01 (n = 84, nine individuals) and 0.11 ± 0.01 (n = 68, nine individuals) in normal and CF HNE cells, respectively.

Figure 3. Response of pH_i to mucosal and serosal additions of H₂DIDS and amiloride in polarized monolayers of normal and CF HNE cells.

For both normal (A and C) and CF (B and D) cell preparations, HNE cells were bilaterally perfused with KBR before adding 4,4′-diisothiocyanatodihydrostilbene-2,2′-disulfonic acid (H₂DIDS; 1.5 mm) and amiloride (500 μm) to the mucosal (M) and serosal (S) compartments as indicated in the tracings. Each trace is representative of seven separate experiments (three different individuals).

The increase in pH_i in response to H₂DIDS could potentially result from two sources: (i) the activity of a Na⁺/H⁺ exchanger during inhibition of AE and/or (ii) the cellular accumulation of HCO₃⁻ during the inhibition of a basolateral AE. To test whether Na⁺/H⁺ exchange contributed to the alkalinization of cells in response to H₂DIDS added to the serosal perfusate, we examined the response of pH_i to amiloride after H₂DIDS treatment. As shown in Fig. 3, the addition of amiloride (500 μm) to the mucosal perfusate failed to alter pH_i in either H₂DIDS-treated normal (Fig. 3A) or CF (Fig. 3B) HNE cells. In contrast to mucosal amiloride addition, serosally applied amiloride (500 μm) following H₂DIDS-elicited alkalinization induced a re-acidification of cell pH back to basal values in both normal (Fig. 3A) and CF (Fig. 3B) cell preparations, consistent with the inhibition of basolateral Na⁺/H⁺ exchange. Furthermore, under basal conditions, mucosal addition of amiloride again failed to alter pH_i, whereas the addition of amiloride to the serosal perfusate caused cell pH to decrease in both normal (Fig. 3C) and CF (Fig. 3D) HNE cells. As noted above, these data are consistent with our previous report that the Na⁺/H⁺ exchanger is restricted to the basolateral membrane in both normal and CF HNE cells (Paradiso, 1997). On average, the magnitude of the decrease of pH_i in response to serosal amiloride was 0.08 ± 0.02 (n = 7, three individuals) and 0.09 ± 0.01 (n = 7, three individuals) in normal and CF HNE cells, respectively.

The absolute change of pH_i in normal and CF cells in response to serosally applied amiloride was not significantly different from the increase in cell pH induced by the serosal addition of H₂DIDS, measured in both cell preparations. Furthermore, serosally applied H₂DIDS following serosal amiloride treatment caused only a small or no change of pH_i in both normal (Fig. 3C) and CF (Fig. 3D) HNE cells, consistent with little accumulation of HCO₃⁻ within cells under basal conditions during the short time period employed in these studies. Taken together, these results indicate that constitutively active basolateral, but not apical, H₂DIDS-sensitive AE and Na⁺/H⁺ exchange contribute to the acute regulation of pH_i in normal and CF airway epithelia under basal conditions in the presence of a CO₂/HCO₃⁻ buffer system.

Effects of mucosal Cl⁻ substitution on pH_i in normal and CF HNE cells

The results thus far indicate that normal and CF airway epithelia lack a H₂DIDS-sensitive pathway for modulating pH_i at their apical membrane (see Fig. 3), suggesting the absence of an apically located AE. To test this notion more rigorously, we investigated whether changes in the luminal Cl⁻ concentration could alter pH_i in normal and CF HNE cells. Our expectation was that if AE were present at the apical membrane of cells, then removal of luminal Cl⁻, i.e. establishing a cell-to-lumen Cl⁻ gradient, should increase pH_i via a lumen-to-cell influx of HCO₃⁻. For these studies, normal and CF HNE cells were perfused initially with bilateral KBR and effects of unilateral replacement of mucosal Cl⁻ with gluconate (constant 5 %P_CO2/25 mm HCO₃⁻; pH 7.4) examined. As depicted in Fig. 4, removal of luminal Cl⁻ (serosal bath constant KBR, pH 7.4) induced a transient alkalinization that returned to baseline in normal HNE cells (Fig. 4A). The mean maximum change (peak-basal value) of pH_i elicited by this mucosal Cl⁻-free treatment was 0.18 ± 0.02 (n = 7, four individuals). In contrast, this manoeuvre failed to alter pH_i in CF airway epithelia (Fig. 4B; n = 8, three individuals).

Figure 4. Effects of mucosal Cl⁻-free Ringer solution on pH_i in polarized monolayers of normal and CF HNE cells.

For both normal (A) and CF (B) cell preparations, HNE cells were initially bilaterally perfused with KBR before changing to mucosal (M) Cl⁻-free (gluconate replacing Cl⁻; constant 5 % P_CO2/25 mm HCO₃⁻, pH 7.4) Ringer solution at the times indicated in the tracings. The serosal perfusate remained KBR throughout these studies. A, representative example of seven separate experiments (four different individuals). B, representative example of eight separate experiments (three different individuals).

Because HNE cells may have at their basolateral border an H₂DIDS-sensitive AE (see Fig. 3), it is possible that any Cl⁻ gradient imposed across the apical membrane was ultimately also imposed across the basolateral membrane by lowering intracellular Cl⁻, resulting in an increase in the serosal-to-cell Cl⁻ gradient across the basolateral membrane. For example, a change in pH_i due to an outward-directed (cell-to-lumen) Cl⁻ gradient coupled to an inward-directed (lumen-to-cell) HCO₃⁻ influx pathway may be reduced (shunted) if a dominant basolateral AE removes HCO₃⁻ entering the cell across the apical membrane. Therefore, to ‘isolate’ the influx of HCO₃⁻ across the apical membrane from the efflux of HCO₃⁻ across the basolateral membrane in response to changes in luminal Cl⁻ concentration, we added H₂DIDS to the serosal perfusate to block any compensatory effects of the putative basolateral AE on HCO₃⁻-dependent changes of pH_i. As shown in Fig. 5, in the presence of serosal H₂DIDS, replacing luminal Cl⁻ with gluconate (imposing a cell-to-lumen Cl⁻ gradient) alkalinized normal cells, consistent with significant lumen-to-cell HCO₃⁻ uptake in normal HNE cells (Fig. 5A). The maximum change of pH_i from steady-state level (induced by H₂DIDS) to peak value (elicited by Cl⁻ removal) was 0.38 ± 0.02 (n = 11, four individuals), which was significantly greater (P < 0.01) than the change in pH_i induced by luminal Cl⁻ removal in non-H₂DIDS-treated normal cells (i.e. 0.18 ± 0.02 pH units; see above). Furthermore, returning Cl⁻ to the luminal perfusate (increasing the lumen-to-cell Cl⁻ gradient) stimulated cells to re-acidify back to basal levels, consistent with cell-to-lumen HCO₃⁻ movement in normal airway epithelia (see Fig. 5A). In contrast, removal of luminal Cl⁻ in the presence of serosal H₂DIDS induced no detectable change of pH_i in CF airway cells (Fig. 5B; n = 8, three individuals). It should be noted that we have previously shown that these Cl⁻-linked changes of pH_i in normal (or CF HNE) cells are absent in a nominally CO₂/HCO₃⁻-free environment (see Fig. 3 of Paradiso, 1992), indicating that the responses of pH_i to altered luminal Cl⁻ concentrations reflect movement of HCO₃⁻ across the apical membrane in normal nasal cells.

Figure 5. Effects of mucosal Cl⁻ substitution on pH_i in polarized monolayers of normal and CF HNE cells exposed to serosal and/or mucosal H₂DIDS.

For both normal (A and C) and CF (B) HNE cell preparations, cells were initially perfused bilaterally with KBR and H₂DIDS (1.5 mm) added to the serosal and/or mucosal compartment before removing Cl⁻ (gluconate replacing Cl⁻; constant 5 % P_CO2/25 mm HCO₃⁻, pH 7.4) or re-adding Cl⁻ to the mucosal perfusate, as shown in the tracings. The serosal perfusate remained KBR throughout the study. A and C, representative examples of eleven separate experiments (four different individuals). B, is representative of eight separate experiments (three different individuals).

Although the changes in cell pH shown in Fig. 5A in response to altering mucosal Cl⁻ could be interpreted as being consistent with the activity of AE at the apical membrane in normal HNE cells, Fig. 5C shows that these Cl⁻-dependent changes of pH_i were not blocked by mucosal H₂DIDS (1.5 mm). On average, in the presence of mucosal H₂DIDS, the magnitude of change of pH_i from steady-state level (induced by serosal H₂DIDS) to peak value obtained in luminal Cl⁻-free was 0.36 ± 0.03 (n = 11, four individuals), which was not significantly different from the mean value determined in the absence of mucosal H₂DIDS.

Effects of forskolin and DPC on pH_i in response to mucosal Cl- substitution in normal HNE cells

Three observations reported here and by others suggest that the change in pH_i in response to the removal/re-addition of luminal Cl⁻ may be mediated by the movement of HCO₃⁻ across the apical membrane via CFTR, which appears to be constitutively active in normal HNE cells. First, unilaterally replacing mucosal Cl⁻ with gluconate induced an alkalinization of pH_i under basal conditions in normal (Fig. 5A), but not CF (Fig. 5B), HNE cells. Second, removal of luminal Cl⁻ under basal conditions causes a major depolarization of the apical membrane electrical potential difference (V_a) in normal, but not CF, HNE cells (Willumsen et al. 1989). Depolarization of V_a would favour conductive entry of HCO₃⁻ into the cell, perhaps via CFTR (Illek et al. 1997), which would lead to cell alkalinization. Finally, recent studies by Huang et al. (2001) reported substantial CFTR channel activity in Calu-3 cells, a human airway cell line, under basal conditions that was approximately half the activity measured after forskolin addition, suggesting that CFTR Cl⁻ channels were half-maximally activated under basal conditions. Moreover, they showed that the constitutive activation of the CFTR Cl⁻ channel reflected release of 5′-adenosine triphosphate (ATP) into the lumen that was metabolically degraded into adenosine. Adenosine activated the apically located adenosine receptors (A_2B-subtype) and adenylyl cyclase present in the apical membrane by means of a G protein (i.e. G_s). Sufficient adenosine 3′,5′-cyclic monophosphate (cAMP) was generated to activate protein kinase A, and hence CFTR, in a diffusionally restricted apical microdomain, without increasing cAMP in other cellular compartments. Similar constitutive release of ATP and generation of adenosine have been reported both in vivo (Donaldson et al. 2000) and in vitro (Watt et al. 1998) in normal and CF HNE cells, and constitutive activation of CFTR (∼half-maximal) has been reported in normal nasal epithelium in vivo (Knowles et al. 1995) under basal conditions.

To investigate the effects of maximal CFTR activation on pH_i, we examined the effects of forskolin on changes in mucosal Cl⁻ concentration in normal HNE cells. As shown in Fig. 6, in the presence of serosal H₂DIDS the addition of forskolin (10 μm) to the mucosal/serosal perfusate and subsequently H₂DIDS to the mucosal compartment elicited no change in pH_i in normal HNE cells perfused bilaterally with KBR. In the presence of forskolin and bilateral H₂DIDS, replacing luminal Cl⁻ with gluconate (imposing a cell-to-lumen Cl⁻ gradient) again alkalinized cells, consistent with significant lumen-to-cell HCO₃⁻ uptake in normal HNE cells (Fig. 6A). The maximum change of pH_i from steady-state level (induced by H₂DIDS) to peak value (elicited by Cl⁻ removal) was 0.37 ± 0.02 (n = 7, three individuals), which was not significantly different from the change in pH_i induced by luminal Cl⁻ removal in non-stimulated normal cells treated with bilateral H₂DIDS (i.e. 0.36 ± 0.03 pH units; also see Fig. 5C). Furthermore, returning Cl⁻ to the luminal perfusate (increasing the lumen-to-cell Cl⁻ gradient) stimulated cells to re-acidify back to basal levels, again consistent with cell-to-lumen HCO₃⁻ movement in normal airway epithelia (see Fig. 6A).

Figure 6. Effects of forskolin and diphenylamine-2-carboxylate (DPC) on pH_i in response to mucosal Cl⁻ substitution in polarized monolayers of normal HNE cells.

A, normal HNE cells were initially perfused with bilateral (BL) KBR. Serosal H₂DIDS (1.5 mm) and bilateral forskolin (F; 10 μm) were added to the perfusate at the times indicated in the tracing, before changing the mucosal perfusate from KBR to Cl⁻-free (gluconate replacing Cl⁻; constant 5 % P_CO2/25 mm HCO₃⁻, pH 7.4) KBR. Cl⁻ was subsequently re-added to the mucosal compartment at the time indicated in the tracing. B, normal HNE cells were bilaterally perfused with KBR and mucosal DPC (100 μm) and serosal H₂DIDS (1.5 mm) added to the perfusate at the times indicated in the tracing. In the presence of DPC, changing the mucosal perfusate from KBR to Cl⁻-free KBR elicited no change in pH_i, whereas removal of DPC induced an alkalinization of the cells. For both traces, the serosal perfusate remained KBR throughout the study. A, representative example of seven separate experiments (three different individuals). B, representative example of seven separate experiments (four different individuals).

In the absence (see Fig. 5C) and presence (see Fig. 6A) of forskolin, the initial rate of alkalinization (ΔpH_i min⁻¹) in normal HNE cells, measured over the pH_i range 7.05 to 7.20 in response to mucosal Cl⁻ removal, was 0.14 ± 0.02 (n = 11; four individuals) and 0.19 ± 0.02 (n = 7; four individuals), respectively. In the presence of forskolin, the initial rate in the change in pH_i (i.e. ΔpH_i min⁻¹ = 0.19) induced by luminal Cl⁻ removal in normal cells was marginally greater (P < 0.05) than the change in pH_i (i.e. ΔpH_i min⁻¹= 0.14) induced by luminal Cl⁻ removal in non-stimulated normal HNE cells. This difference represents an increase in the rate of alkalinization in forskolin-treated normal cells over non-stimulated cells by ∼26 %. Furthermore, in the presence of bilateral H₂DIDS, bilaterally applied forskolin (10 μm) had no effect on steady-state pH_i in the presence or absence of luminal Cl⁻ in CF HNE cells (n = 7; three individuals; data not shown).

Although there are no available blockers that selectively inhibit the CFTR Cl⁻ channel activity without potentially affecting other membrane channels/transporters, DPC has been reported (Stutts et al. 1990; Zhang et al. 2000) to decrease CFTR anion conductance in several cell types, including airway epithelial cells. We utilized DPC to investigate whether the Cl⁻-induced changes in pH_i in response to unilaterally replacing Cl⁻ with gluconate in the mucosal bath were mediated via CFTR. As depicted in Fig. 6B, in normal HNE cells bilaterally perfused with KBR, the addition of DPC (100 μm) to the mucosal perfusate elicited little or no change in basal pH_i, whereas the subsequent administration of serosal H₂DIDS again caused pH_i to increase. However, mucosal DPC attenuated the increase of pH_i in response to mucosal Cl⁻-free Ringer solution (compare Fig. 6B with Fig. 5A). Notably, when DPC was removed from the mucosal compartment, the inhibitory effect of DPC was reversed, and the cells alkalinized. The maximum average change of pH_i from steady-state level (induced by serosal H₂DIDS) to peak value following luminal DPC removal was 0.18 ± 0.02 (n = 7, four individuals).

Effects of mucosal DPC on pH_i in response to serosal Ba²⁺ in normal and CF HNE cells

Previous studies on HNE cells have reported that the V_a also reflects, in part, the basolateral K⁺ conductance, i.e. cell depolarization in response to the blockade of a basolateral K⁺ conductance will be reflected in a depolarization of V_a (Willumsen et al. 1989). Based on this information, we designed protocols to test whether CFTR transports HCO₃⁻ in response to altered V_a by exposing polarized HNE cells to serosal Ba²⁺, a known K⁺ channel blocker (Devor & Frizzell, 1998).

For these studies, we first tested whether Ba²⁺ could alter pH_i in cells perfused with a nominally CO₂/HCO₃⁻ buffer solution. As illustrated in Fig. 7A, when normal HNE cells were bilaterally perfused with Hepes-buffered NaCl Ringer solution, the addition of H₂DIDS (1.5 mm) and subsequently Ba²⁺ (5.0 mm) to the serosal compartment induced no change in basal pH_i. However, when normal (Fig. 7B and C) and CF (Fig. 7D) HNE cells were perfused with bilateral KBR, the addition of H₂DIDS (1.5 mm) to the serosal perfusate again caused pH_i to increase. Moreover, in the presence of bilateral KBR, the subsequent addition of serosally applied Ba²⁺ (5.0 mm) to normal airway epithelia induced a substantial and further increase in pH_i that was completely reversible with the removal of Ba²⁺ from the serosal compartment (Fig. 7B). The magnitude of the change of pH_i from the H₂DIDS-treated to peak values elicited by Ba²⁺ was 0.19 ± 0.03 (n = 6, three individuals). Furthermore, in normal HNE cells, the mucosal addition of DPC (100 μm) reversibly blocked changes of pH_i in response to serosal Ba²⁺ (Fig. 7C). Finally, neither mucosal DPC (100 μm) nor serosal Ba²⁺ (5.0 mm) altered pH_i in CF airway cells (Fig. 7D). Collectively, these data support a role for CFTR in HCO₃⁻ transport across the apical membrane in normal HNE cells.

Figure 7. Effects of mucosal DPC on pH_i in response to serosal Ba²⁺ in polarized monolayers of normal and CF HNE cells.

A, normal HNE cells were bilaterally perfused with Hepes-buffered NaCl Ringer solution before adding H₂DIDS (1.5 mm) and Ba²⁺ (5.0 mm) to the serosal compartment at the times shown in the trace. For both normal (B and C) and CF (D) HNE cell preparations, cells were perfused bilaterally with KBR and H₂DIDS (1.5 mm) administrated to the serosal perfusate. Following H₂DIDS treatment, DPC (100 μm) and/or Ba²⁺ (5.0 mm) were added to and/or removed from the serosal and/or mucosal compartment in both normal (B and C) and CF (D) cell preparations at the times shown in the tracings. Each trace is representative of six separate experiments (three different individuals).

Effects of serosal Cl⁻ substitution on pH_i in normal and CF HNE cells

For these studies, normal and CF HNE cells were initially perfused in bilateral KBR with H₂DIDS (1.5 mm) added to the serosal perfusate to block the putative basolateral AE. As shown in Fig. 8, in the presence of serosal H₂DIDS, changing the serosal perfusate to Cl⁻-free (gluconate replacement) KBR (mucosal bath constant KBR, pH 7.4) caused no change in pH_i in either normal (Fig. 8A) or CF (Fig. 8B) airway cells. However, with serosal Cl⁻-free Ringer solution, i.e. in the presence of a cell-to-serosal Cl⁻ gradient, cell pH increased in both normal and CF HNE cells when H₂DIDS was removed from the serosal compartment. The initial rate of alkalinization (ΔpH_i min⁻¹) measured over the pH_i range 7.05–7.25 in response to serosal Cl⁻ substitution was 0.16 ± 0.04 (n = 10; four individuals) and 0.18 ± 0.03 (n = 10; four individuals) in normal and CF HNE cells, respectively. These initial rates were not significantly different. It should be noted that the maximum absolute change of pH_i in response to serosal Cl⁻ substitution was slightly smaller in normal (ΔpH_i= 0.48 ± 0.02) compared to CF (ΔpH_i= 0.56 ± 0.01) HNE cells. Moreover, unlike CF HNE cells, there was a spontaneous secondary slow re-acidification of pH_i following the peak alkalinization in normal airway cells (Fig. 6A). These two observations for normal and CF HNE cells are addressed later in the study. Furthermore, as shown in Fig. 8, the re-addition of serosal Cl⁻ (establishing a serosal-to-cell Cl⁻ gradient) caused both normal (Fig. 8A) and CF (Fig. 8B) HNE cells to re-acidify rapidly back toward basal pH_i values. Taken together, these results are consistent with the activity of a H₂DIDS-sensitive AE at the basolateral membrane in both normal and CF cell preparations.

Figure 8. Effects of H₂DIDS and serosal Cl⁻ substitution on pH_i in polarized monolayers of normal and CF HNE cells.

For both normal (A) and CF (B) HNE cell preparations, cells were initially perfused bilaterally with KBR and H₂DIDS (1.5 mm) added to the serosal compartment before removing Cl⁻ (gluconate replacing Cl⁻; constant 5 % P_CO2/25 mm HCO₃⁻, pH 7.4) from the serosal perfusate, as shown in the tracings. In the presence of serosal Cl⁻-free KBR, removal of serosal H₂DIDS caused pH_i to alkalinize, whereas the re-addition of serosal Cl⁻ caused cell pH to re-acidify back to basal level. The mucosal perfusate remained KBR throughout the study. Each tracing is representative of ten separate experiments (four different individuals).

Effects of mucosal and serosal Cl⁻ on pH_i in alkalinized normal and CF HNE cells

Because AE can be broadly divided into Na⁺-dependent AE, i.e. requiring external Na⁺ for the cycling of extracellular Cl⁻ for intracellular HCO₃⁻, or Na⁺-independent AE (Boron, 1986), we next tested for Na⁺-dependent and Na⁺-independent AE activity at the apical and basolateral domains of normal and CF HNE cells. For Na⁺-dependent AE, HNE cell preparations were initially perfused bilaterally with KBR, and the perfusate subsequently changed to bilateral Hepes-buffered Cl⁻-free (sodium gluconate; pH 7.4 (nominally CO₂- and HCO₃⁻-free)) Ringer solution. This manoeuvre alkalinized cells due to rapid loss of cell CO₂, thereby establishing a cell-to-extracellular HCO₃⁻ gradient across the apical/basolateral domain of airway cells. Following cell alkalinization, the Cl⁻ dependency of recovery from an alkaline/HCO₃⁻ load in the presence of extracellular Na⁺ was tested by exposing cells to mucosal or serosal Cl⁻ (Hepes-buffered NaCl Ringer solution).

As shown in Fig. 9A, the pattern in the change of pH_i was markedly different between normal and CF HNE cells when the perfusate was changed from symmetrical KBR to symmetrical Hepes-buffered Cl⁻-free Ringer solution. In normal cells, the initial increase of pH_i following bilateral Cl⁻-free treatment was followed by a spontaneous secondary re-acidification back towards the basal value. In contrast to normal cells, CF HNE cells rapidly alkalinized and no secondary re-acidification was noted. Moreover, the magnitude of the increase of pH_i following bilateral Cl⁻-free treatment was significantly (P < 0.01) reduced in normal (ΔpH_i= 0.54 ± 0.04, n = 9; three individuals) compared to CF (ΔpH_i= 0.79 ± 0.03, n = 9; three individuals) HNE cells. In the presence of bilateral Na⁺, the addition of Cl⁻ to the mucosal bath (a condition that established a lumen-to-cell gradient for Cl⁻ and a cell-to-lumen gradient for HCO₃⁻) had little effect on the rate of the secondary re-acidification in normal HNE cells and elicited no change in pH_i in CF cells (Fig. 9A). However, when Cl⁻ was added to the serosal perfusate, pH_i rapidly returned towards baseline in both cell preparations (Fig. 9A).

Figure 9. Effects of mucosal and serosal Cl⁻(± Na⁺) on pH_i in polarized monolayers of alkalinized normal and CF HNE cells.

For both normal and CF HNE cell preparations, cells were initially perfused bilaterally with KBR. In one set of experiments shown in panel A, HNE (normal and CF) cells were exposed to bilateral Hepes-buffered Cl⁻-free (sodium gluconate; no CO₂ or HCO₃⁻, pH 7.4) Ringer solution to alkalinize the cells (due to the rapid loss of cell CO₂). In a second set of experiments, shown in panel B, HNE (normal and CF) cells were exposed to bilateral Hepes-buffered Na⁺- and Cl⁻-free (N-methyl-d-glucamine gluconate replacing Na⁺ and Cl⁻; no CO₂ or HCO₃⁻, pH 7.4) Ringer solution. Following alkalinization of the cells, Cl⁻ was re-added to the mucosal or serosal perfusate at the times depicted in the tracings. Each trace is representative of nine separate experiments (three different individuals).

To test for the Na⁺ dependency of AE activity, HNE cell preparations were initially perfused bilaterally with KBR and the perfusate subsequently changed to bilateral Hepes-buffered Na⁺- and Cl⁻-free (NMG gluconate; pH 7.4 (nominally CO₂- and HCO₃⁻-free)) Ringer solution to alkalinize cells (Fig. 9B). Following cell alkalinization, the Cl⁻ dependency of recovery from an alkaline load was again tested by exposing cells to mucosal or serosal Cl⁻ (Hepes-buffered NMGCl Ringer solution). As shown in Fig. 9B, in normal HNE cells, the initial increase of pH_i following bilateral Na⁺- and Cl⁻-free treatment was followed by a spontaneous secondary re-acidification back towards baseline. Again in contrast to normal cells, CF HNE cells rapidly alkalinized and no secondary re-acidification was detected. Moreover, the maximum absolute increase of pH_i following bilateral Na⁺- and Cl⁻-free treatment was again significantly (P < 0.01) reduced in normal (ΔpH_i= 0.59 ± 0.03, n = 9; three individuals) compared to CF (ΔpH_i= 0.83 ± 0.02, n = 9; three individuals) HNE cells. The addition of Cl⁻ to the mucosal bath (again, a condition that established a lumen-to-cell gradient for Cl⁻ and a cell-to-lumen gradient for HCO₃⁻) slightly decreased or had no effect on the rate of the secondary re-acidification in normal HNE cells and elicited no change in pH_i in CF cells (Fig. 9B). In contrast to the mucosal addition of Cl⁻, when Cl⁻ was added to the serosal perfusate, pH_i rapidly accelerated towards baseline in both airway cell preparations (Fig. 9B). The observation that re-acidification of pH_i in response to serosal addition of Cl⁻ in alkalinized cells was similar in normal and CF epithelia with (see Fig. 9A) or without Na⁺ (see Fig. 9B) suggests that the predominant AE pathway is Na⁺-independent.

Effects of serosal H₂DIDS on pH_i in alkalinized normal and CF HNE cells

We next investigated whether re-acidification of pH_i in response to serosal Cl⁻ could be blocked by H₂DIDS in airway cells challenged with an alkaline/HCO₃⁻ load (Fig. 10). In normal (Fig. 10A) and CF (Fig. 10B) cells bilaterally perfused with KBR, the addition of serosal H₂DIDS (1.5 mm) caused a small increase in pH_i. When the serosal and mucosal compartments were subsequently changed to Hepes-buffered Na⁺- and Cl⁻-free (NMG gluconate; nominally CO₂- and HCO₃⁻-free) Ringer solution to alkalinize cells, no recovery of pH_i was detected when Cl⁻ was added to the serosal compartment in the presence of H₂DIDS. However, removal of H₂DIDS from the serosal compartment in the continual presence of Cl⁻ caused an increase in the rate of pH_i recovery from the alkaline challenge in both normal (Fig. 10A) and CF (Fig. 10B) HNE cell preparations.

Figure 10. Effects of serosal H₂DIDS (± Cl⁻) on changes of pH_i in polarized monolayers of alkalinized normal and CF HNE cells.

For both normal (A) and CF (B) HNE cell preparations, cells were initially perfused bilaterally (BL) with KBR. In the presence of bilateral KBR, H₂DIDS (1.5 mm) was added to the serosal compartment (A and B) at the times shown in the tracings, before changing from bilateral KBR to bilateral Hepes-buffered Na⁺- and Cl⁻-free (N-methyl-d-glucamine gluconate replacing Na⁺ and Cl⁻; no CO₂ or HCO₃⁻, pH 7.4) Ringer solution. Following alkalinization of the cells, serosal H₂DIDS blocked the re-acidification of pH_i when Cl⁻ was re-added to the serosal perfusate in both normal (A) and CF (B) HNE cells, as shown in the tracings. Note that in the presence of serosal Cl⁻, removal of H₂DIDS from the serosal compartment resulted in a rapid re-acidification of pH_i in both normal and CF HNE cells (A and B). Each trace is representative of seven separate experiments (three different individuals).

Spontaneous acidification and effects of mucosal DPC on pH_i in alkalinized normal HNE cells

The spontaneous secondary acidification following an alkaline shift in normal HNE cells could potentially result from: (i) the reversal of a Na⁺/H⁺ exchanger (i.e. a cell-to-extracellular gradient of Na⁺ linked to an extracellular-to-cell gradient of H⁺), which is known to be present on the basolateral, but not apical, membrane of human nasal cells (Paradiso, 1997); (ii) the reversal of a Na⁺-HCO₃⁻ cotransporter (i.e. loss of both Na⁺ and HCO₃⁻ from cells); and/or (iii) loss of cell HCO₃⁻ via CFTR.

With regard to Na⁺/H⁺ exchange, it is possible that during an alkaline shift in pH_i, Na⁺ could exit the cell in exchange for extracellular H⁺ via the Na⁺/H⁺ exchanger, and consequently, cells would acidify. However, this possibility seems unlikely, for three reasons. First, the pattern and rate of the spontaneous secondary re-acidification were similar in the presence and absence of extracellular Na⁺ (see Fig. 9). Second, we have repeated these experiments in the absence of extracellular Na⁺, pretreating cells with serosal amiloride (500 μm; n = 6, three different individuals) to block the activity of the basolateral Na⁺/H⁺ exchanger (Paradiso, 1997). This manoeuvre did not prevent the secondary re-acidification in normal HNE cells (data not shown). Finally, we have previously reported (Paradiso, 1997) that CF HNE cells also have Na⁺/H⁺ exchanger activity at their basolateral border that is pharmacologically and kinetically identical to that of the Na⁺/H⁺ exchanger in normal nasal tissue, yet no spontaneous secondary re-acidification was detected in CF tissue following the initial alkaline shift in pH_i (see Fig. 9).

Although we do not have functional evidence for the presence of a Na⁺-HCO₃⁻ cotransporter at the apical/basolateral aspect of human nasal cells, Devor and co-workers (1999) have functionally identified a basolateral Na⁺-HCO⁻ cotransporter in Calu-3 cells. Thus, it is possible that the spontaneous re-acidification in response to an alkaline shift of pH_i in normal HNE cells resulted from the reversal of the cotransporter. However, like the argument against the involvement of Na⁺/H⁺ exchange in the re-acidification in response to an alkaline shift in pH_i in normal HNE cells, it again seems unlikely that a cell-to-extracellular movement of Na⁺ and HCO₃⁻ via Na⁺-HCO⁻ cotransport could account for the spontaneous recovery following an alkaline challenge, since, as noted above, the pattern of re-acidification was identical in Na⁺-containing (see Fig. 9A) and Na⁺-free (see Fig. 9B) Ringer solutions. Moreover, as noted above in the discussion of Fig. 2, a role for a Na⁺-HCO₃⁻ cotransporter in the regulation of pH_i in surface nasal epithelial cells is not suggested by this study, since basolaterally applied amiloride could completely and reversibly inhibit recovery of pH_i from an acid challenge when airway cells were perfused with KBR.

To elucidate whether the spontaneous secondary re-acidification towards basal values in normal cells challenged by CO₂ removal was mediated via CFTR, we tested whether mucosally applied DPC blocked this process. As shown in Fig. 11A, when normal HNE cells were initially perfused with symmetrical KBR, the addition of mucosal DPC (100 μm) elicited no change in basal pH_i. However, changing the mucosal/serosal perfusate to symmetrical Hepes-buffered Na⁺- and Cl⁻-free Ringer solution resulted in a larger alkalinization in cell pH and markedly reduced the secondary re-acidification in DPC-treated compared to non-DPC-treated normal cells (see Fig. 10A). On average, the maximum absolute increase of pH_i following bilateral Na⁺- and Cl⁻-free treatment was significantly (P < 0.01) increased in DPC-treated cells (ΔpH_i= 0.78 ± 0.02, n = 7; three different individuals) compared to non-treated (i.e. ΔpH_i of 0.59 ± 0.03, see Fig. 10A) normal HNE cells. Moreover, the subsequent removal of DPC from the mucosal bath caused pH_i to return rapidly to baseline (Fig. 11A). In contrast to normal cells, the mucosal addition or removal of DPC (100 μm) had no effect on the magnitude change of pH_i in CF HNE cells when the perfusate was changed from symmetrical KBR to symmetrical Hepes-buffered Na⁺- and Cl⁻-free Ringer solution (Fig. 11B). On average, the maximum absolute increase of pH_i following bilateral Na⁺- and Cl⁻-free treatment was not significantly different in CF cells treated with mucosal DPC (ΔpH_i= 0.85 ± 0.03, n = 7; three individuals) compared to non-DPC-treated CF cells (ΔpH_i= 0.83 ± 0.02; see Fig. 10B).

Figure 11. Effects of mucosal DPC on changes of pH_i in polarized monolayers of alkalinized normal and CF HNE cells.

For both normal (A) and CF (B) HNE cell preparations, cells were initially perfused bilaterally (BL) with KBR before adding DPC (100 μm) to the mucosal perfusate at the times shown in each trace. In the presence of DPC, changing from bilateral KBR to bilateral Hepes-buffered Na⁺- and Cl⁻-free (N-methyl-d-glucamine gluconate replacing Na⁺ and Cl⁻; no CO₂ or HCO₃⁻, pH 7.4) Ringer solution, resulted in a larger alkalinization of pH_i in normal DPC-treated cells (A) compared to normal cells not treated with DPC (see trace A of Fig. 10), and that removal of DPC elicited a rapid re-acidification of pH_i back to base level in normal (A), but not CF (B), HNE cells. Each trace is representative of seven separate experiments (three different individuals).

Discussion

Regulation of basal pH_i in normal and CF HNE cells

In a previous study (Paradiso, 1997) using polarized human nasal epithelial cells in primary culture, we reported that basal pH_i was ∼7.1 in both normal and CF cells bilaterally perfused with nominally CO₂/HCO₃⁻-free Hepes-buffered NaCl Ringer solution. Furthermore, we showed in this earlier study that both normal and CF HNE cells expressed at their basolateral membrane a Na⁺/H⁺ exchanger that was inactive at basal pH_i≥ 7.1 (Paradiso, 1997).

In the present study, when polarized normal and CF HNE cells were bilaterally perfused in CO₂-containing KBR, basal pH_i was 6.94 and 6.95, respectively (also see Fig. 1). We speculate that the lower basal pH_i in normal and CF HNE cells bilaterally perfused with KBR resulted from the inclusion of CO₂ in the perfusate, which lowered pH_i and led to the activation of both a basolateral H₂DIDS-sensitive AE and a Na⁺/H⁺ exchanger. This speculation is based on the following observations generated by this study. First, serosally applied H₂DIDS elicited an alkalinization of pH_i in both normal and CF cell preparations (see Fig. 3A and B), a response predicted for the inhibition of an active basolateral AE. Second, serosally applied amiloride induced an acidification of basal pH_i in both normal and CF cell preparations, a response consistent with the inhibition of an active basolateral Na⁺/H⁺ exchanger (see Fig. 3C and D). Third, the parallel activities of AE and a Na⁺/H⁺ exchanger were revealed by the observation that serosally applied amiloride acidified pH_i following post-H₂DIDS-induced alkalinization of cell pH (Fig. 3A and B), and that the addition of H₂DIDS subsequent to amiloride caused a small increase in pH_i (Fig. 3C and D). Taken together, our data strongly suggest that basal pH_i in airway epithelia exposed to physiological P_CO2 is maintained via the coordinated activities of a Na⁺/H⁺ exchanger and AE at the basolateral membrane in normal and CF HNE cells. The coordinated activities of both exchangers may play an important compensatory function in maintaining cell pH within narrow limits during dynamic shifts in CO₂ tension in the airway lumen during the respiratory cycle, i.e. CO₂ falls (∼0 Torr) during inspiration and rises (∼40 Torr) during expiration.

HCO₃⁻ transport across the apical membrane in HNE cells

In the present study, ion gradients were manipulated across the apical membrane to investigate whether HCO₃⁻ influx/efflux pathways were linked to luminal perfusate Cl⁻ concentrations in normal and CF airway cells. In the presence of serosal H₂DIDS (to block HCO₃⁻ efflux across the basolateral membrane; see Fig. 4 and Fig. 5) and symmetrical HCO₃⁻ (constant 5 %P_CO2), removal of luminal Cl⁻ resulted in cytoplasmic alkalinization in normal HNE cells due to the influx of HCO₃⁻ across the apical membrane (see Fig. 5A; also see Paradiso, 1992). Conversely, when alkalinized normal HNE cells were subsequently exposed to luminal Cl⁻, cells rapidly re-acidified towards basal levels (see Fig. 5A), consistent with the cellular efflux of HCO₃⁻ across the apical membrane. Furthermore, as discussed above, these Cl⁻-dependent changes of pH_i are largely independent of exogenously applied forskolin in normal HNE cells (see Fig. 6A), suggesting that the membrane mechanism responsible for the transport of HCO₃⁻ across the apical membrane is constitutively active under basal conditions.

Although these Cl⁻-dependent changes of pH_i, in response to changes in luminal Cl⁻ appeared to be consistent with an apically located AE in normal HNE cells, we argue that the mechanism mediating these changes in pH_i is the constitutively active CFTR Cl⁻/HCO₃⁻ conductance in the apical membrane in normal airway epithelia, based on a series of observations (also see discussion above relevant to Fig. 6A). Previous microelectrode studies (Willumsen et al. 1989) showed that removal of luminal Cl⁻ caused V_a to depolarize in normal (but not CF) cells by ∼30 mV, whereas the re-addition of luminal Cl⁻ repolarized V_a. As reported above, normal and CF HNE cells exhibit an average pH_i of ∼6.95 when cells are exposed to bilateral KBR. From this mean pH_i value, cell HCO₃⁻ concentrations, calculated from the Henderson-Hasselbach equation (Roos & Boron, 1981), is ∼8 mm. Accordingly, in the presence of luminal KBR, the ratio of HCO₃⁻ concentration across the apical membrane ([25 mm]_out/[∼8 mm]_in) is approximately equal to and opposite from the electrical gradient (V_a∼30 mV; Willumsen et al. 1989) across this barrier. Therefore, we predict no net movement of HCO₃⁻ via CFTR under basal conditions. However, removal of luminal Cl⁻ induces a depolarization of ∼30 mV across the apical membrane, i.e. V_a approaches 0 mV, which now favours HCO₃⁻ influx driven by the lumen-to-cell electrochemical gradient for this anion, and, as seen in Fig. 5A, pH_i increases from ∼6.97 to ∼7.4. The concentration of cellular HCO₃⁻ at pH_i of ∼7.4 will be close to 25 mm, and [HCO₃⁻]_in=[HCO₃⁻]_out, consistent with the absence of an electrical driving potential across the apical membrane. When Cl⁻ is restored to the luminal bath (repolarizing V_a by ∼30 mV), HCO₃⁻ now moves from cell-to-lumen, again driven by the electrochemical gradient across the apical barrier, and the cell re-acidifies.

Several other lines of evidence support the notion of electrogenic HCO₃⁻ transport via CFTR across the apical membrane in normal HNE cells, rather than electroneutral exchange for Cl⁻ via an AE. First, removal of luminal Cl⁻ failed to induce changes of pH_i in CF HNE cells (Fig. 5B), clearly suggesting the requirement for functional CFTR to mediate Cl⁻-associated changes of pH_i in normal airway cells. Second, mucosal H₂DIDS, a known inhibitor of AE (Mastrocola et al. 1998), but not CFTR (Paradiso et al. 2001), failed to block Cl⁻-linked changes of pH_i in normal airway cells in the absence (Fig. 5C) and presence (Fig. 6A) of forskolin. Third, mucosal DPC, an inhibitor of CFTR, reversibly blocked the increase of pH_i in normal airway epithelia in response to mucosal Cl⁻-free Ringer solution (see Fig. 6B). Fourth, both normal and CF cells exposed to bilateral Hepes-buffered Cl⁻- and HCO₃⁻-free Ringer solution acutely alkalinized, whereas only the normal nasal preparations exhibited a secondary re-acidification (see Fig. 9 and Fig. 10). We propose that this re-acidification resulted from loss of cell HCO₃⁻ via CFTR, since it was absent in CF airway epithelia (see Fig. 9 and Fig. 10) and was reversibly inhibited by mucosal DPC in normal airway cells (see Fig. 11). Finally, in normal (but not CF) nasal cells bilaterally perfused with KBR (but not Hepes-buffered Ringer solution), serosally applied Ba²⁺ induced an increase in pH_i that was reversibly blocked by mucosal DPC (see Fig. 7), strongly arguing for the conductive movement of HCO₃⁻ via CFTR.

Further support for the concept that CFTR mediates apical membrane HCO₃⁻ translocation is derived from patch clamp studies by Poulsen et al. (1994), who reported CFTR-mediated HCO₃⁻ conductance in NIH 3T3 cells recombinantly expressing wild-type CFTR. Similar findings have been reported by Linsdell et al. (1999), Illek et al. (1997) and Hogan et al. (1997) in a variety of epithelial cell types.

We would note here that some of our findings differ from those of others. Unlike the finding in mouse pancreatic and submandibular ducts, which have at their apical membrane a CFTR-regulated AE system (Lee et al. 1999a), our findings strongly indicate that CFTR, and not AE, is the primary HCO₃⁻ transport route in normal airway cells and that AE is restricted to the basolateral membrane in both normal and CF HNE cells. Moreover, recent studies by Wheat and co-workers (2000) reported that CFTR induces the mRNA expression of ‘downregulated in adenoma’ (DRA), and that DRA is itself an AE at the apical membrane in a tracheal epithelial cell line (CFT-1) derived from a CF patient. They suggested that the tracheal HCO₃⁻ secretion defect in patients with CF is partly a result of the downregulation of the apically located AE activity mediated by DRA with loss of CFTR function. The apparent discrepancy between their data and ours may be in the selection of tissues used for the studies, i.e. a transformed cell line versus epithelial cells in primary culture.

The concept that CFTR mediates HCO₃⁻ translocation across the apical membrane has implications for regulation of ASL pH as well as pH_i. For example, HCO₃⁻ secretion via CFTR should have an impact on altering pH and HCO₃⁻ content of ASL. Consistent with this notion, there is strong evidence that lack of HCO₃⁻ secretion renders the pH of liquids lining epithelial surfaces abnormally acidic in CF patients. For example, Kaplan and co-workers reported that CF ejaculate is acidic (pH ∼6.6) relative to normal semen (pH > 8.0), consistent with the absence of HCO₃⁻ in seminal fluid (Kaplan et al. 1968). Furthermore, studies by Durie (1989) reported that reduced HCO₃⁻ and Cl⁻ transport within pancreatic ducts could account for deficient fluid secretion in the pancreas of CF subjects. In addition, Kaiser & Drack (1974) reported diminished excretion of HCO₃⁻ from single sweat glands of patients with CF. With relevance to human airways, studies by Smith & Welsh (1992) have reported that cAMP stimulates HCO₃⁻ secretion across normal but not CF nasal epithelia. Importantly, recent studies (Coakley et al. 2000) on well differentiated bronchial epithelial cells have shown that ASL pH is more acidic and exhibits a lower HCO₃⁻ concentration in CF compared to normal airway cultures.

HCO₃⁻ transport across the basolateral border in HNE cells

Utilizing two experimental approaches to assess for AE, our data demonstrate that normal and CF HNE cells express an AE that is restricted to the basolateral domain of cells. In the first approach, changes in pH_i were detected in response to imposed Cl⁻ gradients across the basolateral membrane, and the changes of pH_i were reversibly blocked by serosal H₂DIDS in both normal and CF HNE cells (see Fig. 8). In the second approach, recovery from an alkaline challenge was independent of extracellular Na⁺, accelerated by serosal, but not mucosal Cl⁻, and reversibly inhibited by serosal H₂DIDS in normal HNE cells (see Fig. 9 and Fig. 10). Like normal airway cells, the recovery from an alkaline load in CF cells was again independent of extracellular Na⁺, absolutely dependent on serosal, but not mucosal addition of Cl⁻, and reversibly blocked by serosal H₂DIDS (see Fig. 9 and Fig. 10).

The biological role of the basolateral AE provides several important functions that are interrelated. For example, this exchanger serves as an important mechanism for Cl⁻ entry and HCO₃⁻ exit from the cell and thus functions as an acid-loader during acute cellular alkalosis. In addition to the metabolic role for AE, AE coupled to a Na⁺/H⁺ exchanger will have important influences on the distribution of ions and water between the extracellular space and the cell during volume regulation (Grinstein et al. 1985). Finally, for Cl⁻-driven secretion in airways, the basolateral Na⁺/H⁺ exchanger and Cl⁻/HCO₃⁻ exchanger, operating in parallel, are functionally equivalent to the basolateral Na⁺-K⁺-2Cl⁻ cotransporter and thus can provide an alternative mechanism for the serosal-to-cell uptake of Cl⁻.

Conclusion

We have identified pathways for HCO₃⁻ translocation across the apical and basolateral domains of polarized airway epithelia. Across the apical membrane of normal airway epithelia, HCO₃⁻ appears to translocate via a CFTR-mediated electrodiffusive pathway, with no evidence for AE. In contrast, a H₂DIDS-sensitive AE pathway exists for HCO₃⁻ translocation across the basolateral barrier. The basolateral AE, in concert with the basolateral Na⁺/H⁺ exchanger, appears to control responses to local environmental changes (e.g. P_CO2) that can acutely alter pH_i. We speculate that in normal airway tissues, the transport of HCO₃⁻ via CFTR is central to maintaining ASL pH within physiological levels. Moreover, the secretion of HCO₃⁻ into ASL appears to be balanced by the secretion of H⁺ via a recently reported (Paradiso et al. 2000) non-gastric form of the H⁺,K⁺-ATPase, which is localized at the apical membrane in both normal and CF airway epithelia. Furthermore, it is reasonable to suggest that it is the activity of the H⁺,K⁺-ATPase that, in part, maintains the electrochemical driving force for HCO₃⁻ secretion via CFTR in normal airway cells by lowering ASL pH. In addition, in CF airways, ASL pH is predicted to be more acidic than normal airway epithelia with loss of CFTR-mediated HCO₃⁻ secretion. In support of this later notion, our recent in vitro studies (Coakley et al. 2000) showing that ASL pH is lower (pH 6.13) in CF airway epithelial cells relative to normal cells (pH 6.44) was linked to a reduction in ASL HCO₃⁻ content in CF compared to normal airway cells. We speculate that the lack of CFTR-mediated HCO₃⁻ secretion, coupled to a constitutively active H⁺,K⁺-ATPase, rather than the absence of a regulatory interaction between apical CFTR and AE, accounts for this difference in luminal pH between CF and normal airway epithelia.