Airway epithelial cells have been shown to respond to extracellular nucleotides by activation of a Cl⁻ secretory pathway (Willumsen & Boucher, 1989; Mason et al. 1991; Knowles et al. 1992) and inhibition of Na⁺ absorption (Mall et al. 2000). The regulation of ion transport by external nucleotides has been described to be mediated by purinoreceptor activation and a rapid and transient intracellular Ca²⁺ increase (Mason et al. 1991; Mall et al. 2000; Walsh et al. 2000). However, cellular signals other than calcium are involved in the physiological response to nucleotides. In particular, the pH_i responses to nucleotides are not clearly understood. External ATP stimulates an intracellular acidification in different types of epithelia, but the mechanism involved in the response is still not clear. In human nasal epithelium (Paradiso, 1997), aortic endothelial cells (Kitazono et al. 1989) and cardiac myocytes (Puceat et al. 1991), the acidification produced by external ATP was transient with a pH_i recovery to (or above) the basal value. In the present study, we report that exposure of human bronchial epithelial monolayers to apical ATP produces a sustained intracellular acidification. The external ATP-induced acidification does not involve binding to a suramin-sensitive purinoreceptor, but does require ATP hydrolysis. The ATP-induced acidification is amiloride sensitive, dependent on external Na⁺ and enhanced at low external pH. However, the pH_i response to ATP is independent of Ca²⁺, protein kinase C (PKC) and protein kinase A (PKA) signalling.

16HBE14o- cells were grown in Eagle's minimal essential medium (EMEM), supplemented with 10 % fetal calf serum, 1 % penicillin G, 1 % streptomycin and 1 % l-glutamine, in Corning culture flasks coated with a collagen-fibronectin solution at 37 °C with humidified 5 % CO₂ atmosphere.

When the 16HBE14o- cells reached confluency, they were washed twice with a Hepes-buffered saline solution (20 mm Hepes, 120 mm NaCl, 3 mm KCl, 9 mm glucose and 10 mm Na₂HPO₄) and isolated at 37 °C using a trypsin solution (1 % poly-vinyl pyrrolidone, 0.2 % EGTA and 0.25 % trypsin containing 0.02 % EDTA).

The glass coverslips were mounted on an inverted epi-fluorescence microscope (Diaphot 200, Nikon). Ultraviolet light from a xenon 100 W lamp (Nikon) was filtered through alternating 440 and 490 nm interference filters. The resultant fluorescence was filtered at 510 nm and then collected using an intensified CCD camera system (Darkstar, Photonic Science). A custom-built chopper circuit allowed an electronic adjustment of the duration of the exposure to each wavelength. Images were digitised and analysed using the Starwise Fluo system (Imstar, Paris, France). The ratio of fluorescence signals emitted at each excitation wavelength was converted to pH_i values using the high K⁺-nigericin calibration technique (Thomas et al. 1979). The calibration curve was obtained by permeabilising the cells using the H⁺-K⁺ exchanger ionophore nigericin (10 μg ml⁻¹) in Krebs solution in which NaCl was replaced with KCl (140 mm) and the pH was adjusted to various pH values between 5.8 and 8.4. Addition of nigericin allowed pH_i to be clamped to varying pH_o values. The fluorescence ratio was allowed to reach a steady state after each solution change. The experimental data were analysed by fitting the following equation to the calibration curve:

The frequency of data acquisition was every 10 −15 s.

The basal pH_i (steady-state pH_i) was calculated as the average of the first 4 min of fluorescence acquisitions under control conditions. When a drug was added, an average acquisition period of 4 min was used to determine the mean pH_i value after addition of the tested drug or solution.

Figure 1 Effect of apical ATP (10−4m) on pHi in 16HBE14o- confluent epithelial monolayers (A) and in isolated epithelial cells (B)

In some experiments, the effect of external ATP on the pH_i of isolated 16HBE14o- cells was tested. The basal pH_i of 7.127 ± 0.034 pH units (n = 5) in isolated cells was not significantly different from the basal pH_i measured in 16HBE14o- monolayers. However, in isolated cells, external ATP (10⁻⁴m) produced a transient acidification of 0.175 ± 0.021 pH units (n = 5), followed by a pH_i recovery towards basal levels (Fig. 1B). This response differed from the sustained acidification obtained in monolayers.

Figure 2 Effect of pre-treatment of 16HBE14o- monolayers with suramin (10−4m) on pHi before and during exposure to apical ATP (10−4m).

Figure 3 Intracellular pHi recording during apical pre-treatment of 16HBE14o- monolayers with AMP-PNP (10−4m) and subsequent exposure to apical ATP (10−4m).

Figure 4 Intracellular pH decrease induced by exposure of 16HBE14o- monolayers to apical ATP (10−4m) following pre-treatment with thapsigargin (10−6m), chelerythrine chloride (2 × 10−6m), RpcAMP (200 ×10−6m) and (more ...)

Figure 5 Intracellular pH changes induced by exposure of 16HBE14o- monolayers to apical ATP (10−4m) in the absence of external Na+ (A) and during exposure to 10−3m amiloride (B). C, concentration dependence of the effect of amiloride on the basal (more ...)

The role of Na⁺-H⁺ exchange in the response to apical ATP was tested using amiloride as an inhibitor of Na⁺-H⁺ exchanger activity. As shown in Fig. 5B, apical amiloride (10⁻³m) alone induced a significant diminution of pH_i of 0.195 ± 0.018 pH units from a basal value of 7.146 ± 0.013 (n = 6, P < 0.05). Subsequent apical addition of ATP (10⁻⁴m) did not produce any further pH_i decrease. The inhibition of ATP-induced intracellular acidification by amiloride was concentration dependent (Fig. 5C).

Figure 6 pHi changes in 16HBE14o- monolayers induced by apical ATP (10−4m) during treatment with ZnCl2 (10−4m).

Figure 7 pHi records (A) and mean pHi changes (B) following exposure of 16HBE14o- monolayers to apical ATP (10−4m) at different pHo. *P < 0.05.

In this study, we demonstrate that exposure of human bronchial epithelial cell monolayers to luminal ATP produces a rapid, sustained and reversible intracellular acidification.

In other tissues such as human nasal epithelium, rat cardiac myocytes and bovine aortic endothelial cells, it has been previously reported that external ATP produced a biphasic variation in pH_i. External ATP produced an initial acidification followed by a re-alkalinisation to (or above) the steady-state pH_i value (Kitazono et al. 1989; Puceat et al. 1991; Paradiso, 1997). In nasal epithelium, ATP was applied to the basolateral side of the monolayer (Paradiso, 1997) and in the other tissues ATP was applied to isolated cells (Kitazono et al. 1989; Puceat et al. 1991). The results which we report for 16HBE14o- are consistent with these previous observations. We also found that when ATP was applied to isolated 16HBE14o- cells, the acidification was followed by a recovery phase. However, apart from these comparative experiments in isolated cells, all of the experiments reported here describe the effects of apical ATP in intact 16HBE14o- monolayers to produce a sustained pH_i decrease. The differences in pH_i profiles observed upon basolateral (or isolated cell) or apically restricted ATP exposure might be explained by a polarised purinoreceptor expression or membrane-selective permeability effects of extracellular ATP. Polarised cellular responses involving membrane restricted signalling of Ca²⁺, cAMP or G-proteins have already been described in epithelial cells (Paradiso et al. 1995; Hwang et al. 1996; Wilson et al. 1996; Di Sole et al. 1999). In particular, in renal A6 cells transfected with the Na⁺-H⁺ exchanger isoform NHE-3, apical adenosine inhibited both the apical NHE-3 and the endogenous basolateral NHE. However, basolateral adenosine inhibited the apical NHE-3 and stimulated the basolateral NHE (Di Sole et al. 1999). Polarised pH control mechanisms could also be associated with microdomains in extracellular pH regulation identified in colonic epithelia (Chu & Montrose, 1995).

In nasal epithelium, the mechanism by which basolateral ATP elicited the pH_i decrease was not elucidated (Paradiso, 1997). Intracellular signalling triggered by external ATP is usually transduced via nucleotide binding to a purinoreceptor. Therefore, we tested the involvement of purinoreceptors in the pH_i response to external ATP using a broad-spectrum purinoreceptor inhibitor, suramin. Suramin alone produced an alkalinisation of the cell but the mechanism involved was not further investigated in this paper. However, in contrast to the calcium response to ATP which is completly abolished by suramin (Walsh et al. 2000), the pH_i response to ATP was not significantly affected by suramin. In addition, the total ineffectiveness of a non-hydrolysable form of ATP, AMP-PNP, on pH_i also indicates that a mechanism other than purinoreceptor activation is involved in the pH_i response to ATP (Weiss et al. 1992; Gheber et al. 1995). Some P2X purinoreceptor subtypes, such as P2X₄ and P2X₇, are antagonised only poorly by suramin (IC₅₀ > 300 μM; Buell et al. 1996). However, the sensitivity of the P2X₄ and P2X₇ subtypes to external zinc, protons and Na⁺ excludes their involvement in the pH_i response to ATP (North & Surprenant, 2000). Zinc, which inhibited the pH_i response to ATP, is known to increase the activity of P2X₄ (Wildman et al. 1999; Xiong et al. 1999) and to decrease the activity of P2X₇ (Virginio et al. 1997). Thus, these properties of Zn²⁺ would exclude the involvement of P2X₄. External protons, which enhanced the pH_i response to ATP, are known to decrease the activity of both P2X₄ (Stoop et al. 1997) and P2X₇ (Virginio et al. 1997), thus excluding their involvement in the response to ATP. In addition, external Na⁺ inhibits P2X₇ function (Virginio et al. 1997). As removal of external Na⁺ decreased the pH_i response to ATP, this last observation confirms that P2X₇ cannot be involved in this response.

The basal pH_i values reported here for human bronchial epithelial isolated cells or monolayers are consistent with pH_i values obtained in other epithelial cell types (Boron, 1986) including human nasal epithelial cells (Paradiso, 1992; Willumsen & Boucher, 1992).

The regulation of pH_i in airway epithelia has been shown to be dependent on a variety of acid extruders, including Na⁺-H⁺ exchange (Paradiso, 1992; Lubman & Crandall, 1994), the H⁺-ATPase pump (Lubman et al. 1989) and alkali extrusion via Cl⁻-HCO₃⁻ exchange (Nord et al. 1987; Lubman et al. 1995). However, in our study, all the experiments were performed in Hepes-buffered and HCO₃⁻-free medium, which would greatly attenuate the contribution of Cl⁻-HCO₃⁻ exchange to pH_i regulation.

In this study, we tested a variety of inhibitors of H⁺ extrusion and H⁺ loading transport mechanisms on the pH_i response to external ATP. ZnCl₂ has been used as a plasma membrane H⁺ channel blocker and H⁺-ATPase inhibitor (Thomas & Meech, 1982; Lukacs et al. 1993). Apical exposure to ZnCl₂ (10⁻⁴m) produced a significant decrease in the basal pH_i in 16HBE14o- monolayers. Since the electrochemical gradient for H⁺ favours entry into the cell via channels and the H⁺-ATPase works against this gradient, the initial acidification produced by Zn²⁺ can only be explained by inhibition of a H⁺ pump and not H⁺ channels. This result, therefore, indicates the contribution of an apical membrane H⁺-ATPase pump to the steady-state pH_i in 16HBE14o- cells.

In addition, pre-treatment with ZnCl₂ significantly attenuated the ATP-induced acidification. However, the overall pH_i decrease induced after ZnCl₂ and ATP was not significantly different from the pH_i acidification induced by ATP alone. We interpret the partial inhibition of the ATP-pH_i response by Zn²⁺ to be due to its blocking effect on H⁺ channels, thus preventing ATP activation of these channels and subsequent intracellular acidification due to H⁺ entry down an electrochemical gradient. However, in view of the high pH sensitivity of H⁺ extrusion mechanisms such as Na⁺-H⁺ exchanger activity (Aronson et al. 1982), the pH_i response to ATP in the presence of Zn²⁺ might be the consequence of ATP inhibition of an activated Na⁺-H⁺ exchanger. Since the Na⁺-H⁺ exchanger is activated allosterically by internal H⁺, it is possible that the initial acidification produced by Zn²⁺ brings the pH_i to the threshold for activation of the Na⁺-H⁺ exchanger. After ATP addition the exchanger is inhibited and pH_i falls to a new steady state. The value of this new steady-state pH_i will depend on the activity of any remaining functional acid extrusion/alkali loading transport mechanisms, the metabolic production of acid equivalents and the passive permeability of the membrane to H⁺.

A possible inhibition of the H⁺-ATPase pump by ATP is incompatible with the amiloride experiments. Amiloride completely inhibited the effect of ATP on pH_i, whereas it has been shown previously that amiloride has no effect on epithelial H⁺-ATPase pumps (Ehrenfeld et al. 1985). Since ATP produced a further fall in pH_i in the presence of Zn²⁺, the ATP effect on pH_i may be a combination of the activation of Zn²⁺-sensitive H⁺ channels and the inhibition of some other Zn²⁺-insensitive H⁺ extrusion process (e.g. Na⁺-H⁺ exchange).

We showed that apical application of amiloride (10⁻³m) or Na⁺ removal produced a rapid and significant decrease of the basal pH_i, indicating a contribution of Na⁺-H⁺ exchange (NHE) to the basal pH_i. This result is consistent with previous observations of the role of NHE activity in maintaining the steady-state pH_i in airway epithelia (Paradiso, 1992).

The presence of an apical NHE activity in airway epithelial cells is controversial. The NHE-2 and NHE-3 isoforms are known to be usually expressed in the apical membrane of epithelial cells such as intestinal epithelial cells and in the brush border of renal cells (Amemiya et al. 1995; Hoogerwerf et al. 1996; Biemesderfer et al. 1997). However, NHE-2 and NHE-3 are not expressed in human airways and NHE-1 is the only isoform found and located in the basolateral membrane (Brant et al. 1995; Dudeja et al. 1999). In addition, in human nasal epithelium, and in rat alveolar epithelial cells, removal of apical Na⁺ does not affect the pH_i recovery from an acidification, which is inhibited in the absence of basolateral Na⁺ (Willumsen & Boucher, 1992; Lubman & Crandall, 1994). In contrast, apical NHE activity has been reported in the apical membrane of rat and human lungs (Sano et al. 1988; Oelberg et al. 1993), in the apical membrane of sheep tracheal epithelial cells (Acevedo & Steele, 1993) and in fetal sheep lungs (Shaw et al. 1990).

The complete inhibition of the ATP-induced acidification by apical exposure to amiloride and the partial inhibition of the response by removal of apical Na⁺ indicates the involvement of NHE activity in the pH_i response to apical ATP. However, the action of amiloride on the ATP-induced acidification must involve inhibition of an acid-loading transporter in addition to its inhibitory effect on NHE. If inhibition of NHE by ATP was the sole mechanism for the intracellular acidification one would have expected a complete inhibition of the response in sodium-free solutions. Thus an additional acid-loading pathway must have been inhibited by amiloride for it to have produced complete block of the ATP effect on pH_i. Amiloride has been shown to inhibit proton channel currents across epithelial cell membranes (Gilbertson et al. 1992; DeSimone et al. 2001) and this effect, combined with its inhibition of Na⁺-H⁺ exchange, may explain the complete amiloride inhibition of the ATP-induced intracellular acidification.

The inhibition of Na⁺-H⁺ exchange by external ATP may be via a direct extracellular effect without the involvement of intracellular signalling. Our experiments failed to show the involvement of any classical intracellular signalling pathway (Ca²⁺, PKC, cAMP-PKA) known to regulate Na⁺-H⁺ exchanger activity, in the pH_i response to ATP.

ATP is a known regulator of the cystic fibrosis transmembrane conductance regulator (CFTR) which, besides acting as a chloride ion channel, can also regulate the function of other proteins both at the membrane and in the cytosol. In a previous paper, we reported an autocrine role for CFTR in ATP-induced changes to intracellular calcium in human bronchial epithelium (Walsh et al. 2000). In keeping with the results presented here, the CFTR protein via an autocrine function may also indirectly be responsible for cytosolic pH changes. However, in view of the discussion above on the lack of evidence for a role of intracellular modulators which also affect CFTR activity (PKC, cAMP-PKA), it would appear unlikely that the pH_i effects of ATP involve CFTR.

External ATP stimulates Cl⁻ secretion in airway epithelia (Willumsen & Boucher, 1989; Mason et al. 1991; Knowles et al. 1992) and inhibits Na⁺ absorption (Mall et al. 2000), via a mechanism involving a rapid and transient intracellular Ca²⁺ increase (Mason et al. 1991; Clarke & Boucher, 1992; Mall et al. 2000; Walsh et al. 2000). In addition, the activity of the apical Na⁺-H⁺ exchanger isoforms NHE-2 and NHE-3 have been shown to be inhibited by a rise in intracellular Ca²⁺ (Cohen et al. 1990; Burns et al. 1991). However, we show that, in human bronchial cells, thapsigargin, a Ca²⁺-ATPase inhibitor which induces intracellular Ca²⁺ mobilisation, did not affect the ATP-induced pH_i decrease, suggesting that intracellular Ca²⁺ is not involved in the external ATP effect. The insensitivity of the ATP-induced acidification to intracellular calcium signalling which is classically generated by purinoreceptor stimulation further strengthens the argument against the involvement of purinoreceptor stimulation.

The inhibition of NHE-3 activity by PKC has been described in renal and colonic epithelial cells (Janecki et al. 1998; Di Sole et al. 1999). However, neither PMA (PKC activator), chelerythrine chloride (PKC inhibitor) nor RpcAMP (PKA inhibitor) affected the ATP-induced pH_i decrease, indicating that neither PKC nor PKA activation is involved in the response. This is consistent with the absence of effect of PMA and isoquinoline H7 (PKC inhibitor) on acidification induced by external ATP in human nasal epithelium (Paradiso, 1997).

Inhibition of Na⁺-H⁺ exchange would be expected to produce an intracellular acidification. Besides this direct effect on pH_i, there is the possibility that NHE activity indirectly contributes to the ATP-induced acidification by affecting the local surface membrane pH, which may influence the protonation/permeability of ATP. Alternatively, Na⁺-H⁺ exchange may produce a localised external acid environment which could contribute to proton availability for entry through proton channels which may be opened by external ATP. In both these cases one would expect that the pH_i response to ATP would be influenced by changes in external pH.

We investigated the hypothesis that the acid-base properties of ATP, as other nucleotides, might be directly responsible for the intracellular acidification (Corfu & Sigel, 1991). In particular, we tested the sensitivity of the ATP-induced intracellular acidification to extracellular pH. A pH sensitivity of purinoreceptor activity has been reported; however, we show in this study that the response was not mediated by a purinoreceptor. We observed that reduction of external pH to 6.2, which increases the level of the protonated form of ATP, enhanced the intracellular acidification induced by external ATP. Correspondingly, raising the external pH to 11 inhibited the ATP-induced acidification. However, the changes in external pH did not significantly affect the basal pH_i, which suggests that external ATP activates a quiescent H⁺ diffusion entry pathway. In addition, the fact that the non-hydrolysable ATP analogue did not mimic the ATP effects suggests that hydrolysis of ATP is required to produced the fall in pH_i. Therefore a simple explanation would be that protonated ATP (H-ATP) crosses the plasma membrane and liberates H⁺ into the cytosol during hydrolysis. Alternatively, hydrolysable ATP may increase proton permeability per se across the plasma membrane, possibly through proton channels.

Taken together, we provide evidence for an intracellular acidification of human bronchial epithelial cells induced by apical ATP which does not involve signalling via Ca²⁺, PKC or PKA nor binding to a purinoreceptor. The results suggest that the ATP-induced intracellular acidification results from a combination of the activation of a Zn²⁺-sensitive proton channel, inhibition of amiloride-sensitive Na⁺-H⁺ exchange and diffusion of protonated hydrolysable ATP into the cell.

This work was funded by the Wellcome Trust (055695/Z/98/Z/ CH/TG/JF). The 16HBE14o- cell line was obtained as a gift from D. C. Gruenert (Gene Therapy Core Center, University of California, San Francisco, CA 94143, USA).