Influence of extracellular bicarbonate on the short‐circuit current and intracellular free calcium of human cultured sweat duct cells

Transepithelial short‐circuit current (Iscc) and intracellular free Ca2+ (Ca2+i) was studied in monolayers of cultured human sweat duct cells (CSDCs) in the presence or absence of HCO3‐ (and CO2) in the bathing solutions. Addition of HCO3‐ (and CO2) increased the control Iscc by more than 50%. The effect of HCO3‐ (and CO2) on Iscc was confined to the serosal bath. The HCO3‐ (and CO2) effect was also studied during stimulation with the cholinergic agonist methacholine (MCh), which in CSDC induces a complex response consisting of an initial Iscc and Ca2+i spike, which is independent of extracellular Ca2+, followed by regular Iscc and Ca2+i oscillations, which are absent during Ca(2+)‐free bathing conditions. The sustained Iscc and Ca2+i oscillations, but not the initial Iscc and Ca2+i spike were abolished by the removal of extracellular HCO3‐ (and CO2). It is concluded that the Ca2+ influx and the Iscc in CSDCs are critically influenced by the presence of extracellular HCO3‐ (and CO2) in the bathing solutions.


INTRODUCTION
Ion transport properties of human cultured sweat duct cells (CSDCs) are analogous to those of the intact sweat duct (Quinton, 1987), i.e. a barium-sensitive K+ leak and ouabainsensitive Na+-K+-pumps are situated in the serosal membrane and amiloride-sensitive Na+ channels have been located in the mucosal membrane (Pedersen, 1989(Pedersen, , 1990a. Such asymmetrical distribution of channels and pumps leads to an active transepithelial Na+ transport, which can be measured as amiloride-sensitive short-circuit current (IS,). When stimulated with cholinergics an additional tetraethylammonium (TEA)-sensitive K+ leak is activated in the serosal membrane, leading to an increased amiloride-sensitive I which includes an initial I spike followed by a sustained phase of regular ISeC oscillations (Pedersen, 1990b(Pedersen, , 1991Larsen, Novak & Pedersen, 1990;Brayden, Pickles & Cuthbert, 1991). Preliminary studies, showing that sustained I,,, oscillations following cholinergics were abolished during HCO3--free conditions, motivated the present study, in which the effect on I,,, of extracellular HCO3-during control and cholinergic stimulation was studied.
In addition we measured the effect of HCO3on intracellular free Ca2+ (Ca 2+), which is assumed to be the main intracellular regulator of ISCC in these cells (Pedersen, 1990b). The transepithelial bioelectric properties of CSDCs, grown to confluence on a permeable support, were studied by voltage clamp technique, and Ca 2+ was measured by microfluorescence techniques using the Ca2+ probe, fura-2. Skin samples were obtained from young healthy subjects, who underwent plastic surgery. Cell culture procedures have been described elsewhere (Pedersen, 1984(Pedersen, , 1989. Briefly, sweat glands were isolated and dissected after a period of collagenase treatment, and the coiled part of the reabsorptive duct was transferred to culture flasks supplied with a hormone-supplemented growth medium. After 3 weeks of exponential growth, the cells were subcultured and stored in liquid nitrogen. For electrophysiological measurements the frozen cells were thawed and subcultured in high density on a dialysis membrane which covered a hole in the centre of the bottom of a small plastic cup. When confluence was obtained I measurements were performed at 37°C in Ussing chambers as previously described (Pedersen, 1989). For Ca> measurements cells were subcultured in glass flaskettes (Tecnunc) at low density, i.e. it was intended in these experiments to produce small clusters of cells, in order to get better fura-2 load and facilitate the action of agonists acting fri-om the serosal membrane facing the glass surface. Fura-2 AM was added to CSDCs, bathed in the same Ringer solution as used after incubation, at a final concentration of 20 /(M for 30 min at 37 'C. After loading, cells were washed twice and mounted in a modified Sykes-Moore chamber with a volume of 0 2 ml. This chamber was stationed on the stage of a Zeiss Axiovert-10 microscope equipped for epifluorescence with a 75 W xenon lamp and a 20 x objective (Plan-Neofluar, Zeiss). A continuous gravity-produced flow with fully aerated and prewarmed Ringer solution was applied to the chamber with a perfusion rate of at least 10 ml/min. In comparison with the small volume of the Sykes-Moore chamber, this continuous high flow prevented accumulation of any fura-2 that leaked from the cells during experiments, thus allowing observations of repeated pulses of stimuli to be made. The temperature in the chamber was maintained at 37 'C by regulation of the flow rate. Glass tubes were used in order to avoid gas diffusion from the flow system. Measurements of intracellular Ca'2 were made using a central processor unit (MSP-21 Zeiss), controlling shutters and filters. Through interference filters, barrier filters and a dichroic mirror (all Zeiss) the fluorescence signal from CSDCs was acquired alternatively at excitation wavelengths 340 and 380 nm using an emission wavelength of 520 nm. A pinhole diaphragm situated before the photomultiplier (R928/PMT Zeiss) was used to pass the fluorescence arising from each of the two excitation wavelengths from the central region, representing a collection of approximately ten cells. Background fluorescence values for each wavelength, obtained before each experiment using unloaded cells, were automatically subtracted from the corresponding signals measured in fura-2-loaded cells before taking the ratio (340/380). Readings of both wavelengths (average of 100 meaLsurements each) were done every 5 s, and the shutter was closed between readings. The ratio, corrected for background values, was converted to Ca 2 concentration values using an external calibration standard and the formula derived by Grynkiewitcz, Poenie & Tsien (1985) for dual wavelength measurements: Ca2 +=K(R,-Ro)/ R,-RJ where R0 and R are the ratios at 0 Ca2+ and saturating Ca2 , respectively. Rx is the experimental ratio. K represents K,(F(,1S), where K, (2 25 x 10 m at 37 'C) is the effective dissociation constant for fura-2, and F,, and F and are the fluorescence intensities at 380 nm minus and plus Ca2', respectively.
In the present set-up a F,b/t = 8 8 was measured.

RESULTS
In the presence of extracellular HCO3a highly significant stimulation of I,, is seen in CSDCs. Figure A and B illustrates the effect of a bilateral solution change from HCO3-free to HCO3--buffered Ringer solution. In Fig. 1A   mucosal side during HCO3--buffered conditions, whereas in Fig. 1 B amiloride is added after return to HCO3--free conditions. It can be seen that the major part of the HCO3-induced I,,,increase is amiloride sensitive, i.e. is likely to be carried by Na+. The results from thirty-six experiments are presented in Values, given as means+s.E.M., represent the results from thirty-six different epithelial preparations, derived from six preparations from each of six donors. ** P < 0-001; * P < 0 01. HCO3 . In this situation only a minor increase of the ISC(, was seen (from 53 1 +6 2 to 55-8 + 63 1iA/cm2; n = 8); i.e. less than one-tenth of the increase seen following substitution with HCO3 . Thus, extracellular HCO3exerts a specific effect on ,S(, which cannot be assigned the presence of anions in general. The S((, response to sequential addition of 25 mm HCO3-, first to the mucosal solution and thereafter to the serosal solution, is shown in Fig. 1 C. Serosal exposure to HCO3-Ringer solution is seen to dominate the final steady-state Ie, which is significantly increased. The small inhibition of IC(C following mucosal HCO3addition might reflect either a transepithelial mucosa-toserosa HCO3-flux due to the difference in HCO3-concentrations between the two sides, or an intracellular acidification due to permeation of C02, which on the mucosal side is not counteracted by a simultaneous HCO3entry. The polarized response to HCO3-probably reflects the result of HCO3 entry across the serosal membrane. The nature of such a HCO3-pathway, however, remains unknown, i.e. addition of recognized anion exchange inhibitors, DIDS (4,4'-diisothiocyanatostilbene-2-2'-disulphonic acid, 50 ,iM) or SITS (4-acetamido-4'-isothiocyanatostilbene-2,2'disulphonic acid, 100 aM), had no effect on the HCO3 -induced IS alterations. This, however, does not rule out the presence of an anion exchanger as the disulphonic stilbenes do not always block such pathways (Binder, Foster, Budinger & Hayslett, 1987).
When CSDCs are stimulated with methacholine (MCh) in HCO3--buffered solutions, an abrupt large transient Is( spike is seen followed by a sustained phase of regular Is(, oscillations ( Fig. 2A; Pedersen, 1987Pedersen, , 1990bPedersen, , 1991. During HCO3--free bathing conditions the CSDCs respond to MCh addition with an apparently unaffected initial IS spike, whereas the otherwise seen subsequent IS((} oscillations are absent (Fig.2 B and C). Addition of HCO3to obtain small concentrations (2-5 mM) in the serosal bath, however, immediately revives regular Ise, oscillation ( Fig. 2B; representative of more than 45 similar experiments), whereas HCO3addition to the mucosal bath has no effect (5 out of 5 experiments). To see whether the effect of HCO3on IS(e oscillations was specific for this ion, or rather related to a HCO3--induced increase of intracellular pH (pHi), the effect of NH4C1 addition was studied. One might predict that addition of NH4C1 leads to an initial intracellular alkalinization due to non-ionic diffusion of NH3, which subsequently associates with intracellular H+. Addition of 2 mm NH4C1 to the mucosal bath, during HCO3-free conditions, also restored missing 'S((l oscillations (Fig. 2 C; representative for 9 out of 9 similar experiments) i.e. producing a response similar to that of HCO3addition.
No revival of oscillations was seen when NH4C1 was added to the serosal bath (seen in 4 out of 4 similar experiments). This NH4C1 sidedness, which is similar to the polarized NH -NH4+ permeability demonstrated in renal tubular cells (Kikeri, Sun, Zeidel & Herbert, 1989), might indicate that the serosal membrane is permeable to NH4 , e.g. via the serosal K+ channels (Pedersen, 1990b;Larsen, Novak & Pedersen, 1990). NH4+ entry would then dominate over NH3 permeation in the serosal membrane, leading to intracellular acidification, whereas the mucosal membrane, which notably has no significant K+ permeability, only allows NH3 entry.
Following the addition of MCh, CSDCs produce an initial Ca`+ spike which in the presence of Ca2`in the bathing solutions is followed by sustained Ca`2 oscillations (Fig.   3 A); i.e. a response similar to the ISCC response in these cells (Pedersen & Poulsen, 1991).
The Ca`2 response to MCh (25 ,aM) during HCO3--free bathing conditions (Fig. 3 B) was therefore investigated. It can be seen that the Ca`2 perturbations in Fig. 3 B are similar to the ISCC response illustrated in Fig. 2B, in which an identical experimental condition was used. It is likewise seen that Ca`+ oscillations are induced by subsequent addition of small amounts of HCO3-(5 mM). By analogy with the IS experiment these oscillations were not influenced by further exposure to CO2 (1-2 %), regulating the pH of the Ringer solution to 7 4 (not shown). In order to substantiate an assumed effect of HCO3on the Ca2`influx pathway The role of extracellular Ca2" on the sustained phase is demonstrated in panel A. In panel B the addition of HCO3-(5 mM) was able to revive oscillations. During additional Ca2"-free conditions (panel C) addition of HCO3-(5 mM) failed to evoke oscillatory activity, until extracellular Ca2" (1 2 mM) was added.
(reflected by the oscillatory phase), cells were stimulated with MCh during both Ca2+and HCO3--free conditions (Fig. 3 C). During these conditions HCO3addition alone did not evoke Ca"+ oscillations, which first appeared when Ca2" was added.

DISCUSSION
The present study shows that the presence of HCO3in the serosal bath stimulates both control ISCC and MCh-induced IS,, and Ca"+ oscillations in CSDCs. The ISCC stimulation is evoked from the serosal side only, it requires low HCO3concentrations (i.e. 2-5 mM), and it seems to be independent of CO2. Hence, it seems likely that a HCO3--specific pathway is present in the serosal membrane of CSDCs. In this way the response of CSDCs is comparable with that of the intact tissue, in which the Na+ reabsorption was reduced 30 % following removal of HCO3from the bath (Sato, 1977). One of the important events following addition of HCO3into the bathing solutions includes increase in pHi (Jentsch, Korbmacher, Janicke, Fischer, Stahl, Helbig, Hollwede, Cragoe, Keller & Wiederholt, 1988;Boyarsky, Ganz, Sterzel & Boron, 1988 response when added during HCO3--free bathing conditions. This comparable effect on IS in CSDCs of NH3Cl and HCO3addition might indicate that HCO3acts via an increased pHi. In contrast to the effect of HCO3and NH4Cl, only insignificant alterations of the ISCC were produced when the CSDCs were exposed to variations of the extracellular pH (Pedersen, 1989). This might indicate that the presented HCO3-effects are mediated by sites that are accessible only from the cytoplasm, i.e. the Na+ reabsorption in CSDCs, reflected by I,SC (Pedersen, 1989;Larsen Novak & Pedersen, 1991), might be modified by pHi, as demonstrated in other tissues (Funder, Ussing & Wieth, 1967;Goldfarb, Egnor & Charney, 1988;Harvey, Thomas & Ehrenfeld, 1988). Any of the systems involved in net active Na+ transport in the CSDC might be influenced by pHi variations, i.e. it is known that intracellular acidification (1) reduces Na+-K+ pump activity (Skou & Esmann, 1984;Eaton, Hamilton & Johnson, 1984), (2) decreases Na+ conductance (Palmer, 1985;Harvey et al. 1988), and (3) decreases K+ conductance (Moody, 1984;Keller, Jentsch, Koch & Wiederholt, 1986). However, pHi might also influence the activity of ion transportregulating intracellular messengers such as Ca2l (Siffert & Akkerman, 1987;Baker & Honerjager, 1978). When CSDCs are stimulated with the cholinergic agonist, MCh, an initial ISCC spike is seen in parallel with an initial Cal' spike followed by regular I, and Ca"2 oscillations. The initial response is independent of extracellular Ca2", whereas the subsequent oscillations are absent if extracellular Ca2" is omitted (Pedersen, 1990b(Pedersen, , 1991Pedersen & Poulsen, 1991). During MCh stimulation a substantial effect of HCO3-removal was seen, presumably caused by an increased metabolism and acid production, which would lead to an increased demand for pHi-regulating processes. Although a normal MCh-induced initial I,SC and Ca"2 spike was obtained, the normally seen sustained phase ofIS and Ca"+ oscillations was abolished or highly reduced during the absence of extracellular HCO3-. Oscillatory activity was, however, revived by the addition of HCO3or NH4Cl. Thus, it can be theorized that the Ca2" influx pathway in CSDCs is regulated by pHi. In a variety of other tissues repetitive Ca2" release involves the action of phosphoinositides (Berridge & Gallione, 1988), acting at a specific receptors, which interestingly has been demonstrated to be dependent on pHi (Ives & Daniel, 1987;Enyedi, Brown & Williams, 1989;Siskind McKoy, Chobanian & Schwartz, 1989;Ferris, Huganir, Supattapone & Snyder, 1989). We have not been able to demonstrate a direct action of phosphoinositides in CSDCs. A significant phosphoinositide hydrolysis was, however, demonstrated in CSDCs following stimulation with MCh (Doughney, Pedersen, McPherson & Dormer, 1989). Thus, in CSDCs phosphoinositides might act via such an intracellular messenger in a pH.-dependent way, as suggested above.
The oscillations in CSDCs are assumed to be the result of a repetitive Ca2" discharge from a plasma membrane-associated intracellular Ca2" pool, incorporated as a link between the extracellular space and the cytoplasm (Pedersen, 1990b(Pedersen, , 1991 as also suggested in other tissues (Wakui, Potter & Petersen, 1989). Taken together it therefore seems likely, that pHi acts via an phosphoinositide-dependent repetitive Ca2" discharge pathway from this membrane-associated Ca2+ pool, which, on the other hand, is continuously supplied with Ca2" from the extracellular space.
The close relation between oscillatory activity and the presence of HC03in the bathing solutions demonstrated in this study might have implications for the investigation of oscillatory phenomena in other tissues. Thus inferences derived from studies in the absence of C02-HCO3may not reveal the exact nature of this interesting and significant physiological phenomenon. 31-2