INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DESPITE CONTINUED INVESTIGATION over the past 20 years the mode and regulation of Cl transport across amphibian skin have still not been resolved (for references see Refs. 8, 14, 16). The amphibian skin is a multilayered, heterocellular epithelium, and at least four routes for Cl movement (principal cells, glands, mitochondria-rich cells, and paracellular pathways) must be considered. Principal cells can be excluded with certainty, because experimental data indicate negligible Cl conductance (G_Cl) across either apical or basolateral membranes under control or stimulated conditions (4, 16-18, 24). Glands secrete Cl after stimulation with -adrenergic agonists (15), but they cannot account for the enormous G_Cl that prevails spontaneously under short-circuit conditions (12) or that can be activated by voltage perturbation (13, 18). Mitochondria-rich cells or tight junctions between the cells of the outer living layer are the remaining candidates for this fraction of Cl movement. To refine the search for the mechanism of G_Cl, attempts have been made to identify regulatory factors in amphibian skin. In light of the observation that an inhibitor of phosphodiesterases (PDE), theophylline, stimulates G_Cl, the involvement of cAMP was proposed (10, 11). Such an effect is consistent with many observations in other tissues in which cAMP stimulates Cl channels. Although the idea that cAMP may activate a conductive Cl pathway is supported by more recent data with membrane-permeant analogs of cAMP (9, 25), the question remains as to whether this stimulation occurs in the range of physiologically achieved concentrations of cAMP. Since the observation that procaine, an agent without known effect on PDE, affects G_Cl similarly to theophylline (5), the question has arisen as to whether the stimulation of G_Cl by theophylline, or the more powerful 3-isobutyl-1-methylxanthine (IBMX), actually results from the inhibition of PDE with subsequent increase in cAMP or is due to direct interaction of the xanthines with the Cl pathway. Support for the latter notion comes from the obvious difference in stimulation pattern between cAMP and PDE inhibitors (9). Recently synthesized xanthine derivatives without inhibitory effect on PDE (3) provide an option for addressing this problem more thoroughly. These agents activate the CFTR Cl channel in Calu-3 cells (2). We show here that the particularly effective derivatives from the above study increase the voltage-activated G_Cl in toad skin, thus corroborating the view that xanthines can interfere with Cl pathways without the participation of cAMP.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Experiments were performed on isolated skins of Bufo viridis. Techniques for tissue preparation and analysis of voltage-activated G_Cl are described elsewhere (19). In brief, abdominal skin of the toad was glued with cyanoacryl adhesive glue (Histoacryl; Braun, Melsungen, Germany) to Lucite gaskets on the basolateral side and mounted in a modified Ussing-type chamber. Edge damage was minimized by using gaskets from cured Sylgard 184 (Dow Corning, Midland, MI), which were covered with a thin layer of heavy silicone grease to seal the tissue on the mucosal side.

In experiments for the analysis of serosal effects and for determination of cAMP content, split epithelia devoid of the connective tissue layer and glands were prepared by the collagenase method, as described recently (20). The collagenase solution was supplemented with 0.1 mg/ml trypsin inhibitor to avoid interference from traces of trypsin (21). Both half-chambers were separately and continuously perfused with Ringer solution (in mM: 110 Na⁺, 2.5 K⁺, 1 Ca²⁺, 1 Mg²⁺, 114 Cl, and 3.5 HEPES, pH = 7.5) at 3-5 ml/min. In some experiments, Cl of the mucosal solution was replaced by NO or SO on an equimolar basis. Mucosal perfusion solutions were always supplemented with 2 × 10⁵ M amiloride to eliminate transepithelial Na⁺ transport. The exposed tissue area and the volume of the half-chambers were 0.6 cm² and 0.4 ml, respectively. Ag/AgCl electrodes in 1 M KCl or 1 M NaCl served for measurement of transepithelial voltage (V_t) or for passing transepithelial current (I_t) with a custom-built automatic device, which also provided recordings of transepithelial conductance (G_t) from the change of I_t after brief, small perturbations of V_t (5 mV, 200 ms every 2.5 s) with analog sample-and-hold circuitry. V_t and I_t are referred to the mucosal side. All data were stored on a computer after analog to digital conversion and later plotted with commercial graphics software (Origin; Microcal, Northampton, MA).

For analysis of voltage-activated G_Cl the tissues were alternatively voltage-clamped to 30 or +80 mV, because we (16) and others (14) have shown that G_Cl is essentially inactivated or fully activated at these levels of V_t. The voltage-activated G_Cl can therefore be obtained as the difference between the steady-state value of G_t at the activating holding potential (+80 mV) and baseline G_t at the inactivating clamp potential (30 mV). Because of the long time constants for activation and inactivation of G_Cl (1-3 min), each determination cycle took several minutes.

Voltage-activated G_Cl was determined on the same tissue pieces under control conditions and after stimulation with the xanthines. For analysis of concentration-response dependencies, the tissue was voltage-clamped to 80 mV to activate G_Cl and the xanthines were then added in increasing concentrations. Conductance-voltage (G-V) relationships were determined either by clamping to increasing holding potentials with intermittent inactivation at 30 mV or by imposing sequentially increasing holding potentials. Because it took 3-5 min to approach a steady state at each value of V_t (and the same time for inactivation with the first technique), most G-V relationships were obtained with the second method. The data were not systematically different. The values of G_t at the clamp potentials were plotted and fitted to a sigmoidal regression function (Origin). Determinations under the influence of xanthines were always preceded by measurements under control conditions for individual comparison.

The cAMP content of the epithelial cells was determined in split toad skins glued to Lucite gaskets of 2-cm diameter. Because of the large size of the skins, 6-10 tissue samples could be obtained from each animal, which allowed the determination of cAMP content in duplicate samples from the same animal under different conditions. After preequilibration in Ringer solution for 30 min, the tissues were transferred to the experimental solutions, which always contained 2 × 10⁵ M amiloride, and incubated for 10 min. Pieces (1.5 cm²) were cut with a cork drill, plunged into 12% ice-cooled trichloroacetic acid, and stored at 18°C until analysis. cAMP was determined with an ELISA kit including acetylation (catalog no. DE0450; R&D Systems, Minneapolis, MN). Data are given as matched pairs based on tissue area.

The xanthine derivatives 3,7-dimethyl-1-propyl xanthine (X-32) and 3,7-dimethyl-1-isobutyl xanthine (X-33), kindly made available by Dr. F. Becq (Laboratoire de Physiologie des Régulations Cellulaires, UMR 6558, Université de Poitiers, Poitiers, France), were dissolved in DMSO. Stock solutions were added to Ringer solution to final concentrations of 2-1,000 µM. IBMX (Sigma) was used at concentrations of 10-500 µM. Furosemide (Sigma) was dissolved in an equimolar amount of 0.1 M NaOH and diluted with Ringer solution to the appropriate concentrations after neutralization. Laboratory chemicals were of analytical grade and were purchased from Sigma.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Figure 1 shows the record of voltage-activated G_Cl in an isolated toad skin epithelium (split skin) under control conditions and after serosal or mucosal addition of X-32. The xanthine analog increased G_Cl on application to either side. The stimulation of G_Cl appeared quickly and was complete within 3 min. Maximal stimulation was achieved at ~1,000 µM from either side. It was completely reversible on washout, with a time course similar to that of the onset of action. Because of the use of the split epithelium devoid of corial connective tissue layers, the time course of activation and washout was only slightly slower for serosal than for mucosal application.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Original record of the stimulatory effect of the xanthine derivative 3,7-dimethyl-1-propyl xanthine (X-32) on voltage-activated conductance across split toad skin. The drug was applied from either the serosal (ser) or mucosal (muc) side. Transepithelial potential was clamped to serosa +80 mV (dashed bar along the x-axis) to activate Cl conductance (GCl). The mucosal solution contained 2 × 105 M amiloride to eliminate Na+ transport. Gt, transepithelial conductance.

The concentration-response relationship for the increase in G_Cl after addition of X-32 to the mucosal (n = 7) or serosal (n = 3) side is shown in Fig. 2. Stimulation was not notably different for serosal or mucosal application, and the data were pooled. The sigmoidal regression line indicates that >80% of the effect was exerted at 300 µM; half-maximal stimulation occurred at 108 ± 9 µM. In seven other experiments, the effect of the xanthine derivatives X-32 and X-33 on G_Cl was tested at 200 µM. Because no difference in efficacy between the analogs was detected, data obtained with X-32 and X-33 are pooled and summarized in Table 1. The xanthine derivatives increased the inactivated tissue conductance (at 30 mV) only slightly but elevated the voltage-activated G_Cl (at +80 mV) substantially. For comparison, the effect of 100 µM IBMX was determined in most of the tissue pieces (n = 6). This xanthine derivative increased the baseline and inactivated conductance transiently (not shown) but had also little influence on the steady-state baseline conductance. In relative terms, the stimulation of voltage-activated G_Cl was greater with IBMX than with X-32 and X-33. The latter were, however, used at less than maximally effective concentrations. After replacement of mucosal Cl by NO or SO, under which condition voltage-activated G_Cl is eliminated, the xanthines X-32 and X-33 or IBMX had no effect on G_t at 30 or +80 mV (data not shown).

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Dose-response curve for the stimulatory effect of X-32 on activated conductance across split toad skin. The results of 7 mucosal () and 3 serosal () applications were pooled together for curve fitting.

View this table: [in this window] [in a new window]   Table 1.   Effect of xanthine derivatives X-32 and X-33 or IBMX on baseline conductance and voltage-activated Cl conductance in toad skin

The stimulation of voltage-activated G_Cl by theophylline in the toad skin is associated with a shift of the G-V relationship (9). The present study showed that IBMX and the xanthine derivatives without influence on PDE have similar response patterns. G-V relationships under control conditions and after application of 200 µM X-32 or 50 µM IBMX are depicted in Fig. 3. It is evident that both xanthine analogs, IBMX and X-33, in addition to elevating the maximal level of G_Cl, shift the G-V relationship to less negative potentials, i.e., they facilitate voltage activation. The influence of the xanthine derivatives was readily reversible on washout (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Effect of the xanthine derivatives 3-isobutyl-1-methylxanthine (IBMX) and 3,7-dimethyl-1-isobutyl xanthine (X-33) on the relationship of transepithelial voltage (Vt) and GCl (G-V) across toad skin.

Table 2 shows the mean values from 11 experiments with X-32 (300-1,000 µM on the mucosal side). The data were pooled because no consistent dependence on concentrations >300 µM X-32 was detectable; at lower concentrations, the shift of the G-V relationship was smaller but numerical estimates were not done. On average, the half-maximal voltage for activation of G_Cl decreased significantly by 26.5 mV. The maximal level of G_Cl increased by 53% compared with control conditions. For comparison, results from G-V relationships for IBMX (100-300 µM; n = 5) are included in Table 2. There appears to be no fundamental difference between the xanthine derivatives, regardless of their putative influence on the activity of PDE.

Furosemide, a blocker of the Na-K-2Cl symport, inhibits G_Cl and shifts the G-V relationship in toad skin to higher voltages when applied from the apical side (22). Because this is opposite to the influence of the xanthine derivatives, it was of interest to examine the interaction between these compounds. Figure 4 shows the record of a typical experiment. G-V relationships were first recorded under control conditions, then during exposure to 500 µM X-32 on the serosal side, and finally during additional application of furosemide to the mucosal solution at three different concentrations. It is interesting to note that the magnitude of G_Cl at 80 mV, i.e., the standard activating voltage, was considerably more affected than the maximally activated G_Cl. This is consistent with a shift of the activation curve, which is confirmed by the G-V relationships derived from these data as shown in Fig. 5. The voltage of half-maximal activation was reduced by X-32 from 75.5 to 48.5 mV and increased to 85, 98, and 118 mV under the influence of 300, 600, and 1,000 µM furosemide, respectively. The changes induced by furosemide in the presence of X-32 are comparable to those observed in the sole presence of the inhibitor, although the absolute level is shifted to lower voltages. Similar results were obtained in four other experiments, in which furosemide was applied in the presence of X-32 and X-33 or IBMX or the xanthines were added after inhibition with furosemide.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Original record of the response to mucosal addition of furosemide on the stimulation of Gt by 500 µM X-32 (serosal) in response to stepwise increasing serosa-positive clamping potentials.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   G-V relationships derived from the data in Fig. 4 for the effect of furosemide on the stimulatory effect of X-32 on Gt.

The content of cAMP in split epithelia from five toad skins was analyzed under control conditions, with 500 µM X-32, and with 200 µM IBMX. For comparison, tissue pieces from four animals were also analyzed in the presence of 100 µM forskolin. The results are summarized in Table 3. In contrast to the large scatter of the cAMP content in tissues from different animals as evidenced by the variance of control and experimental data, duplicate samples from the same animal were rather similar. The ratio of experimental vs. control content of cAMP in these matched samples indicates that neither X-32 nor IBMX affected the tissue cAMP content notably. Forskolin, a direct activator of adenylyl kinase, induced a >20-fold increase in cAMP content, thus verifying the responsiveness of the adenylyl cyclase system and the analysis procedure.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Because elevation of cellular cAMP (applied as membrane-permeant analogs) increases G_Cl in toad skin (25), this mechanism suggests itself as the explanation for the stimulation of G_Cl by the PDE inhibitors theophylline and IBMX (7, 10, 12). In a more careful analysis of the response of G_Cl on cAMP or IBMX, however, we have shown that the response patterns after application of these two agents are fundamentally different and not compatible with a common mediation by cAMP (9). Whereas IBMX affects mainly the voltage-activated G_Cl, cAMP analogs elevate a baseline conductance, which is poorly Cl selective and almost insensitive to voltage in the highly conductive state. On the basis of these findings, we have proposed that cAMP converts and fixes the Cl pathway into a permanently open state, whereas xanthine derivatives such as theophylline and IBMX increase the response of G_Cl to voltage, i.e., sensitize a component of the conductive Cl pathway, which we termed "voltage sensor" (9). Concurrent inhibition of PDE, if existing, would then be a side effect of no relevant functional consequence. The present experiments confirm this view. Except for a lower molar efficacy, the effect on the voltage-activated G_Cl of toad skin is identical for IBMX, which is a putative inhibitor of PDE, and the tested xanthine derivatives X-32 and X-33, which do not inhibit PDE in CHO and Calu-3 cells (2, 3).

For all tested xanthine derivatives, stimulation is limited to a conductive pathway that depends on the presence of Cl in the mucosal solution and is activated by serosa-positive voltages. No stimulation by any of the xanthine derivatives is observed when the mucosal Cl is substituted by SO or G_Cl is blocked by NO. The latter interacts with the Cl-dependent activation site, although being readily permeant in the open Cl pathway of amphibian skin (9). These data indicate that the xanthines stimulate a conductive Cl influx but cannot overcome the blockage of the activation site with NO. The time course of the voltage activation is similar to that under control conditions, suggesting that the xanthines stimulate this pathway rather than opening a new route.

In toad skin, neither the xanthine derivative X-32 nor the putative inhibitor of PDE, IBMX, had any substantial effects on the measured cAMP content of the tissue when these agents were applied at concentrations that stimulate G_Cl maximally. Responsiveness of the system was confirmed in the present experiments by the large increase in cAMP content under the influence of the direct activator of adenylyl cyclase, forskolin. Because the data reflect mostly the response of the principal cells, which account for >95% of the epithelial cells, local changes of the cAMP content, e.g., in mitochondria-rich cells, might have evaded detection. This could happen if PDE at these sites were unexpectedly inhibited by the xanthines. We consider this possibility unlikely, at least for X-32, because the agreement between our data and those from two other cell types, CHO and Calu-3 cells (2, 3), suggest a general nonresponsiveness of PDE to these xanthine derivatives. We therefore conclude, as has been done for CHO and Calu-3 cells (2, 3), that the effects of both X-32 and IBMX on G_Cl originate from a direct interaction of the xanthines with the Cl pathway and are unrelated to cAMP. We cannot explain why IBMX does not increase the cAMP content of B. viridis skin. This behavior is unlike the response of frog skin (Rana esculenta), in which another unspecific inhibitor of PDE, theophylline, increased cAMP by >100% (6).

All tested xanthine derivatives elevated the voltage-activated fraction of G_Cl. More specifically, these activators shift the relationship between conductance and V_t to lower voltages, i.e., they facilitate the opening of the Cl pathway. This concentration-dependent shift of the G-V relationship to lower activating voltages can be reversed by furosemide, which inhibits voltage-activated G_Cl in toad skin (22) by interfering with a voltage-sensitive regulator of the conductive Cl pathway. This effect of furosemide, which is observed only after mucosal addition (22), is evidently unrelated to the Na-K-2Cl cotransporter. This symport is localized to the basolateral membranes of amphibian skin, and transport occurs in electroneutral mode. Furthermore, the more powerful inhibitor of the cotransporter, bumetanide, has negligible effect on the voltage-activated G_Cl (22). The gradual concentration dependence of the G-V relationship on both the activation by xanthines and the inhibition by furosemide points to interference with a regulatory site of the same pathway rather than induction/disappearance of new pathways.

The reversal of the stimulation induced by X-32 provides further evidence against involvement of cAMP because an effect of furosemide on the production of cAMP is highly unlikely and furosemide has no inhibitory effect on the G_Cl induced by membrane-permeant analogs of cAMP (22). It is intriguing to speculate that xanthines and furosemide affect the same structure, which controls the voltage sensitivity of the conductance activation, in an opposite direction, although it is not possible to specify the location and the chemical nature of this site or the mode of interference. The questionable role of cAMP in G_Cl regulation is illustrated further by the response of G_Cl to low concentrations of cAMP applied as the membrane-permeant, nonmetabolized analog 8-(4-chlorophenylthio)- adenosine 3',5'-cyclic monophosphate (CPT-cAMP). Although a quantitative numerical correlation is difficult because of potential differences in affinity, it is clear that CPT-cAMP has no stimulatory influence on G_Cl at concentrations comparable to physiological levels of cAMP and even 5-10 times higher (23). At supraphysiological concentrations, however, the voltage-modulated G_Cl is replaced by permanent opening of the conductive Cl pathway. Accordingly, cAMP would affect the same Cl-specific route but via a different signaling chain from that for the voltage-mediated opening.

Similar to IBMX or theophylline, the xanthine derivatives X-32 and X-33 were evidently without influence on the time course of the increase in G_Cl after perturbation of the holding potential from 30 to +80 mV, which leads to activation of Cl current with half-times of 0.3-2 min in control tissues. This behavior is fundamentally different from the near-instantaneous change in Cl current after voltage perturbation in the presence of high concentrations of cAMP. It is not possible at present to specify the kind of interaction, particularly as the location of the Cl pathway (paracellular across the tight junctions or transcellular via mitochondria-rich cells) is unknown (16). Nevertheless, the present findings indicate that agents in the medium can directly modify the Cl-specific sites in this epithelial tissue. Similar conclusions have been derived from experiments on frog skin with the local anesthetic procaine (5). For this agent, partition of the uncharged form of the molecule in the lipid phase of the membrane was proposed to transfer the effect. It is possible that the same mode of action applies to the xanthine derivatives, which also possess lipophilic properties. Direct binding to regulative sites in the Cl pathway as proposed for interference with CFTR Cl channels (3) remains another option. Activation of Cl permeability in Calu-3 cells has been associated with a direct influence of the xanthines on CFTR (2). Although the presence of CFTR in mitochondria-rich cells of toad skin was recently demonstrated by immunohistochemical location and PCR (1), the small quantity of CFTR detected by PCR and its absence from apical membrane areas cast doubt on the involvement of this Cl channel in voltage-activated or cAMP-induced G_Cl.