The Lipid Products of Phosphoinositide 3-Kinase Contribute to Regulation of Cholangiocyte ATP and Chloride Transport*

ATP stimulates Cl− secretion and bile formation by activation of purinergic receptors in the apical membrane of cholangiocytes. The purpose of these studies was to determine the cellular origin of biliary ATP and to assess the regulatory pathways involved in its release. In Mz-Cha-1 human cholangiocarcinoma cells, increases in cell volume were followed by increases in phophoinositide (PI) 3-kinase activity, ATP release, and membrane Cl− permeability. PI 3-kinase signaling appears to play a regulatory role because ATP release was inhibited by wortmannin or LY294002 and because volume-sensitive current activation was inhibited by intracellular dialysis with antibodies to the 110 kDa-subunit of PI 3-kinase. Similarly, in intact normal rat cholangiocyte monolayers, increases in cell volume stimulated luminal Cl− secretion through a wortmannin-sensitive pathway. To assess the role of PI 3-kinase more directly, cells were dialyzed with the synthetic lipid products of PI 3-kinase. Intracellular delivery of phosphatidylinositol 3,4-bisphosphate, and phosphatidylinositol 3,4,5-trisphosphate activated Cl− currents analogous to those observed following cell swelling. Taken together, these findings indicate that volume-sensitive activation of PI 3-kinase and the generation of lipid messengers modulate cholangiocyte ATP release, Cl− secretion, and, hence, bile formation.

Intrahepatic biliary epithelial cells, or cholangiocytes, have a major influence on the volume and composition of bile through absorption and secretion of fluid and electrolytes. Opening of chloride channels in the apical membrane has been identified as one important point for regulation of cholangiocyte secretion and involves the cystic fibrosis transmembrane conductance regulator (CFTR), 1 the protein product of the CF gene. Mutations in this gene inhibit channel function and result in cholestatic liver disease in 17-38% of patients (1)(2)(3)(4). It is unclear why the prevalence of biliary disease in CF is less than that of pancreatic or pulmonary disease. One potential explanation for these clinical differences is that CFTR-independent pathways may also regulate biliary secretion.
Recently, Cl Ϫ channels other than CFTR have been identified in cholangiocytes, and they appear to contribute importantly to modulation of cellular transport to meet changing physiologic demands. In isolated cells, Cl Ϫ currents activated by increases in cell volume are 2-3-fold greater than those activated by increases in cAMP (5)(6)(7)(8). The resulting solute efflux favors movement of water out of the cell and recovery of volume toward basal values, a general process referred to as regulatory volume decrease (9). The cellular mechanisms mediating channel opening and volume regulation in biliary cells have not been identified. However, cell volume increases appear to stimulate ATP release, and either removal of extracellular ATP or blockade of P2 receptors prevents cell volume recovery (10). Consequently, ATP release may serve as a signal activating Cl Ϫ channel opening by binding to apical P2 receptors.
In other cell types, cell volume increases stimulate parallel activation of multiple kinases (11) including phosphoinositide (PI) 3-kinase (12). PI 3-kinase is a heterodimer composed of a 110-kDa catalytic unit and an 85-kDa regulatory unit that are tightly associated (13,14). Upon activation, PI 3-kinase phosphorylates phosphatidylinositol and is capable of producing three lipid products: phosphatidylinositol 3-phosphate (PtdIns-3-P), phosphatidylinositol 3,4-bisphosphate (PtdIns-3,4-P 2 ), and phosphatidylinositol 3,4,5-trisphosphate (PtdIns-3,4,5-P 3 ). In liver cells, physiologic increases in cell volume have been shown to be a potent stimulus for PI 3-kinase activation (12), and PI 3-kinase has been shown to play a role in vesicle trafficking and bile formation in the isolated perfused rat liver (15). Based on the observations that cell volume and extracellular ATP are potent stimulators of cholangiocyte secretion, the purpose of these studies was to assess the cellular origin of biliary ATP and the potential role of PI 3-kinase as a signal modulating cholangiocyte Cl Ϫ secretion.

EXPERIMENTAL PROCEDURES
Cell Culture-Studies in isolated cells were performed in Mz-Cha-1 cells, originally isolated from human adenocarcinoma of the gallbladder (16). Studies in polarized monolayers were performed utilizing normal rat cholangiocytes (NRC) in culture (17). Each model system expresses phenotypic features of differentiated biliary epithelium including receptors, signaling pathways, and ion channels similar to those found in primary cells (6,18,19). Moreover, increases in Mz-Cha-1 (8) and NRC (data not shown) cell volume are followed by opening of membrane K ϩ and Cl Ϫ channels. Mz-Cha-1 cells were passaged at weekly intervals and maintained in culture at 37°C in a 5% CO 2 incubator in HCO 3 Ϫcontaining CMRL-1066 media (Life Technologies, Inc.) supplemented with 10% heat-inactivated fetal bovine serum, 100 IU/ml penicillin, and 100 g/ml streptomycin (8). NRC cells were cultured as described previously (17).
Cell Size Measurements-Mean cell volume was measured in Mz-Cha-1 cell suspensions by electronic cell sizing (Coulter Multisizer, Accucomp software version 1.19, Hialeah, FL) using an aperture of 100 m. Cells in subconfluent culture were harvested with 0.05% trypsin, suspended in cell culture media, centrifuged for 1 min at ϳ1000 ϫ g, resuspended in 3 ml of isotonic buffer, and incubated with gentle agitation for 30 -45 min. Aliquots (ϳ500 l) of cell suspension were added to 20 ml of isotonic or hypotonic (40% less NaCl) buffer. Measurements of ϳ20,000 cells at specified time points after exposure to isotonic or hypotonic buffer were compared with basal values (time 0). Changes in values are expressed as relative volume normalized to the basal period.
Whole Cell Measurement of Cl Ϫ Currents-Membrane Cl Ϫ currents were measured using whole cell patch clamp techniques (20,21) in Mz-Cha-1 cells. Cells on a coverslip were mounted in a chamber (volume, ϳ400 l) and perfused at 4 -5 ml/min with a standard extracellular solution containing 140 mM NaCl, 4 mM KCl, 1 mM CaCl 2 , 2 mM MgCl 2 , 1 mM KH 2 PO 4 , 10 mM glucose, and 10 mM HEPES/NaOH (pH 7.40). The standard intracellular (pipette) solution for whole cell recordings contained 130 mM KCl, 10 mM NaCl, 2 mM MgCl 2 , 10 mM HEPES/ KOH, 0.5 mM CaCl 2 , and 1 mM EGTA (pH 7.32), corresponding to a free [Ca 2ϩ ] of ϳ100 nM (22). Patch pipettes were pulled from Corning 7052 glass and had a resistance of 3-7 M⍀. Recordings were made with an Axopatch ID amplifier (Axon Instruments, Foster City, CA) and were digitized (1 kHz) for storage on a computer and analyzed using pCLAMP version 6.0 programs (Axon Instruments, Burlingame, CA) as described previously (21,23). Pipette voltages (V p ) are referred to the bath. Current-voltage relations were measured between Ϫ120 mV and ϩ100 mV in 20-mV increments (400-ms duration, 2 s between test potentials). In the whole cell configuration, V p corresponds to the membrane potential, and upward deflections of the current trace indicate outward membrane current. Changes in membrane Cl Ϫ permeability were assessed at a test potential of Ϫ80 mV (E k ) to minimize any contribution of K ϩ currents and values were reported as current density (pA/pF) to normalize for differences in cell size as recently described (24).
Transepithelial Cl Ϫ Transport Measurements-NRC cells were utilized to study vectorial Cl Ϫ movement across monolayers. Cells were grown to confluency on collagen-treated polycarbonate filters with a pore size of 0.4 m (Costar, Cambridge, MA) until a resistance of Ͼ1,000 ⍀ cm 2 was achieved as measured by an epithelial tissue voltmeter (EVOHM; World Precision Instruments, Sarasota, FL). Cells were mounted in a Trans-24 mini-perfusion system for tissue culture cups (Jim's Instrument Manufacturing Inc., Iowa City, Iowa). All experiments were carried out at 37°C, and basolateral and apical (luminal) sides were perfused continuously and independently in a closed system with the standard extracellular buffer solution (as described above) by bubbling O 2 through air-lift circulators. Transepithelial voltage (V t ) was clamped to 0 mV and short circuit current (I sc ) was recorded through agar bridges (3% agar in 3 M KCl) connected to Ag-AgCl electrodes (cartridge electrodes; World Precision Instruments). Previous studies indicate that I sc is a reflection of electrogenic Cl Ϫ secretion from the basolateral to the apical chamber (19). Experimental results were compared with control studies (basal and swelling-induced I sc ) performed on the same day to minimize any potential effects of day-to-day variability in current amplitude.
ATP Bioluminescence Assay-ATP release into media was detected by means of a bioluminescence assay as described previously (25,26). NRC cells were grown to confluence in a 35-mm dish, washed twice with phosphate-buffered saline, and incubated with OptiMEM-I reduced serum medium plus luciferase-luciferin reagent (lyophilized reagent, Calbiochem, La Jolla, CA). The dish was placed on a platform, lowered into the recording chamber of a Turner model TD20/20 luminometer, and studied immediately in real time. Because background luminescence (cells and medium without luciferase-luciferin reagent) is less than 0.1 arbitrary light unit (ALU), ATP released from cells into the media catalyzes the luciferase-luciferin reaction. Bioluminescence was measured in continuous 15-s photon collection intervals. To induce cell volume increases, the extracellular buffer was diluted 20 -40% by adding water. In control studies, an equal volume of isotonic buffer was added to assess possible ATP release because of mechanical stimulation (27). The small changes in bioluminescence associated with isotonic exposures were Ͻ10% of values associated with hypotonic exposure (data not shown).
Statistics-Results are presented as the means Ϯ S.E., with n representing the number of cells for patch clamp studies and the number of culture plates or repetitions for other assays. Student's paired or unpaired t test was used to assess statistical significance as indicated, and p values Ͻ 0.05 were considered to be statistically significant.

Inhibition of PI 3-Kinase Delays
Cholangiocyte Volume Recovery from Swelling-Exposure of cells to hypotonic buffer (40% decrease in NaCl, ϳ205 mOsm), caused a rapid initial increase in relative volume to 1.20 Ϯ 0.01 (n ϭ 5, p Ͻ 0.001) within 3 min. The increase was followed by gradual recovery toward basal values despite the continued exposure to hypotonic buffer (Fig. 1). To evaluate whether PI 3-kinase contributes to cell volume recovery, analogous studies were performed in the presence of wortmannin. Hypotonic exposure after preincubation with wortmannin (50 nM) resulted in a similar initial increase to 1.18 Ϯ 0.01 at 3 min. However, volume recovery was significantly delayed. The relative volume of 1.15 Ϯ 0.01 at 20 min and 1.14 Ϯ 0.01 at 30 min in the presence of wortman- nin significantly exceeded control values of 1.10 Ϯ.01 and 1.07 Ϯ 0.01, respectively (n ϭ 5 for each, p Ͻ 0.01; Fig. 1). These findings indicate that inhibition of PI 3-kinase impairs recovery from cell swelling.
Effect of PI 3-Kinase Inhibition on Volume-sensitive Cl Ϫ Currents-In Mz-Cha-1 cells, cell volume recovery from swelling depends upon opening of Cl Ϫ channels in the plasma membrane. To assess whether PI 3-kinase contributes to channel regulation, whole cell currents were measured under basal conditions and following cell volume increases induced by hypotonic exposure.

Effect of PI 3-Kinase Inhibition on Transepithelial Cl
Ϫ Secretion-To determine whether the effects of PI 3-kinase on volume-stimulated Cl Ϫ currents are relevant to transepithelial transport, parallel studies were performed in NRC monolayers mounted in Ussing chambers. NRC cells plated on collagencoated filters form polarized monolayers, and increases in I sc are due in large part to increases in apical Cl Ϫ conductance (19). Inserts were mounted in the recording chamber and allowed to equilibrate with the standard extracellular buffer, and basal I sc was recorded (7.41 Ϯ 0.29 A/cm 2 , n ϭ 8). Simultaneous perfusions of the apical and basolateral chambers with hypotonic buffer (30% less NaCl, ϳ217 mOsm), to increase cell volume, resulted in a transient increase in the I sc to 14.01 Ϯ 1.32 A/cm 2 (n ϭ 4), consistent with opening of conductive pathways (see Fig. 4A). The peak I sc response occurred rapidly and tended to return toward basal values over 10 min despite the continued presence of hypotonic buffer. In the presence of wortmannin (50 nM) the I sc response was significantly attenuated (9.12 Ϯ 0.26 A/cm 2 , n ϭ 4, p Ͻ 0.05; Fig. 3). These results Although both wortmannin and LY294002 are thought to be selective inhibitors of PI 3-kinase, the potential for inhibition of other kinases cannot be fully excluded. Consequently, three alternative strategies were utilized to assess the specificity of PI 3-kinase in cholangiocyte volume-stimulated ATP and Cl Ϫ transport as outlined in the following sections.
Increases in Cell Volume Activate PI 3-Kinase-PI 3-kinase activation from multiple stimuli (mitogens, chemotactic peptides, and osmotic changes) consistently result in phosphorylation of Akt (protein kinase B), its downstream effector, in all cell types studied (32)(33)(34). To determine whether increases in cholangiocyte cell volume activate PI 3-kinase as observed in other cell types (12), an assay measuring Akt phosphorylation was utilized. In Mz-Cha-1 cells, hypotonic exposure (40% decrease in NaCl) increased Akt Thr 308 phosphorylation 20 -50% by 6 -12 min in all trials (n ϭ 4). There was no increase in total Akt or actin levels (Fig. 4). This study demonstrates that increases in cholangiocyte cell volume increase PI 3-kinase activity as measured by Akt phosphorylation. Additionally, the rapid activation of PI 3-kinase following hypotonic exposure is consistent with an early role in the response to cell volume increases.
Intracellular Dialysis with a Specific Antibody to PI 3-Kinase Inhibits Volume-activated Cl Ϫ Currents-To determine whether the inhibitory effects of wortmannin and LY294002 on volumeactivated Cl Ϫ channel activity represents a specific effect on PI 3-kinase, an alternative strategy utilizing intracellular dialysis with specific antibodies to the 110-kDa catalytic subunit of PI 3-kinase was evaluated. These antibodies have been shown to inhibit growth factor-stimulated PI 3-kinase activity in cultured fibroblasts (31). For these studies, the antibodies were delivered to the cell interior by inclusion in the patch pipette. Intracellular dialysis with anti-PI 3-kinase antibody (5 g/ml) completely inhibited Cl Ϫ currents in response to hypotonic exposure with a maximal average current density of Ϫ2.9 Ϯ 0.4 pA/pF (n ϭ 5). In contrast, currents measured during intracellular dialysis with either heatinactivated antibodies (5 g/ml) or antibodies to unrelated proteins (␤-galactosidase, 5 g/ml) were similar to controls (Ϫ36.7 Ϯ 5.9 pA/pF and Ϫ39.4 Ϯ 12.2 pA/pF respectively, n ϭ 5 for each, p Ͻ 0.005, Fig. 5). These findings support a specific role of PI 3-kinase in volume-sensitive Cl Ϫ channel activation.
Intracellular Dialysis with the Lipid Products of PI 3-Kinase Activates Cl Ϫ Currents-PI 3-kinase phosphorylates the D3 position of the inositol ring forming three lipid products: Pt-dIns-3-P, PtdIns-3,4-P 2 , and PtdIns-3,4,5-P 3 . PtdIns-3-P is constitutively present in unstimulated cells and changes little upon stimulation. In contrast, PtdIns-3,4-P 2 and PtdIns-3,4,5-P 3 are undetectable under resting conditions but increase rapidly upon stimulation, suggesting that they function as secondary messengers, translating extracellular stimuli to cellular responses (35)(36)(37). The development of synthetic lipid products of PI 3-kinase has made the direct study of the PI 3-kinase signaling pathways possible. These synthetic phosphoinositides have been shown to be capable of mediating PI 3-kinase-dependent cellular processes including membrane ruffling and chemotaxis (38) and cell survival and gluconeogenesis through Akt regulation (32,39).

FIG. 4. Increases in cholangiocyte cell volume result in Akt phosphorylation.
Under isotonic conditions, Western blotting of Mz-Cha-1 cells with Thr 308 antibody (T308), specific against the phosphorylated form of Akt, readily detected phosphorylated Akt. Hypotonic exposure (40% decrease in NaCl) resulted in increased Akt phosphorylation after 9 min of exposure. Constant actin and total Akt levels reflect equivalent sample loading. The increase in Akt phosphorylation is indicative of PI 3-kinase activation. the resulting current was not statistically greater than the individual lipids alone (Ϫ16.2 Ϯ 1.7 pA/pF, n ϭ 5, p Ͻ 0.001 (compared with PIP-2); Fig. 6, A and C). These currents demonstrated characteristics of volume-stimulated Cl Ϫ currents, with reversal near 0 mV (E ClϪ ), outward rectification, and time-dependent inactivation at depolarizing potentials above ϩ60 mV (Fig. 6, B and D). Moreover, the current response generated by intracellular dialysis with PtdIns-3,4-P 2 and Pt-dIns-3,4,5-P 3 was inhibited in the presence of extracellular apyrase to remove any ATP released from cells (Ϫ2.6 Ϯ 0.1 pA/pF, n ϭ 4, p Ͻ 0.05, Fig. 6, A and B). These results are consistent with a model where the lipid products of PI 3-kinase contribute to direct modulation of membrane ATP permeability.
PI 3-Kinase Modulates Swelling-activated ATP Release-In other cell types, increases in cell volume result in ATP release, stimulation of purinergic (P2) receptors, and Cl Ϫ channel activation (40). In cholangiocytes, activation of P2 receptors by extracellular ATP represents a potent stimulus for Cl Ϫ secretion (19). Thus, PI 3-kinase could potentially modulate current activation through stimulation of ATP release, modulation of P2 receptors, or coupling receptor binding to channel opening. To assess the site of action of PI 3-kinase, two strategies were utilized. First, the effect of wortmannin on the current response to exogenous ATP was assessed. For these studies, Mz-Cha-1 cells in hypotonic buffer were exposed to exogenous ATP in the presence of wortmannin (50 nM). If PI 3-kinase modulates P2 receptors or couples receptor binding to Cl Ϫ channel opening, wortmannin would be expected to inhibit ATP-dependent current activation. In the presence of wortmannin, hypotonic exposure failed to activate Cl Ϫ currents. However, subsequent addition of ATP (10 M) to the perfusate resulted in instantaneous activation of Cl Ϫ currents (representative trace, Fig. 7A), increasing current density from Ϫ6.1 Ϯ 2.5 pA/pF to Ϫ25.5 Ϯ 7.3 pA/pF (n ϭ 5, p Ͻ 0.05; Fig.  7B). These findings indicate that PI 3-kinase is likely to function more proximally in the signaling pathway by modulating local ATP concentrations outside of the cell.
To assess this possibility more directly, the effects of PI 3-kinase inhibition on ATP efflux were evaluated utilizing a luciferin-luciferase assay. NRC cells in isotonic buffer demonstrated a basal release of ATP (0.53 Ϯ 0.08 ALU). Exposure to wortmannin or LY294002 decreased basal ATP release (wortmannin 0.21 Ϯ 0.02 ALU, LY294002 0.20 Ϯ 0.07 ALU, n ϭ 6 for each, p Ͻ 0.001; Fig. 6). In control cells, exposure to hypotonic perfusate stimulated ATP efflux (an increase of 0.73 Ϯ 0.07 ALU, at 20% dilution, and an increase of 0.47 Ϯ 0.06 at 40% dilution), as shown in Fig. 8. The increases in swelling-induced ATP release were inhibited by both wortmannin (0.19 Ϯ 0.01 ALU at 20% and 0.19 Ϯ 0.01 ALU at 40%) and LY294002 (0.12 Ϯ 0.01 ALU at 20% and 0.17 Ϯ 0.02 ALU at 40%, n ϭ 6 for each, p Ͻ 0.001; Fig. 8). In all studies, the addition of apyrase (2 units/ml) eliminated bioluminescence consistent with ATP scavenging (data not shown). These findings indicate that PI 3-kinase activity contributes to regulation of both basal and swelling-activated ATP release. DISCUSSION ATP exerts potent regulatory effects on many epithelial cells by binding to one or more purinergic receptors in the plasma membrane. In most epithelia, however, little is known regarding the mechanisms responsible for modulation of nucleotide release. These studies of a model Cl Ϫ secretory cell indicate that volume-sensitive changes in PI 3-kinase activity contribute to changes in cellular ATP efflux rates and suggest that the D3 products of phosphoinositol phosphorylation exert the principal biological effects. Thus, PI 3-kinase may contribute to rapid coordination of cellular ATP release and membrane ion permeability and to modulation of biliary secretion and bile flow in response to changing physiologic demands.
The formation of bile by the liver depends upon complementary interactions between hepatocytes, which transport bile acids and other organic solutes into the canalicular space and cholangiocytes, which line the lumen of intrahepatic bile ducts and are responsible for electrolyte and water transport. Extracellular ATP has been proposed to function as a signal coordinating the separate transport functions of hepatocytes and cholangiocytes because (a) ATP is present in bile where it has direct access to the apical membrane of cholangiocytes (41) and (b) nanomolar concentrations of ATP stimulate cholangiocyte secretion by binding to P2Y 2 receptors in the apical membrane (19,42). Both hepatocytes and cholangiocytes may be capable of electrodiffusional release of ATP (41,43). However, the polarity of release and the signals responsible for its regulation have not been defined. With these issues in mind, several points merit emphasis.
First, these studies provide additional evidence that cholangiocytes themselves are capable of regulated release of ATP. Moreover, significant amounts of extracellular ATP are detectable under basal conditions (constitutive release), and the FIG. 5. Intracellular dialysis with antibodies to PI 3-kinase prevents volume-dependent current activation. Antibodies were delivered to the cell interior by inclusion in the patch pipette. Cell volume increases were stimulated by exposure to hypotonic buffer (20% decrease in NaCl) as indicated by the bar. A, representative whole cell current tracing. Compared with control antibody (␤-galactosidase, 5 g/ml, upper tracing), intracellular dialysis with an antibody to the 110-kDa subunit of PI 3-kinase (5 g/ml) inhibited Cl Ϫ current activation by hypotonic exposure (lower tracing). B, average currents stimulated by exposure of cells to hypotonic buffer during intracellular dialysis with heat-inactivated (100°C, 30 min) p110 antibody (5 g/ml) or control antibody (␤-galactosidase, 5 g/ml, n ϭ 5 for each) versus intracellular dialysis with p110 antibody, which inhibited the current response (n ϭ 5, p Ͻ 0.05). amount increases in parallel with changes in cell volume (volume-sensitive release). ATP was always detectable in the media from cultures of both biliary cell models. In NRC monolayers, which form high resistance junctions between cells, the luciferin reaction mixture was added to the apical chamber only. Because luminometric readings reflect movement of ATP across the apical membrane, these findings indicate that increases in cell volume represent a potent stimulus for apical ATP release into bile.
Second, experimental results from several model systems indicate that apical ATP permeability is regulated in part by changes in endogenous PI 3-kinase activity. Utilizing a specific assay, which measures the phosphorylation of the PI 3-kinase effector Akt, we demonstrated that cell volume increases are associated with an increase in PI 3-kinase activity. Additionally, both constitutive and volume-sensitive ATP release was inhibited by the PI 3-kinase inhibitors wortmannin and LY294002. These findings suggest that PI 3-kinase modulates ATP release, but other sites of action are possible as well. It is notable that these inhibitors did not inhibit Cl Ϫ channel opening stimulated by exogenous ATP. Although the addition of ATP overcomes the inhibition of the volume-stimulated current activation by wortmannin, the magnitude of the current is less than that observed with hypotonic challenge. Several explana-FIG. 6. The lipid products of PI 3-kinase activate Cl ؊ currents. The synthetic lipids, PtdIns-3-P (10 M), PtdIns-3,4-P 2 (10 M), and PtdIns-3,4,5-P 3 (10 M), were delivered to the cell interior by inclusion in the patch pipette individually or in combination. A, the average maximum Cl Ϫ current density for all cells, measured at Ϫ80mV, is shown under basal conditions and during intracellular dialysis with control lipid (PIP-2), PtdIns-3-P, PtdIns-3,4-P 2 , PtdIns-3,4,5-P 3 , combined PtdIns-3,4-P 2 and PtdIns-3,4,5-P 3 together, and the combined lipids in the presence of apyrase. B, the average current-voltage relation of whole cell currents measured under basal conditions (q) and during intracellular dialysis with the lipid products (PtdIns-3,4-P 2 and PtdIns-3,4,5-P 3 ) (E). Cl Ϫ currents activated during intracellular dialysis with the lipids were characterized by outward rectification and reversal near 0 mV. In the presence of apyrase (2 units/ml) the lipids failed to activate currents (ƒ). In C and D, representative whole cell recordings are shown during intracellular dialysis with the combined lipid products, PtdIns-3,4-P 2 and PtdIns-3,4,5-P 3 , in equimolar amounts. C, currents at Ϫ80mV (downward deflection of the current tracing) correspond to I ClϪ . Intracellular dialysis with PtdIns-3,4-P 2 and PtdIns-3,4,5-P 3 resulted in characteristic current activation (top tracing). In contrast, intracellular dialysis with control lipid, PIP-2 (10 M), had no effect. D, whole cell currents were measured at test potentials between Ϫ120 mV and ϩ100 mV in 20 mV increments. I-buff, intracellular buffer; E-buff, extracellular buffer (see text). Under basal conditions no current activity is observed (top tracing). Cl Ϫ currents increased markedly with hypotonic exposure (20% decrease in NaCl) and demonstrated time-dependent inactivation at depolarizing potentials above ϩ60 mV (second tracing). PtdIns-3,4-P 2 -and PtdIns-3,4,5-P 3 -activated currents had similar biophysical properties (third tracing). Intracellular dialysis with control lipid, PIP-2 (10 M), had no effect (bottom tracing).
tions may exist for this observation, including: (a) changes in cell volume may activate additional signaling pathways responsible for channel activity than those involved with purinergic binding alone; (b) the type and/or number of ATP binding sites may change with increases in cell volume; and (c) the pathway(s) coupling receptor binding to channel activation may in fact have some degree of PI 3-kinase dependence.
The D3-phosphorylated products of phosphatidylinositol are thought to contribute importantly to the PI 3-kinase signaling pathway, coupling external stimuli to cellular metabolism. To assess their role in coupling cell volume changes to membrane transport, additional studies were performed using intracellular dialysis with the synthesized lipids PtdIns-3-P, PtdIns-3,4-P 2 , and PtdIns-3,4,5-P 3 . Intracellular delivery of these lipids through the patch pipette permits careful comparison among cells under conditions where key components of the membrane regulatory apparatus (i.e. receptors, cytoskeleton, and channels) remain largely intact. Intracellular dialysis with PtdIns-3,4-P 2 or PtdIns-3,4,5-P 3 resulted in spontaneous activation of currents over several minutes as the lipids diffused into the cell interior. Both lipids were individually capable of activating currents with a similar magnitude, which was not statistically larger when both were delivered simultaneously. This is consistent with the finding that PtdIns-3,4-P 2 and PtdIns-3,4,5-P 3 appear to mediate more complex signaling pathways in higher eukaryotes (36,44). PtdIns-activated currents measured under isotonic conditions were analogous to those activated by cell volume increases. PtdIns-3-P and the control lipid, PIP-2, in the same concentrations had no effect. Moreover, the response to intracellular D3-phospholipids was eliminated by removal of extracellular ATP, again consistent with an effect on cellular ATP release.
The regulatory role of PI 3-kinase in ATP transport and Cl Ϫ channel activity is of particular interest to cholangiocyte secretion because PI 3-kinase has been shown to be an important modulator of bile flow. In the isolated perfused rat liver model, for example, inhibition of PI 3-kinase by wortmannin decreases both bile salt and phospholipid transport and decreases basal bile flow by 25% (15). The present studies suggest that PI 3-kinase might also play an important physiologic role in the regulation of cholangiocyte Cl Ϫ secretion, which is thought to account for ϳ40% of human bile formation (45). In NRC in monolayer culture, increases in cell volume (a) stimulate apical ATP release and (b) increase I sc , the electrophysiologic signature of Cl Ϫ secretion in these cells. These effects appear to be related because the secretory response is decreased or eliminated by wortmannin to inhibit ATP release, apyrase to remove ATP from the apical solution, and suramin or reactive blue-2 to inhibit P2 receptor activation (10). Thus, it is attractive to speculate that PI 3-kinase represents a critical intermediary signal coupling changes in cholangiocyte cell volume to secretion of Cl Ϫ through effects on local ATP concentrations.
Assuming that PI 3-kinase is one of the primary signals regulating volume-sensitive ATP release, several additional points merit further investigation. First, the elements responsible for translating increases in cholangiocyte volume to generation of PI 3-kinase lipid products are yet to be determined. Second, because the molecular identities of the ATP transporting protein and the volume-sensitive Cl Ϫ channel(s) are not established, the cellular site(s) of action of the D3 phosphorylated lipid messengers is not clear. Lastly, functional interactions between PI 3-kinase and other kinases are likely to be operative, and the sequence of action and relative importance of these kinases has not been established.
Taken together, these findings indicate that there are dynamic functional interactions between cell volume, PI 3-kinase activity, ATP release, and Cl Ϫ secretion that contribute to ductular bile formation. Similar results in a hepatocyte cell line (25) suggest that PI 3-kinase may represent a general signal involving both cell types in the bile secretory unit. Consequently, cell volume may represent an important signal involved in "hepato-biliary coupling" through release of ATP into the lumen of intrahepatic ducts. Further characterization of FIG. 7. Exogenous ATP overcomes volume-stimulated Cl ؊ current inhibition by PI 3-kinase. A, representative whole cell current tracing. In the presence of wortmannin (50 nM), the Cl Ϫ current response to hypotonic exposure (20% decrease in NaCl) is inhibited. Subsequent addition of ATP (10 M), however, results in Cl Ϫ current activation (downward deflection of the tracing). B, average peak current response recorded under basal conditions, during hypotonic exposure in the presence of wortmannin, and during ATP exposure in the presence of wortmannin (wort).
FIG. 8. Volume-stimulated ATP release. Increases in cell volume enhanced ATP release from NRC monolayers as detected using luminometry (q). Cell swelling was stimulated by diluting media 20 and 40% by adding water (arrows at 4 and 8 min). In all studies, isotonic buffer was added at 2 min to account for any mechanical stimulation of ATP release. Values on the y axis represent luminescence and are recorded in arbitrary light units. Inhibition of PI 3-kinase with wortmannin (Ⅺ) or LY294002 (‚) markedly attenuated volume-dependent ATP efflux. Cells were preincubated with wortmannin or LY294002 for 10 min. the mechanisms involved may provide novel strategies for stimulation of ductular secretion and bile flow in cholestatic liver diseases.