K+ channels participate in the regulatory volume decrease (RVD) accompanying hepatocellular nutrient uptake and bile formation. We recently identified KCNQ1 as a molecular candidate for a significant fraction of the hepatocellular swelling-activated K+ current (IKVol). We have shown that the KCNQ1 inhibitor chromanol 293B significantly inhibited RVD-associated K+ flux in isolated perfused rat liver and used patch-clamp techniques to define the signaling pathway linking swelling to IKVol activation. Patch-electrode dialysis of hepatocytes with solutions that maintain or increase phosphatidylinositol 4,5-bisphosphate (PIP2) increased IKVol, whereas conditions that decrease cellular PIP2 decreased IKVol. GTP and AlF4− stimulated IKVol development, suggesting a role for G proteins and phospholipase C (PLC). Supporting this, the PLC blocker U-73122 decreased IKVol and inhibited the stimulatory response to PIP2 or GTP. Protein kinase C (PKC) is involved, because K+ current was enhanced by 1-oleoyl-2-acetyl-sn-glycerol and inhibited after chronic PKC stimulation with phorbol 12-myristate 13-acetate (PMA) or the PKC inhibitor GF 109203X. Both IKVol and the accompanying membrane capacitance increase were blocked by cytochalasin D or GF 109203X. Acute PMA did not eliminate the cytochalasin D inhibition, suggesting that PKC-mediated IKVol activation involves the cytoskeleton. Under isotonic conditions, a slowly developing K+ current similar to IKVol was activated by PIP2, lipid phosphatase inhibitors to counter PIP2 depletion, a PLC-coupled α1-adrenoceptor agonist, or PKC activators and was depressed by PKC inhibition, suggesting that hypotonicity is one of a set of stimuli that can activate IKVol through a PIP2/PKC-dependent pathway. The results indicate that PIP2 indirectly activates hepatocellular KCNQ1-like channels via cytoskeletal rearrangement involving PKC activation.
- patch clamp
- phosphatidylinositol 4,5-bisphosphate
- regulatory volume decrease
liver cell volume maintenance in the face of nutrient, bile acid, and xenobiotic uptake is mediated by a regulatory volume decrease (RVD) involving K+ and anion efflux. Swelling-activated K+ efflux and the ensuing choleresis can be induced by exposure of the liver or isolated hepatocytes to hypotonic solutions (4, 27, 43, 44). However, little is known about either the molecular candidate(s) responsible for passive K+ flux or the signaling pathways involved in RVD-induced bile formation. Work in our laboratory (27) recently defined conditions for assaying a swelling-activated K+ current (IKVol) in rat hepatocytes and identified KCNQ1/KCNE3 as molecular candidates for IKVol. KCNQ1 belongs to the voltage-dependent, outwardly rectifying KCNQ channel family. Functionally, it conducts K+ current to increase anion secretion in intestinal cells (38) and is involved in gastric acid secretion in parietal cells (12).
In heterologous expression systems, KCNQ1 is activated by plasma membrane phosphatidylinositol 4,5-bisphosphate (PIP2) (15, 29, 30, 36, 49), likely through a direct electrostatic interaction with positively charged amino acids in this protein (18, 36). However, it is not yet clear whether this stimulatory effect of PIP2 on KCNQ1 is valid for native KCNQ1-like channels. For instance, exogenously applied PIP2 inhibited KCNQ1-like currents in guinea pig atrial myocytes (6). Cardiac KCNQ1 channels also are activated by other factors, including protein kinase C (PKC) (41, 48) and cell swelling (13, 25). Hepatocellular swelling-activated K+ efflux involves cytoskeletal changes mediated by p38MAPK (44) and Src (1) signaling pathways, although further details of the signaling pathway linking volume increase to K+ channel activation remain unclear. In this respect, we recently found that hepatocellular IKVol required cytoplasmic Mg-ATP (27). Mg-ATP is necessary for KCNQ channel activity (39), probably through PIP2 synthesis catalyzed by phosphatidylinositol (PI) 4-kinase and PI5-kinase (17). It is not clear, however, whether PIP2 is responsible for activation of KCNQ1 by cell swelling.
Although strong evidence suggests that PIP2 functions as an activator of resting KCNQ1 channels, its effect on swelling-activated KCNQ1 is not yet known. Neither can we exclude the possibility that product(s) of phospholipase C (PLC)-mediated PIP2 hydrolysis, such as diacylglycerol or inositol 1,4,5-trisphosphate exert such a role. Interestingly, PLC is activated by hepatocellular swelling (1, 32). The present study was therefore designed to explore the role of the PIP2/PLC signaling pathway in the modulation of resting and swelling-activated KCNQ1-like K+ currents in short-term cultured rat hepatocytes and to determine whether these channels contribute to RVD-induced whole organ K+ flux in the intact liver. Our results demonstrate that KCNQ1 channels indeed participate in the RVD-induced K+ efflux in the intact liver. Furthermore, PIP2 indirectly regulates IKVol through a PLC-dependent process involving PKC activation and cytoskeletal rearrangement.
MATERIALS AND METHODS
Animals and materials.
Female Sprague-Dawley rats (200–225 g) were obtained from Charles River (Montreal, QC) and housed under a 12:12-h light-dark regime with access to water and rat chow ad libitum. All experimental protocols were approved by the Queen’s University/Canadian Council on Animal Care. Unless otherwise noted, chemicals were purchased from Sigma-Aldrich (St. Louis, MO) or British Drug Houses (Toronto, ON, Canada) and were of the highest grade available. Chromanol 293B [trans-6-cyano-4-(N-ethylsulfonyl-N-methylamino)-3-hydroxy-2,2-dimethylchromane] was a gift from Aventis Pharma (Frankfurt, Germany).
Rat liver perfusion and whole organ K+ flux.
Livers from rats weighing 200–210 g were perfused via the portal vein with Krebs-Henseleit bicarbonate-buffered saline, using a single-pass, constant-flow perfusion system (21). The saline solution was warmed to 37°C, saturated with 95% O2-5% CO2 (vol/vol), and perfused at a flow rate of 2.7 ± 0.4 ml·min−1·g liver−1. Tissue viability was achieved by maintaining portal pressure (average 15–17 cmH2O), O2 supply, temperature, and buffer pH (7.35–7.40) throughout the perfusion. The Krebs-Henseleit solution contained (in mM) 116.8 NaCl, 25 NaHCO3, 1.4 NaH2PO4, 1 CaCl2, 1.2 MgSO4, and 0.9 mM K+ to increase the driving force for K+ efflux. Livers were perfused for 30 min with isotonic Krebs before the introduction of hypotonic Krebs solution, in which the concentration of NaCl was reduced by 40 mM. The established inhibitor of KCNQ1 channels, chromanol 293B, or the vehicle (dimethyl sulfoxide; DMSO) was perfused for 2 min during the hypotonic challenge. Chromanol 293B (1 mM) in DMSO was diluted to 50 μM in the Krebs perfusate.
Effluent perfusate K+ was monitored continuously during liver perfusion with a K+-selective electrode (Ionplus; Orion Products, Thermo Electron, Beverly, MA). The electrode was calibrated before the start of the perfusion with two standard K+ solutions (0.5 and 5 mM K+), and set in-line immediately downstream of the hepatic vein. Effluent perfusate K+ was monitored and acquired for off-line analysis with the use of a CB-405 connector box and Datacan V acquisition and analysis software from Sable Systems International (Las Vegas, NV). Net K+ flux was calculated as the difference between the influent concentration and the effluent and expressed as micromoles per gram of liver per 10-min hypotonic challenge.
Isolation and culture of rat hepatocytes.
Hepatocytes were isolated using an enzymatic dissociation procedure as described previously (20). Livers were perfused via the portal vein with nominally Ca2+-free Krebs solution containing 0.35 mg/ml collagenase (Liberase; Roche Biochemicals, Montreal, QC, Canada) for 7–10 min. After the enzymatic treatment, hepatocytes were dissociated and cultured on glass coverslips in a humidified atmosphere of 95% air-5% CO2 at 37°C. The culture medium contained DMEM salts plus 0.15% NaHCO3, 10 mM HEPES, 10% (vol/vol) fetal calf serum (Invitrogen Canada, Burlington, ON, Canada), 2 mM glutamine, 5 μg/ml insulin, 1 μM dexamethasone, 100 U/ml penicillin, and 100 μg/ml streptomycin (pH adjusted to 7.4 with NaOH). Electrophysiological experiments were carried out on cells cultured between 18 and 48 h.
Current and capacitance recording.
Single hepatocytes were voltage clamped using the whole cell configuration of the patch-clamp technique with an Axopatch 200A amplifier (Axon Instruments, Foster City, CA) and Clampex 7 software as described previously (20). Patch pipettes were made from borosilicate glass (no. G85165T-4; Warner Instruments, Hamden, CT) and had a resistance of 2–4 MΩ when filled and immersed in K+-containing solutions. All experiments were performed at 20°C. Membrane capacitance was monitored using Membrane Test software (Clampex 7). Briefly, pipette capacitance was cancelled in cell-attached configuration before membrane rupture. The time course of capacitance development was monitored from a holding potential of 0 mV by repetitively applying voltage steps of 0.1 s in duration to 20 mV. The whole cell membrane capacitance was calculated from the output as described by Lindau and Neher (28). The sampled data were plotted using Origin 6.0 software (Origin Lab, Northampton, MA). The time course of swelling-induced current development was monitored at 0 mV. Currents were mainly K+ currents under this condition (27). Whole cell currents were digitized (Digidata 1200B) at 5 kHz, and sampled data were analyzed using Origin 6.0 software. To minimize swelling-activated anion currents, we carried out whole cell recordings using low Cl−-containing pipette and bath solutions, and 0.1 mM 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid (Toronto Biochemicals, Toronto, ON), a blocker of volume-activated Cl− channels in rat hepatocytes (26, 31), was included in the bath solution. Cells were bathed in a solution containing (in mM) 140 Na-gluconate, 5 K-gluconate, 5 glucose, 5 HEPES, 1 CaCl2, and 1 MgCl2 (pH adjusted to 7.4 with NaOH). The isotonic pipette solution contained (in mM) 1 MgCl2, 1 CaCl2, 5 HEPES, 4 ATP, 11 EGTA, and 140 K-gluconate, or 96 mM K-gluconate when FVPP (a mixture of 4 KF, 3 Na3VO4, and 10 K4P2O7) was added (pH adjusted to 7.2 with KOH). Either 20 or 80 mM raffinose was included in the pipette solution to induce hypotonic stress to activate volume-sensitive K+ currents (27). The concentrations of free Ca2+ and Mg2+ in the pipette solution were calculated to be about 13 nM and 37 μM, respectively (27). Chromanol 293B, phorbol 12-myristate 13-acetate (PMA), cytochalasin D, aminoalkyl-bisindolylmaleimide I (GF 109203X), phenylephrine, or phentolamine was added to the bath solution. Unless otherwise noted, wortmannin, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY-294002), ML-7, PIP2, U-73122, U-73343, neomycin, AlCl3, poly-l-lysine, diacylglycerol (DAG) analog 1-oleoyl-2-acetyl-sn-glycerol (OAG), arachidonic acid (AA), GTP, PMA, or 100 μM AlF4− (a mixture of 100 μM AlCl3 and 10 mM NaF) was added to the pipette solution and was present throughout the recording period. In some experiments, cytochalasin D, PMA, and GF 109203X were added to the culture medium. Chromanol 293B, cytochalasin D, phenylephrine, phentolamine, wortmannin, LY-294002, ML-7, U-73122, U-73343, AA, GF 109203X, and PMA were prepared as 5–100 mM stock solutions in DMSO. In preliminary experiments, we confirmed that DMSO alone did not have any appreciable effect on K+ currents at concentrations of up to 0.5%. Neomycin, AlCl3, GTP, AlF4, and poly-l-lysine were stored as 10–100 mM stock solutions in distilled water. PIP2 was dissolved in CCH4-CH3OH-H2O (5:5:1, vol/vol/vol) and diluted to 10 μM in the pipette solution before use.
Time courses for each experimental condition in which the pattern of time courses was significantly different from that of controls were generated from pooled data from three rats or at least four cells and are presented as means ± SE (patch-clamp data) or means ± SD (perfusion studies) with the number of experiments stated. Histograms were generated from the pooled current or capacitance values reached after 20 min of dialysis with the pipette solution following the initiation of whole cell recording. Statistical comparisons were made using either Student’s paired or unpaired t-tests as appropriate, and differences were considered to be significant at P < 0.05.
KCNQ1 is involved in RVD-induced K+ flux in intact liver.
A previous study in our laboratory (27) showed that the swelling-induced K+ current IKVol was partially inhibited by chromanol 293B, a KCNQ1 inhibitor (3). To address the possibility that KCNQ1 also plays a role during swelling-induced RVD in vivo, we examined the effect of chromanol 293B on the K+ release during hypotonic challenge in isolated perfused rat livers. Reduction of perfusate osmolarity by removing 40 mM NaCl led within 2 min to a transient increase of the effluent perfusate K+ concentration, which peaked within 5 min (Fig. 1A). When 293B (73 ± 9.2 μM) was perfused just before the peak (3 min after the onset of hypotonic exposure), the rate of K+ release was decreased at all sampling times and baseline was reached about 1 min earlier than in the control perfusions. As shown in Fig. 1B, under control conditions, net K+ flux during the 10-min hypotonic challenge was 343.2 ± 38.7 μmole K+/g liver, whereas 293B reduced the volume regulatory K+ release by ∼48% to 180.6 ± 19.4 μmole K+/g liver (n = 3 for each condition, P < 0.05).
PIP2 activates IKVol.
Earlier work in our laboratory (27) identified KCNQ1 as a candidate for IKVol in hepatocytes. PIP2 activates KCNQ1 in heterologous expression systems (15, 29, 30, 36, 49) but inhibits endogenous KCNQ-like currents in guinea pig atrial myocytes (6). To determine whether PIP2 modulates hepatocellular K+ currents under iso- and hypotonic conditions, the latter of which would induce a significant KCNQ-like current, we dialyzed short-term cultured rat hepatocytes with a solution containing 10 μM PIP2 in the absence or presence of 80 mM raffinose. Under isotonic conditions, outward K+ currents at 0 mV decayed from 137 ± 63 pA (n = 7) immediately after attainment of the whole cell configuration to 12 ± 3 pA at 20 min of dialysis with control intracellular solution (Fig. 2, A and C), indicating slow loss of a channel-activating substance. Inclusion of PIP2 in the pipette solution activated a slowly developing K+ current, IK (Fig. 2A), that reached 416 ± 92 pA at 20 min of dialysis (Fig. 2C; n = 4, P < 0.01). Our preliminary experiments showed that 10 μM represents a close-to-saturating PIP2 concentration, because 50 and 100 μM PIP2 generated the same current amplitude, whereas at <1 μM, PIP2 did not affect hepatocellular K+ currents (data not shown). Intracellular dialysis with FVPP, a mixture of phosphatase inhibitors (10) that activates KCNQ2/3 (49) and other channels (18) by decreasing the rate of PIP2 depletion, also stimulated a similar current (Fig. 2A) that reached 978 ± 165 pA at 20 min of dialysis (Fig. 2C; n = 8, P < 0.01), indicating that rat hepatocytes express a high phosphatase activity. Addition of the KCNQ channel inhibitor chromanol 293B (30 μM) to the bath 5–15 min before whole cell attainment significantly inhibited PIP2- and FVPP-activated currents at 20 min to 163 ± 52 pA (n = 3, P < 0.05) and 347 ± 85 pA, respectively (Fig. 2, A and C; n = 4, P < 0.01). Both PIP2- and FVPP-activated currents responded to changes in the K+ equilibrium potential. These results indicate that a significant fraction of both currents was produced by KCNQ1. Under hypotonic conditions induced by dialysis with 80 mM raffinose (27), 10 μM PIP2 accelerated the development of IKVol (Fig. 2B), from 472 ± 58 to 824 ± 110 pA at 20 min (Fig. 2C; n = 4, P < 0.05). Together, these data suggest that a tonic level of PIP2 is necessary for maintaining IK activity in resting rat hepatocytes and stimulating IKVol in hypotonically challenged cells.
PIP2 resynthesis by PI4/5-kinase is necessary for IKvol activation.
We next examined whether IKVol activation involved PI4/5-kinase-catalyzed synthesis of PIP2. At micromolar concentrations, wortmannin inhibits the activity of most PI kinases (33), resulting in a significant decrease of PIP2 content (33, 49). In rat hepatocytes, pretreatment with 10 μM wortmannin reduced resting PIP2 content by 25% and eliminated recovery from vasopressin-stimulated PIP2 metabolism (34). We found that addition of 10 μM wortmannin to the pipette solution containing 80 mM raffinose significantly inhibited the development of IKVol (Fig. 3A), reaching 252 ± 49 pA at 20 min of dialysis (Fig. 3B; n = 5, P < 0.05). Inclusion of 10 μM PIP2 in the pipette solution reduced the inhibitory effect of 10 μM wortmannin, resulting in 437 ± 32 pA of current at 20 min, not significantly different from the hypotonic control (Fig. 3B; n = 4).
Wortmannin at micromolar concentrations also blocks myosin light chain kinase (MLCK) and PI3-kinase (35, 39). If the effect of 10 μM wortmannin reflected an involvement of MLCK rather than PI4-kinase, the MLCK inhibitor ML-7 should also attenuate the development of IKVol in a similar manner as wortmannin. This was not the case. ML-7 (10 μM) significantly stimulated the development of IKVol, reaching 715 ± 65 pA at 20 min (Fig. 3B). Despite this, 10 μM wortmannin still suppressed IKVol by 45% in the presence of ML-7 (data not shown). The pronounced inhibition of IKVol seen with 10 μM wortmannin thus does not appear to result from MLCK inhibition.
At submicromolar concentrations (e.g., 50 nM), wortmannin inhibits PI3- but not PI4-kinase (35). Micromolar LY-294002 also specifically blocks PI3-kinase activity (42). Pretreatment of rat hepatocytes (46) or human hepatoma cells (9) with either 50 nM wortmannin or 10 μM LY-294002 significantly inhibited PI3-kinase activity induced by hypotonic solutions. Inclusion of 50 nM wortmannin in the pipette solution significantly stimulated IKVol (Fig. 3A), so that at 20 min of dialysis, mean current at 0 mV reached 767 ± 73 pA (Fig. 3B). Similarly, dialysis with 10 μM LY-294002 accelerated IKVol development (919 ± 132 pA at 20 min; Fig. 3B) compared with the hypotonic control. These results confirm and extend those obtained using the intact liver in which net K+ efflux in response to hypotonic challenge was not affected by nanomolar wortmannin or LY-294002 (44). The combined results suggest that the decrease in IKVol in the presence of 10 μM wortmannin was primarily mediated by reducing PIP2 synthesis by inhibition of PI4-kinase, although additional nonspecific effects cannot be excluded.
PIP2 chelators do not inhibit IKVol.
Plasma membrane PIP2 can directly modulate many ion channels and transporters by an electrostatic interaction with these proteins (18). To further address the possibility that PIP2 regulates IKVol indirectly, as opposed to a direct electrostatic effect at the level of the swelling-activated K+ channel, we examined the effect on IKVol of internal application of the following polyvalent cation chelators of PIP2: neomycin (10 μM), AlCl3 (10 μM), or poly-l-lysine (30 μg/ml). All of these substances interfere with the electrostatic interaction of PIP2 and channels (6, 49). Neither AlCl3 nor neomycin significantly affected IKVol over a 30-min dialysis, reaching 550 ± 120 and 604 ± 123 pA, respectively, at 20 min (Fig. 4A; n = 6, P < 0.05). Surprisingly, poly-l-lysine significantly increased IKVol by about 2.5-fold to 1,258 ± 159 pA at 20 min (Fig. 4, A and B; n = 8, P < 0.05). Poly-l-lysine activates PLC in rat hepatocytes (14), raising the possibility that poly-l-lysine stimulates IKVol through PLC activation. To test this possibility, we added 10 μM U-73122 to the pipette solution containing poly-l-lysine. U-73122 significantly decreased poly-l-lysine-stimulated IKVol by 44% to 821 ± 81 pA at 20 min of dialysis (Fig. 4, A and B; n = 6, P < 0.05). These results suggest that PIP2 stimulation of IKVol does not occur through a direct electrostatic interaction with this channel; rather, this occurs through a process involving PIP2 metabolism.
PLC activation is required for IKVol.
Cellular hydrolysis of PIP2 is mainly carried out by PLC and PI3-kinase. Hepatocellular PLC is activated by cell swelling (32). We hypothesized that inhibition of the swelling-activated PLC and consequent inhibition of PIP2 metabolism should inhibit the development of IKVol in rat hepatocytes. To test this idea, we dialyzed cells with the PLC inhibitor U-73122 in the presence of 80 mM raffinose. U-73122 (10 μM) reduced the development of IKVol (Fig. 5A) to 154 ± 41 pA at 20 min of whole cell recording (Fig. 5B). Inclusion of 10 μM PIP2 in the pipette solution containing 10 μM U-73122 increased IKVol to 344 ± 80 pA, about 42% of the control (Fig. 5, A and B; n = 5, P < 0.05), suggesting that U-73122 reduced IKVol by inhibiting PIP2 metabolism and that the latter could be partially recovered by exogenously applied PIP2. Inhibition by U-73122 was specifically due to PLC blockade, because U-73343, a structural analog of U-73122 lacking PLC inhibitory activity, did not significantly affect IKVol (Fig. 5B). These results show that decreasing the rate of PIP2 hydrolysis catalyzed by PLC does not enhance but, rather, reduces IKVol, implying that PIP2 regulates IKVol through a PLC-dependent process.
Enhancement of PLC activity stimulates IK and IKVol.
To further test the proposal that increased PLC activity favors the development of IKVol, we assessed the effect of PLC-linked G protein activation, either directly or via the G protein-coupled α1-adrenergic receptor. Initially, we recorded the effects of G protein activators on resting K+ current (IK) under isotonic conditions. Intracellular application of the G protein substrate GTP (0.1 mM) blocked the decay of IK and induced a slowly developing outward current (Fig. 6A) that reached 465 ± 63 pA at 20 min of dialysis (Fig. 6C; n = 7, P < 0.01). This GTP-enhanced IK was inhibited by the intracellular application of 10 μM U-73122 (Fig. 6A), decreasing IK to 195 ± 39 pA at 20 min after initiation of whole cell recording (Fig. 6C; n = 4, P < 0.05), implying that PLC activation by G proteins mediates IK. Although GDP (0.1 mM) also significantly increased IK from basal to 144 ± 30 pA, this was significantly less than the response to the same concentration of GTP (Fig. 6C; n = 4, P < 0.05). We previously reported that the α1-adrenoceptor ligand phenylephrine stimulated K+ release from the isolated, perfused rat liver (19). To determine whether G protein-coupled receptor activation would further enhance PLC (and hence IK), we perfused 10 μM phenylephrine into the bath solution 0.5 min after whole cell configuration was established. Cells were dialyzed with GTP to provide a reservoir of G protein substrate. Phenylephrine further increased the GTP-induced IK to 774 ± 83 pA at 20 min of dialysis (Fig. 6C; n = 4, P < 0.05). Incubation of the cells with phentolamine (10 μM), an α1-adrenoceptor antagonist, before phenylephrine inhibited the agonist-induced response by 70% (260 ± 33 pA at 20 min) (Fig. 6C; n = 7, P < 0.05), indicating that phenylephrine stimulated IK via its specific α1-adrenoceptor activity. Furthermore, the α1-adrenergic response was decreased to 421 ± 25 pA when the cells were dialyzed with U-73122 (Fig. 6C; n = 4, P < 0.05). These results suggest that PLC activation by G proteins and G protein-coupled receptors stimulates hepatocellular IK.
The effect of G protein-mediated PLC activation on IKVol was investigated in rat hepatocytes dialyzed with 80 mM raffinose. Intracellular application of GTP (0.1 mM) significantly enhanced the rate and extent of IKVol activation (Fig. 6B), reaching 885 ± 83 pA at 20 min of dialysis (Fig. 6C; n = 10, P < 0.05). Dialysis with the same concentration of GDP did not significantly affect the time course of IKVol development (559 ± 110 pA at 20 min) (Fig. 6C; n = 7, P < 0.05). In addition, GTP-stimulated IKVol was further increased to 1,197 ± 81 pA at 20 min by the intracellular application of AlF4− (0.1 mM) (Fig. 6C; n = 7, P < 0.05), which directly reacts with G proteins to increase enzyme activity (2). These results show that enhancement of G protein activity contributes to the development of IKVol in the rat hepatocyte. To assess whether PLC is involved in GTP-activated IKVol, we included U-73122 in the pipette solution containing GTP. U-73122 decreased the GTP-activated IKVol to control levels (560 ± 68 pA at 20 min of dialysis) (Fig. 6, B and C; n = 6). These results indicate that enhancement of PLC activity by G proteins contributes to IKVol in rat hepatocytes.
DAG-dependent PKC is involved in IK and IKVol activation.
PLC hydrolyzes PIP2 to DAG and inositol 1,4,5-trisphosphate (IP3). Earlier observations in our laboratory (27) showed that increased cytosolic Ca2+ is not involved in IKVol, suggesting that IP3 does not play a major role in channel activation. DAG activates PKC and hence may activate IK and IKVol via generation of downstream products of PLC activity. Because DAG can alter the gating of some cation channels, including KCNQ1 (23), we dialyzed rat hepatocytes with the water-soluble DAG analog OAG to determine whether DAG stimulates IK and IKVol. OAG (10 μM) increased IK to 459 ± 70 pA at 20 min (Fig. 7; n = 5, P < 0.01), indicating that OAG can activate KCNQ1-like channels in the rat hepatocytes. To test the possibility that OAG directly associates with and activates the channel, we dialyzed hepatocytes in the isotonic bath solution with OAG in the absence of ATP to eliminate any residual kinase activity. Under these conditions, outward current at 20 min (98 ± 13 pA) was significantly lower than that recorded in the presence of OAG and ATP (Fig. 7; n = 4, P < 0.01). These results suggest that OAG activates IK in an ATP-dependent process that does not involve direct association of OAG with the channel. To determine whether the stimulatory effect of OAG on IK was due to the activation of PKC, we pretreated cells with GF 109203X, a highly selective inhibitor of multiple PKC subtypes (40). The addition of GF 109203X (2 μM) to the bath solution inhibited the OAG-induced increase in IK (139 ± 26 pA; Fig. 7; n = 4, P < 0.05). These results suggest that the OAG stimulation of IK may be a consequence of PKC activation.
In contrast to IK, neither 10 nor 50 μM OAG significantly affected IKVol induced by 80 mM raffinose (420 ± 33 or 594 ± 84 pA, respectively, at 20 min; Fig. 7; n = 6 and 5). This raised the possibility that cell swelling induced with 80 mM raffinose is associated with significant, perhaps saturating, endogenous DAG synthetic activity so that exogenous activators have no additional effect. To test this proposal, we applied 20 mM raffinose, rather than 80 mM, to decrease the level of cell swelling. At 20 mM, raffinose increased IKVol to 340 ± 77 at 20 min of dialysis, or 70% of the response to 80 mM raffinose (Fig. 7; n = 4, P < 0.05). OAG further increased the 20 mM raffinose-induced IKVol to 525 ± 20 pA at 20 min of dialysis (Fig. 7; n = 4, P < 0.05).
AA, a product of DAG lipase-catalyzed hydrolysis of DAG, is involved in the activation of Drosophila light-sensitive transient receptor potential channels (5). To establish whether DAG activates the hepatocellular IKVol in a similar process, we added 10 μM AA to pipette solutions containing 80 mM raffinose. IKVol induced by either 80 (Fig. 7; n = 5, P < 0.05) or 20 mM raffinose (data not shown) was not significantly affected by AA, indicating that AA is not necessary for IKVol development in rat hepatocytes.
Involvement of PKC in the OAG-induced stimulation of IK and IKVol was further evaluated with an activator and an inhibitor of PKC. Intracellular application of the PKC activator PMA (10 μM) slowly activated IK with a time course mimicking that observed with OAG. Pretreating the hepatocytes for at least 20 min with GF 109203X (2 μM) significantly inhibited the PMA response (Fig. 8A). PMA-stimulated IK (421 ± 70 pA at 20 min; n = 4, P < 0.01 compared with isotonic control) was decreased to 105 ± 19 pA at 20 min (n = 4, P < 0.05) in GF 109203X treated cells (Fig. 8C). PKC is activated and transferred from the cytosol to the plasma membrane by cell swelling and metabolic stress in hepatoma cells and a cholangiocarcinoma line (37, 45). To test whether PKC mediates cell swelling-induced IKVol, we included 10 μM PMA in the pipette solution with 80 mM raffinose. This acute exposure to PMA did not significantly affect IKVol (502 ± 27 pA at 20 min; Fig. 8C; n = 6), similar to the observations by Grunnet et al. (13) in Xenopus oocytes.
PKC is activated by acute exposure to micromolar concentrations of PMA, whereas longer exposures (18 h) inactivate the kinase (37). Furthermore, chronic exposure to PMA decreased cell swelling- and metabolic stress-induced PKC translocation in hepatoma and cholangiocarcinoma cells (37, 45). We used this method to confirm the involvement of PKC in IKVol development in rat hepatocytes. Cells were pretreated with PMA (0.1 μM) or its vehicle, DMSO, for at least 18 h before whole cell recording. Pretreatment of cells with PMA did not have any significant effect on IK (data not shown). In contrast, PMA significantly attenuated IKVol (Fig. 8B), attaining only 45% (221 ± 42 pA; Fig. 8C; n = 8, P < 0.05) of the control amplitude at 20 min of dialysis. Exposure to GF 109203X had a similar inhibitory effect on IKVol as the 18-h PMA incubation, with mean current reaching 252 ± 31 pA at 20 min (Fig. 8C; n = 4, P < 0.05 relative to hypotonic control). These data show that PKC plays an important role in the development of IKVol.
IKvol development involves cytoskeleton and PKC.
As shown earlier, IKVol in rat hepatocytes develops slowly over minutes of dialysis with 80 mM raffinose, suggesting that channel recruitment to the plasma membrane occurs in response to an as yet only partially defined series of events. To investigate whether the rat hepatocyte IKVol involves a vesicular-to-plasma membrane flux of membrane constituents containing K+ channels, we measured outward K+ current at 0 mV (Fig. 9A) and cell capacitance as a monitor of plasma membrane area (Fig. 9B). Pooled data from these experiments at 20 min of dialysis are illustrated in Fig. 9C. Under isotonic conditions membrane capacitance slowly declined about 4 pF over 20 min from 23 ± 0.9 to 18 ± 1.6 pF (Fig. 9B). Conversely, membrane capacitance did not significantly change from 25 ± 0.6 pF over 20 min of dialysis with 80 mM raffinose. Thus hypotonic conditions caused a 4-pF increase in cell capacitance relative to isotonic solutions.
PKC is involved in the recruitment of a volume-sensitive vesicular pool to a readily releasable state in a human cholangiocarcinoma cell line (11), and Ca2+-independent PKC isoforms mediate trafficking of the bile salt export protein to the canalicular membrane in HepG2 cells and rat hepatocytes (24). Stimulation of PKC activity by acute exposure to 10 μM PMA did not significantly affect capacitance (25 ± 0.8 pF at 20 min; Fig. 9C; n = 6), similar to the observations by Gatof et al. (11) in cholangiocytes. However, both chronic exposure to 0.1 μM PMA and perfusion with GF 109203X reduced raffinose-induced capacitance and current increase (PMA, 22 ± 0.5 and 221 ± 42 pA; GF 109203X, 22 ± 0.8 and 252 ± 31 pA at 20 min; Fig. 9C; P < 0.05 relative to hypotonic control), indicating that PKC is involved in channel recruitment.
Grunnet et al. (13) reported that activation of KCNQ1 by hypotonic solutions was inhibited by the microfilament polymerization blocker cytochalasin D. Pretreatment of the cultured cells with cytochalasin D (2 μM) for 2–6 h significantly depressed IKVol (Fig. 9A) to a mean current of 251 ± 49 pA at 20 min (Fig. 9C; n = 4, P < 0.05). Cytochalasin D also blocked the raffinose-induced capacitance increase (Fig. 9B), reaching 21 ± 0.7 pF at 20 min (Fig. 9C), which was not significantly different from isotonic conditions. Acute exposure of hepatocytes to 10 μM PMA did not rescue the cells from the inhibitory effects of cytochalasin D (22 ± 0.6 and 241 ± 56 pA at 20 min; Fig. 9C). These results suggest that the swelling-induced development of IKVol involves an F-actin-dependent recruitment process in addition to PKC-dependent activation.
The present study identified PIP2 as a second messenger in the intracellular pathway leading to the activation of KCNQ1-like currents by cell swelling in short-term cultured rat hepatocytes. Our results suggest that PIP2 (maintained through PI kinase synthetic activities) hydrolysis mediated by PLC sustains the PKC activity requisite for the development of IKVol (Fig. 10). This conclusion is based on the observations that stimulation of PLC and PKC activities increased the amplitude and rate of development of IKVol, whereas manipulations that would depress PLC and/or PKC inhibited IKVol. Additional results from membrane capacitance measurement and whole organ K+ flux studies indicate that cytoskeletal reorganization, potentially mediating a channel recruitment process, is requisite for PKC activation of hepatocellular IKVol and that KCNQ1 contributes significantly to this K+ flux in the intact liver.
PIP2 is necessary for IKVol.
PIP2 is reported to either stimulate human KCNQ1 when heterologously expressed (15, 30, 36, 49) or inhibit KCNQ1-like currents in guinea pig atrial myocytes (6). The present results demonstrate that an increase in cellular PIP2 concentration stimulates both resting and swelling-activated KCNQ1-like K+ currents in primary cultures of rat hepatocytes. Specifically, procedures designed to increase intracellular PIP2 increased both IK and IKVol (dialysis with PIP2 or inhibitors of lipid phosphatases or PI3-kinase) (Figs. 2 and 3), whereas conditions that decrease cellular PIP2 synthesis (inhibition of PI4-kinase) decreased IKVol (Fig. 3). PI4-kinase-mediated PIP2 synthesis is reported to be necessary for KCNQ2 and KCNQ3 activities (39, 49), in addition to other cation channels and transporters (18). Supporting the proposal that PIP2 synthesis is also required for hepatocellular IKVol activation, we showed that these currents are blocked by micromolar but not nanomolar concentrations of wortmannin (Fig. 3). A 15-min exposure of intact rat hepatocytes to 10 μM wortmannin depresses steady-state plasma membrane PIP2 by ∼35% (34). The inhibitory effects of micromolar wortmannin on IKVol do not reflect inhibition of PI3-kinase or MLCK, because blockers of these enzymes, LY-294002, 50 nM wortmannin, or ML-7, did not mimic the effect of 10 μM wortmannin (Fig. 3). In addition, inhibition of IKVol by 10 μM wortmannin was attenuated by intracellular application of PIP2 (Fig. 3). The combined results suggest that PI4-kinase-catalyzed PIP2 synthesis supports the development of IKVol. An explanation for the noted difference in PIP2 effects on KCNQ1 in hepatocytes and multiple heterologous expression systems compared with guinea pig atrial myocytes (6) may reside in a species and/or cell type in which the channel is expressed.
PIP2 activates IKVol indirectly through a PLC-sensitive process.
Because PIP2 can be hydrolyzed by PLC, IKVol activation by PIP2 may occur through either a direct electrostatic interaction with the hepatocellular KCNQ-like protein or a product of its hydrolysis by PLC. Our results (Fig. 4) show that IKVol was not inhibited by molecules (neomycin, AlCl3, poly-l-lysine) that interfere with the electrostatic interaction of PIP2 with KCNQ/KCNE channels (6, 49). Furthermore, plasma membrane PIP2 concentration was not significantly affected by G protein-coupled receptor-linked PLC activity in both rat hepatocytes due to the resynthesis of PIP2 by PI4-kinase (34) or tobacco pollen cells exposed to hypotonic solutions (50). Therefore, it seems unlikely that PIP2, at least as a signaling molecule, plays a direct role in the activation of IKVol. It is noted that these experiments do not exclude the possibility that PIP2 may be tightly bound to the IKVol channels, making them insensitive to both PIP2 chelation during whole cell recording and PIP2 concentration fluctuations in vivo.
Hypotonic swelling of rat hepatoma cells is associated with U-73122-sensitive PLC activation (32). We found that U-73122, but not its inactive isomer U-73343, inhibited IKVol development in primary cultures of rat hepatocytes (Fig. 5). In addition, U-73122 inhibited both PIP2-stimulated resting and swelling-induced K+ currents (Fig. 5). These results imply that a threshold concentration of PIP2 is required to sustain PLC activity to activate KCNQ1-like channels. Supporting the proposal, GTP stimulated the rate and extent of IKVol development; this was enhanced with AlF4−, and these activities were attenuated by U-73122 (Fig. 6). In resting cells, a slowly developing K+ current similar to IKVol was activated by dialysis with 0.1 mM GTP or the PLC-coupled adrenoceptor agonist phenylephrine (Fig. 6), supporting our hypothesis that this conductance appears in response to conditions that stimulate PLC activity.
PIP2 stimulates hepatocellular K+ currents through a PKC-dependent process.
Activation of PLC results in accumulation of IP3 and the PKC activator DAG in rat hepatocytes. IP3 (and increased cytosolic Ca2+) is not responsible for RVD-induced K+ flux in the perfused rat liver (43) or activating IKVol in isolated rat hepatocytes (27). Although DAG can be generated by PLA2, this pathway is not likely involved in IKVol development, because inhibition of this phospholipase does not affect RVD-induced K+ efflux from the perfused rat liver (43). AA and other DAG metabolites also are not involved in activating IKVol in rat hepatocytes or RVD-induced K+ flux in the perfused rat liver (43). Several of our observations support an important role for PKC in the process of K+ current activation in rat hepatocytes. First, the DAG analog OAG stimulated K+ currents in isotonically bathed cells (IK) in an ATP-dependent and PKC inhibitor-sensitive manner and enhanced IKVol induced by weakly hypotonic solutions (Fig. 7). Second, acute exposure to the PKC activator PMA activated IK, and activation was inhibited by pretreatment with GF 109203X (Fig. 8). And last, long-term treatment of cells with PMA to downregulate endogenous PKC activity and GF 109203X attenuated hypotonically induced IKVol (Fig. 8). This PKC is likely to be a Ca2+-insensitive subtype, because human cardiac KCNQ1 was activated by Ca2+-insensitive PKC (48), and the development of IKVol in rat hepatocytes (27) and RVD-induced K+ flux in rat liver (43) are Ca2+ independent.
The events underlying PKC induction of hepatocellular IKVol may include either a direct activation via channel phosphorylation or indirect activation by way of F-actin disassembly and recruitment of channels to the plasma membrane. Supporting the first possibility, KCNQ1 can be directly activated by PKC protein, because KCNQ1 has conserved PKC phosphorylation sites in its COOH terminus (23, 41), OAG stimulates human KCNQ1/KCNE1 (23), and PMA activated the resting IK in the present study (Fig. 8). Nevertheless, changes in the structure of the F-actin cytoskeleton are a prerequisite for swelling-induced ion channel activity (1, 7, 8, 13), probably through a process of vesicle-mediated insertion and retrieval of proteins (7). This process is enhanced by PMA (11) and impaired by mutation of PKC activity (16), implying that PKC is a critical regulatory element for efficient cytoskeleton-mediated membrane protein sorting. Our results further suggest that PKC-induced IKVol activation is dependent on recruitment of channels to the plasma membrane, because PMA was unable to eliminate the inhibitory effect of cytochalasin D on the development of IKVol and the swelling-induced membrane capacitance increase (Fig. 9).
Several types of K+ channels have been proposed as the molecular entities mediating the swelling-activated K+ current associated with RVD in different tissue, and work in our laboratory recently identified KCNQ1 as a significant participant of this current in isolated rat hepatocytes (27). In the present study, we extended earlier results by demonstrating that KCNQ1 channels provide a significant fraction of the K+ flux associated with RVD in the intact perfused rat liver (Fig. 1). To our knowledge, this is the first report that provides evidence for the role of KCNQ1 in RVD-induced K+ flux in vivo. RVD and whole organ K+ flux have been correlated with an increase in bile flow and bile acid excretion (4). RVD dysfunction induced by K+ flux disruption is associated with cholestasis (4), toxin-induced hepatocellular injury (45), and ischemia related to hypoperfusion or organ preservation (22). We speculate that disturbance of the normal development of the hepatocellular volume-regulated K+ current may play a fundamental role in the pathophysiological consequences of liver dysfunction.
In summary, we have presented functional evidence for KCNQ1-like channels as significant contributors to RVD-induced K+ efflux in the intact perfused rat liver. Pharmacological and electrophysiological evidence reported supports the view that PIP2 indirectly mediates the development of hepatocellular IKVol through a PKC-dependent process and that PKC stimulation of IKVol requires cytoskeletal rearrangement resulting in increased numbers of functional K+ channels in the plasma membrane (Fig. 10). Recent evidence demonstrates that the normal trafficking to the plasma membrane of KCNQ1 is disrupted in channels bearing mutations associated with disease in nonhepatic tissue (47). The molecular identification of KCNQ1/KCNE3 in rat and human liver (27) and characterization of the signaling pathways involved in IKVol activation reported should permit discovery of ways in which to manipulate channel activity and ameliorate specific liver dysfunctions.
This work was supported by grants from the Canadian Institutes of Health Research (CIHR) and the Natural Sciences and Engineering Research Council (NSERC) (to C. E. Hill). A CIHR Training Grant in Digestive Sciences partially supported W. Z. Lan and P. Y. T. Wang. P. Y. T. Wang was the recipient of an NSERC Postgraduate Scholarship.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society