Volume-sensitive Chloride Channel Activity Does Not Depend on Endogenous P-glycoprotein*

To determine whether endogenous P-glycoprotein, the MDR1 gene product that functions as a drug transport pump, is a volume-sensitive Cl (cid:50) channel molecule or a protein kinase C-mediated regulator of the Cl (cid:50) channel, whole-cell patch-clamp and molecular biologi- cal experiments were carried out in a human small intestinal epithelial cell line. Endogenous expression of P-glycoprotein was confirmed by Northern blot analysis, reverse transcription-polymerase chain reaction, Western blot analysis, and immunostaining. The P-gly-coprotein expression was abolished by the antisense (but not sense) oligonucleotide for the MDR1 gene, whereas the magnitude of the Cl (cid:50) current activated by osmotic swelling was not distinguishable between both antisense- and sense-treated cells. The volume-sensitive Cl (cid:50) currents were not specifically affected by the anti-P-glycoprotein monoclonal antibodies, MRK16, C219, and UIC2. An inhibitor of P-glycoprotein-mediated pump activity, verapamil, was found to never affect the Cl (cid:50) current. A substrate for the P-glycoprotein-medi-ated drug pump, vincristine or daunomycin, did not pre- vent swelling-induced activation of the Cl (cid:50) current. Fur-thermore, the Cl (cid:50) current was not affected by an activator of protein kinase C (12- O -tetradecanoylphor- bol-13-acetate or 1-oleoyl-2-acetyl- sn -glycerol). Thus, it is concluded on coverslips tissue culture chambers Chamber Slide; Nunc, Naperville). The cells fixed in phosphate-buffered saline (PBS) containing 1% paraformaldehyde for 30 min at 4 °C. They were washed with PBS, and blocking was made with 10% normal goat serum in PBS for 30 min at room temperature. They were then incubated overnight at 4 °C with 0.3 ml of PBS containing 2.7 (cid:109) g of monoclonal antibody (MRK16) and then washed and incubated for 2 h at room temperature with 0.3 ml of biotinylated sheep antimouse Ig (Amersham) diluted 1:200 with PBS. They were then washed and incubated for 2 h at room temperature with 0.3 ml of streptavidin Texas Red (diluted 1:300 with PBS, Amersham). After a third wash with PBS, the cells were mounted on slides and examined with a fluorescence microscope (model BH-2; Olympus, Tokyo). Antisense Study— The antisense and sense phosphorothioate oligonucleotides corresponding to position (cid:50) 15 to (cid:49) 5 (relative to the trans-lation initiation site) of the human MDR1 coding regions (38, 39) were synthesized and purified by high pressure liquid chromatography (Bio-Service Corp.; Saitama, Japan). Confluent Intestine 407 cells were seeded in 16-well culture chambers at 1 (cid:51) 10 4 cells/200 (cid:109) l of the culture medium and cultured for 3 days. The oligonucleotides (10 (cid:109) M ) were then added into the culture medium and renewed every 24 h for 4 days before immunostaining and patch-clamping. All experiments using sense or antisense oligonucleotides were repeated three times, and the results obtained were essentially reproducible.

Cell volume regulation is one of the most fundamental functions of living cells. A variety of ion channels and transporters are known to be activated upon osmotic cell swelling and involved in volume regulation (1)(2)(3). Among them, volume-sensitive Cl Ϫ channels have recently been studied in a variety of cell species (4 -16). However, the volume-sensitive Cl Ϫ channel molecule has not as yet been identified. P-glycoprotein, the MDR1 gene product, which is an ATPdependent transporter responsible for multidrug resistance to many anti-neoplastic agents (17), was recently proposed as a candidate molecule of the volume-sensitive Cl Ϫ channel (18,19). Higgins and his collaborators (18) suggested that P-glycoprotein (P-gp) 1 is bifunctional, switching between a drug pump upon exposure to its substrates and a Cl Ϫ channel upon osmotic swelling, based on the following observations. 1) Volume-sensitive Cl Ϫ conductances correlated with P-gp expression induced by transfection with MDR1 (18); 2) inhibitors of the drug pump inhibited the volume-sensitive Cl Ϫ conductance (18,20); 3) the volume-sensitive Cl Ϫ conductance was abolished by treatment with antisense oligonucleotides against the 5Ј-end of the MDR1 gene (18); and 4) intracellular administration of a substrate for the P-gp-mediated drug pump prevented swelling-induced activation of the Cl Ϫ conductance but failed to inhibit the preactivated volume-sensitive Cl Ϫ conductance (19,20).
Subsequently, many laboratories have failed to find any correlation between the size of volume-sensitive Cl Ϫ currents and the P-gp expression level in multidrug-resistant or MDR1transfected cells and their control counterparts (21)(22)(23)(24)(25). Moreover, volume-sensitive Cl Ϫ currents in many cell types have been found to be insensitive to blockers of P-gp-mediated drug transport (12, 22-24, 26, 27). Higgins and his collaborators (28) then proposed an alternative hypothesis that P-gp regulates volume-sensitive Cl Ϫ channels via protein kinase C (PKC)mediated phosphorylation of P-gp.
The purpose of the present study is to directly examine whether endogenous P-gp is a volume-sensitive Cl Ϫ channel molecule or a PKC-mediated regulator of the Cl Ϫ channel. We employed a human small intestinal epithelial cell line (Intestine 407) as a model system for the following reasons. (a) Small intestinal epithelial cells endogenously express sizable P-gp (29,30); (b) cell volume regulation is a prerequisite for the normal function of enterocytes in which accumulation of osmotically active solutes is inevitably produced by their vigorous active transport (31,32); (c) in Intestine 407 cells prominent activation of volume-sensitive Cl Ϫ currents is actually induced by osmotic swelling (4), and the channel properties have been well characterized (4,33); and (d) in this cell line the PKC activity has been known to play a significant role in receptormediated cell responses (34). The present results clearly indicated that the endogenous activity of volume-sensitive Cl Ϫ channel is totally independent of endogenously expressed P-gp in this cell line.

EXPERIMENTAL PROCEDURES
Cells-Human small intestinal epithelial cell line (Intestine 407) was cultured, as described previously (4). A human hepatoma cell line (Hep G2), which is known to endogenously express MDR1 (35), was cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum.
mRNA Analysis-Cultured cell total RNA was isolated by the gua-nidinium isothiocyanate lysis method and sedimentation in CsCl. For the Northern blotting total RNAs (10 g) isolated from Intestine 407 or Hep G2 cells were electrophoresed in a 1% agarose-formaldehyde gel and were transferred to a Hybond-N ϩ membrane (Amersham, Little Chalfont, United Kingdom). The probe used was a human MDR1 cDNA, which includes the entire coding region and was 32 P-labeled by random priming. Hybridization was performed, as described previously (36). The filter was exposed to an x-ray film at Ϫ70°C for 14 days with an intensifying screen. Amplification and detection of human MDR1 and MDR2 mRNA in Intestine 407 and Hep G2 cells were achieved by reverse transcriptionpolymerase chain reaction (RT-PCR), as described previously (35). MDR1-specific sequences were amplified by using the sense strand primer CTTCAGGGTTTCACATTTGGC (residues 2314 -2334) and the antisense strand primer CTCATGAGTTTATGTGCCACC (residues 2889 -2909), which yield a 596-base pair (bp) PCR product. MDR2specific sequences were CTTTCAAGCAGTAGCCACG (residues 585-603; sense strand) and TGTGCCCTCATCAGGGT (residues 1340 -1356; antisense strand). The resulting PCR product using these amplimers was 772 bp.
P-glycoprotein Analysis-Crude membrane fractions were prepared from Intestine 407 and Hep G2 cells, as described previously (37). These fractions (30-g proteins) were electrophoresed in 6% SDS-polyacrylamide gel and transferred to an Immobilon-P membrane (Millipore, Bedford). Western blot analysis was made with a monoclonal antibody C219 using the ECL detection system (Amersham, Little Chalfont, United Kingdom).
For the immunofluorescence detection of P-gp Intestine 407 cells were cultured on glass coverslips or tissue culture chambers (16-well Chamber Slide; Nunc, Naperville). The cells were fixed in phosphatebuffered saline (PBS) containing 1% paraformaldehyde for 30 min at 4°C. They were washed with PBS, and blocking was made with 10% normal goat serum in PBS for 30 min at room temperature. They were then incubated overnight at 4°C with 0.3 ml of PBS containing 2.7 g of monoclonal antibody (MRK16) and then washed and incubated for 2 h at room temperature with 0.3 ml of biotinylated sheep antimouse Ig (Amersham) diluted 1:200 with PBS. They were then washed and incubated for 2 h at room temperature with 0.3 ml of streptavidin Texas Red (diluted 1:300 with PBS, Amersham). After a third wash with PBS, the cells were mounted on slides and examined with a fluorescence microscope (model BH-2; Olympus, Tokyo).
Antisense Study-The antisense and sense phosphorothioate oligonucleotides corresponding to position Ϫ15 to ϩ5 (relative to the trans-lation initiation site) of the human MDR1 coding regions (38,39) were synthesized and purified by high pressure liquid chromatography (Bio-Service Corp.; Saitama, Japan). Confluent Intestine 407 cells were seeded in 16-well culture chambers at 1 ϫ 10 4 cells/200 l of the culture medium and cultured for 3 days. The oligonucleotides (10 M) were then added into the culture medium and renewed every 24 h for 4 days before immunostaining and patch-clamping. All experiments using sense or antisense oligonucleotides were repeated three times, and the results obtained were essentially reproducible.
Electrophysiology-Suspensions of spherical Intestine 407 cells were prepared by detaching from the plastic substrate and culturing with agitation for 10 -120 min. The cells placed in a chamber (0.3 ml) were perfused at 5 ml/min. A hypotonic challenge was made by switching the perfusate from an isotonic to hypotonic solution.
The pipettes were fabricated from borosilicate glass capillaries (Hilgenberg, Malsfeld, Germany) using a puller (P-97; Sutter Instrument, Novato). Wide-tipped patch pipettes with a tip resistance of around 1.5 M⍀ when filled with the pipette solution were employed in order to reduce the voltage drop across the residual series resistance and to efficiently introduce a monoclonal antibody into the cytosol.
The following agents were added to the bath solution: verapamil, nifedipine, cyclosporin A, vincristine, 1,9-dideoxyforskolin (DDFSK), tamoxifen, 12-O-tetradecanoylphorbol-13-acetate (TPA), and 1-oleoyl-2acetyl-sn-glycerol (OAG). When necessary, vincristine or daunomycin was added to the pipette solution. All the agents except for cyclosporin Statistical Analysis-Statistical differences of the data were evaluated by Student's t test and considered significant at p Ͻ 0.05. Values are given as mean Ϯ S.D.

RESULTS
Endogenous Expression of P-glycoprotein mRNA in Intestine 407 Cells-As shown in Fig. 1A, Northern blot hybridization with a 4.2-kilobase pair MDR1 cDNA probe suggested endogenous expression of MDR1 mRNA in Intestine 407 cells. However, the MDR1 cDNA probe might have detected MDR2 mRNA, the product of which is not a P-gp isoform conferring multidrug resistance. Therefore, we used the RT-PCR technique to study the precise MDR1 mRNA expression. Messenger RNA from Hep G2 cells was used as a positive control (35). As shown in Fig. 1B Fig. 1C illustrates the result of a Western blot experiment performed using a monoclonal antibody against P-gp, C219. A band of apparent molecular weight of 170,000 was clearly detectable in the lanes loaded with crude cell membrane protein from Hep G2 cells (lane 1) and Intestine 407 cells (lane 2), indicating the endogenous P-gp expression at the protein level.
For further confirmation of the P-gp expression indirect immunofluorescence was carried out using MDR1-specific monoclonal antibody, MRK16. As shown in Fig. 2A, Intestine 407 cells were found to highly express P-gp.
Absence of Effect of Antisense Inhibition of P-glycoprotein Expression on Volume-sensitive Cl Ϫ Currents-As shown in Fig. 2, the cells treated with antisense oligonucleotides (D) did not exhibit clear fluorescence above the background (B), whereas the fluorescence was clearly detectable in cells treated with sense oligonucleotides (C). Thus, it appears that P-gp expression was virtually abolished by 4-day incubation of Intestine 407 cells with the antisense oligonucleotides.
In both cells treated with the sense and antisense oligonucleotides, as shown in Fig. 3 (A and B, left traces), activation of whole-cell Cl Ϫ currents, essentially to the same extent, was consistently observed in association with osmotic swelling. These results are in contrast to a previous antisense study in NIH-3T3/MDR cells (18). Voltage dependence of the current profile (Fig. 3, A and B, right traces) and the mean amplitudes of peak current density (Fig. 3C) were indistinguishable between the cells treated with the sense and antisense oligonucleotides. These results clearly indicate that the activity of the volume-sensitive Cl Ϫ channel does not require the P-gp expression.
Absence of Specific Effect of Monoclonal Antibodies to Pglycoprotein on the Volume-sensitive Cl Ϫ Currents-The effect of a monoclonal antibody raised against an intracellular epitope of P-gp, C219 (42), was investigated by introducing the antibody into the pipette (intracellular) solution, since C219 has been reported to inhibit ATP binding to P-glycoprotein (43) and to suppress volume-sensitive Cl Ϫ currents in the multidrug-resistant human breast cancer cell line (44). A hypotonic challenge brought about activation of Cl Ϫ currents even in the presence of C219 (10 g/ml; Fig. 4A). As shown in Fig. 4, B and C, voltage dependence and peak amplitudes of the Cl Ϫ currents recorded in C219-loaded cells were essentially similar to those observed in the absence of C219.
Activation of Volume-sensitive Cl Ϫ Currents Even in the Presence of a Substrate for P-glycoprotein-Vincristine and daunomycin, which are transported by P-gp, have previously been shown to prevent activation of volume-sensitive Cl Ϫ channels in NIH-3T3/MDR cells (19,20). In the presence of vincristine or daunomycin within Intestine 407 cells, however, volume-sensitive Cl Ϫ currents could be fully activated, as shown in Fig. 6  (A and B). The mean amplitude of peak Cl Ϫ current density (recorded at ϩ80 mV: Fig. 6C) and voltage dependence of the current profile (data not shown) in the cells loaded with vincristine or daunomycin were not significantly different from those in the control cells. Extracellular application of vincristine also never prevented swelling-induced activation of the Cl Ϫ current and never significantly suppressed preactivated Cl Ϫ currents (data not shown).
Varying Effects of Inhibitors for P-glycoprotein on the Volume-sensitive Cl Ϫ Currents-Verapamil is known to reverse multidrug resistance by inhibiting P-gp-mediated drug pump (47) and also inhibit volume-sensitive Cl Ϫ currents in NIH-3T3 cells exhibiting overexpression of P-gp (18,20). As shown in   Fig. 7B) partially inhibited the Cl Ϫ currents by their application after or prior to osmotic swelling. Similar partial inhibition was also induced by extracellular cyclosporin A (10 M), which was reported to be ineffective in NIH-3T3/MDR cells (20). In contrast, pretreatment with extracellular DDFSK (100 M, Fig.  7C, lower trace) or tamoxifen (1 M) almost completely prevented swelling-induced activation of the Cl Ϫ current. Bath application of tamoxifen largely suppressed preactivated volume-sensitive Cl Ϫ currents. However, preactivated volumesensitive Cl Ϫ currents were not affected by bath application of DDFSK (Fig. 7C, upper trace). This fact is in contrast with the observation in NIH-3T3/MDR cells (20). As summarized in Fig.  8, reversers of multidrug resistance produced varying effects on volume-sensitive Cl Ϫ currents depending on drug species and application method.
Absence of Effect of Activators of Protein Kinase C on the Volume-sensitive Cl Ϫ Currents-Recently, activation of the volume-sensitive Cl Ϫ channel has been suggested to be regulated by PKC-mediated phosphorylation of P-gp based on the observation that prior treatment with TPA (30 nM, 2 min) prevented activation of volume-sensitive Cl Ϫ currents (28). Therefore, we examined the effects of PKC activators, TPA and OAG, on the volume-sensitive Cl Ϫ current in Intestine 407 cells. As shown in Fig. 9A, application of TPA (30 nM) before or after hypotonic challenge did not affect the Cl Ϫ currents. A higher concentration of TPA (100 nM) was also without effect (data not shown). A diacylglycerol analogue, OAG (50 M), which was shown to actually activate PKC-mediated cell responses in this cell line (34), was also little effective (Fig. 9B). DISCUSSION Recently, there has been an upsurge of interest in the possible relation between the drug pump P-gp and the Cl Ϫ channel activity because of structural similarity of P-gp to the cystic fibrosis transmembrane conductance regulator (CFTR) Cl Ϫ channel, both of which belong to the ABC (ATP-binding cassette) superfamily of transporters. In this connection, two intriguing hypotheses have been proposed by the same group: "pump/channel bifunctional hypothesis" (18,19) and "PKCmediated channel regulator hypothesis" (28). In the first hypothesis, P-gp is assumed to be itself a volume-sensitive Cl Ϫ channel molecule (18,19). However, virtually complete inhibition of P-gp expression by antisense oligonucleotides against the 5Ј-end of the MDR1 gene failed to affect swelling-induced activation of the Cl Ϫ channel in Intestine 407 cells (Figs. 2 and  3). This result is in good agreement with an increasing number of recent papers, which report lack of correlation between P-gp expression and the magnitude of the volume-sensitive Cl Ϫ conductance (21)(22)(23)(24)(25). Also, in the present study four types of monoclonal antibodies directed against P-gp (C219, MRK16, MRK17, and UIC2) were all found to be ineffective in inhibiting the activity of the volume-sensitive Cl Ϫ channel in Intestine 407 cells (Figs. 4 and 5). According to the bifunctional model, FIG. 6. Effects of P-gp substrates on the volume-sensitive Cl Ϫ currents. A and B, representative traces of whole-cell Cl Ϫ currents recorded during application of alternating 2-s pulses (from 0 to Ϯ40 mV) and 5-s step pulse from Ϫ80 to ϩ20, ϩ40, ϩ60, and ϩ80 mV (at asterisk) before and after hypotonic challenge (at horizontal lines) in Intestine 407 cells dialyzed for over 10 min with intrapipette vincristine (250 M, A) or daunomycin (250 M, B). Arrowheads are the same as those in Fig.  3. C, the mean current density of the peak Cl Ϫ current at ϩ80 mV in the presence of either substrate was not statistically different from that in the absence of substrate (control). binding of the substrate to the drug transporter P-gp should inhibit activation of Cl Ϫ currents associated with osmotic swelling (19,20). However, as shown in Fig. 6, Cl Ϫ currents of normal size were activated upon osmotic swelling of Intestine 407 cells loaded with the substrate (vincristine or daunomycin). Sensitivity of swelling-induced Cl Ϫ currents to prior treatment with inhibitors for P-gp-mediated drug transport (such as verapamil) has been thought to be additional support for the bifunctional hypothesis (18,20). In Intestine 407 cells, verapamil never suppressed the volume-sensitive Cl Ϫ current (Fig. 7). This observation is definitely incompatible with the bifunctional hypothesis. However, there remains the possibility that P-gp and the volume-sensitive Cl Ϫ channel bear some structural similarity or some functional correlation with each other, since many other inhibitors for the drug pump function of P-gp (DDFSK, tamoxifen, nifedipine, and cyclosporin A) suppressed, to a varying extent, the volume-sensitive Cl Ϫ current in Intestine 407 cells (Figs. 7 and 8).
The second "PKC-mediated regulator hypothesis" is based on the observation that swelling-induced Cl Ϫ currents were abolished by pretreatment with a PKC activator, TPA (30 nM, 2 min), in NIH-3T3 cells transfected with human MDR1 but not in the parental cells lacking detectable P-gp expression (28), and also by the fact that the P-gp pump function is likely to be modulated by phosphorylation due to PKC activation (48,49). In the present study, however, both TPA and OAG failed to prevent swelling-induced activation of the Cl Ϫ currents in Intestine 407 cells (Fig. 9). This fact clearly rules out the possibility that endogenous P-gp is involved in the modulation of the volume-sensitive Cl Ϫ channel through PKC-mediated phosphorylation of P-gp. Also, our previous studies showed that the activity of the volume-sensitive Cl Ϫ channel in Intestine 407 cells was not affected by treatment with PKC inhibitors, polymyxin B (4) and H-7 (33). Therefore, it appears that PKCmediated phosphorylation of P-gp (or the Cl Ϫ channel) is not involved in the mechanism of activation or regulation of volume-sensitive Cl Ϫ channels in the human epithelial cells.
The present study clearly showed that P-gp, an ABC transporter, is not the volume-sensitive Cl Ϫ channel molecule nor its PKC-mediated regulator. CFTR, which also belongs to the ABC superfamily, is not responsible for the volume-sensitive Cl Ϫ channel molecule, because the volume-sensitive Cl Ϫ channel is not activated by cyclic AMP (4), has a halide selectivity distinct from that for CFTR (4), and was found to be insensitive to FIG. 8. Varying effects of a number of P-gp inhibitors of volume-sensitive Cl Ϫ currents. A, effects of bath application of verapamil (100 M, n ϭ 9), nifedipine (100 M, n ϭ 7), cyclosporin A (10 M, n ϭ 4), tamoxifen (1 M, n ϭ 4), or DDFSK (100 M, n ϭ 6) on the preactivated Cl Ϫ currents. The percent of inhibition was evaluated from the peak Cl Ϫ current densities (at ϩ40 mV) before (control, 130 -150 pA/pF) and after application of the compound. B, effects of pretreatment with extracellular verapamil (100 M, n ϭ 4), nifedipine (100 M, n ϭ 4), cyclosporin A (10 M, n ϭ 4), tamoxifen (1 M, n ϭ 4), or DDFSK (100 M, n ϭ 4) on the swelling-induced activation of Cl Ϫ currents. The percent of inhibition was evaluated from the peak Cl Ϫ current densities (at ϩ40 mV) before and after (control, 80 -100 pA/pF) washout of the compound. Essentially similar results were obtained for the peak Cl Ϫ current observed at ϩ80 mV (n ϭ 2-4).
FIG. 9. Effects of TPA or OAG on the volume-sensitive Cl Ϫ currents. Representative traces of whole-cell Cl Ϫ currents recorded during application of alternating 2-s pulses (0 to Ϯ40 mV) in the absence or presence of TPA (30 nM, A) or OAG (50 M, B). The PKC activator was applied to the bath after (a) or before (b) hypotonic challenge. Arrowheads and horizontal lines above the current traces are the same as those in Fig. 7. The mean values of preactivated volume-sensitive Cl Ϫ current density (at ϩ40 mV) before and after compound application were 157.4 Ϯ 20.7 and 175.6 Ϯ 29.9 pA/pF (n ϭ 6) for TPA (A, a) as well as 186.7 Ϯ 22.6 and 178.9 Ϯ 18.4 pA/pF (n ϭ 5) for OAG (B, a), respectively. Those activated in the presence of TPA (A, b) and OAG (B, b) were 157.9 Ϯ 13.4 (n ϭ 5) and 183.7 Ϯ 38.3 pA/pF (n ϭ 3) at ϩ40 mV, respectively. glibenclamide (30 M), 2 which is known to inhibit CFTR Cl Ϫ channels (50,51). Another cloned non-ligand-gated Cl Ϫ channel is CLC (52). Among the CLC members, CLC-2 was shown to exhibit volume sensitivity (53). However, the volume-sensitive Cl Ϫ channel in Intestine 407 cells is distinct from CLC-2 in the voltage sensitivity, halide selectivity, and sensitivity to a stilbene derivative Cl Ϫ channel blocker (4,33). Thus, the molecular identification of the volume-sensitive Cl Ϫ channel is still open to investigation.