Abstract
Ethanol inhibits L-type Ca++ channels, but little is known about its effect on other voltage-gated Ca++ channels. To examine non-L-type channels we used nerve growth factor-differentiated PC12 cells treated with the L channel blocker nifedipine. Using selective Ca++ channel antagonists, we found that N-type and P/Q -type channels mediate most of the remaining depolarization-evoked Ca++ rise. Ethanol (10–150 mM) inhibited depolarization-induced rises in intracellular Ca++ with maximal inhibition of 46% achieved using 50 mM ethanol. Inhibition was time dependent, requiring at least 8 min to develop fully. Ethanol did not alter Ca++ mobilization, sequestration, extrusion or capacitative entry. Sp-adenosine cyclic 3′,5′-phosphorothioate, a specific activator of protein kinase A (PKA), blocked inhibition by ethanol, whereas the protein kinase C activator phorbol 12-myristate, 13-acetate did not. Okadaic acid, an inhibitor of protein phosphatases type-1 and type-2A, also blocked inhibition by ethanol with an IC50 of 3 nM. This was prevented by inhibiting PKA, indicating that the action of okadaic acid was due to increased PKA-mediated phosphorylation. These results indicate that ethanol can inhibit N-type and P/Q-type channels and this is antagonized by activating PKA. The findings suggest the sensitivity of these channels to ethanol is regulated by a phosphoprotein that is a substrate for PKA and protein phosphatase type-2A.
Voltage-gated Ca++ channels mediate Ca++entry into neurons and regulate neurotransmitter release, firing patterns, gene expression and differentiation (McClesky, 1994; Oliveraet al., 1994; Ghosh and Greenberg, 1995). Several types of Ca++ channels have been identified with distinct electrophysiological and pharmacological properties (Zhang et al., 1993; Randall and Tsien, 1995). T channels activate at low voltage, inactivate quickly, and are blocked by low concentrations of Ni++. L channels are activated by high voltage, inactivate slowly and are blocked by dihydropyridines. N, P and Q channels are activated by high voltage and blocked by the peptide neurotoxins ω-conotoxin GVIA (N), ω-agatoxin IVA (P and Q) and ω-conotoxin MVIIC (N, P and Q). Other channels have also been described that are resistant to organic Ca++channel blockers (G1-G3 and R-type channels) (Forti et al., 1994; Randall and Tsien, 1995).
Several manifestations of ethanol intoxication and dependence appear due to changes in Ca++ channel function (Messing and Diamond, 1997). In nerve terminals from rat neurohypophysis and in NGF-differentiated PC12 cells, brief exposure to intoxicating concentrations (10–50 mM) of ethanol inhibits L-type channels by decreasing open channel probability (Wang et al., 1994) and promoting channel inactivation (Mullikin-Kilpatrick and Treistman, 1995). In N1E-115 neuroblastoma and NG108–15 neuroblastoma-glioma cells, high concentrations of ethanol (100–300 mM) reduce T-type currents by 15 to 20% (Twombly et al., 1990), and in rat neurohypophysis, 50 to 100 mM ethanol reduces N-type current by 30 to 40% (Wang et al., 1991). In rat Purkinje neurons P-type currents appear insensitive to ethanol (Hall et al., 1994). The effect of ethanol on other types of Ca++channels is not known.
Protein phosphorylation regulates ion channels (Nestler and Greengard, 1984) and ethanol can alter phosphorylation through actions on signal transduction pathways that regulate protein kinases, particularly cAMP-dependent PKA and PKC (Messing and Diamond, 1997). In some cells (Nagy et al., 1989; Rabin et al., 1993) ethanol stimulates cAMP formation and activates PKA. In rodent hepatocytes (Hoek et al., 1987) and human platelets (Rubin et al., 1988), ethanol activates phospholipase C to generate the PKC activator diacylglycerol (Nishizuka, 1992). PKA activation appears to be required for inhibition of adenosine transporters by ethanol (Coeet al., 1996), whereas PKC activation is important for inhibition of AMPA/kainate receptor currents (Dildy-Mayfield and Harris, 1995) and enhancement of mouse and bovine GABAA receptor function (Wafford and Whiting, 1992) by ethanol. No studies have identified a role for phosphorylation in the regulation of voltage-gated Ca++ channels by ethanol.
In this study, we examined the effect of ethanol on non-L-type Ca++ channels, by measuring depolarization-induced rises in [Ca++]i in NGF-differentiated PC12 cells treated with the L channel blocker nifedipine. Our findings indicate that ethanol can inhibit N-type and P/Q-type channels by a mechanism that is antagonized by PKA.
Methods
Materials.
trk-PC12 6–24 cells were a gift from David Kaplan, (Frederick Cancer Center, Frederick, MD). ω-agatoxin IVA was generously provided by Michael Adams (University of California, Riverside, CA), Research Biochemicals International as part of the Chemical Synthesis Program of the National Institute of Mental Health (Contract N01 MH3003), and by Pfizer (Groton, CT). Cyclosporin A was supplied by Sandoz Pharmaceuticals (East Hanover, NJ).125I-ω-conotoxin GVIA was purchased from Amersham (Arlington Heights, IL) and unlabeled ω-conotoxin GVIA was obtained from Peptides International (Louisville, KY). NGF was purchased from Collaborative Biomedical Products (Bedford, MA). Okadaic acid (Na+-salt), Sp-cAMPS and deltamethrin were from LC Laboratories (Woburn, MA). Ionomycin and ryanodine were purchased from Calbiochem (San Diego, CA). Ouabain, monensin, poly-l-ornithine and laminin were from Sigma (St. Louis, MO). Fura-2 AM, fura pentapotassium salt and 4-bromo-A23187 were purchased from Molecular Probes (Eugene, OR).
Cell culture.
trk-PC12 cells were maintained in plastic tissue culture flasks at 37°C in a humidified atmosphere of 90% air and 10% CO2 in complete medium containing DMEM supplemented with 10% heat-inactivated horse serum, 5% fetal bovine serum, 2 mM glutamine, 50 U/ml of penicillin, 50 μg/ml of streptomycin and 200 μg/ml of G418 (geneticin). For Ca++ imaging studies, 5 × 105 cells were plated on 22-mm square glass coverslips (Warner Instruments, Hamden, CT) that had been treated for 30 min with 10% HCl in ethanol, washed in PBS, incubated with 0.1 mg/ml of poly-l-ornithine for 30 min, and then coated with laminin (30 μg/ml) overnight at 37°C. The cells were grown for 24 to 48 hr in complete medium supplemented with 50 ng/ml of NGF. After this time, approximately 85 to 90% of the cells had extended neurites more than one cell body in length.
ω-Conotoxin GVIA binding.
Binding of125I-ω-conotoxin GVIA to cells was measured by a modification of a previously described method (Williams et al., 1992). Cells (1.5 × 105) were plated in poly-l-ornithine treated 24-well plates in complete medium. The after day cells were rinsed with buffer A containing 140 mM NaCl, 5 mM KCl, 12 mM glucose, 10 μM CaCl2, 1 mg/ml of BSA and 10 mM HEPES (pH 7.4). Cells were then incubated in 0.4 ml of buffer A containing 80 pM125I-ω-conotoxin GVIA for 1 hr at 37°C. This concentration was chosen because it approximates theKd (60 pM) for equilibrium saturation binding of 125I-ω-conotoxin GVIA to brain membranes (Wagneret al., 1988). After aspiration of the buffer, the wells were rapidly rinsed four times with 1 ml of buffer B (160 mM choline chloride, 1.5 mM CaCl2, 1 mg/ml of BSA and 5 mM HEPES, pH 7.4) at 4°C. After aspiration of all liquid, 1 ml of 1N NaOH was added to each well. Plates were incubated at 37°C overnight and radioactivity contained in 0.8 ml samples taken from each well was determined by liquid scintillation counting. Specific binding of 125I-ω-conotoxin GVIA was calculated as the difference between binding measured in the absence and the presence of 500 nM ω-conotoxin GVIA and accounted for 77 ± 2% of total binding. Protein concentrations were measured by the Lowry method with BSA standards (Lowry et al., 1951).
Measurement of intracellular Ca++.
Cells attached to coverslips were incubated in DMEM containing 25 mM HEPES (pH 7.4) and 5 μM fura-2 AM for 25 min at 37°C. Cells were rinsed twice with 5 mM KCl buffer (85 mM NaCl, 5 mM KCl, 45 mM choline Cl, 2 mM CaCl2, 5 mM glucose, 25 mM HEPES, pH 7.4). The coverslip was mounted onto a perfusion chamber (model RC-21B, Warner Instruments, Hamden, CT) and perfused with 5 mM KCl buffer. All experiments were conducted at 27°C.
An Olympus IMT-2 inverted light microscope fitted with a Nikon UV-F 40X oil immersion objective and a 150W xenon lamp was used for fluorescence measurements. After paired excitation at 350 and 380 nm, fluorescence emission intensities at 510 nm were detected using a liquid-cooled CCD camera (Photometrics Ltd., Tucson, AZ) fitted with a Thompson 7883 chip (384 × 576 pixels). Exposure times were 0.05 to 0.08 sec and 2 × 2 binning was used to enhance image intensity and reduce the time required to transfer data from the chip to the computer. Paired 350 and 380 nm images were separated by less than 0.3 sec.
The [Ca++]i in cell bodies was calculated using the program BDS Image (Oncor Imaging Systems, Gaithersburg, MD) from the ratio of emission intensities at 350 and 380 nm after subtracting for background fluorescence in regions devoid of cells. A mask image was created to identify each cell body as a region of interest for analysis. A mean [Ca++]i value was calculated from values for all pixels within each region of interest (Grynkiewicz et al., 1985). TheKd of fura-2 for Ca++ was estimated to be 190 nM (Williams and Fay, 1990). Rmax was measured by incubating cells for 15 min in high Ca++ buffer, which was similar in composition to 5 mM KCl buffer but contained 10 mM CaCl2, 5 μM 4-bromo-A23187, 5 μM monensin and 0.4 mM ouabain (Williams and Fay, 1990). To measure Rmin, cells were incubated for five min in Ca++ free-buffer identical in composition to high Ca++ buffer except that CaCl2 was replaced by 30 mM EGTA, pH 8.7. Rmax values ranged from 4.68 to 7.20 and Rmin values from 0.49 to 0.85. Background fluorescence in cells not loaded with fura-2 was less than 1% of fluorescence in cells containing fura-2. Ethanol (100 mM) did not alter the affinity of fura-2 for Ca++ in a cell-free calibration assay (Molecular Probes, Eugene, OR) using fura-2 pentapotassium salt and 0–10 mM CaCl2. Ethanol also did not alter leakage of fura-2 from the cells.
Addition of drugs.
Because dihydropyridines preferentially bind to L channels on depolarized cells (Greenberg et al., 1986), we first incubated cells with 50 mM KCl buffer containing 1 μM nifedipine or 10 μM nimodipine. The 50 mM KCl buffer was identical in composition to 5 mM KCl buffer except that KCl was substituted for choline chloride. Cells were then incubated in 5 mM KCl buffer containing 1 μM nifedipine or 10 μM nimodipine for 5 min to allow [Ca++]i to return to resting levels. Cells were then depolarized in 50 mM KCl buffer in the continued presence of nifedipine or nimodipine, and fluorescence images were recorded. This predepolarization step resulted in maximal inhibition of subsequent depolarization-induced [Ca++]i rises by these dihydropyridines.
Cells were preincubated with ω-agatoxin IVA in 5 mM KCl buffer for 5 min before depolarization. Lysozyme (1 mg/ml) was added to buffers containing ω-agatoxin IVA to prevent absorption of the toxin to plastic tubing and containers. Because millimolar concentrations of Ca++ inhibit binding of ω-conotoxin GVIA to N-type channels (Rosenberg et al., 1989), cells were preincubated with ω-conotoxin GVIA in buffer C containing 140 mM NaCl, 5 mM KCl, 12 mM glucose, 10 μM CaCl2, 1 mg/ml of BSA and 10 mM HEPES (pH 7.4) for 25 min at 27°C. Cells were then equilibrated in 5 mM KCl buffer containing ω-conotoxin GVIA for at least 5 min before depolarization. This long incubation period in 10 μM Ca++produced maximal inhibition by ω-conotoxin GVIA of subsequent [Ca++]i rises. Incubation in buffer C without ω-conotoxin GVIA did not alter subsequent depolarization-evoked rises in [Ca++]i.
The effect of caffeine on Ca++ release from internal stores was studied in cells first incubated in 50 mM KCl buffer for 75 sec to stimulate filling of caffeine-sensitive Ca++ stores (Reber and Reuter, 1991). Cells were then incubated in 5 mM KCl buffer for 5 min to allow [Ca++]i to fall to resting levels before 30 mM caffeine was added and images were recorded.
The effect of ryanodine was examined in cells preincubated with 10 μM ryanodine for 20 min in 5 mM KCl buffer. Caffeine (30 mM) was then added for 5 min to stimulate use-dependent blockade of ryanodine receptor/Ca++ release channels by ryanodine (Reber et al., 1993). Caffeine was then removed and, after 5 min, cells were depolarized with 50 mM KCl buffer in the continued presence of ryanodine while images were recorded.
All drugs were added by gravity driven perfusion for 20 sec using an eight-valve perfusion device (AutoMate Scientific, Oakland, CA). Switching solutions produced a lag time of less than 1 sec while the solution passed through tubing to the chamber. The entire volume of the chamber was replaced in less than 12 sec.
Analysis of data.
Results are expressed as mean ± S.E. In most experiments control and treatment values were measured using the same cells. This was possible because the magnitude of the rise in [Ca++]i in cells treated with nifedipine alone was similar (97 ± 4%) for up to three successive depolarizations, when each was separated by at least 5 min. In contrast, repeated stimulation with bradykinin elicited progressively smaller Ca++ rises when separated by 5-min intervals. Therefore, data from bradykinin experiments were derived from different batches of cells in control and ethanol-exposed conditions. Differences between means were analyzed by two-tailed Student’s t tests or by ANOVA, and where P < .05, multiple comparisons were evaluated by the Scheffe F-test.
Results
Identification of Ca++ channels in trk-PC12 cells.
PC12 cells express N-type, L-type and P/Q-type channels, and treatment with NGF for several days increases expression of N-type channels (Plummer et al., 1989; Liu et al., 1996). To facilitate our studies of non-L-type channels, we used the PC12 clone trk-PC12 6–24, which overexpresses the tyrosine kinase NGF receptor trkA and undergoes rapid neuronal differentiation after brief NGF treatment (Hempstead et al., 1992). We found that binding of the N channel antagonist125I-ω-conotoxin GVIA (80 pM) to undifferentiated trk-PC12 6–24 cells (13.44 ± 1.44 fmol/mg; n = 6) was 2.6-fold higher than binding to wild-type PC12 cells (5.14 ± 0.45 fmol/mg; n = 9; P < .05). Treatment with 50 ng/ml of NGF for 48 hr increased binding in trk-PC12 cells to 20.29 ± 0.98 fmol/mg (n = 4; P < .05 compared to untreatedtrk-PC12 cells). This indicates that N channel expression is markedly increased in this PC12 clone especially after treatment with NGF. Therefore we used NGF-differentiated trk-PC12 6–24 cells for our studies.
To determine whether trk-PC12 cells express functional Ca++ channels, we measured depolarization-induced rises in [Ca++]i in NGF-differentiated cells. In the absence of channel antagonists, depolarization stimulated a rapid rise in [Ca++]i from a resting value of 70 ± 3 nM to a peak level of 472 ± 25 nM after 10 sec (fig. 1A). Over the next 50 sec [Ca++]i declined to a plateau of 297 ± 13 nM that persisted for at least another two min (data not shown). Treatment with 5 mM ethylene glycol bis(β-aminoethyl ester)-N, N′-tetraacetic acid (EGTA) completely blocked depolarization-induced rises in [Ca++]i, indicating that Ca++ influx is required for this response. Pretreatment with the L channel antagonist nifedipine only slightly delayed and reduced the peak rise in [Ca++]i, but markedly reduced the plateau phase. No further reduction was observed in cells treated with 10 μM nimodipine (data not shown). In contrast, addition of the N channel blocker ω-conotoxin GVIA delayed and markedly inhibited the peak rise (213 ± 10) but did not reduce the plateau further. These findings indicate that trk-PC12 cells express functional L-type and N-type channels, and that dihydropyridine-insensitive channels are particularly important for generating the peak rise in [Ca++]i.
To examine dihydropyridine-insensitive channels, we focused the remainder of our studies on the peak rise in [Ca++]i. To reduce a small contribution from L channels to our results, we conducted subsequent experiments in the presence of 1 μM nifedipine. Incubation with 1 μM ω-conotoxin GVIA reduced the peak rise in [Ca++]i in nifedipine-treated cells by approximately 55% (fig. 1 B), confirming a major role for N-type channels in generating this response. PC12 cells express mRNA for α1ACa++ channel subunits (Starr et al., 1991), which appear to mediate P- and Q-type Ca++currents (Stea et al., 1994; Randall and Tsien, 1995). The P/Q channel antagonist agatoxin IVA reduced the peak [Ca++]i rise by approximately 6% at a concentration of 50 nM and by 38% at 300 nM (fig. 1 B). No further inhibition was observed using 1 μM ω-agatoxin IVA. These findings suggest that in addition to N-type channels, P/Q-type channels contribute to the peak rise in [Ca++]i.
PC12 cells also express T type channels, and although no high-affinity, specific antagonists are available, T channels can be inhibited by 100 μM nickel (Soong, 1993). Addition of 100 μM nickel did not significantly increase inhibition achieved by a combination of ω-conotoxin GVIA and ω-agatoxin IVA (fig. 1 B), suggesting that T type channels do not contribute to the peak rise in [Ca++]i. The combination of ω-conotoxin GVIA, ω-agatoxin IVA, and nickel blocked 91% of the response, suggesting that other, unidentified channels that are insensitive to organic Ca++ channel blockers and 100 μM nickel play a minor role in generating the peak rise in [Ca++]i in these cells.
Ethanol inhibits K+-evoked [Ca++]i rises.
In depolarized, nifedipine-treated cells, ethanol inhibited the peak [Ca++]i rise, but had little effect on the plateau phase of the response (fig.2A). Inhibition of the peak was concentration-dependent and was maximal with 50 mM ethanol (fig. 2B). The effect of ethanol was not immediate, but required at least 8 min of preincubation to develop completely (fig. 2C). After washout of ethanol, inhibition reversed slowly and only partially after 20 min (fig. 2C).
To determine whether [Ca++]i rises mediated by N-type and P/Q-type channels are inhibited by ethanol, we studied cells treated with individual Ca++ channel antagonists. If inhibition by ethanol was specific for only one channel subtype, then blockade of that channel with a selective antagonist should prevent further inhibition by ethanol. In cells treated with nifedipine alone, ethanol inhibited peak [Ca++]i rises by over 40%; addition of ω-conotoxin GVIA or ω-agatoxin IVA significantly reduced inhibition by ethanol (fig. 2D). This indicates that ethanol inhibits [Ca++]i rises mediated by both N-type and P/Q-type channels.
Ethanol does not inhibit Ca++-induced Ca++ release.
Voltage- or receptor-mediated increases in [Ca++]i stimulate Ca++ release from intracellular stores by activating ryanodine receptor/Ca++ release channels present in endoplasmic reticulum (Meissner, 1994). These channels are activated by caffeine and are inhibited by micromolar concentrations of ryanodine. Because Ca++-induced Ca++ release contributes to depolarization-induced Ca++ rises in PC12 cells (Reber and Reuter, 1991), we examined whether ethanol reduces depolarization-evoked [Ca++]i rises by inhibiting Ca++ release channels. Caffeine stimulated a peak rise in [Ca++]i that was approximately 3.5-fold the resting level (fig.3A). Preincubation with 50 mM ethanol did not alter resting [Ca++]ior inhibit the response to caffeine. Ryanodine reduced depolarization-evoked [Ca++]i rises, but 50 mM ethanol still inhibited the remaining [Ca++]i rise by 53 ± 11% (fig. 3B). This percentage of inhibition by ethanol was similar to that observed in cells depolarized without ryanodine (41 ± 4%; P = .175). These results indicate that ryanodine receptor/Ca++ release channels contribute to depolarization-induced [Ca++]i rises intrk-PC12 cells, but these channels are not inhibited by 50 mM ethanol.
Ethanol does not impair capacitative Ca++entry.
Release of Ca++ from intracellular stores stimulates influx of extracellular Ca++into the cytoplasm by a process termed capacitative Ca++ entry (Berridge, 1995). This Ca++ current is carried by Ca++ release-activated Ca++channels and sustained elevations in [Ca++]i due to calcium entry through these channels can be stimulated by treating cells with agents that deplete intracellular Ca++ stores, such as thapsigargin (Berridge, 1995). In thapsigargin-treatedtrk-PC12 cells, [Ca++]i rose slowly over 3 min to 316 ± 10 nM and remained elevated for at least 5 min thereafter (fig. 4A). Addition of 50 mM ethanol did not reduce [Ca++]i, suggesting that ethanol does not inhibit Ca++ release-activated Ca++ channels in trk-PC12 cells.
Ethanol does not alter Ca++ extrusion or Na+-Ca++ exchange.
The level of [Ca++]i is tightly regulated by mechanisms for Ca++extrusion and sequestration that serve to restore the resting level of [Ca++]i after depolarization. These mechanisms include Na+-Ca++ exchange and active Ca++ transport across the plasma membrane and into endoplasmic reticulum through the action of Ca++-ATPases (Clapham, 1995). To examine whether ethanol inhibits depolarization-induced rises in [Ca++]i by promoting Na+-Ca++ exchange we incubated cells in Na+-free buffer to prevent Ca++ efflux through Na+/Ca++ exchangers. In Na+-free buffer, resting [Ca++]i was higher (172 ± 10 nM, n = 20, P = .0001) than in buffer containing Na+ (96 ± 8 nM,n = 48), reflecting inhibition of Ca++ efflux through Na+-Ca++ exchangers in Na+-free buffer. However, depolarization increased [Ca++]i to a maximum of 449 ± 32 nM (n = 20) that was similar to that observed in cells depolarized in Na+-containing buffer (465 ± 36 nM,N = 48, P = .08). Moreover, in cells depolarized in Na+-free buffer, ethanol inhibited the rise in [Ca++]i by 37 ± 6% (n = 20), which is similar to the level of inhibition observed in cells depolarized in Na+-containing buffer (41 ± 4%,n = 45, P = .62). These findings indicate that Na+-Ca++ exchange does not modulate the peak rise in [Ca++]i during depolarization and that ethanol does not reduce depolarization-induced [Ca++]i rises by enhancing Na+-Ca++exchange.
We next examined whether ethanol stimulates other mechanisms for Ca++ sequestration or extrusion by exposing cells to the Ca++ ionophore ionomycin in Na+-free buffer containing 2 mM CaCl2 for 25 sec. The buffer was then rapidly replaced by Na+-free buffer containing 400 nM CaCl2 to reduce the Ca++gradient across the plasma membrane and decrease Ca++ influx through ionomycin ionophores. Images were recorded every 10 sec to measure the rate of decline in [Ca++]i. As shown in figure 4B, similar rates were observed in cells treated with or without ethanol. These findings suggest that ethanol does not promote Ca++ sequestration or extrusion.
Ethanol enhances [Ca++]i rises stimulated by bradykinin.
In PC12 cells, bradykinin activates B2 bradykinin receptors coupled to inositol 1,4,5-trisphosphate production leading to a rapid rise in [Ca++]i (Fasolatoet al., 1988). The response to bradykinin is biphasic with an initial peak rise due to IP3-mediated Ca++ release, and a subsequent plateau phase due to Ca++ influx through membrane channels that are insensitive to L and N channel antagonists (Fasolato et al., 1988; Sher et al., 1988). If ethanol reduces depolarization-induced [Ca++]i rises by promoting sequestration or extrusion of Ca++, then ethanol should also reduce bradykinin-stimulated [Ca++]i rises. However, in contrast to depolarization-evoked [Ca++]i rises, the peak rise induced by bradykinin was potentiated by ethanol (fig.5). This suggests that inhibition by ethanol is specific for depolarization-evoked [Ca++]i rises mediated by voltage-gated Ca++ channels. These results provide additional evidence that ethanol does not promote Ca++ sequestration or extrusion, reduce Ca++ mobilization, or inhibit other mechanisms for Ca++ entry in trk-PC12 cells.
PKA regulates inhibition of [Ca++]i rises by ethanol.
Because protein phosphorylation regulates ion channel function (Nestler and Greengard, 1984) and, in some cells, ethanol stimulates second messenger cascades that lead to activation of PKA (Nagy et al., 1989; Rabin et al., 1993) or PKC (Hoek et al., 1987; Rubin et al., 1988) we examined whether ethanol inhibits depolarization-induced [Ca++]i rises by activating these kinases. If ethanol acts by stimulating PKA, then treatment with a PKA agonist should also inhibit depolarization-induced [Ca++]i rises. However, exposure to Sp-cAMPS at a concentration (30 μM) that produces near maximal stimulation of PKA-mediated glycogenolysis in rat hepatocytes (Rothermel et al., 1983) enhanced depolarization-induced [Ca++]i rises (fig.6A). In addition, inhibition by ethanol was not prevented by treating cells with the PKA antagonist Rp-cAMPS at a concentration (30 μM) that maximally inhibits PKA activation by glucagon in hepatocytes (Rothermel et al., 1984). The PKC activator PMA (Nishizuka, 1992), like ethanol, inhibited depolarization-evoked rises in [Ca++]i (fig. 6A). However, treatment with a maximally effective-concentration of PMA did not prevent further inhibition by ethanol, indicating that PMA and ethanol have different mechanisms of action. Therefore ethanol does not appear to inhibit depolarization-induced [Ca++]i rises intrk-PC12 cells by activating PKA or PKC.
Because phosphorylation appears to regulate the sensitivity of GABAA receptor-gated ion channels to ethanol (Wafford and Whiting, 1992), we investigated whether inhibition of dihydropyridine-insensitive Ca++ channels by ethanol was regulated by phosphorylation. As shown in figure 6A, activation of PKA with Sp-cAMPS completely prevented inhibition of [Ca++]i rises by ethanol. Okadaic acid, which inhibits type-1 (PP-1) and type-2A (PP-2A) protein phosphatases (Bialojan and Takai, 1988), also prevented inhibition by ethanol, whereas cyclosporin A (fig. 6A) and deltamethrin (data not shown), which inhibit protein phosphatase type-2B (PP-2B, calcineurin) (Liu et al., 1991; Enan and Matsumura, 1992), had no effect. Half maximal inhibition was observed with 3 nM okadaic acid (fig. 6B). The PKA inhibitor Rp-cAMPS restored sensitivity to ethanol in cells treated with okadaic acid (fig. 6A), demonstrating that the effect of okadaic acid was mediated by increased PKA-mediated phosphorylation. These results indicate that activation of PKA prevents inhibition of dihydropyridine-insensitive Ca++ channels by ethanol.
Discussion
The major findings of this study are that 1) NGF-differentiatedtrk-PC12 cells express L-type, N-type and P/Q-type channels as determined by responses to selective Ca++channel antagonists, 2) ethanol can inhibit N-type and P/Q-type channels in these cells and 3) inhibition by ethanol is antagonized by PKA. Using NGF-differentiated trk-PC12 cells as a model system to study dihydropyridine-insensitive Ca++channels, we found that K+ depolarization evoked biphasic [Ca++]i rises in these cells, with a peak phase mediated mainly by N-type channels sensitive to ω-conotoxin GVIA and P/Q-type channels sensitive to ω-agatoxin IVA. The effect of ω-agatoxin IVA was seen mainly at high concentrations (>200 nM), indicating that these channels are probably Q-type (Randall and Tsien, 1995). Ethanol inhibited the peak [Ca++]i rise with maximal inhibition of approximately 45% at 50 mM ethanol. Ethanol did not act by inhibiting ryanodine receptor/Ca++ release channels or capacitative Ca++ entry, or by enhancing Ca++ sequestration or extrusion. [Ca++]i rises induced by bradykinin, which stimulates IP3 mediated Ca++ mobilization and receptor-gated Ca++ influx (Fasolato et al., 1988), were increased by ethanol, indicating that inhibition of [Ca++]i rises by ethanol is specific for depolarization-evoked Ca++ entry through voltage-gated channels. Both ω-conotoxin GVIA and ω-agatoxin IVA reduced inhibition by ethanol, demonstrating that ethanol inhibited [Ca++]irises mediated by N-type and P/Q-type channels.
This study is the first demonstration of inhibition of an ω-agatoxin IVA-sensitive channel by ethanol. Using nerve terminals from rat neurohypophysis, Wang et al. (1991) demonstrated inhibition of ω-conotoxin VIA-sensitive N-type channels by ethanol. However, investigators from the same laboratory found that N-type channels in NGF-differentiated PC12 cells are not inhibited by 50 mM ethanol (Mullikin-Kilpatrick and Treistman, 1995). The discrepancy between their findings and ours may reflect the use of different cell lines and techniques. One key difference may lie in the use of Ba++ rather than Ca++ as the charge carrier. In the earlier study of N-type channels in neurohypophyseal terminals, Ca++ was the charge carrier (Wang et al., 1991), whereas in the study using PC12 cells, Ba++ was used (Mullikin-Kilpatrick and Treistman, 1995). Although ethanol inhibits both Ca++ and Ba++ influx through L-type channels in PC12 cells (Mullikin-Kilpatrick and Treistman, 1994), inhibition of N-type channels by ethanol could be specific for Ca++. A similar result was observed with AMPA/kainate receptors expressed in Xenopus oocytes, where inhibition by ethanol was reduced when Ba++was substituted for extracellular Ca++(Dildy-Mayfield and Harris, 1995). If true, this would suggest that inhibition of dihydropyridine-insensitive channels by ethanol involves Ca++-activated processes such as phosphorylation by Ca++-sensitive kinases or Ca++-mediated channel inactivation.
We found that inhibition of depolarization-induced [Ca++]i rises by ethanol required several min of ethanol exposure to become maximal and was long-lasting, only partly reversing 20 min after washout of ethanol. The slow onset and persistence of the effect may reflect the action of metabolic events such as phosphorylation that regulate channel function. Subunits of N and P/Q channels are phosphorylated by PKA and PKC (Hell et al., 1994; Sakurai et al., 1995). PKA activation stimulates N channels in rat nodose ganglion cells (Gross et al., 1990) but inhibits N channels in mouse dorsal root ganglion cells (Gross and MacDonald, 1989) and rat neostriatal neurons (Surmeier et al., 1995). Similarly, PKC activation increases N channel activity in superior cervical ganglion cells (Bernheim et al., 1991), rat hippocampal CA3 and cortical pyramidal cells (Swartz, 1993; Swartz et al., 1993) and several peripheral neurons (Swartz, 1993), but decreases N channel function in freshly dissociated hippocampal neurons (Doerner et al., 1990). PKC potentiates Q channel-mediated synaptic transmission at CA3-CA1 synapses in rodent hippocampus (Wheeleret al., 1994) and activation of PKC or PKA enhances Q-type currents (Randall and Tsien, 1995) expressed by Xenopusoocytes injected with mRNA from cerebellar granule cells (Fournieret al., 1993). However, in intact cerebellar granule cells, activation of PKA slightly decreases non-L- and non-N-type current, half of which is mediated by Q-type channels (Randall and Tsien, 1995). Therefore, phosphorylation appears to regulate N-type and P/Q-type channels, but whether it increases or decreases function may depend on cell-specific splice variants of channel subunits or other proteins that modulate channel function.
Because phosphorylation regulates Ca++ channels, we considered whether ethanol inhibits channel function by altering PKC or PKA activity. We found that treatment with the cAMP analog Sp-cAMPS enhanced depolarization-induced [Ca++]i rises whereas the PKA inhibitor Rp-cAMPS had no effect, indicating that ethanol could not inhibit [Ca++]i rises by altering PKA activity. Moreover, although activation of PKC with PMA inhibited depolarization-evoked rises in [Ca++]i, the actions of PMA and ethanol were additive, indicating different mechanisms of action. These findings demonstrate that ethanol does not inhibit depolarization-induced [Ca++]i rises by modulating PKA or PKC activity. Whether ethanol acts by regulating other kinases requires further investigation.
Brief exposure to 100 to 150 mM ethanol for 10 min has been reported to increase cAMP levels by 9 pmol/106 cells in NG108–15 cells (Nagy et al., 1989) and by 3 pmol/mg protein in PC12 cells (Rabin et al., 1993). Because ethanol increases cAMP levels and Sp-cAMPS prevented inhibition of [Ca++]i rises by ethanol, it seems paradoxical that ethanol inhibited the response to depolarization. One might expect that ethanol-induced increases in cAMP would antagonize the inhibitory effect of ethanol on depolarization-induced [Ca++]i rises. However, ethanol-induced increases in cAMP are extremely small and are only about 1% of increases stimulated by forskolin (10 μM) or the adenosine agonist 2-chloroadenosine (1 μM) in PC12 cells (Rabinet al., 1993). Moreover, ethanol does not increase cAMP levels in N1E-115 neuroblastoma cells (Stenstrom and Richelson, 1982), some clones of PC12 cells (Rabe et al., 1990), or in all studies involving NG108–15 cells (Gordon et al., 1986). Because Rp-cAMPS had no effect on [Ca++]i rises in ethanol-treated trk-PC12 cells (fig. 6), ethanol may not increase cAMP levels in these cells, or if it does, the increase is too small to alter the response to depolarization or prevent inhibition of [Ca++]i rises by ethanol
In addition to regulating ion channel function, phosphorylation can alter the sensitivity of certain membrane proteins to ethanol. For example, in NG108–15 cells, ethanol-sensitive adenosine transporters require PKA activation to be inhibited by ethanol (Coe et al., 1996), whereas mouse and bovine GABAAreceptors appear to require PKC phosphorylation of a splice-variant long form of the γ2 subunit for enhancement of receptor function by ethanol (Wafford and Whiting, 1992). Our findings suggest that the sensitivity of non-L-type Ca++ channels to ethanol may also be modulated by phosphorylation since we found that activation of PKA with the selective agonist Sp-cAMPS completely prevented inhibition of depolarization-evoked [Ca++]i rises by ethanol.
The phosphatase inhibitor okadaic acid also prevented inhibition by ethanol and this was reversed by treatment with the specific PKA antagonist Rp-cAMPS, indicating that the effect of okadaic acid was due to preservation of PKA-phosphorylated sites on proteins that regulate channel function. Unlike okadaic acid, which inhibits PP-1 and PP-2A, inhibitors of PP-2B (calcineurin), did not alter inhibition by ethanol. Okadaic acid reduced the effect of ethanol with an IC50 value of approximately 3 nM, close to its IC50 value for inhibition of PP2-A (1 nM) (Bialojan and Takai, 1988). These findings suggest that inhibition of non-L-type channels by ethanol is regulated by a phosphoprotein that is a substrate for PKA and PP-2A. Further work is needed to identify the phosphoproteins that interact with ethanol to regulate channel function.
N-type and P/Q-type channels are expressed in dendrites and presynaptic terminals (Westenbroek et al., 1992, 1995) and regulate synaptic transmission at several synapses (Takahashi and Momiyama, 1993; Wheeler et al., 1994). Thus, inhibition of N-type and P/Q-type channels by intoxicating concentrations of ethanol could depress neurotransmitter release in several brain regions. Such inhibition could underlie inhibition of electrically stimulated release of acetylcholine and other neurotransmitters by ethanol in rat neocortex (Carmichael and Israel, 1975) and contribute to sedation and cognitive impairment observed during ethanol intoxication (Messing and Diamond, 1997). In addition, mutations in the gene encoding the α1A subunit of P/Q-type channels are linked to the tottering and leaner phenotypes in mice (Fletcher et al., 1996) and to familial hemiplegic migraine, episodic ataxia type 2 and spinocerebellar ataxia 6 in humans (Ophoff et al., 1996; Zhuchenko et al., 1997). Ataxia is a common feature of these disorders, supporting a role for ethanol-induced inhibition of P/Q-type channels in the induction of ataxia. Further studies of brain N-type and P/Q-type channels will be required to establish whether ethanol inhibits these channels in multiple regions of the mammalian central nervous system.
Acknowledgments
The authors thank S. Finkbeiner, J. Lansman and A. Gordon for helpful discussions, and I. Diamond for critical reading of this manuscript.
Footnotes
-
Send reprint requests to: Dr. Robert O. Messing, Building 1, Room 101, 1001 Potrero Avenue, San Francisco, CA 94110.
-
↵1 This work was supported by grants from the National Institute on Alcohol Abuse and Alcoholism and the Alcoholic Beverage Medical Research Foundation (R.O.M.).
- Abbreviations:
- PKA
- protein kinase A
- PKC
- protein kinase C
- GABA
- γ-aminobutyric acid
- AMPA
- α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid
- Sp-cAMPS
- Sp-adenosine cyclic 3′, 5′-phosphorothioate
- Rp-cAMPS
- Rp-adenosine cyclic 3′, 5′-phosphorothioate
- DMEM
- Dulbecco’s modified Eagle’s medium
- cAMP
- cyclic adenosine monophosphate
- EGTA
- ethylene glycol bis(β-aminoethyl ester)-N, N′-tetraacetic acid
- HEPES
- N-[2-hydroxyethyl]piperazine-N′-[2-ethanesulfonic acid]
- [Ca2+]i
- intracellular calcium concentration
- PP-1 protein phosphatase type-1
- PP-2A, protein phosphatase type-2A
- PP-2B
- protein phosphatase type-2B
- Received March 4, 1997.
- Accepted May 23, 1997.
- The American Society for Pharmacology and Experimental Therapeutics