Introduction

Top Abstract Introduction Methods Results Discussion References

For some time, we have been interested in the ability of growth factors, such as EGF, to modulate the contractile activity of a variety of smooth muscle systems (Berk et al., 1985, 1986; Berk and Alexander, 1989; Muramatsu et al., 1985; 1988; Hollenberg et al., 1989; Hollenberg, 1994, 1994a). In gastric LM the contractile action of EGF, which can be blocked by indomethacin, is caused by the diacylglycerol lipase-catalyzed release from diacylglycerol of a presumed arachidonate precursor that yields a contractile cyclooxygenase product (Yang et al., 1991). Thus, EGF acts in a sense indirectly via a cyclooxygenase product to cause its contractile effect. Because diacylglycerol might be produced via a concurrent phospholipase D/phosphatidate phosphatase reaction, we wondered if phospholipase D might play a role in the contractile action of EGF; and we reasoned that the addition of ethanol to the LM tissue might, via the formation of phosphatidyl ethanol, attenuate the contractile action of EGF by reducing the conversion of phosphatidic acid to diacylglycerol. It was, therefore, our working hypothesis that ethanol either alone or in combination with EGF might modulate gastric smooth muscle contractility. Somewhat to our surprise, in preliminary experiments we observed that ethanol on its own (20-500 mM), in a guinea pig gastric LM preparation, caused a contractile response that appeared to parallel the responsiveness of the tissue to EGF. Further, like the EGF response, the contractile effect of ethanol was blocked by the tyrosine kinase inhibitor, genistein. In view of these preliminary observations, the primary objectives of the work we describe in this report were: 1) to use a variety of signal pathway inhibitor probes for a comparative study of the contractile actions of EGF and ethanol in the guinea pig gastric LM preparation and 2) to compare the effects of ethanol with the potencies of other alcohols. A comparison was also made between the ethanol-mediated response of the gastric LM preparation and the ethanol-mediated contractile response of the pharmacologically distinct CM preparation obtained from the same gastric tissue. The data for the CM tissue were obtained to provide for a contrast between the two preparations coming from the same organ.

	Methods

Top Abstract Introduction Methods Results Discussion References

Tissue preparations for bioassay. The guinea pig gastric LM and CM preparations were obtained as described previously (Muramatsu et al., 1988; Hollenberg et al., 1989) from male albino guinea pigs (about 350 g) cared for in accordance with the Canadian Council on Animal Care. After sacrifice, the animals were exsanguinated from the common carotid arteries, and stomach tissue was isolated. The LM and CM strips (3 × 10 mm) were prepared by cutting the gastric tissue (middle two thirds of stomach) either along or at a right angle to the LM muscle bundles. This procedure permits a measurement of the contraction of either the LM or CM elements in strips derived from the same tissue. Each strip was mounted vertically in a disposable plastic cuvette containing 4 ml of Krebs-Henseleit buffer of the following composition (mM): NaCl, 118; KCl, 4.7; CaCl₂, 2.5; MgCl₂, 1.2; NaHCO₃, 25; KH₂PO₄, 1.2; and glucose, 10 in distilled deionized water. The bath medium was maintained at 37°C and was gassed with 95% O₂/5% CO₂ to maintain the pH at 7.4. A resting tension of 1.0 g was applied initially to the tissue, which was allowed to equilibrate for about 1 h. Contractile responses were recorded isometrically with Statham (UTC 2) or Grass force-displacement transducers. Routinely, the responsiveness of each tissue was assessed by exposure to either 50 mM KCl (1.8 ± 0.3 g tension, average ± S.E.M. for n = 10) or 1 µM carbachol. For the construction of concentration-response curves for ethanol and other alcohols, contractile responses were expressed as a percentage (% KCl) of the contractile response to 50 mM KCl. Data in the figures were expressed as mean values ± S.E.M. for the number of determinations recorded in the figure legends. Differences between mean values were assessed for significance using a Student's t test.

Western blot analysis. Tissue strips to be used for Western blot assay were prepared exactly as for the LM bioassay and were exposed to contractile agonists (EGF, 17 nM; EtOH, 170 mM) for a time corresponding to the peak of tissue contraction (from 1 to 5 min). Tissue was frozen immediately on a solid CO₂-cooled plexiglass plate, chopped with a scalpel and immediately solubilized in immunoprecipitation buffer (1% v/v NP40, in Tris-HCl, pH 7.4, containing 1 mM sodium orthrovanadate and 1 µM each of the protease inhibitors phenylmethylsulfonyl fluoride and leupeptin). Tissue extracts were clarified by centrifugation at 15,000 × g for 20 min at 4°C. With the use of tissue extracts containing the same amount of protein (Folin reagent assay), phosphotyrosyl proteins were harvested (overnight at 4°C) with Sepharose bead-coupled monoclonal antiphosphotyrosine antibody (6D9) (50-60 µg/ml packed beads) that had been prepared according to Glenney et al. (1988). Bead-bound protein was washed three times with this buffer by centrifugation, and protein in the bead pellet was solubilized in 30 µl of boiling sample buffer (Laemmli, 1971) in preparation for polyacrylamide gel electrophoresis (80 mm × 50 mm × 1.5 mm, 10% gel) and transfer to nitrocellulose (0.45 µm; BioRad, Richmond, CA) for Western blot detection of protein. Phosphotyrosyl proteins were detected by use of horseradish perioxidase-coupled monoclonal antiphosphotyrosine antibody (6D9), along with chemiluminescence (ECL) detection (Amersham, Oakville, Ontario, Canada). Molecular weight markers were from BioRad (Mississauga, Ontario, Canada).

Chemicals and other reagents. Ethanol (reagent grade) was from Commercial Alcohols Inc. (Brampton, ON); reagent grade methanol, 1-propanol and butanol were from Fisher Scientific (Fair Lawn, NJ). Human epidermal growth factor and human TGF- were from Upstate Biotechnology Inc. (Lake Placid, NY); mepacrine, indomethacin, carbachol, ketoconazole, nifedipine, 4-methylpyrazole, atropine, nordihydroguaiaretic acid, prazosin and yohimbine were from Sigma Chemical Co. (St. Louis, MO). Genistein was from ICN (Costa Mesa, CA); tyrphostin-47 (also designated AG213) was from Calbiochem (La Jolla, CA). U57, 908 was obtained from Upjohn (Kalamazoo, MI). PD153035 was from Parke Davis (Ann Arbor, MI), as was PD98059, obtained with the assistance of Dr. A. Saltiel. GF109203X, LY294002 and chelerythrine were from Biomol (Plymouth Meeting, PA). Pervanadate was prepared as described previously (Kadota et al., 1987) by mixing stock solutions of sodium orthovanadate with an equimolar or molar excess of H₂O₂. After a 15-min period, the reaction was terminated by the addition of catalase (400 U/ml) to metabolize unreacted H₂O₂.

	Results

Top Abstract Introduction Methods Results Discussion References

Contractile action of ethanol in the LM and CM preparations: effects of tyrosine kinase inhibitors, indomethacin and inhibitors of nerve-released agonists. In our initial work, we assessed the action of ethanol in both the LM and CM tissues. Ethanol caused contractile responses in both the LM (fig. 1, left-hand tracings A-C) and CM (fig. 1, right-hand tracings G-I) preparations. The ethanol-induced contractions did not desensitize and were reproducible during a 5- to 6-h period (not shown). Contractions caused by 170 mM ethanol were equivalent to those caused by concentrations of EGF or TGF- (17 nM) that were at the plateau of the EGF/TGF- concentration effect curves (not shown; Hollenberg et al., 1989). TGF- was used for comparison in the CM preparation, because its contractile action is less desensitizing than that of EGF (Hollenberg et al., 1989). The contractile actions of ethanol in the LM and CM preparations were unaffected by 1 µM of the following agents: tetrodotoxin, atropine, prazosin and yohimbine (not shown). As for EGF (fig. 1, tracings E and F), the contractile action of ethanol in the LM preparation was blocked both by the tyrosine kinase inhibitor, genistein, and by the cyclooxygenase inhibitor, indomethacin (fig. 1, tracings B and C; fig. 2). The tyrosine kinase inhibitor, tyrphostin-47 also blocked ethanol-induced contractions in the LM preparation (fig. 2 and data not shown). Genistein and tyrphostin-47 also attenuated ethanol-induced contractions in the CM preparation, albeit less so than in the LM preparation (fig. 1, tracing H and fig. 2). The two tyrosine kinase inhibitors did not affect contractions caused by carbachol in the LM and CM tissues (not shown). The concentration-effect curves for the ability of genistein and tyrphostin-47 to inhibit contractions caused by ethanol (170 mM) in the LM and CM preparations are shown in figure 2. Although a greater than 90% inhibition of the contractile response by both tyrosine kinase inhibitors was possible in the LM preparation, the ethanol-induced contractile response in the CM preparation appeared less affected by genistein and tyrphostin (60% inhibition) (fig. 2). In contrast with the inhibitory action of indomethacin in the LM tissue, a robust contractile response to both ethanol and TGF- was observed in the presence of indomethacin in the CM tissues (fig. 1, tracings I and L). The epoxygenase inhibitor, ketoconazole (5 µM), and the lipoxygenase inhibitor, nordihydroguaiaretic acid (30 µM), had no effect on ethanol-induced contractions in the LM and CM preparations (not shown). To rule out any contribution of cyclooxygenase products to the contractile response of the CM preparation, all further experiments with this tissue were done in the presence of 3 µM indomethacin.

View larger version (K): [in this window] [in a new window]   Fig. 1.   Contractile actions of ethanol and EGF or TGF- in gastric LM and CM strips: effects of genistein (GS, ) and indomethacin (INDO, ). Either LM (left-hand panel) or CM (right-hand panel) strips were first exposed to ethanol (A, B, C, G, H, I: , 170 mM), or EGF (D, E, F: , 17 nM) or TGF- (J, K, L: , 17 nM) to measure a control contractile response, followed by washing the tissues (W, arrows). The preparations were again challenged with either ethanol (), EGF (), or TGF- () either without (A, D, G, J) or after a 20-min pretreatment with either genistein (B, E, H, K: , 8 µM) or indomethacin (C, F, I, L: , 3 µM). The scale for time and tension is shown beside tracing B; each tracing (A to L) shows the responses to a single tissue strip, as indicated by the breaks (//) in the recorded trace. The data in each tracing are representative of experiments done with four to six individual tissue strips taken from three or more separate animals.

View larger version (K): [in this window] [in a new window]   Fig. 2.   Inhibition of ethanol-stimulated contractions in LM (, ) and CM (, ) tissues by either genistein (GS, , ) or tyrphostin-47 (, ): concentration-effect curves. A control contractile response was first monitored by exposing each tissue to ethanol (170 mM) followed by washing. A second response to ethanol was then measured after incubating tissues for 20 min with increasing concentrations of either genistein (GS, , ) or tyrphostin-47 (TP, , ). The contractile response in the presence of each concentration of inhibitor was expressed as a percentage (% control) of the contractile response observed before the addition of either genistein or tyrphostin-47. Data points represent the means ± S.E.M. (bars) for observations made with three to six individual tissue strips taken from two or more different animals.

Actions of other alcohols and concentration-effect curves. The ethanol-induced responses were not affected by 4-methylpyrazol (50 µM: Li and Theorell, 1969), which indicated that the metabolism of ethanol by alcohol dehydrogenase did not play a role in the contractile effect and suggested that other alcohols might be active. Like ethanol, methanol and propanol also caused contractile responses in both the LM and CM tissues (fig. 3). In contrast with these alcohols, butanol (50-150 mM) caused a relaxation of the LM and CM tissue rather than a contraction (not shown). In the LM preparation, the order of alcohol potencies as estimated from the concentrations causing a contractile response 50% of maximum was: ethanol = propanol > methanol. In the CM preparation, the potency order was propanol > ethanol > methanol (fig. 3). In both the LM and CM preparations, the maximum contraction in response to ethanol and methanol was comparable at the plateau of their respective concentration-response curves. However, propanol, at the plateau of its concentration-response curve, caused a much smaller contractile response than that caused by either ethanol or methanol (fig. 3). All of the continuing work was focused only on the action of ethanol in the LM and CM preparations.

View larger version (K): [in this window] [in a new window]   Fig. 3.   Contractile responses of LM (upper) and CM (lower) muscle strips to methanol (), ethanol () and propanol (): concentration-effect curves. The responsiveness of each tissue strip was first monitored by exposure to 50 mM KCl, followed by washing. Tissues were then exposed to increasing concentrations of methanol () ethanol () or propanol () followed by washing. The contractile responses to the alcohols were expressed as a percentage (% KCl) of each tissue's response to 50 mM KCl. Each data point represents the mean ± S.E.M. for four to six measurements made with individual tissue strips taken from two to four different animals.

Role of extracellular calcium. In the absence of extracellular calcium, ethanol failed to cause a contractile response in either the LM or CM preparations; replenishing the medium with calcium in the continued presence of ethanol resulted in a contraction (fig. 4). An antagonist of the voltage-sensitive calcium channel, nifedipine (1 µM), as expected, completely inhibited contractions caused by depolarizing the LM tissue with 50 mM KCl (fig. 5). This concentration of nifedipine also markedly attenuated (90 ± 5% inhibition, average ± S.E.M. for n = 6) the contractile action of EGF in the LM assay (fig. 5). In contrast, 1 µM nifedipine had only a modest inhibitory effect (30 ± 10% inhibition, average ± S.E.M. for n = 10) on ethanol-induced contractions; upon adding the receptor-operated calcium channel blocker, SKF96365 (30 µM) to a nifedipine-treated LM preparation, it was possible to block completely the ethanol-induced contraction (fig. 5). SKF96365 alone did not completely block the contractile response (not shown). Similar effects of nifedipine and SKF96365 were observed for ethanol in the CM preparation (not shown).

View larger version (K): [in this window] [in a new window]   Fig. 4.   Role of extracellular calcium. Either longitudinal (A, upper) or circular (B, lower) muscle strips were first exposed to ethanol (170 mM) followed by washing (W, arrow) and preincubation for 20 min in a calcium-free Krebs-Henseleit buffer containing 0.2 mM ethyleneglycol-bis(-aminoethyl ether)-N,N,N,N-tetraacetic acid. Tissues were again challenged with ethanol, followed by replenishing the buffer with 2.5 mM CaCl2 (+). The scale for time and tension is shown to the right of tracing B. The data are representative of experiments done with three or more tissue strips taken from at least two different animals.

View larger version (K): [in this window] [in a new window]   Fig. 5.   Role of calcium influx in the contractile actions of ethanol and EGF in longitudinal muscle strips: effects of nifedipine (NIF) and SKF96365 (SKF). Control contractile responses in individual tissue strips were first monitored in response to KCl (50 mM), EGF (17 nM) or ethanol, followed by washing each tissue and preincubation for 20 min with either nifedipine (1 µM, hatched bars) or with nifedipine (1 µM) combined with 30 µM SKF96365 (solid bar), followed by rechallenging the tissues with KCl, EGF and EtOH. Data represent observations with 6 to 10 individual tissue strips taken from three or more different animals. *Response of nifedipine-treated tissues smaller than control (P < .05).

Effects of U57,908 and mepacrine. The diacylglycerol lipase inhibitor, U57,908, which at a concentration of 20 µM effectively and selectively inhibits diacylglycerol lipase activity in the guinea pig gastric smooth muscle tissue without affecting either diacylglycerol kinase or phospholipase A₂ activity (Yang et al., 1991), was able to inhibit (85 ± 5%, average ± S.E.M. for n = 6) ethanol-induced contractions in the LM preparation (fig. 6, tracing A and lower panel). The same concentration of U57,908 did not significantly inhibit ethanol-induced contractions in the CM preparation (fig. 6, tracing C and lower panel). The phospholipase A₂ inhibitor, mepacrine (3 µM), which was found previously to attentuate angiotensin-II-induced contractions in the gastric preparations (Yang et al., 1993), had no effect on ethanol-induced contractions in either the LM or CM preparation (fig. 6, tracings B and D).

View larger version (K): [in this window] [in a new window]   Fig. 6.   Effects of inhibitors of phospholipase A2 and diacylglycerol lipase on ethanol-induced contractions in LM and CM tissue strips. Upper: a control response to ethanol (, 170 mM) was first monitored in either LM (A, B: left-hand panel) or CM (C, D: right-hand panel) strips followed by washing and preincubation of each tissue for 20 min with either the diacylglycerol lipase inhibitor, U57,908 (, 20 µM: A, C), or the phospholipase A2 inhibitor, mepacrine (MP , 3 µm: B, D). Tissues were again challenged with ethanol in the continued presence of each inhibitor. The tracings are representative of three or more experiments with individual tissue strips taken from two or more different animals, as summarized (lower) by the histograms below the tracings, wherein the inhibitory action of U57,908 in the LM preparation is compared with the lack of inhibition in the CM.

Potential roles of protein kinase C, phosphatidylinositol 3-kinase and MEK. In the LM preparation, it was possible to monitor a reproducible contraction caused by the kinase C activator, PDBu (fig. 7, tracing A). The kinase C antagonist, chelerythrine, was not able to block PDBu-induced contractions completely (not shown) and we therefore turned to the use of the kinase C antagonist GF which at a concentration of 1 µM was able to abolish completely PDBu-induced contractions (fig. 7, tracing A). At this concentration of GF, contractions elicited by EGF in the LM preparation were attenuated (70 ± 19% inhibition: average ± S.E.M. for n = 6), whereas contractions caused by ethanol were unaffected (fig. 7, tracings B and C).

View larger version (K): [in this window] [in a new window]   Fig. 7.   Differential effect of inhibiting kinase C on the contractile responses in LM tissue caused by EGF. A control LM contractile response was first monitored for PDBu (, 0.1 µM: tracing A), EGF (, 17 nM: tracing B) and ethanol (, 170 mM: tracing C) followed by a tissue wash (W, arrow). All tissues were then preincubated for 20 min with kinase C inhibitor, GF (, 1 µM) and were rechallenged with PDBu (), EGF () and ethanol (). The tracings are representative of experiments done with six independent tissue strips coming from two separate animals. The scale for time and tension is shown to the right of tracing C.

View larger version (K): [in this window] [in a new window]   Fig. 8.   Effects of inhibitors of PI-3-kinase and MEK on longitudinal strip contractions caused by EGF and ethanol. After monitoring a control contractile response to either ethanol (, 170 mM: tracings A and C) or EGF (, 17 nM: tracings B and D), followed by washing (W, arrows), the tissues were preincubated for 20 min with either the PI-3-kinase inhibitor, Wortmannin (WMN, *, 0.1 µM), or the MEK inhibitor, PD98059 (, 1 µM) and were then rechallenged with either ethanol (tracings A and C) or EGF (tracings B and D). The scale for time and tension is shown to the right of tracing C. The tracings are representative of experiments done with 3 to 12 independent tissue strips.

Role of EGF receptor kinase activation in the contractile action of ethanol. Because ethanol, depending on its concentration, had been shown to stimulate or inhibit the kinase activity of the EGF receptor in A431 membranes (Thurston et al., 1992), one possibility that had to be considered was that the contractile activity of ethanol in the gastric preparations might be caused by trans-activation of the EGF receptor. To evaluate this possibility we made use of the potent selective EGF receptor kinase inhibitor, PD153035 (Fry et al., 1994). This inhibitor at 1 µM completely abolished the contractile action of EGF in the LM and of TGF- in the CM preparation, but had no effect on the contractile action of ethanol in these tissues (fig. 9 and data not shown). The effects of PD153035, along with the actions of all of the agents used to probe the actions of ethanol and EGF in the LM and CM tissues, are summarized in table 1.

View larger version (K): [in this window] [in a new window]   Fig. 9.   Cross-activation of the EGF receptor does not account for ethanol-induced longitudinal muscle contractions. After monitoring control contractions caused by either ethanol () or EGF () in the same tissue strip, the preparation was washed (W, arrow) and preincubated for 20 min with the EGF receptor kinase inhibitor, PD153035 (, 1 µM). The tissue was then challenged again sequentially with EGF (, 17 nM) and then ethanol (, 170 mM). The tracing is representative of three independently conducted experiments. The scale for time and tension is shown on the right.

Tyrosine phosphatase and phosphotyrosyl proteins. In keeping with the ability of genistein and tyrphostin-47 to inhibit the contractile action of ethanol, the tyrosine phosphatase inhibitor, pervanadate (1 µM), which did not cause a contractile response on its own at this concentration, potentiated the contractile action of ethanol in the LM preparation (fig. 10, upper tracing). The potentiation, which resulted in leftward shift of the ethanol concentration-effect curve, was most prominent at low ethanol concentrations (fig. 10, lower panel); the maximal contractile action of ethanol was not enhanced by pervanadate. To determine whether the contractile action of ethanol was accompanied by changes in the phosphotyrosyl protein content of the tissue, LM preparations were collected from the organ bath between 1 and 5 min after exposure to either ethanol or EGF during the development of muscle tension and were extracted for Western blot analysis. A small, but reproducible increase in tyrosine phosphorylation of several proteins was detected (constituents A to E, fig. 11). The protein(s) for which tyrosine phosphorylation was increased appeared to differ for the EGF and ethanol-induced responses. For instance, bands A, B, C, D and E were increased by treatment with EGF; but only bands B, C and E appeared to be increased by ethanol. There was also a marked EGF-mediated increase for a constituent that barely entered the separating gel; ethanol did not appear to increase the phosphorylation of this constituent. Densitometry of the bands showed that there was an approximately 2.2-fold increase for constituent D for EGF; but no increase for ethanol (fig. 11). On the other hand, both EGF and ethanol caused a reproducible increase in the tyrosine phosphorylation of constituent B (about 2.2-fold for EGF; 1.5-fold by densitometry for ethanol: fig. 11, lanes 2 and 3). EGF also caused an increase in the phosphorylation (1.5-fold by densitometry) of a constituent (position A, fig. 11) migrating at a region that might be anticipated for the EGF receptor (160-180 kdaltons).

View larger version (K): [in this window] [in a new window]   Fig. 10.   Potentiation of ethanol-induced contractions in longitudinal muscle by pervanadate. Upper: the responsiveness of the tissue to a comparatively low concentration of ethanol (, 34 mM) was first monitored followed by washing (W, arrow). After reequilibration, the tissue was first exposed to a noncontractile concentration of pervanadate (PV, , 1 µM) followed in 5 min by the addition of the previously monitored concentration of ethanol (, 34 mM). Lower: the effect on the ethanol concentration-effect curve caused by prior exposure of the tissue to 1 µM pervanadate as shown in the upper tracing was evaluated, as outlined in the legend to figure 3, for preparations exposed to ethanol either in the absence () or presence () of 1 µM pervanadate. Data points represent the mean ± S.E.M. (bars) for measurements done with three to six individual tissue strips. The scale for time and tension is shown beside the top tracing. *Pervanadate-treated, significantly greater than control (P < .05).

View larger version (K): [in this window] [in a new window]   Fig. 11.   Stimulation of protein tyrosine phosphorylation by EGF and ethanol in longitudinal muscle strips. Tissues mounted in an organ bath as for a bioassay were either untreated (control) or were exposed for 1 min to either EGF (17 nM) or ethanol (170 mM). During the course of tension development, tissues were harvested, quick-frozen and prepared for immunoabsorption and Western blot analysis as outlined under "Methods." Equal amounts of protein extract were processed for each gel lane; the luminescent bands observed were eliminated by preincubation of the antiphosphotyrosine antibody with 25 mM p-nitrophenyl phosphate (not shown). Molecular mass (KD) marker positions are shown on the left; constituents (A to E), for which increased luminescence was observed for either EGF (lane 2) or ethanol-treated (lane 3) tissues, compared with control tissues (lane 1) are denoted (right) with a dot. The increase in phosphorylation of a constituent that may represent the EGF receptor is observed at position A in lane 2. Migration toward the anode  is indicated by the arrow.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The main finding of our study was that ethanol caused a contractile response in guinea pig gastric LM and CM preparations that in many ways reflected the contraction caused in these tissues by EGF, and that pointed to a role for a tyrosine kinase pathway for the action of ethanol in the LM and CM tissue. Differences in the sensitivity of EGF-mediated contractions to several inhibitors (e.g., genistein, indomethacin, U57,908) in the distinct LM and CM preparations were also paralleled by differences in the sensitivity of ethanol-induced contractions to the same inhibitors. The actions of ethanol were mimicked by other alcohols (methanol, propanol) and were not caused by the metabolism of ethanol by alcohol dehydrogenase. The concentration range over which ethanol caused its contractile effect (20-500 mM) is at a level that could be achieved either in gastric tissues or in blood during in the course of moderate ethanol consumption by humans (Wilkinson, 1980). Although the order of alcohol potencies in the CM tissue (propanol > ethanol > methanol) reflected the oil/water partition coefficients for these alcohols, the potency order in the LM preparation (ethanol = propanol > methanol) did not correspond to lipid solubility. Further, the anomalous actions of butanol (relaxation instead of contraction) suggest that the alcohols may affect the tissues via multiple mechanisms, only some of which may be related to the alcohol carbon chain length.

Contractions elicited by either EGF or ethanol required extracellular calcium and were blocked by the tyrosine kinase inhibitors, genistein and tyrphostin-47. Further, in the LM preparation, both ethanol and EGF-induced contractions were blocked by the cyclooxygenase inhibitor, indomethacin, and by the diacylglycerol lipase inhibitor, U57,908; whereas in the CM preparation, the contractile actions of EGF and ethanol persisted in the presence of these two enzyme inhibitors. These similarities between the actions of ethanol and EGF could not have been caused by the trans-activation of the EGF receptor by ethanol (Thurston et al., 1992), because the potent and selective EGF receptor kinase inhibitor, PD153035 (Fry et al., 1994), did not affect ethanol-induced contractions (fig. 9). In a number of ways, the contractile action of ethanol in the LM preparation (blocked by indomethacin, genistein and U57,908) also reflected the contractile action of the G-protein-coupled agonist, angiotensin-II, which is blocked by the same set of enzyme inhibitors in this tissue (Yang et al., 1993); but the action of ethanol in the LM preparation was quite distinct from the effect of other G-protein-coupled agonists, such as carbachol, which was not blocked by indomethacin, genistein or U57,908 in the gastric tissue (Yang et al., 1991, 1992). Like EGF (Yang et al., 1991) and angiotensin-II (Yang et al., 1993), the substrate (presumed to be arachidonate) released by ethanol action that yields a contractile metabolite via cyclooxygenase, would appear to arise from the metabolism of diacylglycerol by diacylglycerol lipase, rather than via the activation of phospholipase A₂ (fig. 6). Unfortunately, it is not yet possible to identify the cyclooxygenase product responsible for the contractile responses caused by EGF, angiotensin-II and ethanol.

A role for a tyrosine kinase pathway in the contractile action of ethanol was supported by three lines of evidence. First, the contractile effect in both the LM and CM preparations was blocked by two structurally different tyrosine kinase inhibitors that act via different mechanisms (Akiyama et al., 1987; Levitzki, 1992; Levitzki and Gazit, 1995). The advantages and liabilities of using these tyrosine kinase inhibitors to assess a role for tyrosine kinase pathways in biological processes are well appreciated by us (Laniyonu et al., 1994; Wolbring et al., 1994) and by others (Levitzki, 1992; Young et al., 1993). We have documented elsewhere the ability of the inhibitors to block tyrosine phosphorylation and tyrosine kinase activity in extracts of smooth muscle tissues (Yang et al., 1992; Laniyonu et al., 1994). Nonetheless, because of the ability of genistein and tyrphostin to affect enzymes other than tyrosine kinases (Young et al., 1993), the data obtained with these reagents must be interpreted with caution. A second result pointing to a role for a tyrosine kinase pathway was that the tyrosine phosphatase inhibitor, pervanadate, at concentrations that did not alone affect muscle tension, potentiated the contractile action of ethanol in the LM preparation (fig. 10). Finally, a contractile concentration of ethanol on its own was able to increase the phosphotyrosyl protein content of the gastric tissue (fig. 11). The identities of the phosphotyrosyl proteins involved in this process remain to be determined.

In tissues such as liver and platelets, phospholipase C has been identified as a target for ethanol, as indicated by ethanol-induced increases in phosphoinositide turnover, increases in intracellular calcium and the breakdown of phosphatidyl choline (Pittner and Fain, 1992; Rubin and Hoek, 1998a, b; Hoek et al., 1987, 1992). These effects of ethanol on phospholipase C activation have been attributed to a modulation by ethanol of G-protein function (Hoek et al., 1992; Rubin and Hoek, 1988a, b). In part, this ability of ethanol to activate either phosphatidylcholine or phosphoinositide-specific phospholipase C directly may account for some of our data. Whether phospholipase D plays a role in the contractile response still remains an open question. However, in view of our results implicating a tyrosine kinase pathway in the contractile action of ethanol in gastric tissue, and in keeping with recent data pointing to the ability of the G-protein -subunit to regulate cell function via an as yet unidentified tyrosine kinase pathway (Touhara et al., 1995), an alternative hypothesis to consider, apart from a phosphotyrosyl-induced activation of phospholipase C-, via SH2-domain interactions (Pawson, 1995), is that ethanol, by liberating G-protein -subunits, may activate a tyrosine kinase pathway that could regulate calcium influx (Laniyonu et al., 1994; Hollenberg 1994a; Lee et al., 1993), thereby causing some of ethanol's cellular effects.

Despite the several parallels between the actions of ethanol and EGF in the gastric contractile preparations, the use of several signal pathway probes in addition to the tyrosine kinase inhibitors, indomethacin and U57,908, pointed to significant differences in the signal transduction pathways activated by the two agonists. The signal pathway mediators evaluated were: calcium, kinase C, MEK and PI-3-kinase, all of which are believed to play roles in the action of growth factors and G-protein-coupled agonists. For instance, both ethanol and EGF required the presence of extracellular calcium to cause a contractile effect (fig. 4). This result suggested that the combined action of released inositol tris phosphate and diacylglycerol (potentially generated by phospholipase C-) to elevate intracellular calcium and activate kinase C (Berridge, 1993) were evidently not sufficient to cause the contractile response. However, the mode of calcium entry triggered by EGF and ethanol appeared to differ, because the antagonist of the voltage-regulated calcium channel, nifedipine, caused a substantial inhibition (90%) of EGF-induced contractions, but only partially (about 30% inhibition) attenuated ethanol-induced contraction; the antagonist of the "receptor-operated" calcium channel, SKF96365, was required in addition to nifedipine to block completely contractions elicited by ethanol (fig. 5). Differences in the role of kinase C were also observed, because ethanol-induced contractions in the LM tissue were refractory to the inhibitor of the -, - and -isoforms of kinase C, GF109203X, whereas this kinase C antagonist inhibited EGF-induced contractions by about 70%. Finally, our work revealed that the MEK inhibitor, PD98059 (Dudley et al., 1995), and the PI-3-kinase inhibitors, Wortmannin (Ui et al., 1995) and LY294002, selectively inhibited the contractile response to EGF in the LM preparation without affecting contractions caused by ethanol (fig. 8). Thus, the signal pathway activated by ethanol to cause a contractile response in the LM preparation would appear not to involve the MAPK/PI-3-kinase pathways that have been documented for growth factors such as insulin and EGF.

In the CM preparation, neither the PI-3-kinase inhibitors nor the MEK inhibitor blocked contractions caused by either EGF or ethanol. Thus, in keeping with our previous observations with EGF and angiotensin-II (Yang et al., 1992: 1993), the CM tissue appears to use signal transduction pathways, both for ethanol and for EGF, that differ from the pathways mediating a contractile response in the LM tissue. The lack of effect of the PI-3-kinase inhibitors and the MEK inhibitor in the CM tissue highlight the specificity of action of these inhibitors in the LM tissue. Overall, the observed similarities and distinct differences between the signal pathways activated by ethanol and EGF, both of which cause a contractile response that can be blocked by tyrosine kinase inhibitors, merit further investigation. Our data raise the possibility that ethanol may also act in part via a tyrosine kinase pathway for some of its effects on target tissues other than gastric smooth muscle. In this regard, one aspect of the action of ethanol that can be singled out for attention is its apparent requirement for an influx of extracellular calcium (figs. 4 and 5). Because ethanol can affect intracellular calcium concentrations in a variety of organs, it is important to point out that tissues such as heart and brain possess a calcium-sensitive cytoplasmic tyrosine kinase that can affect ion channel function (Lev et al., 1995; Sasaki et al., 1995). Thus, some of the pathophysiologic effects of ethanol might be caused by the activation of a similar nonreceptor tyrosine kinase pathways in target tissues. Whether or not ethanol may activate tyrosine kinase pathways in human gastric tissue or in other human organs, to play a role in the pathophysiological action of ethanol in humans, is an intriguing question that merits further study.