- Email: [email protected]
Influence of vagotomy on mucosal protection against alcohol-induced gastric damage in the rat
GASTROENTEROLOGY
1984:87:903-8
Influence of Vagotomy on Mucosal Protection Against Alcohol-Induced Gastric Damage in the Rat JULIA
M.
HENAGAN,
and THOMAS
A.
GREGORY
S. SMITH,
We examined the role of the vagus nerve in mediating the protective effects of subcutaneous 16,16dimethyl prostaglandin E, (16,16-dmPGE,) or of the mild irritant 30% ethanol (topically) against gastric mucosal injury induced by concentrated solutions of ethanol. Anesthetized rats underwent either truncal vagotomy or sham truncal vagotomy and were studied acutely or 7 days later. Under acute conditions, in rats with in tact vagi, oral saline followed by 1000/, ethanol produced severe gastric hemorrhagic and necrotic lesions throughout the glandular gastric mucosa. Oral 30% ethanol or pretreatment with 16,16-dmPGE2 (5, 10, or 25 pgikg) before giving oral significantly
reduced
the magnitude
of injury
when mucosa was subsequently exposed to 100% ethanol. In animals with truncal vagotomy. the protective effect of 16,16-dmPGE, or 30% ethanol was not observed. Similar results were noted when animals were studied 7 days after sham or truncal vagotomy. In other studies, the ability of 16,16dmPGE, to prevent gastric injury induced by 50% ethanol or 80 mM aspirin in acid solution (160 mM HCJ) with and without prior vagotomy was compared. Although 16,16-dmPGE2 (5 or 25 pglkg) pretreatment significantly reduced the degree of gastric damage induced by both agents in the nonvagotomized state, and by aspirin under vagotomized conditions, only partial protection by 16,16-dmPGE, against ethanol injury was observed in the vagotomized state. These results suggest that the mechanisms
whereby
tive effects
prostaglandins
against
aspirin
mediate
and
R. SEIDEL,
MILLER
Department of Surnerv of Physiology ._ and Department Medical School at Houston, Houston, Texas
saline
EDWARD
their protec-
ethanol
may
be
Received October 27. 1983. Accepted May 8, 1984. Address requests for reprints to: Thomas A. Miller. M.D.. Department of Surgery, The University of Texas Medical School, 6431 Fannin. Room 4266, Houston, Texas 77030. This work was supported by Research Grant AM25838 from the National Institutes of Health. C 1984 bv the American Gastroenterological Association . - .._ IX b+._~ ^,.
and Cell Biology,
different ability
The University
and that the vagus
of 16,16-dmPGE2
ethanol to prevent the rat stomach.
of Texas
influences
the
and of the mild irritant
nerve
30%
alcohol-induced
gastric injury in
It is well known that prostaglandins (PGs) possess the unique ability to prevent gastric mucosal injury under a variety of experimental conditions independent of their antisecretory properties, a process called “cytoprotection” (l-4). It has also been shown that a mild irritant, such as dilute alcohol, when topically applied to gastric epithelium, can prevent gastric injury on subsequent exposure of the mucosa to a necrotizing agent, such as absolute alcohol (5,6). Because available evidence suggests that this latter process may be mediated by endogenous PGs, this phenomenon has been labeled “adaptive cytoprotection” (5,6). Whereas such observations strongly implicate a role for PGs in mucosal defense, the mechanism by which exogenously administered or endogenously released PGs prevent gastric injury remains unknown. Previous studies have shown that direct synthesis and release of endogenous PGs in the rat stomach may be under vagal control (7-g). Further, bicarbonate secretion induced by exogenous PG in the canine stomach can be inhibited by atropine and tetrodotoxin (10). These studies suggest that if adaptive cytoprotection is indeed mediated by endogenous PGs, the vagus nerves could play a major role in the release or effect of these substances. They also suggest that exogenously administered PGs may, in part, mediate their effects through cholinergic pathways. The present study was undertaken to ascertain the role of the vagus nerves in mediating mucosal proAbbreviations 16,16-dmPGE,, din.
used in this 16,16-dimethyl
paper: 2DG, prostaglandin
2-deoxy-o-glucose: E,; PG. prostaglan-
904
HENAGAN
ET AL.
GASTKOENTEKOI,O(;‘I’
tection against alcohol-induced gastric injury and to determine whether a centrally mediated vagal stimulant, such as 2-deoxy-u-glucose (2DG) (11). also possesses the ability to prevent gastric mucosal damage. Because macroscopic criteria were used in this study to determine the presence or absence of gastric injury. the term “cytoprotection” will not be used in this report as it assumes that cells have been examined
microscopically
age has been
and
their
protection
from
dam-
demonstrated.
Materials
and Methods
Female Sprague-Dawley rats averaging 200 g in weight were used for all studies and housed in wire mesh bottom cages to prevent coprophagia. After a 24-h fast in which water but not food was allowed. animals were administered ether anesthesia and underwent laparotomy through a vertical abdominal incision: they received either a sham or truncal vagotomy. In sham-operated animals, the gastroesophageal junction was manipulated and the incision was closed. In vagotomized animals, a subdiaphragmatic bilateral truncal vagotomy was performed. All animals were allowed a 2-h recovery from operation. at which time rats from both groups received either a Z-ml bolus of saline or the mild irritant 30% ethanol (volivol) in saline, orally, via a metal oroesophageal tube. In pilot studies, 30% ethanol alone failed to produce any macroscopic evidence of damage in the rat stomach, either with or without concomitant vagotomy. Thirty minutes later, all animals received a 2-ml oral bolus of 100% ethanol. Some or animals also received 2DG (200 mgikg intraperitoneally) 16,16-dimethyl prostaglandin EZ (16,16-dmPGE,) (5, 10. or 25 pg/kg subcutaneously) 30 min before receiving the initial oral bolus of saline. Rats in all groups were killed 1 h after administration of absolute ethanol, their abdomens were opened, stomachs were clamped at the gastroesophageal and pyloric junctions and then removed, and residual volumes were collected and measured. Stomachs were opened along the greater curvature and laid flat, and the gastric mucosal surface was examined for evidence of necrotic and hemorrhagic mucosal lesions by an observer unfamiliar with the experimental protocol. The damaged area was expressed as the percentage of surface epithelium showing gross damage on macroscopic inspection of the mucosa. In a second set of experiments, the protocol was identical to that already described except that the sham or truncal vagotomy was performed 1 wk before the day of study. Because vagal denervation has the capability of altering antral motility, surgical treatment varied in this group of rats in that they also received a Heineke-Mikulicz type pyloroplasty to overcome potential gastric stasis. Sham-operated animals also received a pyloroplasty. In animals pretreated with 16.16-dmPGE,. only a 5-pgikg dose was administered subcutaneously. In a third set of experiments, rats were again studied acutely and were surgically prepared as first described. Treatment consisted of a 2-ml oral bolus of either 50%) ethanol (volivol) or 80 mM aspirin in an HCl solution in
Vol. 87. No. -I
which the final concentration was 160 mM HCl. One hour after exposure to either agent, rats were given another 2 ml of 160 mM HCl orally, and 1 h after that. a second 2-ml bolus of acid alone. The treatment regimen used here was employed because previous studies have shown that 80 mM aspirin in such an acid solution consistently elicits damage in the rat stomach (12). One hour after the final oral dose of acid alone, animals were killed and their stomachs were removed and examined as before. In this series of experiments, some animals also received 16.16dmPGEz pretreatment (5 or 25 pg/kgsubcutaneously) 30 min before administration of the damaging agent. The 2DG used in experiments was obtained from Sigma Chemical Company, St. Louis, MO. The 16,16-dmPGE2 was a gift from Dr. John Pike of the Upjohn Company, Kalamazoo, Mich. All results are expressed as the mean + SEM. Data from corresponding groups were analyzed using the unpaired Student’s t-test and differences were considered significant when a p value of ~0.05 was obtained.
Results In all experimental groups in which macroscopic evidence of mucosal injury was observed, damage was confined to the glandular portion of stomach with sparing of the forestomach. Damage was generallv limited to the oxyntic mucosa although, on occasion. hemorrhagic and necrotic lesions were also observed in the antrum. All animals receiving saline followed by 100% ethanol under acute conditions consistently showed mucosal injury involving on the average about 44% of the glandular gastric epithelium. Damage in vagotomized animals was slightly, but significantly. more pronounced when compared with sham-operated controls (Figure 1). In rats with intact vagi. prior exposure to 30% ethanol dramatically reduced the area of injury. However, in rats receiving truncal vagotomy, attenuation of damage was not observed, with the degree of injury not being significantly different from that achieved with lOO?% ethanol alone. If 16,16-dmPGE, was given before alcohol exposure in the saline-100% ethanol group, protection against macroscopic damage was again seen in sham-operated rats. but in those animals subjected to prior vagotomy. exogenously applied PC: was unable to reverse the incidence of necrotic and hemorrhagic lesions (Figure 1). Although the higher doses of 16.16-dmPGE, (10 and 25 pgikg) generally conferred more complete protection against alcohol injury in the nonvagotomized state, all three doses were unable to prevent damage in the vagotomized state (Table 1). Because these observations suggested that intact vagal innervation was necessary for mucosal protection to occur against damage induced by absolute ethanol, it followed that protection might also be induced by a known vagal stimulant such as 2DG
0c:tober
‘1981
USham
p
Vagotomy p
greater than in the corresponding group with intact vagi (Figure 2). Residual volumes in both acute and chronic studies varied greatly among the different experimental groups, but generally tended to be less in the vagotomized state when compared with animals not receiving prior vagotomy. This difference was particularly pronounced in the animals receiving 2DG under acute conditions in which the residual volume was 2.9 + 0.2 ml in nonvagotomized rats and 0.5 + 0.2 ml (p < 0.001) in vagotomized rats. Such findings raised the question of whether dilution of gastric contents may have contributed to the protection observed in the nonvagotomized state, but not in animals receiving prior vagotomy. To examine this possibility. three additional experiments were performed. In the first set of experiments, 6 animals each were subjected to either sham or truncal vagotomy, allowed to recover from operation for 2 h and then given 2 ml of normal saline orally. Thirty minutes later, they were killed. their stomachs were removed, and the resiclual volumes were measured. In sham-operated animals, the residual volume averaged 0.1 ml. This is in contrast to an average volume of 0.6 ml in vagotomized animals (p < 0.001 compared to sham). These findings are the reverse of what one would expect if dilution was responsible for our results. To ascertain the possible ) that 2DG secretory effects (and hence dilution might elicit in the nonvagotomized state, a second set of experiments were performed in which animals were subjected to sham or truncal vagotomy, allowed to recover from operation, and given 2DG intraperitoneally (200 mgikg). Thirty minutes later, they were killed, their stomachs were removed, and residual volumes were measured. Residual volumes (n = 6) averaged 0.3 ml for vagotomized rats in contrast to 0.03 ml (4 of 6 animals had no measurable residual volume) for sham-operated animals (11 = 6). Although these differences in mean volumes did not quite achieve statistical significance, they also were the reverse of what one would expect if dilution accounted for our results. 111 a third set of experiments, the 2DG experiments were repeated as originally described (i.e., a vagotomized and nonvagotomizecl group], but in addition, another vagotomy
Vagotomy
p
p
2 :: 607 I, r 2 40. 7 m :;
20-
Y5 60
Figure
1
o-
SAL-100
PGEl + 2DG + 30-100 SAL-100 SAL-100
Effects of 16.16-dimethyl
prostaglandin E, (PW,), Zdeoxy-u-glucose (2DG). and 30% ethanol on gastric, damage to the glandular mucosa induced by lOO%, ethanol in acutely vagotomized and sham-vagotomized (nonvagotomized) rats. Numbers within bars or adjacent to bars represent the n values for corresponding groups. The dose of 16.16-dm PGE, used was 5 pLg/kg subcutaneously. The SAI.-100 equals normal saline before administration of 100c% ethanol orally. The SO100 equals 30% ethanol before administration of 100% ethanol orally.
in animals having intact vagi (11). Thus. in the last set of acute animals in this series. sham-operated and vagotomized rats were pretreated with ZIG 30 min before receiving the oral saline-100% ethanol combination, and, as shown in Figure 1, significant protection was observed in the intact vagus group. whereas virtually none was seen after vagotomy. In the chronic studies of animals in which sham or truncal vagotomy was performed 7 days before study, the magnitude of ulceration was less in all experimental groups when compared with those obtained under acute conditions, but the results of the acute study were confirmed. In rats with sham vagotomy, pretreatment with both 30% ethanol and 16,16-dmPGE, (5 pgikg subcutaneously] significantly reduced the degree of gross damage induced by 100% ethanol. whereas necrotic and hemorrhagic: lesions in vagotomized animals were consistently present and riot significantly altered by any pretreatment regimen (Figure 2). Again. as demonstrated in the acute group, the degree of injury in vagotomized rats exposed to saline alone before 100% ethanol was
Table
1.
lti.l&Dimethyl
Prostaglandin
Ez Pretreatment
and
Gastric
Damtrge
by 100ryO Ethanol
‘!o of glandular ..______. Pretreatment
.~_______
43.7
Saline 5 /q/kg 16.1~.tlmPGE, 10 &kg
16.16-dmPGE,
25 @g/kg 1ti.16-cimPGE,
lfi.lfj-dntPGE,.
16.16-dimethyl
prostaglandin
mucosa
damaged
Nonvagotomized
Vagotomized
-t 5.0 (I1 = 25)”
60.8
f
16.9 + 4.1 (11 == 17)” 5.2 f 1.8 (II = 61”
58.3
T? 6.0 (I1 =
55.0
t 6.3 (1-l 7= 7)
10.1
53.3
t 6.8 (n = 81
f
3.3
(I1
E,. ‘I i’ I’. 0.01: ” p CI ~.0005
=
7)”
compared
with
c:orrrspontliu~
\aagottrmizd
4.5 (11 7 19)
animals.
12)
906
HENAGAN
B2
loo-
E
80.
cl
GASTROENTER’JLOGY
ET AL.
Vagotomy
Pyloroplasty
+ Pyloroplasty
~~0.025
p
PGEp + SAL-100
30-100
pco.01
z 8
60-
zz 3
40.
: 5
20.
?I s9
Figure
O-
SAL-100
2. Effects of 16,16-dimethyl prostaglandin E, (PGE,) or 30% ethanol on gastric damage to the glandular mucosa induced by 100% ethanol in chronically vagotomized (vagotomy and pyloroplasty) or nonvagotomized (pyloroplasty alone) rats. Numbers within bars or adjacent to bars represent the n values for corresponding groups. The dose of 16.16dmPGEZ used was 5 &kg subcutaneously. The SAL-100 equals normal saline before administration of 100% ethanol orally. The 30-100 equals 300/U ethanol before administration of 100% ethanol orally.
group was included in which extra saline was administered orally to overcome the differences in residual volume observed in our initial studies. In this additional experimental group, 2 ml of saline was administered orally as before, but just prior to giving the Z-ml oral bolus of 100% ethanol another 2.5 ml of saline was given orally to ensure that the residual volume in this group was not significantly different from that in nonvagotomized animals. As shown in Table 2, our initial results were confirmed with respect to the degree of mucosal damage under vagotomized and nonvagotomized conditions, and the additional group receiving the extra saline demonstrated just as much damage as the vagotomized group not receiving this treatment modality. From these various experiments, we concluded that dilution was not a factor in explaining the differences in the magnitude of gastric injury observed in vagotomized and nonvagotomized animals. In those studies that were designed to test the effect of vagotomy on aspirin-mediated damage and Table
2.
2-Deoxy-o-glucose
Pretreated
cl 60-
Sham
Vagotomy
Truncal
Nonvagotomized (n = 7) Vagotomized (n = 8) Dilute vagotomized (n = 12)”
~~0025 -T
---p~0.005 -
_,,,: m NS
NS
,,I
7.7 + 1.3 78.1 k 4.7” 71.3 + 4.1”
2.2 +- 0.5 1.6 f 0.2 2.2 t 0.3
” p < 0.001 compared with corresponding mals. ” Extra saline given orally before ethanol. See text for details.
Vagotomy
p~oooo5
14
Residual volume
without mE2
(ml)
nonvagotomized aniadministration of 100%
No. 4
the damage induced by a lower concentration of alcohol (50% ethanol), both damaging agents were observed to injure the gastric mucosa under both sham and truncal vagotomy conditions, but the degree of injury with aspirin was significantly less than that observed with ethanol (Figure 3). Animals undergoing vagotomy before exposure to 50% ethanol demonstrated an increase in gastric damage as compared to the corresponding nonvagotomized group. In animals with intact vagi, the protective effect of 16,16-dmPGE, was again demonstrated against ethanol-induced injury [Figure 3). This protective effect was more pronounced in animals pretreated with the 25-pglkg dose of 16,16-dmPGE2 than with the 5-pugi kg dose. In contrast, when vagotomized rats exposed to ethanol were pretreated with 16,16-dmPGEz, protection against alcohol damage was significantly less pronounced (Figure 3). In animals pretreated with the 5-pug/kg dose of 16,16-dmPGE2 in the vagotomized state, the magnitude of injury was significantly less (p < 0.005) when compared with vagotomized controls, but was not different from the degree of damage in nonvagotomized controls. Protection was much more pronounced if animals received 25 pgikg of 16,16-dmPGEZ in the vagotomized state, but even in this circumstance, the degree of injury after ethanol exposure was greater than that observed in animals receiving this dose of 16,16-dmPGE, in the nonvagotomized state (Figure 3). In animals exposed to acidified aspirin, the degree of injury in the vagotomized state was not greater than in the shamoperated group (Figure 3). Furthermore, 16,16-
Rats
% glandular mucosa damaged
87.
Vol.
Figure
3
with PGEl 15rUb%‘5rk!/k) 50% Ethanol
?
2;
without KE2
80mM
with PGE2 i5pg/kgl
Aspirin
Effect of 16.16-dimethyl prostaglandin E, (PGE,] on gastric damage to the glandular mucosa induced by 50% ethanol or 80 mM aspirin in acid solution in acutely vagotomized and sham-vagotomized (nonvagotomized rats. Numbers within bars or adjacent to bars represent the n values for corresponding groups. NS equals not significant.
October
1984
dmPGE, (5 r_Lg/kg)afforded significant and comparable protection in all animals receiving this treatment whether or not previous vagotomy had been performed.
Discussion Previous studies by Robert and associates (2,5,6) demonstrated that both exogenous PGs and mild irritants (such as dilute alcohol) were capable of preventing the damaging effects of a necrotizing agent such as 100% ethanol when topically applied to the rat stomach. The present study confirmed these findings and further demonstrated that surgical vagotomy markedly lessened the protective effects of 16,16-dmPGE, and of the mild irritant 30% ethanol against gastric mucosal injury induced by 100% ethanol. In addition, vagotomy also exacerbated the magnitude of gastric injury in mucosa exposed to oral saline-100% ethanol acutely and 1 wk after vagal transection, and in mucosa exposed to 50% ethanol in acid medium under acute conditions when compared with corresponding mucosa in which vagal innervation was intact. Furthermore, the findings that in the vagotomized stomach even high doses of 16,16-dmPGE, could not overcome the damaging effects of 100% ethanol and were only partially effective in overcoming damage induced by a lower concentration of alcohol (i.e., 50% ethanol) suggest that some degree of vagal innervation is necessary to adequately protect the gastric epitheliurn against alcohol injury. The mechanism underlying the adverse effects of vagotomy on mucosal protection is not known. The fact that such effects were demonstrable under both acute and chronic experimental conditions militates against the acute traumatic effects of surgery as being responsible for our results. Further, the finding that vagotomized rats had greater residual volumes in contrast to nonvagotomized rats before exposing the gastric epithelium to 100% ethanol is the opposite of what one would expect if dilution was the explanation for our results. In addition, the observation that the volume of gastric contents before ethanol administration in sham-operated animals was not influenced by 2DG and that administration of extra saline to vagotomized animals treated with 2DG failed to obliterate the damaging effects of 100% ethanol are further evidence that dilution did not play a role in mediating our findings. Although it has been shown that vagotomy can diminish gastric mucosal blood flow for varying periods of time after vagal transection (13--151, it is unlikely that alterations in gastric perfusion can explain our results. Because mucosal ischemia is known to aggravate the damaging effects of bile salts (16,17), one would expect that both
VAGOTOMY
AND GASTRIC
MUCOSAL
DAMAGE
907
aspirin and ethanol damage would be similarly influenced by vagotomy if ischemia was the mechanism whereby vagal denervation mediated its effect. This, however, was not the case. Recent evidence (5,6) indicates that mild irritants may mediate their protective effects through endogenous PG formation and release. This is based on the finding that indomethacin, an inhibitor of PG synthesis (18), can block the protective effect of mild irritants (5), and that a known mild irritant, such as NaOH (0.01-0.1 M), can directly stimulate PG synthesis in doses observed to prevent injury by 100% ethanol (6). If indeed mild irritant-induced protection is mediated by endogenous PGs, our findings with topical 30% ethanol before exposing the gastric mucosa to 100% ethanol suggest that intact vagal innervation may be a necessary prerequisite for PG synthesis to occur. That a link between PC synthesis and the vagus nerve may exist has been previously observed (7-g). Using an in vitro rat whole stomach preparation, Coceani et al. (7) observed that vagal stimulation elicited the release of PGE and PGF. Bennett et al. (8), using a similar whole stomach preparation from the rat, noted that electrical stimulation of vagal nerve fibers in its wall elicited the production of a substance resembling PGEr which could be abolished with tetrodotoxin. In other studies, Singh (9) was able to elicit the release of a PGF2,like product in the rat stomach in response to both parasympathetic stimulation and acetylcholine infusion which ceased when atropine was administered concomitantly with the stimulus. Finally, the observation in the present study that the vagal stimulant 2DG can prevent alcohol injury only when the vagus nerve is intact suggests that 2DG-induced protection may likewise be mediated through a vagally dependent PG synthetic pathway. Our findings would also suggest that the vagus plays some role in mediating the protective effects of PG independent of any direct action on PG synthesis. In gastric epithelium exposed to both 50% and 100% ethanol, vagotomy markedly altered the ability of exogenous PGEz to confer protection against alcohol injury. This alteration was much more pronounced in mucosa exposed to 100% ethanol than with 50% ethanol, but even with 50% ethanol a high dose of PGEz (25 Fg/kg) was not as effective in rendering protection in the vagotomized state as it was when vagal innervation was intact. Such observations are consistent with the hypothesis that vagaldependent mechanisms are responsible, at least in part, for the ability of PG to prevent alcohol damage in the rat stomach. Further, if protection by 30% ethanol and 2DG is indeed mediated by endogenous PGs, the lack of protection after vagotomy may not be entirely related to inhibition of PG synthesis, but
GASTROENTEROLOGY
908 HENAGAN ET AL.
conceivably a portion of it could be a reflection of the inability of endogenous PGs to mediate their protective effects as a result of the vagal denervation. How vagal integrity enhances the protective effects of PGs (whether exogenously administered or endogenously released) against alcohol-induced injury is unknown and must await further study. It is of interest in this regard that in addition to our findings with 16,16-dmPGE,, Mozsik et al. (19) noted that the ability of prostacyclin to prevent alcohol damage in the rat stomach was also adversely influenced by surgical vagotomy. Of equal interest is the recent finding in our laboratory that PG-induced alkaline secretion in the canine stomach can be abolished by concomitant treatment with atropine or tetrodotoxin (lo), further suggesting that PGs may, in part, mediate their effects through cholinergic pathways. Despite an apparent relationship between the ability of gastric mucosa to resist alcohol-induced injury and the status of vagal innervation, a similar association could not be demonstrated with mucosa exposed to aspirin. The ability of mucosa to withstand aspirin damage was not influenced by prior vagotomy either when mucosa was exposed to aspirin alone or when pretreated with 16,16-dm PGE2 before such exposure. The explanation for the different responses of mucosa to alcohol and aspirin in the presence of vagotomy is uncertain. That such diverse effects were observed, though, suggests that the mechanism of injury of the two damaging agents is not identical and that the means by which PGs mediate their protective effects against these two agents is also different. In summary, we have observed that the ability of PG to prevent gastric mucosal injury induced by alcohol, but not aspirin, is dependent, at least in part, upon vagal innervation in the rat stomach. This adverse effect of vagal denervation in alcohol-exposed mucosa does not seem to be related to the acute traumatic effects of surgery because it can also be demonstrated under more chronic conditions. As PG is protective against injury induced by aspirin and alcohol in the nonvagotomized state, but only partially capable of rendering protection against alcohol injury in the vagotomized state, in contrast to that induced by aspirin, it would appear that the mechanism of damage and the means by which PGs prevent injury in response to the two damaging agents may be different.
Vol.87. No. 4
References 1. Robert A. Cytoprotection by prostaglandins. Gastroenterology 1979:77:761-7. 2. Robert A, Nezamis JE. Lancaster C. Hanchar AJ. Cqtoprotection by prostaglandins in rats: prevention of gastric: necrosis produced by alcohol, HCI, NaOH. hypertonic N&i. ant1 thermal injury. Gastroenterology 1979;77:433-43. 3. Miller TA, Jacobson ED. Gastrointestinal cytoprotection by prostaglandins. Gut 1979:20:75-87. 4. Miller TA. Protective effects of prostaglandins against gastric mucosal damage: current knowledge and proposed mechanisms. Am J Physiol 1983:245(Gastrointest Liver Physiol 8):G601-23. 5. Chaudhury TK, Robert A. Prevention by mild irritants of gastric necrosis produced in rats by sodium taurocholate. Dig Dis Sci 1980;25:830-6. 6. Robert A, Nezamis JE, Lancaster C. Davis JP, Field SO. Hanchar AJ. Mild irritants prevent gastric nwrosis through “adaptive cytoprotection” mediated by prostaglandins. Am J Physiol 1983;245(Gastrointest Liver Physiot 8):Gll3-21. 7. Coceani F, Pace-Asciak C. Volta F. Wolfe LS. Effect of nerve stimulation on prostaglandin formation and release from the rat stomach. Am J Physiol 1967;213:1056-64. 8. Bennett A. Friedmann CA, Vane JR. Release of prostaglandin E, from the rat stomach. Nature 1967;216:873-6. 9. Singh J. Prostaglandin release from rat stomach following vagal stimulation or administration of acetylcholine. Em J Pharmacol 1980:65:39-48. 10. Miller TA. Henagan JM. Watkins LA, 1.0~ TM. Prostaglandin induced bicarbonate secretion in the canine stomach: characteristics and evidence for a cholinergic mechanism. J Surg Kes 1983:35:106-12. 11. Hirschowitz B, Sachs G. Vagal gastric secretory stimulation by 2-deoxy-o-glucose. Am J Physiol 1965:209:452-60. 12. Carmichael HA, Nelson LM, Russell RL. Cimetidine and prostaglandin: evidence for different modes of action on the rat gastric mucosa. Gastroenterologg 1978:74:1229-32. 13. Hunter GC. Goldstone J, Villa R, Way LW. Effect of vagotomy upon intragastric redistribution of microvascular flow. J Surg Res 1979:26:314-9. 14. Seifert J, Lenz J, Bruckner W. Brendel L’v. HolIe F. Are 1-~iootl flow Lhanges after selective proximal vagotomy responsible for a necrosis in the gastric wall? Res Exp Med 1980:177:2636. 15. Bell PRF, Battersby C. Effect of vagotomy on gastric mucosal blood flow. Gastroenterology 1968:54:1032-7. 16. Ritchie WP Jr. Acute gastric mucosal damage induced by bile salts, acid and ischemia. Gatroenterology 1975;68:699-707. 17. Ritchie WP Jr, Shearburn EW, III. Influence of isoproterenol and cholestyramine on acute gastric mucosal ulcerogenesis. Gastroenterology 1977:73:62-5. 18. Vane JR. Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs. Nature (New Biol) 1971:231:232-5. 19. Mozsik GY. Moron F. Javor T. Cellular mechanisms of the development of gastric mucosal damage and of gastrocytoprotection induced by prostacyclin in rats. A pharmacological study. Prostaglandins Leukotrienes Med 1982:9:71-84.
1984:87:903-8
Influence of Vagotomy on Mucosal Protection Against Alcohol-Induced Gastric Damage in the Rat JULIA
M.
HENAGAN,
and THOMAS
A.
GREGORY
S. SMITH,
We examined the role of the vagus nerve in mediating the protective effects of subcutaneous 16,16dimethyl prostaglandin E, (16,16-dmPGE,) or of the mild irritant 30% ethanol (topically) against gastric mucosal injury induced by concentrated solutions of ethanol. Anesthetized rats underwent either truncal vagotomy or sham truncal vagotomy and were studied acutely or 7 days later. Under acute conditions, in rats with in tact vagi, oral saline followed by 1000/, ethanol produced severe gastric hemorrhagic and necrotic lesions throughout the glandular gastric mucosa. Oral 30% ethanol or pretreatment with 16,16-dmPGE2 (5, 10, or 25 pgikg) before giving oral significantly
reduced
the magnitude
of injury
when mucosa was subsequently exposed to 100% ethanol. In animals with truncal vagotomy. the protective effect of 16,16-dmPGE, or 30% ethanol was not observed. Similar results were noted when animals were studied 7 days after sham or truncal vagotomy. In other studies, the ability of 16,16dmPGE, to prevent gastric injury induced by 50% ethanol or 80 mM aspirin in acid solution (160 mM HCJ) with and without prior vagotomy was compared. Although 16,16-dmPGE2 (5 or 25 pglkg) pretreatment significantly reduced the degree of gastric damage induced by both agents in the nonvagotomized state, and by aspirin under vagotomized conditions, only partial protection by 16,16-dmPGE, against ethanol injury was observed in the vagotomized state. These results suggest that the mechanisms
whereby
tive effects
prostaglandins
against
aspirin
mediate
and
R. SEIDEL,
MILLER
Department of Surnerv of Physiology ._ and Department Medical School at Houston, Houston, Texas
saline
EDWARD
their protec-
ethanol
may
be
Received October 27. 1983. Accepted May 8, 1984. Address requests for reprints to: Thomas A. Miller. M.D.. Department of Surgery, The University of Texas Medical School, 6431 Fannin. Room 4266, Houston, Texas 77030. This work was supported by Research Grant AM25838 from the National Institutes of Health. C 1984 bv the American Gastroenterological Association . - .._ IX b+._~ ^,.
and Cell Biology,
different ability
The University
and that the vagus
of 16,16-dmPGE2
ethanol to prevent the rat stomach.
of Texas
influences
the
and of the mild irritant
nerve
30%
alcohol-induced
gastric injury in
It is well known that prostaglandins (PGs) possess the unique ability to prevent gastric mucosal injury under a variety of experimental conditions independent of their antisecretory properties, a process called “cytoprotection” (l-4). It has also been shown that a mild irritant, such as dilute alcohol, when topically applied to gastric epithelium, can prevent gastric injury on subsequent exposure of the mucosa to a necrotizing agent, such as absolute alcohol (5,6). Because available evidence suggests that this latter process may be mediated by endogenous PGs, this phenomenon has been labeled “adaptive cytoprotection” (5,6). Whereas such observations strongly implicate a role for PGs in mucosal defense, the mechanism by which exogenously administered or endogenously released PGs prevent gastric injury remains unknown. Previous studies have shown that direct synthesis and release of endogenous PGs in the rat stomach may be under vagal control (7-g). Further, bicarbonate secretion induced by exogenous PG in the canine stomach can be inhibited by atropine and tetrodotoxin (10). These studies suggest that if adaptive cytoprotection is indeed mediated by endogenous PGs, the vagus nerves could play a major role in the release or effect of these substances. They also suggest that exogenously administered PGs may, in part, mediate their effects through cholinergic pathways. The present study was undertaken to ascertain the role of the vagus nerves in mediating mucosal proAbbreviations 16,16-dmPGE,, din.
used in this 16,16-dimethyl
paper: 2DG, prostaglandin
2-deoxy-o-glucose: E,; PG. prostaglan-
904
HENAGAN
ET AL.
GASTKOENTEKOI,O(;‘I’
tection against alcohol-induced gastric injury and to determine whether a centrally mediated vagal stimulant, such as 2-deoxy-u-glucose (2DG) (11). also possesses the ability to prevent gastric mucosal damage. Because macroscopic criteria were used in this study to determine the presence or absence of gastric injury. the term “cytoprotection” will not be used in this report as it assumes that cells have been examined
microscopically
age has been
and
their
protection
from
dam-
demonstrated.
Materials
and Methods
Female Sprague-Dawley rats averaging 200 g in weight were used for all studies and housed in wire mesh bottom cages to prevent coprophagia. After a 24-h fast in which water but not food was allowed. animals were administered ether anesthesia and underwent laparotomy through a vertical abdominal incision: they received either a sham or truncal vagotomy. In sham-operated animals, the gastroesophageal junction was manipulated and the incision was closed. In vagotomized animals, a subdiaphragmatic bilateral truncal vagotomy was performed. All animals were allowed a 2-h recovery from operation. at which time rats from both groups received either a Z-ml bolus of saline or the mild irritant 30% ethanol (volivol) in saline, orally, via a metal oroesophageal tube. In pilot studies, 30% ethanol alone failed to produce any macroscopic evidence of damage in the rat stomach, either with or without concomitant vagotomy. Thirty minutes later, all animals received a 2-ml oral bolus of 100% ethanol. Some or animals also received 2DG (200 mgikg intraperitoneally) 16,16-dimethyl prostaglandin EZ (16,16-dmPGE,) (5, 10. or 25 pg/kg subcutaneously) 30 min before receiving the initial oral bolus of saline. Rats in all groups were killed 1 h after administration of absolute ethanol, their abdomens were opened, stomachs were clamped at the gastroesophageal and pyloric junctions and then removed, and residual volumes were collected and measured. Stomachs were opened along the greater curvature and laid flat, and the gastric mucosal surface was examined for evidence of necrotic and hemorrhagic mucosal lesions by an observer unfamiliar with the experimental protocol. The damaged area was expressed as the percentage of surface epithelium showing gross damage on macroscopic inspection of the mucosa. In a second set of experiments, the protocol was identical to that already described except that the sham or truncal vagotomy was performed 1 wk before the day of study. Because vagal denervation has the capability of altering antral motility, surgical treatment varied in this group of rats in that they also received a Heineke-Mikulicz type pyloroplasty to overcome potential gastric stasis. Sham-operated animals also received a pyloroplasty. In animals pretreated with 16.16-dmPGE,. only a 5-pgikg dose was administered subcutaneously. In a third set of experiments, rats were again studied acutely and were surgically prepared as first described. Treatment consisted of a 2-ml oral bolus of either 50%) ethanol (volivol) or 80 mM aspirin in an HCl solution in
Vol. 87. No. -I
which the final concentration was 160 mM HCl. One hour after exposure to either agent, rats were given another 2 ml of 160 mM HCl orally, and 1 h after that. a second 2-ml bolus of acid alone. The treatment regimen used here was employed because previous studies have shown that 80 mM aspirin in such an acid solution consistently elicits damage in the rat stomach (12). One hour after the final oral dose of acid alone, animals were killed and their stomachs were removed and examined as before. In this series of experiments, some animals also received 16.16dmPGEz pretreatment (5 or 25 pg/kgsubcutaneously) 30 min before administration of the damaging agent. The 2DG used in experiments was obtained from Sigma Chemical Company, St. Louis, MO. The 16,16-dmPGE2 was a gift from Dr. John Pike of the Upjohn Company, Kalamazoo, Mich. All results are expressed as the mean + SEM. Data from corresponding groups were analyzed using the unpaired Student’s t-test and differences were considered significant when a p value of ~0.05 was obtained.
Results In all experimental groups in which macroscopic evidence of mucosal injury was observed, damage was confined to the glandular portion of stomach with sparing of the forestomach. Damage was generallv limited to the oxyntic mucosa although, on occasion. hemorrhagic and necrotic lesions were also observed in the antrum. All animals receiving saline followed by 100% ethanol under acute conditions consistently showed mucosal injury involving on the average about 44% of the glandular gastric epithelium. Damage in vagotomized animals was slightly, but significantly. more pronounced when compared with sham-operated controls (Figure 1). In rats with intact vagi. prior exposure to 30% ethanol dramatically reduced the area of injury. However, in rats receiving truncal vagotomy, attenuation of damage was not observed, with the degree of injury not being significantly different from that achieved with lOO?% ethanol alone. If 16,16-dmPGE, was given before alcohol exposure in the saline-100% ethanol group, protection against macroscopic damage was again seen in sham-operated rats. but in those animals subjected to prior vagotomy. exogenously applied PC: was unable to reverse the incidence of necrotic and hemorrhagic lesions (Figure 1). Although the higher doses of 16.16-dmPGE, (10 and 25 pgikg) generally conferred more complete protection against alcohol injury in the nonvagotomized state, all three doses were unable to prevent damage in the vagotomized state (Table 1). Because these observations suggested that intact vagal innervation was necessary for mucosal protection to occur against damage induced by absolute ethanol, it followed that protection might also be induced by a known vagal stimulant such as 2DG
0c:tober
‘1981
USham
p
Vagotomy p
greater than in the corresponding group with intact vagi (Figure 2). Residual volumes in both acute and chronic studies varied greatly among the different experimental groups, but generally tended to be less in the vagotomized state when compared with animals not receiving prior vagotomy. This difference was particularly pronounced in the animals receiving 2DG under acute conditions in which the residual volume was 2.9 + 0.2 ml in nonvagotomized rats and 0.5 + 0.2 ml (p < 0.001) in vagotomized rats. Such findings raised the question of whether dilution of gastric contents may have contributed to the protection observed in the nonvagotomized state, but not in animals receiving prior vagotomy. To examine this possibility. three additional experiments were performed. In the first set of experiments, 6 animals each were subjected to either sham or truncal vagotomy, allowed to recover from operation for 2 h and then given 2 ml of normal saline orally. Thirty minutes later, they were killed. their stomachs were removed, and the resiclual volumes were measured. In sham-operated animals, the residual volume averaged 0.1 ml. This is in contrast to an average volume of 0.6 ml in vagotomized animals (p < 0.001 compared to sham). These findings are the reverse of what one would expect if dilution was responsible for our results. To ascertain the possible ) that 2DG secretory effects (and hence dilution might elicit in the nonvagotomized state, a second set of experiments were performed in which animals were subjected to sham or truncal vagotomy, allowed to recover from operation, and given 2DG intraperitoneally (200 mgikg). Thirty minutes later, they were killed, their stomachs were removed, and residual volumes were measured. Residual volumes (n = 6) averaged 0.3 ml for vagotomized rats in contrast to 0.03 ml (4 of 6 animals had no measurable residual volume) for sham-operated animals (11 = 6). Although these differences in mean volumes did not quite achieve statistical significance, they also were the reverse of what one would expect if dilution accounted for our results. 111 a third set of experiments, the 2DG experiments were repeated as originally described (i.e., a vagotomized and nonvagotomizecl group], but in addition, another vagotomy
Vagotomy
p
p
2 :: 607 I, r 2 40. 7 m :;
20-
Y5 60
Figure
1
o-
SAL-100
PGEl + 2DG + 30-100 SAL-100 SAL-100
Effects of 16.16-dimethyl
prostaglandin E, (PW,), Zdeoxy-u-glucose (2DG). and 30% ethanol on gastric, damage to the glandular mucosa induced by lOO%, ethanol in acutely vagotomized and sham-vagotomized (nonvagotomized) rats. Numbers within bars or adjacent to bars represent the n values for corresponding groups. The dose of 16.16-dm PGE, used was 5 pLg/kg subcutaneously. The SAI.-100 equals normal saline before administration of 100c% ethanol orally. The SO100 equals 30% ethanol before administration of 100% ethanol orally.
in animals having intact vagi (11). Thus. in the last set of acute animals in this series. sham-operated and vagotomized rats were pretreated with ZIG 30 min before receiving the oral saline-100% ethanol combination, and, as shown in Figure 1, significant protection was observed in the intact vagus group. whereas virtually none was seen after vagotomy. In the chronic studies of animals in which sham or truncal vagotomy was performed 7 days before study, the magnitude of ulceration was less in all experimental groups when compared with those obtained under acute conditions, but the results of the acute study were confirmed. In rats with sham vagotomy, pretreatment with both 30% ethanol and 16,16-dmPGE, (5 pgikg subcutaneously] significantly reduced the degree of gross damage induced by 100% ethanol. whereas necrotic and hemorrhagic: lesions in vagotomized animals were consistently present and riot significantly altered by any pretreatment regimen (Figure 2). Again. as demonstrated in the acute group, the degree of injury in vagotomized rats exposed to saline alone before 100% ethanol was
Table
1.
lti.l&Dimethyl
Prostaglandin
Ez Pretreatment
and
Gastric
Damtrge
by 100ryO Ethanol
‘!o of glandular ..______. Pretreatment
.~_______
43.7
Saline 5 /q/kg 16.1~.tlmPGE, 10 &kg
16.16-dmPGE,
25 @g/kg 1ti.16-cimPGE,
lfi.lfj-dntPGE,.
16.16-dimethyl
prostaglandin
mucosa
damaged
Nonvagotomized
Vagotomized
-t 5.0 (I1 = 25)”
60.8
f
16.9 + 4.1 (11 == 17)” 5.2 f 1.8 (II = 61”
58.3
T? 6.0 (I1 =
55.0
t 6.3 (1-l 7= 7)
10.1
53.3
t 6.8 (n = 81
f
3.3
(I1
E,. ‘I i’ I’. 0.01: ” p CI ~.0005
=
7)”
compared
with
c:orrrspontliu~
\aagottrmizd
4.5 (11 7 19)
animals.
12)
906
HENAGAN
B2
loo-
E
80.
cl
GASTROENTER’JLOGY
ET AL.
Vagotomy
Pyloroplasty
+ Pyloroplasty
~~0.025
p
PGEp + SAL-100
30-100
pco.01
z 8
60-
zz 3
40.
: 5
20.
?I s9
Figure
O-
SAL-100
2. Effects of 16,16-dimethyl prostaglandin E, (PGE,) or 30% ethanol on gastric damage to the glandular mucosa induced by 100% ethanol in chronically vagotomized (vagotomy and pyloroplasty) or nonvagotomized (pyloroplasty alone) rats. Numbers within bars or adjacent to bars represent the n values for corresponding groups. The dose of 16.16dmPGEZ used was 5 &kg subcutaneously. The SAL-100 equals normal saline before administration of 100% ethanol orally. The 30-100 equals 300/U ethanol before administration of 100% ethanol orally.
group was included in which extra saline was administered orally to overcome the differences in residual volume observed in our initial studies. In this additional experimental group, 2 ml of saline was administered orally as before, but just prior to giving the Z-ml oral bolus of 100% ethanol another 2.5 ml of saline was given orally to ensure that the residual volume in this group was not significantly different from that in nonvagotomized animals. As shown in Table 2, our initial results were confirmed with respect to the degree of mucosal damage under vagotomized and nonvagotomized conditions, and the additional group receiving the extra saline demonstrated just as much damage as the vagotomized group not receiving this treatment modality. From these various experiments, we concluded that dilution was not a factor in explaining the differences in the magnitude of gastric injury observed in vagotomized and nonvagotomized animals. In those studies that were designed to test the effect of vagotomy on aspirin-mediated damage and Table
2.
2-Deoxy-o-glucose
Pretreated
cl 60-
Sham
Vagotomy
Truncal
Nonvagotomized (n = 7) Vagotomized (n = 8) Dilute vagotomized (n = 12)”
~~0025 -T
---p~0.005 -
_,,,: m NS
NS
,,I
7.7 + 1.3 78.1 k 4.7” 71.3 + 4.1”
2.2 +- 0.5 1.6 f 0.2 2.2 t 0.3
” p < 0.001 compared with corresponding mals. ” Extra saline given orally before ethanol. See text for details.
Vagotomy
p~oooo5
14
Residual volume
without mE2
(ml)
nonvagotomized aniadministration of 100%
No. 4
the damage induced by a lower concentration of alcohol (50% ethanol), both damaging agents were observed to injure the gastric mucosa under both sham and truncal vagotomy conditions, but the degree of injury with aspirin was significantly less than that observed with ethanol (Figure 3). Animals undergoing vagotomy before exposure to 50% ethanol demonstrated an increase in gastric damage as compared to the corresponding nonvagotomized group. In animals with intact vagi, the protective effect of 16,16-dmPGE, was again demonstrated against ethanol-induced injury [Figure 3). This protective effect was more pronounced in animals pretreated with the 25-pglkg dose of 16,16-dmPGE2 than with the 5-pugi kg dose. In contrast, when vagotomized rats exposed to ethanol were pretreated with 16,16-dmPGEz, protection against alcohol damage was significantly less pronounced (Figure 3). In animals pretreated with the 5-pug/kg dose of 16,16-dmPGE2 in the vagotomized state, the magnitude of injury was significantly less (p < 0.005) when compared with vagotomized controls, but was not different from the degree of damage in nonvagotomized controls. Protection was much more pronounced if animals received 25 pgikg of 16,16-dmPGEZ in the vagotomized state, but even in this circumstance, the degree of injury after ethanol exposure was greater than that observed in animals receiving this dose of 16,16-dmPGE, in the nonvagotomized state (Figure 3). In animals exposed to acidified aspirin, the degree of injury in the vagotomized state was not greater than in the shamoperated group (Figure 3). Furthermore, 16,16-
Rats
% glandular mucosa damaged
87.
Vol.
Figure
3
with PGEl 15rUb%‘5rk!/k) 50% Ethanol
?
2;
without KE2
80mM
with PGE2 i5pg/kgl
Aspirin
Effect of 16.16-dimethyl prostaglandin E, (PGE,] on gastric damage to the glandular mucosa induced by 50% ethanol or 80 mM aspirin in acid solution in acutely vagotomized and sham-vagotomized (nonvagotomized rats. Numbers within bars or adjacent to bars represent the n values for corresponding groups. NS equals not significant.
October
1984
dmPGE, (5 r_Lg/kg)afforded significant and comparable protection in all animals receiving this treatment whether or not previous vagotomy had been performed.
Discussion Previous studies by Robert and associates (2,5,6) demonstrated that both exogenous PGs and mild irritants (such as dilute alcohol) were capable of preventing the damaging effects of a necrotizing agent such as 100% ethanol when topically applied to the rat stomach. The present study confirmed these findings and further demonstrated that surgical vagotomy markedly lessened the protective effects of 16,16-dmPGE, and of the mild irritant 30% ethanol against gastric mucosal injury induced by 100% ethanol. In addition, vagotomy also exacerbated the magnitude of gastric injury in mucosa exposed to oral saline-100% ethanol acutely and 1 wk after vagal transection, and in mucosa exposed to 50% ethanol in acid medium under acute conditions when compared with corresponding mucosa in which vagal innervation was intact. Furthermore, the findings that in the vagotomized stomach even high doses of 16,16-dmPGE, could not overcome the damaging effects of 100% ethanol and were only partially effective in overcoming damage induced by a lower concentration of alcohol (i.e., 50% ethanol) suggest that some degree of vagal innervation is necessary to adequately protect the gastric epitheliurn against alcohol injury. The mechanism underlying the adverse effects of vagotomy on mucosal protection is not known. The fact that such effects were demonstrable under both acute and chronic experimental conditions militates against the acute traumatic effects of surgery as being responsible for our results. Further, the finding that vagotomized rats had greater residual volumes in contrast to nonvagotomized rats before exposing the gastric epithelium to 100% ethanol is the opposite of what one would expect if dilution was the explanation for our results. In addition, the observation that the volume of gastric contents before ethanol administration in sham-operated animals was not influenced by 2DG and that administration of extra saline to vagotomized animals treated with 2DG failed to obliterate the damaging effects of 100% ethanol are further evidence that dilution did not play a role in mediating our findings. Although it has been shown that vagotomy can diminish gastric mucosal blood flow for varying periods of time after vagal transection (13--151, it is unlikely that alterations in gastric perfusion can explain our results. Because mucosal ischemia is known to aggravate the damaging effects of bile salts (16,17), one would expect that both
VAGOTOMY
AND GASTRIC
MUCOSAL
DAMAGE
907
aspirin and ethanol damage would be similarly influenced by vagotomy if ischemia was the mechanism whereby vagal denervation mediated its effect. This, however, was not the case. Recent evidence (5,6) indicates that mild irritants may mediate their protective effects through endogenous PG formation and release. This is based on the finding that indomethacin, an inhibitor of PG synthesis (18), can block the protective effect of mild irritants (5), and that a known mild irritant, such as NaOH (0.01-0.1 M), can directly stimulate PG synthesis in doses observed to prevent injury by 100% ethanol (6). If indeed mild irritant-induced protection is mediated by endogenous PGs, our findings with topical 30% ethanol before exposing the gastric mucosa to 100% ethanol suggest that intact vagal innervation may be a necessary prerequisite for PG synthesis to occur. That a link between PC synthesis and the vagus nerve may exist has been previously observed (7-g). Using an in vitro rat whole stomach preparation, Coceani et al. (7) observed that vagal stimulation elicited the release of PGE and PGF. Bennett et al. (8), using a similar whole stomach preparation from the rat, noted that electrical stimulation of vagal nerve fibers in its wall elicited the production of a substance resembling PGEr which could be abolished with tetrodotoxin. In other studies, Singh (9) was able to elicit the release of a PGF2,like product in the rat stomach in response to both parasympathetic stimulation and acetylcholine infusion which ceased when atropine was administered concomitantly with the stimulus. Finally, the observation in the present study that the vagal stimulant 2DG can prevent alcohol injury only when the vagus nerve is intact suggests that 2DG-induced protection may likewise be mediated through a vagally dependent PG synthetic pathway. Our findings would also suggest that the vagus plays some role in mediating the protective effects of PG independent of any direct action on PG synthesis. In gastric epithelium exposed to both 50% and 100% ethanol, vagotomy markedly altered the ability of exogenous PGEz to confer protection against alcohol injury. This alteration was much more pronounced in mucosa exposed to 100% ethanol than with 50% ethanol, but even with 50% ethanol a high dose of PGEz (25 Fg/kg) was not as effective in rendering protection in the vagotomized state as it was when vagal innervation was intact. Such observations are consistent with the hypothesis that vagaldependent mechanisms are responsible, at least in part, for the ability of PG to prevent alcohol damage in the rat stomach. Further, if protection by 30% ethanol and 2DG is indeed mediated by endogenous PGs, the lack of protection after vagotomy may not be entirely related to inhibition of PG synthesis, but
GASTROENTEROLOGY
908 HENAGAN ET AL.
conceivably a portion of it could be a reflection of the inability of endogenous PGs to mediate their protective effects as a result of the vagal denervation. How vagal integrity enhances the protective effects of PGs (whether exogenously administered or endogenously released) against alcohol-induced injury is unknown and must await further study. It is of interest in this regard that in addition to our findings with 16,16-dmPGE,, Mozsik et al. (19) noted that the ability of prostacyclin to prevent alcohol damage in the rat stomach was also adversely influenced by surgical vagotomy. Of equal interest is the recent finding in our laboratory that PG-induced alkaline secretion in the canine stomach can be abolished by concomitant treatment with atropine or tetrodotoxin (lo), further suggesting that PGs may, in part, mediate their effects through cholinergic pathways. Despite an apparent relationship between the ability of gastric mucosa to resist alcohol-induced injury and the status of vagal innervation, a similar association could not be demonstrated with mucosa exposed to aspirin. The ability of mucosa to withstand aspirin damage was not influenced by prior vagotomy either when mucosa was exposed to aspirin alone or when pretreated with 16,16-dm PGE2 before such exposure. The explanation for the different responses of mucosa to alcohol and aspirin in the presence of vagotomy is uncertain. That such diverse effects were observed, though, suggests that the mechanism of injury of the two damaging agents is not identical and that the means by which PGs mediate their protective effects against these two agents is also different. In summary, we have observed that the ability of PG to prevent gastric mucosal injury induced by alcohol, but not aspirin, is dependent, at least in part, upon vagal innervation in the rat stomach. This adverse effect of vagal denervation in alcohol-exposed mucosa does not seem to be related to the acute traumatic effects of surgery because it can also be demonstrated under more chronic conditions. As PG is protective against injury induced by aspirin and alcohol in the nonvagotomized state, but only partially capable of rendering protection against alcohol injury in the vagotomized state, in contrast to that induced by aspirin, it would appear that the mechanism of damage and the means by which PGs prevent injury in response to the two damaging agents may be different.
Vol.87. No. 4
References 1. Robert A. Cytoprotection by prostaglandins. Gastroenterology 1979:77:761-7. 2. Robert A, Nezamis JE. Lancaster C. Hanchar AJ. Cqtoprotection by prostaglandins in rats: prevention of gastric: necrosis produced by alcohol, HCI, NaOH. hypertonic N&i. ant1 thermal injury. Gastroenterology 1979;77:433-43. 3. Miller TA, Jacobson ED. Gastrointestinal cytoprotection by prostaglandins. Gut 1979:20:75-87. 4. Miller TA. Protective effects of prostaglandins against gastric mucosal damage: current knowledge and proposed mechanisms. Am J Physiol 1983:245(Gastrointest Liver Physiol 8):G601-23. 5. Chaudhury TK, Robert A. Prevention by mild irritants of gastric necrosis produced in rats by sodium taurocholate. Dig Dis Sci 1980;25:830-6. 6. Robert A, Nezamis JE, Lancaster C. Davis JP, Field SO. Hanchar AJ. Mild irritants prevent gastric nwrosis through “adaptive cytoprotection” mediated by prostaglandins. Am J Physiol 1983;245(Gastrointest Liver Physiot 8):Gll3-21. 7. Coceani F, Pace-Asciak C. Volta F. Wolfe LS. Effect of nerve stimulation on prostaglandin formation and release from the rat stomach. Am J Physiol 1967;213:1056-64. 8. Bennett A. Friedmann CA, Vane JR. Release of prostaglandin E, from the rat stomach. Nature 1967;216:873-6. 9. Singh J. Prostaglandin release from rat stomach following vagal stimulation or administration of acetylcholine. Em J Pharmacol 1980:65:39-48. 10. Miller TA. Henagan JM. Watkins LA, 1.0~ TM. Prostaglandin induced bicarbonate secretion in the canine stomach: characteristics and evidence for a cholinergic mechanism. J Surg Kes 1983:35:106-12. 11. Hirschowitz B, Sachs G. Vagal gastric secretory stimulation by 2-deoxy-o-glucose. Am J Physiol 1965:209:452-60. 12. Carmichael HA, Nelson LM, Russell RL. Cimetidine and prostaglandin: evidence for different modes of action on the rat gastric mucosa. Gastroenterologg 1978:74:1229-32. 13. Hunter GC. Goldstone J, Villa R, Way LW. Effect of vagotomy upon intragastric redistribution of microvascular flow. J Surg Res 1979:26:314-9. 14. Seifert J, Lenz J, Bruckner W. Brendel L’v. HolIe F. Are 1-~iootl flow Lhanges after selective proximal vagotomy responsible for a necrosis in the gastric wall? Res Exp Med 1980:177:2636. 15. Bell PRF, Battersby C. Effect of vagotomy on gastric mucosal blood flow. Gastroenterology 1968:54:1032-7. 16. Ritchie WP Jr. Acute gastric mucosal damage induced by bile salts, acid and ischemia. Gatroenterology 1975;68:699-707. 17. Ritchie WP Jr, Shearburn EW, III. Influence of isoproterenol and cholestyramine on acute gastric mucosal ulcerogenesis. Gastroenterology 1977:73:62-5. 18. Vane JR. Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs. Nature (New Biol) 1971:231:232-5. 19. Mozsik GY. Moron F. Javor T. Cellular mechanisms of the development of gastric mucosal damage and of gastrocytoprotection induced by prostacyclin in rats. A pharmacological study. Prostaglandins Leukotrienes Med 1982:9:71-84.