Evidence that adaptive cytoprotection in rats is not mediated by prostaglandins

GASTROENTEROLOGY

1988;94:948-54

Evidence That Adaptive Cytoprotection in Rats Is Not Mediated by Prostaglandins C. J. HAWKEY, R. T. KEMP, R. P. WALT, N. K. BHASKAR, J. DAVIES, and B. FILIPOWICZ Department

of Therapeutics,

University

Hospital, Nottingham,

United Kingdom

(15) and it has been proposed that adaptive cytoprotection is due to stimulation of endogenous prostaglandin synthesis (g-12). The main pieces of evidence for this proposition are first, that some mild irritants can be shown to stimulate endogenous prostaglandin synthesis (9,10,12) and second, that adaptive changes do not occur, at least to the same extent, after indomethacin pretreatment (g-12). There is, however, conflicting evidence. Mild irritants that are protective have not always been shown to stimulate endogenous prostaglandin synthesis (12,16,17). Reversal of adaptive changes by indomethacin has not been universally reported and has often only been partial (7,9,10). Such observations raise doubts about whether adaptive cytoprotection is critically dependent on prostaglandin synthesis. Several studies have shown that concentrations of ethanol in the range of 40%70% cause extensive desquamation of surface epithelial cells, with formation of a gelatinous mass of mucus, plasma, fibrin, and trapped exfoliated cells (13,18,20). This layer seems to be important for both rapid restitution of the surface epithelium (20) and enhanced protection here is now abundant evidence that the gastric against a second challenge (13), as neither process mucosa can increase its resistance to injury occurs after removal of the gelatinous mass. Concenwhen challenged repeatedly with potentially harmtrations of ethanol that are in the cytoprotective ful agents (1-14). This property has been shown with range have also been shown to disrupt the surface many agents, including ethanol, acid, alkali, hyperepithelium (21-24), and it is possible that this too osmolar solutions, bile acids, and nonsteroidal leads to a covering layer over the epithelial surface antiinflammatory drugs, and over periods of time that could be responsible for all or part of the process ranging from a few minutes to several weeks. of adaptive cytoprotection. Much of the work concerning acute adaptive We have therefore investigated whether adaptive changes has been carried out by Robert and his cytoprotection can occur when endogenous prostacoworkers (7,10,12). They have shown that exposure glandin synthesis is depressed profoundly by treatto mild irritants (e.g., 10°&250~ ethanol, 0.05-0.075 ment with indomethacin. We have also observed the M NaOH, 2%--5% NaCl, and 0.1-0.35 N HCl) and effects of cytoprotective concentrations of ethanol 5-20 mM taurocholate leads to a reduction in the macroscopic signs of mucosal injury caused by subsequent exposure to necrotizing concentrations of these agents (7,10,12). Macroscopically this process, Abbreviation used in this paper: PGE,, prostaglandin E2 described as adaptive cytoprotection, resembles the 0 1988 by the American Gastroenterological Association 0016-5085/88/$3.50 protection conferred by prostaglandin pretreatment

We investigated the proposition that adaptive cytoprotection (enhanced gastric mucosal integrity) caused by exposure to 20% ethanol in rats is mediated by endogenous prostaglandin synthesis. Experiments were performed both without and with indomethacin sufficient to inhibit ex vivo release of prostaglandin Ez by 56% -C9% (2.5mg/kg) or 86% k 4% (10mg/kg). Twenty percent ethanol significantly reduced both macroscopic and histologic signs of injury caused by subsequent exposure to 100% ethanol whether or not the rats were pretreated with indomethacin. Twenty percent ethanol itself caused extensive superficial disruption of the surface epithelium, so that about half of the mucosa was covered by desquamated debris, but did not stimulate prostaglandin release. We conclude that adaptive cytoprotection still occurs after doses of indomethacin sufficient to cause a profound depression of prostaglandin E2 release. The possibility that it could be due to formation of a protective covering of surface debris is worth further investigation.

T

ADAPTIVE CYTOPROTECTION

April 1988

both on the release of prostaglandin histologic appearances.

E2 (PGE,) and on

Materials and Methods The investigation

was carried

out in two main

parts.

Demonstration

of Adaptive

Cytoprotection

Male Wistar rats, housed in individual cages, were fasted for 20 h and deprived of water for 4 h before oral dosing by gavage with indomethacin (2.5 or 10 mg/kg)in 1 ml of 0.9% sodium chloride containing Tween 80 (1 drop per 20 ml) or with this vehicle alone. After 1 h a priming dose of 1 ml of 20% (vol/vol] ethanol (mild irritant) or water (control] was introduced into the stomach by gavage and followed, 15 min later, by a challenge dose of 1 ml of 100% ethanol or water. After a further hour, the rats were killed by COz narcosis. The stomach was removed and opened along the greater curvature, and injury to the gastric corpus was quantified. Macroscopic changes. A coded color photograph of each stomach was projected onto paper; necrotic red lesions and macroscopically normal areas of the gastric corpus were delineated, cut out, and weighed. Macroscopic necrosis was expressed as a proportion of the total area of the gastric corpus. A full-thickness strip of each Histologic changes. stomach wall was taken from a standard site, along the maximum length of the greater curvature, from fundus to antrum regardless of macroscopic appearances. After fixing in formalin, coded paraffin sections were stained with hematoxylin and eosin and simple morphometric techniques (25) were used to measure the total mucosal height, the height of the mucosa appearing histologically normal, and the proportion of the surface mucosa showing discontinuity or disruption. The total mucosal height was quantified using a graticule to measure the perpendicular distance along continuous mucosa (regardless of cellular appearance) from muscularis to surface (lo-25 measurements for each section). Obviously cross-cut mucosa was avoided. Similar measurements from muscularis to the first point of histologic abnormality were used to quantify the height of surviving mucosa. The mucosa was considered histologically normal up to the point where the first cell wjth a pyknotic nucleus or glassy cytoplasm was encountered [unless this was an isolated abnormality) or where there was intramucosal hemorrhage or epithelial disruption. The proportion of the surface epithelium showing discontinuity or disruption was also quantified by graticule measurement. Mechanism

of Adaptive

Cytoprotection

In these experiments rats were prepared as above. They were killed 15 min after receiving 20% ethanol to assess the effect of this cytoprotective maneuver on ex vivo release of PGEz and on histologic appearances. Ex vivo prostaglandin release. The gastric corpus was rapidly dissected and small fragments (3-4 mg each,

949

total 60-120 mg) were incubated for 15 min at 37°C in a sealed tube containing 700 ~1 of gassed Krebs’ solution (95% 02-5% COz) rotating at 30 rpm on a mixing wheel. Indomethacin (10 pg/ml, 28 PM final concentration) was added to the supernatant at the end of the experiments and PGEz was measured by radioimmunoassay (26). Effects on histologic appearance. In these experiments only the highest dose of indomethacin (10 mg/kg)or vehicle alone were used for pretreatment. The rats were killed 15 min after receiving 20% ethanol or water. Coded sections of gastric mucosa were prepared and assessed as described above.

Statistical

Methods

Experiments were conducted to a balanced design so that each treatment condition under evaluation was represented on each experimental day. The significance of differences was assessed by analysis of variance.

Results Demonstration

of Adaptive

Cytoprotection

Macroscopic changes. After 100% ethanol, + 8% (mean and SEM) of the gastric corpus showed macroscopic evidence of mucosal hemorrhage and necrosis similar to that previously reported by Robert and others (9,10,15). Prior application of 20% ethanol reduced the extent of these lesions by 71% -+ 8%, to 12% + 5% of the surface area (n = 10, p < 0.001, Figure la). This protection by 20% ethanol also occurred in rats pretreated with 37%

indomethacin. Priming with 20% ethanol led to a 75% ? 4% reduction in macroscopic damage in the presence of 2.5 mg/kg of indomethacin (n = 10, p < 0.001) and to a 41% + 16% reduction in the presence of 10 mg/kg of indomethacin (n = 10, p < 0.001, Figure la). These adaptive changes were highly significant regardless of indomethacin pretreatment and there was no difference in the extent of mucosal damage attributable to indomethacin pretreatment. Histologic changes. After exposure to 100% ethanol, 69% ? 5% of the surface epithelium was disrupted or discontinuous and the total mucosal height was reduced from 676 ? 20 to 555 +- 29 pm (n = 9, p < 0.001, Figures lb, lc, and 2). The height of the mucosa appearing histologically normal was reduced from 651 2 27 to 407 2 37 pm (n = 9, p < 0.001). Prior exposure to 20% ethanol substantially reduced the depth of damage caused by 100% ethanol, so that the total height of the surviving mucosa was 607 + 28 pm (n = 9, p < 0.001) and the height of the mucosa appearing histologically normal was 545 2 38 pm (n = 9, p < 0.001, Figure lb). There was a slight but significant reduction in the extent of surface epithelial disruption, from 69% ? 9% to 52% 2 9% of the surface length (p < 0.05, Figure lc). In rats pretreated with 2.5 or 10 mg/kg of indomethacin,

950

HAWKEY

MACROSCOPIC @I-

ET AL.

GASTROENTEROLOGY

the extent of damage caused by 100% ethanol not significantly different from that seen in rats given indomethacin. In these rats, prior exposure 20% ethanol also resulted in a partial abrogation the histologic changes (Figures lb. lc), which not significantly different from that seen in rats pretreated with indomethacin.

SIGNS OFNECROSIS

a

**

??

tS

??

**

Vol. 94, No. 4

??

40 2 8 e Y dp

Possible

2u-

Mechanisms

was not to of was not

of Adaptive

Cytoprotection

0

ETHANOL 0 0

0 loo*

INDOMETHACIN

ZD% loo*

0 0

0 IO.3

E

#3co

i

d%

m 103

10MGlKC

2.5MClKG

NIL

MIJCOSALDEPTH 7ca

20% ICm

Top

- total

r b

I-

t J OJ,

A.

Ill ETHANOC 0 0 0 lam INDOMETHACIN

20% ml%

III 0 0 0 lcm

III 20% IOm

oom 0 ml%

1m

IOMGlKG

2.5MClKG

NIL

SURFACE DISRUPTION

rc

100

I

01 ETHANOL0

I

I

0

0 1m

INDONYTHACIN

Figure

?? **

”

I

d

20% lan

i

NIL

1. Quantification

II 1Ok

Ex vivo prostaglandin release. Preliminary experiments showed that release of PGE, was approximately linear over 15 min and was inhibited by 75% I 5% (n = 5) by 1 mgiml of indomethacin (2.8 pm) in vitro. Ex vivo release of PGEp after prior exposure to 20% ethanol (293 2 50 pg/mg . 15 min) was not significantly different from control values (301 2 55 pgimg 15 min). After 2.5 mg/kg of indomethacin it was reduced to 136 2 35 pg/mg 15 min (56% * 9% inhibition, n = 9, p < 0.001) and after 10 mg/kg of indomethacin to 60 ? 22 pg/mg 15 min (86% ‘4% inhibition, p < 0.001, n = 9). Similar reductions were seen with rats not exposed to 20% ethanol (Figure 3). Histologic changes. In control rats 10% * 5% of the surface epithelium showed superficial disruption. Priming with 20% ethanol increased this to 50% t 9% (n = 6, p < O.OOl), with formation of a covering of desquamated cellular debris and amorphous material over 53% ? 8% of the mucosal surface (Figures 4 and 5). In rats pretreated with 10 mgikg of indomethacin, 20% ethanol increased surface epithelial disruption from 20% 2 5% to 55% + 11% so that 47% 2 9% of the surface was covered by debris (Figure 5). In these experiments the total mucosal height after exposure to 20% ethanol was 602 + 12 pm compared with control values of 629 + 15 pm (NS). Values for mucosa appearing histologically normal were 543 2 23 and 605 -C 22 pm, respectively (NS, 0.05 < p < 0.1).

III 1%

2.5MGiKG

0 0

0 lan

20% 100%

lOMG/KG

of changes induced by 100% ethanol with and without prior exposure to 20% ethanol. Each point is the mean (with SEM) of nine observations. a. Macroscopic changes: 100% ethanol caused extensive necrosis; this was significantly reduced by priming with 20% ethanol regardless of indomethacin pretreatment. b. Mucosal depth [both “total” and “normal” mucosa) was reduced by 100% ethanol; this was partially reversed by prior exposure to 20% ethanol regardless of indomethacin pretreatment. c. Slight reduction in the surface desquamation caused by the 100% ethanol, after priming with 20% ethanol. *p < 0.05, **p < 0.01, ***p < 0.001 compared with preceding value.

Discussion These results show that priming of the gastric mucosa by exposure to 20% ethanol leads to a reduction in the extent of mucosal injury subsequently caused by 100% ethanol. The protective changes observed were a reduction in the macroscopic signs of deep mucbsal hemorrhage and necrosis, and greater preservation of the height of the remaining mucosa. Protected mucosa also showed a slight increase in the length that was covered by intact surface epithelium at the end of the experiment, although this could have been due to more

April 1988

ADAPTIVE CYTOPROTECTION

951

Figure 2. Effects of 100% ethanol, with and without prior exposure 100% ethanol; gross surface disruption, reduced height histologically normal mucosa (clear arrows). Right: effect of ethanol; the height of surviving mucosa (black arrows) Magnification, x62.

to 20% ethanol, on histologic appearances. Left: control. Middle: of surviving mucosa (black arrows), and reduced height of priming with 20% ethanol on subsequent damage caused by 100% and of histologically normal mucosa [clear arrows) is greater.

effective restitution after injury. Twenty percent ethanol caused superficial epithelial disruption, with formation of a covering of desquamated debris, but it did not stimulate PGEz release. It induced cytoprotective changes even after treatment with indomethacin sufficient to cause profound depression of PGEz release, suggesting that a mechanism other than stimulation of prostaglandin synthesis or release was involved. Our data showing that 20% ethanol does not stimulate ex vivo release of PGEz appear to conflict

with earlier studies showing increased prostaglandin synthesis and release after cytoprotective doses of ethanol (9), sodium hydroxide (10,12), or sodium chloride (12). However, other studies have found cytoprotective concentrations of hydrochloric acid to have no effect on or even cause inhibition of prostaglandin release from the fundus (10,16). Moreover, sodium taurocholate in concentrations shown to be cytoprotective has been demonstrated to depress in vivo PGEz release (17) into the gastric lumen in humans. While some of these discrepancies may

952 HAWKEYETAL.

GASTROENTEROLOGY Vol. 94. No. 4

PROSTACLANDINSYMHESIS ?? Noethanol

. After 2% ethanol

2.5MGlKG

IOMCIKG

Figure 3. Effect of indomethacin and 20% ethanol on PGEZ synthesis. Each point is the mean [with SEM) of nine measurements. Prostaglandin Ez release is inhibited by 56% 2 9% and by 66% f 4% by indomethacin (2.5 and 10 mgkg, respectively) but is not influenced by prior exposure to 20% ethanol. ***p < 0.001.

be due to methodologic differences, there is currently no coherent body of data to support the suggestion that stimulation of prostaglandin synthesis or release is a common feature of agents capable of activating adaptive cytoprotection.

Figure 4. Effects of 20% ethanol alone on morphology. There is extensive, shallow mucosal disruption over much of the mucosa with formation of a layer of cellular debris (arrows). Magnification, x250.

Our demonstration of adaptive cytoprotection in the face of reduced prostaglandin synthesis also appears to conflict with earlier studies, but here too the discrepancies are more apparent than real. In the earlier studies the reversal by indomethacin only occurred with relatively high doses (i',10) and in For exammany instances was only partial (9-11). ple, in rats dosed with indomethacin and then treated with 20% ethanol before 100% ethanol, the level of mucosal damage was still only about half of that shown for rats not protected with 20% ethanol even though prostaglandin release was depressed (9). Similarly, in experiments using sodium hydroxide or hydrochloric acid, the level of injury in animals pretreated with indomethacin and then subjected to a cytoprotective protocol was intermediate between that seen in unprotected animals and in protected animals not pretreated with indomethacin (l&12). Using mucosal potential difference as an index of injury, Miller and Henagan (11) have shown ethanol to have a greater effect at first challenge in rats pretreated with indomethacin than in controls. However, adaptive changes were evident with subsequent challenge in both groups. As a whole, these observations do not support the notion that the degree of prostaglandin synthesis has a precise rela-

ADAPTIVE CYTOPROTECTION

April 1988

*r

*

SURfAcEDISRUPTION

01

ETHANOL 0 INDMTNACIN

NIL

m

0

zos

10 MC/KG

Figure 5. Quantification of changes induced by 20% ethanol. Top: surface disruption. Bottom: cellular debris. Left: no indomethacin. Right: 10 mg/kg of indomethacin 1 h prior to ethanol. Twenty percent ethanol caused epithelial disruption with formation of a layer of surface debris regardless of indomethacin pretreatment. *p < 0.05, **p < 0.01, ***p < 0.001 compared with preceding value.

tionship to the extent of adaptive cytoprotection. If prostaglandin synthesis were critical, indomethacin pretreatment should be associated with enhanced injury rather than partial or even complete reversal of cytoprotective changes. The apparent partial reversal by indomethacin of adaptive cytoprotection may also arise spuriously. Indomethacin can be shown to enhance the initial mucosal injury caused by sodium taurocholate (271, sodium chloride, ethanol, and sodium hydroxide (9-11). Enhanced injury by either the irritant or challege doses, masquerading as reversal of adaptive changes, may thus account in whole or part for earlier findings. Taken in the context of such earlier findings, our data make it most unlikely that prostaglandin synthesis mediates the adaptive changes observed in acute protocols using mild irritants. Whether prostaglandins play any part in chronic adaptive changes occurring as a result of exposure to ethanol (61, bile acids (5,8), aspirin (l-4), or indomethacin (14) can only be guessed at, but it would be hard to believe that an absolute increase in prostaglandin levels over baseline values could be instrumental in the adaptive changes that occur with aspirin or indomethacin. If increases in prostaglandin synthesis do not mediate the changes of adaptive cytoprotection, there must be one or more alternative mechanisms. Our experiments do not determine conclusively what mechanism is involved but allow some discussion of the possibilities. In the protected stomachs

953

the height of the residual mucosa was greater and the extent of macroscopic changes was reduced compared with unprotected stomachs. Macroscopic changes reflect the formation of areas of deep mucosal hemorrhage and necrosis. Although the pathogenesis of these deep lesions is incompletely understood, they are thought to arise because of stasis of blood flow through superficial capillaries. It is believed that prostaglandins may exert some direct “vasculo-protective” influence on these capillaries in the superficial layers of the mucosa (28). Both aspects of the protection observed in our study may thus involve a common mechanism operating at a superficial level in the gastric mucosa. One possibility is that 20% ethanol activates the same processes as are affected by exogenous prostaglandins, either directly, or by a non-prostaglandin-dependent pharmacologic mechanism. An alternative possibility is that the layer of surface debris formed after exposure to 20% ethanol acts as a physical barrier, reducing the rate of diffusion of 100% ethanol from lumen to mucosa at subsequent challenge, so that the concentration of ethanol in the mucosa fails to reach a critical level sufficient to cause either vascular stasis or further substantial cell death or exfoliation. This hypothesis receives some support from similarities to previous work with different models of gastric mucosal injury, where exposure to concentrations of ethanol higher than used in the present study have resulted in formation of a gelatinous mass that was apparently protective to the gastric mucosa (13,18,20). Lacey (13) has shown that exposure of the rat stomach to 70% ethanol for 30-45 s results in extensive surface injury with formation of a gelatinous surface layer. There was much less injury after a second exposure to the same stimulus, but this enhanced resistance to injury was lost after mechanical removal of the gelatinous covering layer. Other evidence for the protective function of such a layer includes a reduced fall in mucosal potential difference with rechallenge and an interference with epithelial restitution when the layer is removed (20). Our data suggest that a similar process may occur at lower concentrations of ethanol and might explain the adaptive changes induced by them. Our experiments have not thus far identified the physicochemical nature of the debris resulting from exposure to 20% ethanol or its relationship to that demonstrated by previous authors after exposure to higher concentrations of ethanol. Further work to investigate directly whether it can act as a barrier to diffusion will be necessary to test our hypothesis that adaptive changes occurring in the face of profoundly depressed prostaglandin synthesis may be due to physical rather than pharmacologic mechanisms.

954

HAWKEY ET AL.

References 1.

Hurley JW, Crandall LA. The effect of salicylates upon the stomachs of dogs. Gastroenterology 1964;46:36-43. 2. Lev R, Siegel HI, Jerzy Glass GB. Effects of salicylates on the canine stomach: a morphological and histochemical study. Gastroenterology 1972;62:970-80. 3. St John DJB, Yeomans ND, McDermott FT, De Boer WGRM. Adaptation of the gastric mucosa to repeated administration of aspirin in the rat. Dig Dis 1973;18:881-5. 4. Bolton JP, Cohen MM. Effect of repeated aspirin administration on the gastric mucosal barrier and cell turnover. J Surg Res 1977;23:251-6. 5. Deregnacourt J, Code CF. Increased resistance of the gastric mucosa to barrier breakers in the rat. Gastroenterology 1979;77:309-12. 6. Ivey KJ, Tarnawski A, Stachura J, Werner H, Mach T. Burks M. The induction of gastric mucosal tolerance to alcohol by chronic administration. J Lab Clin Med 1980;96:922-32. 7. Chaudhury TK, Robert A. Prevention by mild irritants of gastric necrosis produced in rats by sodium taurocholate. Dig Dis Sci 1980;25:830-6. 8. Scheurer VC, Schlegel JF, Kelly DG, Code CF. Chronic bile exposure increases resistance of canine gastric mucosa to bile. Stand J Gastroenterol 1981;16(Suppl 67):205-10. 9. Konturek SJ, Brzozowski T, Piaztucki I, Radecki T, Dembinski A, Dembinska Kiec A. Role of locally generated prostaglandins in adaptive gastric cytoprotection. Dig Dis Sci 1982; 27:967-71. 10. Robert A, Nezamis JE, Lancaster C, Davies JP, Field SO, Hanchar AJ. Mild irritants prevent gastric necrosis through “adaptive cytoprotection” mediated by prostaglandins. Am J Physiol 1983;245:G113-21. 11. Miller TA, Henagan JM. Indomethacin decreases resistance of gastric barrier to disruption by alcohol. Dig Dis Sci 1984; 29:141-g. 12. Robert A. Cytoprotection and adapted cytoprotection. In: Melis GF, Boeve J, Misiewicz JJ, eds. Peptic ulcer disease: basic and clinical aspects. Dordrecht, Boston, Lancaster, U.K.: Martinus Nijhoff, 1985:290-316. 13. Lacy ER, Gastric mucosal resistance to a repeated ethanol insult. Stand J Gastroenterol 1985;2O(Suppl 110):63-72. 14. Kuwayama J, Eastwood G, Miyaka S, Furukawa H. Adaptation of gastric mucosa to repeated indomethacin administration in the rat (abstr). Gastroenterology 1986;90:1506. 15. Robert A, Nezamis JE, Lancaster C, Hanchar A. Cytoprotection by prostaglandins in rats. Prevention of gastric necrosis

GASTROENTEROLOGY Vol. 94, No. 4

produced by alcohol, HCl, NaOH, hypertonic NaCl and thermal injury. Gastroenterology 1979;77:43343. 16. Konturek SJ, Piastucki I, Brzozowski T, et al. Role of prostaglandins in the formation of aspirin-induced gastric ulcers. Gastroenterology 1981;80:4-9. 17. Rees WDW, Gibbons LC, Warhurst G, Turnberg LA. Studies of bicarbonate secretion by the normal human stomach in vivo. Effect of aspirin, sodium taurocholate and prostaglandin E2. In: Allen A, Flemstrom G, Garner A, Silen W, Turnberg LA, eds. Mechanism of gastric mucosal protection in the upper gastrointestinal tract. London: Raven, 1984:119-24. 18. Morris GP. Harding RK, Wallace JL. A functional model for extracellular gastric mucus in the rat. Virchows Arch [B], 1984;46:239-51. 19. Ito S, Lacy ER. Morphology of rat gastric mucosal damage. defense, and restitution in the presence of luminal ethanol. Gastroenterology 1985;88:250-60. 20. Wallace JL, Whittle BJR. Role of mucus in the repair of gastric epithelial damage in the rat. Inhibition of epithelial recovery by mucolytic agents. Gastroenterology 1986;91:603-11. 21. Davenport HW. Ethanol damage to canine oxyntic glandular mucosa. Proc Sot Exp Biol Med 1967;126:657-62, changes of the 22. Dinoso VP, Ming SC, McNiff J. Ultrastructural canine gastric mucosa after topical application of graded concentrations of ethanol. Am J Dig Dis 1976;21:626-32. 23. Eastwood GL, Kirchner JP. Changes in the fine structure of mouse gastric epithelium produced by ethanol and urea. Gastroenterology 1974;67:71-84. studies on 24. Blom H, Helander HF. Quantitative ultrastructure parietal cell regeneration in experimental ulcers in rat gastric epithelium. Gastroenterology 1981;80:334-43. London: Edward 25. Aherne WA, Dunhill MS. Morphometry. Arnold, 1982. is inhibitory to the 26. Hawkey CJ. Evidence that prednisolone cyclooxygenase activity of human rectal mucosa. Prostaglandins 1982;23:397+10. 27. Whittle BJR. Mechanisms underlying gastric mucosal damage induced by indomethacin and bile salts and the actions of prostaglandins. Br J Pharmacol 1977;60:455-60. 28. Guth PH. Gastric blood flow in ethanol injury and prostaglandin cytoprotection. Stand J Gastroenterol 1986; 21(Suppl 125):86-91.

Received December 23, 1986. Accepted November 16, 1987. Address requests for reprints to: Dr. C. J. Hawkey, Department of Therapeutics, University Hospital, Nottingham NC7 2UH, United Kingdom.