Gastrodudenal ulceration following active immunization with prostaglandin E2 in dogs. Role of gastric acid secretion

PROSTAGLANDINS

GASTRODUODENAL ULCERATION FOLLOWING ACTIVE IMMUNIZATION WITH PROSTAGLANDIN E2 IN DOGS. ROLE OF GASTRIC ACID SECRETION Jan Stephen Redfern, Ph.D., Albert J. Blair, III, M.D., Fred J. Clubb, Jr., D.V.M., Ph.D., Edward Lee, M.D., and Mark Feldman, M.D.

From the Departments of Internal Medicine, Surgery, and Pathology, University of Texas Health Science Center at Dallas and the VA Medical Center, Dallas, TX.

ABSTRACT In this study we present evidence to suggest that gastroduodenal mucosal gastric fistula dogs actively immunized with occur in defects may One of four PGE2-immunized dogs developed a PGE2-thyroglobulin conjugate. chronic pyloroduodenal ulcer with penetration into the pancreas and the other three had endoscopic evidence of gastric and/or duodenal erosions. In contrast, no gastroduodenal mucosal defects were seen in control dogs immunized with thyroglobulin alone. Occurrence of gastroduodenal ulcers or erosions was temporally related to formation of specific antibody to PGE2 suggesting that PGE2 antibody may be responsible for lesion formation. An increase in gastric acid secretion was not observed in PGE2-immunized dogs. Thus, it is likely that mucosal defects occur as a result of an impairment of PGE2-mediated mucosal Since gastroduodenal lesions can be visualized by defense mechanisms. endoscopy, the dog may prove to be useful in studying the role of endogenous PG in ulcer diseases. INTRODUCTION Active irmmnization of rabbits with prostaglandin E2-thyroglobulin conjugate induces gastrointestinal ulcers within 6 to 24 weeks (1,2). Ulcers in the gastrointestinal tract of PGE2-immunized rabbits often perforate with resultant peritonitis and death. Ulcer formation in immunized rabbits is most likely a result of anti-PGE2 antibodies since transfusion of unimnunized rabbits with plasma containing high-titer PGE2 antibodies produces acute gastric ulcers within 9 days (2,3). How PGE2 antibodies result in gastrointestinal ulceration in rabbits is uncertain. By binding to endogenous PGEP and negating its physiologic actions, PGE2 antibody~ conceivably could increase-gastric- acid secretion. It is known, for example, that inhibition of gastric mucosal prostaglandin synthesis by indomethacin in humans or dogs increases gastric acid secretion, presumably by removing the inhibitory effect of endogenous antisecretory prostaglandins (4-6). Thus, it is conceivable that binding of antibody to endogenous antisecretory prostaglandins such as PGE2 may also increase gastric acid secretion. However, it is difficult to measure acid secretion -in vivo in rabbits; thus, it was not possible to determine whether immunization with PGE2 leads to gastric acid hypersecretion followed by ulcer formation in that species. In the present experiment, we imnunized gastric fistula dogs with PGE2-thyroglobulin conjugate to try to answer the following questions: (1) will imnunization with PGE2 induce specific, high-titer antibodies to PGE2, as it does in rabbits; (2) will antibody formation in PGE2-imnunized dogs result in an increase in basal or pentagastrin-stimulated acid secretion; and (3) will gastroduodenal ulcers and/or erosions, assessed by serial fiberoptic gastroduodenoscopy, develop in PGE2-immunized dogs?

NOVEMBER

1987 VOL. 34 NO. 5

623

PROSTAGLANDINS

METHODS Materials PGE2, 6-keto PGFIo, PGD2, 15-keto PGE2 and thromboxane 62 (TxB2) were obtained from Upjohn Diagnostics, Kalamazoo, MI. [5,6,8,11,12,14,15 - 3H] PGE2 (specific activity [s.a.] 160 Ci/mnol) was obtained from Amersham, Arlington Heights, IL. Norit A charcoal was obtained from Pfansteihl Labs., Warlhegan, IL and Dextran T70 from Pharmacia Fine Chemicals. New Market, NJ. Xvlene-based scintillation cocktail (Aquasol) was purchased from New England Nuclear, Boston, MA. Xylazine (Rompun) was obtained from Bayvet Div., Miles Labs., Inc., Shawnee, Kansas, and ketamink (Ketaset) from Bristol-Myers Co., Syracuse, N.Y. Prostaglandin B2, thyroglobulin (Type II, porcine), Freund's complete adjuvant and all other reagents (analytical grade) were purchased from Sigma Chemical Co., St.Louis, MO. Imnunization Protocol Six adult male mongrel dogs weighing 18 to 28 kg were used. Four weeks prior to exoerimentation. each doa was orepared with a vaaallv-innervated aastric fistula drained by a Thomas cannula '(7)'to permit subsequent, serial determinations of gastric acid secretion (see below). Then, all 6 dogs were simultaneously entered into a Is-week study protocol similar to the one previously used by us in rabbits (2). The study began with an initial immunization with either 0.5 mg PGE2-thyroglobulin conjugate or 0.5 mg thyroglobulin, emulsified in Freund's complete adjuvant and injected at multiple intradermal sites. Four dogs received PGE2-thyroglobulin conjugate (PGE2 group) and two dogs received thyroglobulin alone (control group). Booster immunizations of PGE2-thyroglobulin or thyroglobulin alone in Freund's complete adjuvant were given subcutaneously at multiple sites 8, 12 and 16 weeks later. Determination of Plasma Antibody Titers to PGE2 in Response to Immunization A 5 ml blood samole from a foreoaw vein was collected in a heoarinized tube at the beginning of 'the experiment ‘and 1 wk after booster injections (i.e., at wk 9. 13 and 17). Plasma was removed and stored at -20°C until PGE:, antibody titers were measured in vitro by incubating overnight at 4OC serial diiutions of plasma (1:3 to 1:32,0m)with 20 fmol 3H-PGE2 in phos hate-buffered saline (pH 7.4) (8). After incubation, bound and free labeled sH-PGE2 were separated by addition of dextran-coated charcoal. After centrifugation at 2000 x g for 20 min. bound counts were determined by liquid scintillation spectroscopy. PGE2 antibody titer was expressed as the reciprocal of that dilution of plasma required to bind 50% of labeled PGE2. Antibody titers were determined without knowledge of the clinical course of the animal. The extent to which antibodies that developed in response to imnunization with PGE2 reacted with PGE2 and with other PGs and TXB2 was determined in each imnunized dog. In these studies, plasma was diluted so that approximately 50% of 20 fmol 3H-PGE2 would ordinarily be bound after overnight incubation at 4°C (see above). Diluted plasma was then incubated overnight at 4°C with 20 fmol of 3H-PGE2 and either unlabeled PGE2, 6-keto PGFIo, PGF2d, PGD2, PGB2, 15-keto PGE2 or TxB2 over a wide range of concentrations (0.2 nM to 1.4 PM). Bound and free 3H-PGE2 was then determined by addition of dextran-coated charcoal as described above. A series of parallel curves was then generated by lotting the log concentration of unlabeled PG standard against the ratio of sH-PGE2 bound at each concentration of standard (B) to 3H-PGE2 bound at zero concentration of standard (Bo). From these curves, the molar concentration of unlabeled PG reducing the B/B0 ratio of antibody binding of 3H-PGE2 from 1.0 to 0.5 was determined (IC50). Cross-reactivity was subsequently calculated by dividing the IC5O for unlabeled PGE2 by the IC5D for each of the other unlabeled PGs or TxB2, multiplied by 1001.

624

NOVEMBER

1987 VOL. 34 NO. 5

PROSTAGLANDINS

Determination of the effective affinity constant for binding of PGE2 to PGE2 antibodv was calculated usinq the Scatchard olot (9). The ratio of bound to free 3H-PGE2-was plotted against the amount of PGE2 bound. Linear regression analysis of the initial slope of this -line yielded the effective affinity constant which was expressed in units of ~-1. Lesion Formation Gastroduodenoscopy using an Olympus PGF endoscope was performed prior to imnunization, at wk 7 after the initial immunization, and 3 weeks after each booster immunization (i.e., at weeks 11, 15 and 19). After an 18-h fast, dogs were sedated with an intramuscular injection of 10 mg/kg xylazine and the esophagus, stomach, and duodenum were examined by two investigators who were Lesions were considered present only if blinded with respect to study group. both observers concurred; all lesions were photographed. Erosions were diagnosed when a defect in the mucosa was present without obvious depth to the lesion. Ulcers were diagnosed when a mucosal defect with definite depth was present. Biopsies of the mucosa were not taken through the endoscope to avoid trauma to the mucosa which could provide a potential nidus for subsequent ulcer formation. At the end of the 19-week study, post-mortem examination of the entire gastrointestinal tract of each dog was performed afte; euthanasia by administration of 75 mglkg intravenous sodium pentobarbital. The pathologist performing the autopsy was unaware of study group assignment. Measurement of Gastric Acid Secretion Acid secretory studies were performed prior to imnunization (baseline) and also at wk 7. 11. 15 and 19. After an 18-h fast, doqs were Dlaced in a Pavlov slinq. and basal secretions were collected from the gastric fistula by gravity drainage for four 15-min periods. Six ug/kg pentagastrin (Peptavlon, Ayerst Labs., New York, NY) was then injected subcutaneously and acid secretion measured for four additional 15-min oeriods. This dose of oentaaastrin is close to maximal in doas (10). The volume of each 15-min sample of gastric juice was measured and bH determined using a glass pH electrode (PHM 82 Standard pH Meter, Radiometer America, Cleveland, Ohio). .pH was converted to hydrogen ion concentration (11) and acid output was calculated by multiplying hydrogen ion concentration by volume. Peak acid output was defined as the sum of the two highest consecutive 15-min outputs after pentagastrin, multiplied by 2 to express results in mnol/h. One of the 4 dogs in the PGE2 group (dog no. 1) developed a large gastric juice leak around his fistula at wk 9 which required surgical closure of the fistula at wk 11. Therefore, acid secretion results in this dog were available only at baseline and wk 7. Another dog in the PGE2 group (dog no. 3) developed a large leak around the gastric cannula at 17 wks which produced severe excorciation of the abdominal wall. This dog was sacrificed at 18 wk, and therefore acid secretion results were available only at baseline and at wk 7, 11, and 15. RESULTS E2 Antibody Production. Plasma PGE2 antibody titers in individual dogs are shown in Table 1 and mean titers are depicted in Fiaure 1. PGE:, antibodv titers gradually increased in PGE2-imnunizid dogs (dog; l-4); dog; 1 and 2 developed PGE2-antibody titers of 9900 and 1050 at wk 17, respectively. In the two control dogs immunized with thyroglobulin, PGE2 antibody titers remained undetectable (<3) throughout the study (Table 1 and Figure 1). * One dog was autopsied at week 18 (see below).

NOVEMBER

1987 VOL. 34 NO. 5

625

PROSTAGLANDINS

TABLE 1.

PGE2 ANTIBODY

BASAL ACID OUTPUT

TITER, ENOOSCOPIC APPEARANCE

(BAD) AND PEAK ACID OUTPUT

OF GASTROOUOOENAL

(PAO) In mmol/h

7

AND

IN PGE2-It+lUNIZEO AND CONTROL

TIME (WEEKS) AFTER BEGINNING DOG t

MUCOSAEa,

DOGS

IHJNIZATION

0

5

PGE2 TITER

t3

<3

ENOOSCOPY

NL

NL

NL

b

OU

0.0138.2

0.0132.4

b

b

b

9

11

13

15

17

19

PGE,-IMMUNIZED

BAOlPAO

21

150

PGE2 TITER

<3

ENOOSCOPY

NL

NL

NL

4.7126.4

0.0117.4

0.9121.4

GAO/PA0

<3

135

180

1050 GE

GE

0.0117.6

0.0121.2

PGE2 TITER

t3

ENOOSCOPY

NL

NL

DE

NL

c

0.0138.0

5.6143.6

0.8135.8

2.5162.4

c

BAOIPAO

t3

15

PGE2 TITER

<3

ENOOSCOPY

NL

NL

0.0132.0

0.0140.8

t3

630

450

24

315 GE,OE

150 GE.OE

0.0134.6

NL

0.0132.8

0.0139.4

CONTROLS 5

PGE2 TITER

<3

ENOOSCOPY

NL

NL

NL

NL

NL

0.0115.0

0.0124.6

0.0129.2

0.0128.2

2.0117.0

GAO/PA0 6

t3

<3

PGE2 TITER

<3

ENOOSCOPY

NL

NL

NL

1.0128.6

0.0142.2

0.0136

BAOIPAO

NL = normal; OU - duodenal

<3

<3

<3

ulcer; GE = gastric erosions:

<3

t3

DE = duodenal

t3 NL

NL 9.3147.2

0.0151.6

erosions.

Dog 1 had gastric fistula

removed at wk 11; therefore, BAO and PA0 were not measured

ter and endoscopy

at week

15 was omitted to allow fistula site time to heal.

Dog 3 had gastric

leak at wk 17 with severe skin excorciatlon

and was sacrificed;

at wk 11 and thereaf-

therefore,

endoscopy.

and PA0 were not done at wk 19.

626

NOVEMBER

1987 VOL. 34 NO. 5

GAO

PROSTAGLANDINS

3000’ [I e i=

300-

1

30-

2 g 0 2 w

I

CONTROL

94 3-

0

WEElk TIME

I

l3EGi AFTER

IMM”d:ZATION NNlNG

17 -

Figure 1. Mean plasma PGE2 antibody titers after beginning imnunization of dogs with PGE2-thyroglobulin conjugate or thyroglobulin alone (control). The extent to which plasma PGE2 antibodies cross-reacted with other PGs and TxB2 is shown in Table 2. For the most part, PGE2 antibodies cross-reacted
::; ;::

DOG #2
DOG #3

DOG # 4

10.4 14.9 0.8 0.1


:::

* Percentage cross-reaction was calculated from molar concentrations producing 50% inhibition of 3H-PGE2 binding to plasma antibody (see text).

NOVEMBER

1987 VOL. 34 NO. 5

627

PROSTAGLANDINS

Lesion Formation (Table 1). In PGE2-imnunized dogs, dog 1 developed a large ulcer at wk 19. The ulcer was thought by the endoscopists to be in the proximal duodenal bulb. Subsequent histological examination of the ulcer at autopsy revealed the ulcer to begin at the pylorus and extend 1.3 cm distally into the duodenum. The ulcer was 0.3 cm deep with slightly raised margins. On cut section (Figure 2), the ulcer had penetrated into the pancreas. No ulcer had been seen in this dog at wks 7 or 11.

Figure 2. Histological appearance of a chronic pyloroduodenal ulcer in a dog 19 wk after beginning immunization with PGE2. The ulcer had penetrated into the pancreas. None of the other 3 dogs imnunized with PGE2 developed endoscopically visible gastric or duodenal ulcers per se, but each dog had endoscopic evidence of gastric erosions, duodenal erosions or both at wk 11 or thereafter (Table 1). Dog 2 had endoscopically visible gastric erosions at weeks 15 and 19, although they were not obvious at the time of autopsy. Dog 3 had duodenal erosions at week 11 but a normal endoscopic examination at week 15; the stomach was also grossly normal at autopsy at week 18. Dog 4 had transient gastric and duodenal erosions between weeks 11 and 15 but a normal endoscopic examination at week:l9, as well as a normal autopsy. In control dogs immunized with thyroglobulin alone, no lesions were visible through the endoscope at any time during the 19 wk study period (Table 1). Acid Secretion. Immunization with PGE2 did not affect basal or peak acid output compared to baseline values prior to beginning immunization. Furthermore, there was no significant difference in mean basal or pentagastrin-stimulated acid output between PGE2-irnnunized dogs and control dogs (Table 1, Figure 3).

628

NOVEMBER

1987 VOL. 34 NO. 5

PROSTAGLANDINS

CONTROL

40 35

&

+

30

,‘

25 I/

P/IO

P % ‘“1 E 15 s

10

CONTROL

PGE,

5

d

0 0

>SAO 19

Figure 3. Mean basal acid output (BAO) and peak acid output (PAO) before and after beginning immunization of dogs with PGEZ-thyroglobulin conjugate or thyroglobulin alone (control). DISCUSSION

To our knowledge this study presents the first experimental evidence that gastroduodenal mucosal defects may occur in dogs actively immunized with PGEp-thyroglobulin conjugate. One of four PGEZ-imnunized dogs (and the one with the highest anti-PGEZ antibody titer) developed a pyloroduodenal ulcer and the other three PGEZ-immunized dogs had endoscopic evidence of gastric and/or duodenal erosions, even though deep, chronic ulcers per se did not develop. Gastroduodenal erosions were usually transient in nature and were first seen endoscopically 11 weeks after beginning immunization (i.e., 3 weeks after the first booster immunization). In contrast to the damage to the gastroduodenal mucosa dogs with seen in immunized PGEZ-thyroglobulin conjugate, no gastroduodenal ulcers or erosions were seen in control dogs imnunized with thyroglobulin alone. Therefore, neither the thyroglobulin moiety of the PG-thyroglobulin conjugate nor the immunization procedure per se was associated with the formation of gastroduodenal mucosal defects. The endoscopic, gross pathologic and histopathologic features of the chronic pyloroduodenal ulcer which occurred In one of our dogs imnunized with PGE2 was similar to chronic peptic ulcers in humans. For example, the ulcer had penetrated to underlying pancreas, and at the base and margins of the ulcer the four characteristic histologic zones were apparent: acute inflamnatory exudate, fibrinoid necrosis, granulation tissue and scar tissue. In view of the resemblance in appearance of chronic gastric ulcers in both PGEZ-imnunized dogs and rabbits to gastric ulcer disease in man, it is possible that iimiunization with

NOVEMBER

1987 VOL. 34 NO. 5

629

PROSTAGLANDINS

PGE2 may provide a new experimental model for peptic ulceration which closely resembles the human disease. Furthermore, since gastroduodenal lesions in dogs inznunized with PGE2 could be visualized by endoscopy without sacrificing the animal, this species may prove to be useful in studying the pathophysiology of ulcer formation and evaluating agents which might prevent ulcer formation. The mechanism of ulcer formation in one of the dogs actively imnunized with PGE2 is not clear. However, the development of gastroduodenal mucosal lesions in dogs itnnunizedwith PGE2 was temporally related to the production of antibody For example, the gastroduodenal mucosa was normal in appearance at to PGE2. seven weeks when PGE2 antibody titers were undectable or very low, but at 11 weeks and thereafter when PGE2 antibody titers were increasing, gastroduodenal lesions were often present. These findings in dogs confirm previous studies in rabbits showina aastrointestinal ulcer formation followina active irnnunization with PGs (1,2): -Antibodies to PGE2 in dogs following itnmiization were similar in specificity and affinity to antibodies to PGE2 produced in rabbits, although olasma antibodv titers to PGE:,were. on the averaae, considerablv lower in doas than in rabbits. It is likely, therefore, that the-gastroduodenal ulcer in one of the dogs immunized with PGE2 occurred as a result of production of antibody to endogenous PGE2 within the gastroduodenal mucosa. Evidence to support the hypothesis that PGE2 antibody per se was responsible for ulcer formation is the observation that passive imnunizaEon of rabbits with PGE2-hyperimnune plasma induced gastric ulcers within g days (2,3). Theoretically, antibody to PGs may have induced ulcer formation by binding to endogenous PGs which normally inhibit gastric acid secretion. There are several studies suggesting that inhibition of gastric mucosal PG synthesis by nonsteroidal antiinflammatory compounds such as indomethacin increases basal and maximal acid secretion by removing the inhibitory effect of endogenous antisecretory PGs (4-6). However, the role of endogenous PGs in the physiologic regulation of gastric acid secretion in dogs is unclear since inhibition of PG synthesis by indomethacin has also been shown to decrease gastric acid secretion (12). In our experiments using chronic gastric fistula dogs, neither basal nor pentagastrin-stimulated acid output increased during itnnunizationwith PGE2. It seems unlikely, therefore, that hypersecretion of acid in PGE2-inunized dogs was responsible for the development of gastroduodenal mucosal defects, although It is apparent that gastric gastric acid may have played a permissive role. acid is not exclusivelv reauired for ulcer formation since immunized rabbits often develop ulcers in-the jejunum, ileum and colon, especially when anti-PGE2 titers are high (> 3000) (2). The lack of effect of PGE;, immunization on gastric acid secretion in dogs-suggests that endogenous PGE2 may not be a physiologic regulator of acid secretion in this species. However, it is conceivable that higher PG antibody titers may be required before a putative inhibitory effect of endogenous gastric mucosal PGs on gastric acid secretion is negated. An alternative mechanism to explain the formation of an ulcer in one dog after imnunization with PGE2 is that antibodies, after binding endogenous PGE2 within the gastrointestinal mucosa, may have impaired the ability of endogenous PGE2 to mediate mucosal defense mechanisms (13-15). For example, binding of endogenous PGE2 by anti-PGE2 antibody may have impaired gastroduodenal mucous and bicarbonate secretion or may have reduced the nonwettable, hydrophobic lining of the gastrointestinal epithellum, effects which could have impaired the ability of the mucosa to limit the influx of luminal acid-pepsin into the mucosa and which could have subsequently led to erosion and ulceration. Furthermore, is a vasodilator of the gastric microcirculation and anti-PGE2 may have PGE2 resulted in vasoconstriction with reduced delivery of nutrients and oxygen to mucosal cells, thereby decreasing the capability of the mucosa to buffer and remove luminal acid entering the mucosa (16).

630

NOVEMBER

1987 VOL. 34 NO. 5

PROSTAGLANDINS

The occurrence of gastric and duodenal ulcers in dogs and rabbits imnunized against PGE2 raises the possibility that a potential mechanism for peptic ulceration in man may be the spontaneous occurrence of specific antibodies directed against endogenous PGs. However, in a previous study in which this hypothesis was evaluated. we found that serum from 45 ulcer oatients (34 duodenal and 11 gastric) had -a low degree of binding of radiolabeled PGE2; PGF2o and 6-keto PGFIo which was not statistically different from PG binding to serum from control subiects (171. Furthermore. this low deoree of bindina of labeled PGs to serum of-peptic ulcer patients and control subjects did not-represent specific binding to very low concentration of serum antibodies to PG since PG binding to serum could not be displaced by addition of an excess of corresponding unlabeled PG. Thus, there is no evidence as yet that gastroduodenal ulcer disease in humans is caused by spontaneous circulating antibodies against endogenous prostaglandins. ACKNOWLEDGEMENTS This work was supported by grants from the American Society for Gastrointestinal Endoscopy, the National Institutes of Health (AM 16616), and the Veterans Administration. The authors wish to thank M. Dale Prince, Teresa Harrell and Julie Oliver for technical assistance, Vicky Robertson for manuscript preparation and Pat Ladd for illustrations. REFERENCES 1.

Olson, G.A., C.W. Leffler, and A.M. Fletcher. Gastroduodenal ulceration in rabbits producing antibodies to prostaglandins. Prostaglandins 29:475. 1985.

2.

Redfern, J.S., A.J. Blair, E. Lee, and M. Feldman. Gastrointestinal ulcer formation in rabbits immunized with prostaglandin E2. Gastroenterology (in press).

3.

Redfern, J.S., E. Lee, and M. Feldman. Induction of acute gastric ulcers in rabbits by passive transfer of PGE2-hyperirnnune plasma. Gastroenterology z:1591 (abstract). 1987.

4.

Levine, R.A., and E.H., Schwartzel. Effect of indomethacin on basal and histamine stimulated human gastric acid secretion. Gut g:715. 1984.

5.

Feldman, M, and T.J. Colturi. Effect of indomethacin on gastric acid and bicarbonate secretion in humans. Gastroenterology g:1339. 1964.

6.

Gerkens, J.F., D.G. Shand, C. Flexner, A.S. Nies, J.A. Oates, and J.L. Data. Effect of indomethacin and aspirin on gastric blood flow and acid secretion. J. Pharm. Exp. Ther. 203:646. 1977.

7.

Redfern, J.S., R. Thirlby, M. Feldman, and C.T. Richardson. Effect of pentagastrin on gastric mucosal histamine in dogs. Am. J. Physiol. 248:6369. 1965. -

a. Redfern, J.S., E. Lee, and M. Feldman. mucosal prostaglandins Gastroenterology g:966.

in humans: 1967.

NOVEMBER 1987 VOL. 34 NO. 5

Effect of indomethacin on gastric correlation with mucosal damage.

631

PROSTAGLANDINS

9.

Scatchard, G. The attractions of proteins for small molecules Ann. N.Y. Acad. Sci. 51:660. 1949.

and ions.

10. Thlrlby, R.C., and M. Feldman. Effect of chronic sham feeding on maximal gastric acid secretion in the dog. J. Clin. Invest. 2:566. 1984. 11. Moore, E.W., and R.W. Scarlata. The determination of gastric acidity by the glass electrode. Gastroenterology e:178. 1965. 12. Nicoloff, D.M. Indomethacin. Effect on gastric secretion, parietal cell population, and ulcer provocation in the dog. Arch. Surg. x:809. 1968. 13. Allen, A., S. t&Queen, and A. Garner. Gastroduodenal mucus: the effect of prostaglandins and role in mucosal protection. In: Biological Protection with Prostaglandins, Vol. 2. (M.M. Cohen, ed.) CRC Press, Inc., 8oca Raton, Florida, 1986. p. 77. 14. Konturek, S., J. Tasler, J. Bilski, and J. Kania. Prostaglandins alkaline secretion from oxyntic, antral and duodenal mucosa of the dog. J. Physiol. 245:6539. 1983.

and Am.

15. Lichtenberger, L.M., J.E. Hills. Richards, and B.A. Effect of 16,16-dimethyl prostaglandin E2 on the surface hydrophobicity of aspirintreated canine gastric mucosa. Gastroenterology g:308. 1985. 16. Miller, T.A. Protective effects of prostaglandins against gastric mucosal damage: current knowledge and proposed mechanisms. Am. J. Physiol. 245:G601. 1983. 17. Redfern, J.S., and M. Feldman. Lack of specific binding of prostaglandin E2, prostaglandin F2o. and 6-keto prostaglandin Flo to serum in patients with peptic ulcer disease and in healthy subjects. Gastroenterology 9l_:71. 1986. Editor: A. Dubois

632

Received: 6-10-87

NOVEMBER

Accepted: 9-13-87

1987 VOL. 34 NO. 5