European Journal oJClinical Investigation ( 1980) 10,48 1-485

Human gastric mucosal adenylate cyclase activity: effects of various cytoprotective prostaglandins

BERND SIMON & HORST KATHER, Medizinische Universitatsklinik Heidelberg, Gastroenterologische Abteilung, Heidelberg, F.R.G.

Received 14 April 1980 and in revised form 24 June 1980

Abstract. Several prostaglandins prevent ulcer forma- tion (called cytoprotection) by a mechanism other than inhibition of gastric acid secretion. One suggestion is that they increase cyclic AMP in non-parietal cells.

A variety of prostaglandins with potent cytoprotec- tive properties were tested for their capacity to modu- late adenylate cyclase activity in homogenates of human gastric mucosa. Prostaglandin € 2 , prostacyclin (PGI2) and 1 5(S)-methyl-PGE2 stimulated the cyclase in human gastric mucosal biopsy specimens in a dose- dependent manner. Cytoprotective prostaglandins without antisecretory properties such as PGF2P were also able to activate the enzyme system dose-depen- dently. In contrast, cytoprotective prostaglandins such as PGD2, the PGEl-analogue, SC-29333, and the pro- staglandin-like compound CSI did not stimulate human gastric adenylate cyclase. Whereas PGDz did not modulate enzyme activity at all, SC-29333 and Cgj, at concentrations greater than 10 pmol/l, inhibited basal and PGE2-stimulated enzyme activities. These studies suggest that cyclic AMP is not directly related to the cytoprotective effect of prostaglandins, at least in human gastric mucosa.

Key words. Human gastric mucosa, adenylate cyclase activity, cytoprotection, cyclic AMP, prostaglandins, gastric acid secretion.

Introduction Several prostaglandins have been shown in animals to inhibit gastric acid secretion resulting from a variety of stimuli [1-3]. Similar activity has been demonstrated in humans after stimulation with histamine [4], pentagas- trin [5] and food [6, 1, as well as under basal condi- tions.

In addition to their antisecretory effects, certain pro- staglandins of the E-type have been shown to protect

Correspondence: Dr Bernd Simon, Medizinische Univenitatsk- linik Heidelberg. Gastroenterologische Abteilung. Bergheimentr. 58.6900 Heidelberg, F.R.G. 00l4-2972/80/1200-0481502.00 0 1Y80 Blackwell Scientific Publications

rat stomach against erosions and bleeding produced by the administration of either non-steroidal anti-inflam- matory compounds (aspirin, indomethacin) or agents such as absolute ethanol and 0-6 mol/l HCI which produce extensive gastric necrosis [8,9]. This property of Prostaglandins has been termed ‘cytoprotection’.

The mechanism of cytoprotection is still unknown; however, it is separate from the antisecretory effect of prostaglandins, as it may be observed with certain prostaglandins which are not antisecretory (e.g., PGFzb) or at doses which are not antisecretory [lo].

Most of the actions of prostaglandins in biological systems have been shown, at the molecular level, to be mediated by changing levels of cyclic nucleotides. In the following study we have compared the effects of various cytoprotective prostaglandins on the activity of adenylate cyclase in homogenates of human gastric mucosa.

Materials U-[’~P]ATP and []HICAMP were obtained from Radio- chemical Centre, Amersham, Bucks., U.K. Prosta- glandin €2, prostaglandin D2, prostacyclin (PGI2) and 1 5(S)-methyl-prostaglandin E2 were from Upjohn Company, Kalamazoo, Michigan, U.S.A. and Scher- ing AG, Berlin, F.R.G., respectively.

SC - 29333 [( +) - 15 - deoxy - 16a,/.l- hydroxy - 16- methyl]-PGEl methyl ester was obtained from G. D. Searle Company, Chicago, Illinois, U S A . The pro- staglandin-like compound Cg3, 2-hexyl-5-hydroxycyc- lo-pentanheptanoic acid sodium salt (molecular for- mula: CIS HI] 01 Na), was a gift from Dr Maddaus Company, Kiiln-Merheim, F.R.G. The structural formula of Cg) is described elsewhere [ 1 11.

Prostaglandin E2, prostaglandin D2, 1 S(S)-methyl- prostaglandin Ez and SC-29333 were diluted from an ethanolic stock solution (10 mg/ml) to yield the same amount ofethanol in all samples (2.5%) which was also added to the respective controls. Ce3 was dissolved in Tris-HC1 buffer, pH 7.5 (25 mmol/l Tris).

Prostacyclin (as sodium salt) was dissolved in Tris-HC1 (20 mmol/l), pH 9.2. The compound was added immediately before the assay was started.

48 1

482 BERND SIMON & HORST KATHER

All other chemicals and reagents were of the highest grade commercially available.

Methods Stomach tissue was obtained by subtotal gastric resec- tion from six patients suffering from peptic ulcer dis- ease. Scrapings of the gastric mucosa (approximately 2-4 g) were homogenized in a Teflon glass homo- genizer (Zell-Homogenisator, Colora-Messtechnik GmbH, Lorch, Wiirttemberg, West Germany) after addition of 20 ml ofTris-HCI buffer ( 5 mmol/l Tris, pH 7.4), 3 mmol/l MgC12 and 3 mmol/l mercaptoethanol. The homogenate was filtered through a nylon mesh and assayed for enzyme activity. Only freshly prepared homogenates were used.

Full thickness biopsy specimens of the distal antrum were endoscopically obtained from six patients suffer- ing from upper abdominal complaints. Tissue was im- mediately chilled in ice-cold Tris-HCI buffer, pH 7.4, with 3 mmol/l MgCl2 and 3 mmol/l mercaptoethanol (0-2"C), homogenized and assayed for enzyme acti- vity, as described above.

Adenylate cyclase activity was determined by the method of Salomon e f af. [I21 at 30°C. as previously described [ I 3, 141. The incubation mixture contained Tris-HCI (25 mmol/l Tris, pH 8. l ) , 5 mmol/l MgC12,20 mmol/l creatine phosphate, 100 U of creatine kinase per ml. 1 mmol/l cyclic AMP and I mmol/l d 2 P - A T P (40-50 cpm/pmol). The reaction was initiated by adding 20 pI of stopping solution composed of 2"/, (w/v) lauryl sulphate,. 1 mmol/l STP.

2 ooc C .- E

5 1500 n

L n -

C

4J

0

.- c

c a

E 0 I

1000 . 3 E a 501

0

mmol/l cyclic AMP and 40

1 1 1 I I I 1

c '1x10-7 1~10-6 1x10-5 1~10-4 P r o s t a g l a n d i n C o n c .

(moi I I )

The activity of adenylate cyclase was proportional to the amount of protein added up to at least 80 pg per sample. The time courses of basal and prostaglandin- stimulated enzyme activities were linear for up to 20 min. Cyclic "P-CAMP was purified by column chro- matography using Dowex AG 50 W-X 4 and neutral alumina [ 121.

The protein content of the samples was determined according to Lowry et al. [I51 using bovine serum albumin as standard. The activity of adenylate cyclase is given as pmol CAMP formed per mg protein per 15 min.

Adenylate cyclase activities in preparations from antral mucosal biopsies were assayed in triplicate on a paired basis without (basal activity) or with the addi- tion of stimulants. The coefficient of variation aver- aged 3 4 % .

Statistical analysis was done by the Wilcoxon test for paired samples [ 161.

Results Fig. 1 (left part) shows dose-response curves for pro- staglandin €2, prostaglandin FzP and prostaglandin Dr using homogenates of gastric mucosa from six sub- jects.

Basal enzyme activity (Billroth I1 resection material) averaged 500 pmoles of CAMP per mg of protein per 15 min. PGE2 and PGFtp, the prostaglandin compound without any antisecretory property [ 101, caused a dose- dependent increase of enzyme activity with maxima of about 250% and 125% of basal, respectively, at concen- trations of 0.1 mmol/I. The stimulation of enzyme

2 000

1500

1000

500

0

B a s a l

'-0

L 11 1 I 1 I c " 1x10-7 1~10-6 lx1o5 1~10-4

P r o s t a g l a n d l n D2 ( m o l l I I

Fieurc 1. (A) Left part. Dose-response curves for PGEz, PGF2,g and PGD2. Values arc means of six individual experiments carried out with gastric mucosal homogenata from different subjccts. (B) Right part. Effects of increasing concentrations of PGD2 on prostaglandin E2-stirnulated adenylate cyclase activity in human gastric mucosa. Values are means of seven individual experi- ments c a m 4 out with homogcnata from different subjects. The prostaglandin Ezsoncentrations were 1 prnol/l, 10 pmolll respectively.

HUMAN GASTRIC MUCOSAL ADENYLATE CYCLASE ACTIVITY 483

activity by both prostaglandins was half-maximal at about 5 pmol/l in these experiments. None of the PGs produced an inhibition of adenylate cyclase activity, even when tested over a wide concentration range (0.1 nmol/l to 0. I mmol/l).

Prostaglandin D2, however, was ineffective in acti- vating the gastric cyclase, when tested over the same concentration range. In addition, PGD2 (0.1 pmol/l to 300 prnol/l) did not inhibit the stirnulatory effects of prostaglandin E2 on the human enzyme system. The dose-response relationship for PGE2 was identical in the absence and presence of different concentrations of PGDz (right part of Fig. 1).

Fig. 2 (left part) illustrates the effects of PGE2, the PGE,-analogue SC-29333, and the prostaglandin-like compound Cs3 on adenylate cyclase from human gas- tric mucosa. The three compounds were tested at con- centrations varying from 0.1 pmol/l up to 300 pmol/l. In contrast to PGE2, no dose-related stimulation of human gastric adenylate cyclase was observed over this concentration range in the presence of SC-29333 and c83. At concentrations above 10 pmol/l, both compounds inhibited basal enzyme activity. On aver- age about 40-50% inhibition occurred at 200 pmol/l C,U and 100 pmol/l SC-29333.

As opposed to PGD2, SC-29333 and c 8 3 were able to inhibit prostaglandin E2-stimulated adenylate cyclase in a dose-related manner, when tested from 0.1 pmol/l to 300 pmol/l (Fig. 2). Inhibition was detected at 10 pmol/l was was more than 70-80% at 300 pmol/l. Similar inhibiition of prostacyclin-stimulated adeny- late cyclase was observed (not shown).

The guanine nucleotide GTP has been shown to enhance or to facilitate hormonal stimulation of variety of adenylate cyclases [17, 181. The actions of

cn : l o o o t E 1

Table 1. Effect of maximally effective concentrations of various prostaglandins or prostaglandin-like compounds on adenylate cyclase activity in human antral mucosal biopsies in the presence and absence of guanosine triphos-

phate (GTP)

Prostaglandin compounds (0.3 mmol/l)

None PGEz PGI2 1 S(S)-me-PGE? PGF2p PGD2 SC-29333 C81

Adenylate cyclase activity (prnol cAMP/mg protein per I5 rnin)

Without GTP With GTP (0.1 mrnol/l)

260 k 75 250 f 70 650k 100' 64Of 105' 620k 110' 630f 110' 570 k 90' 565 k 90' 540*95* 550f95' 260 k 75 255 f 70 40+ 10' 45+1' 35 f 8' 40+9*

Means kSEM of Seven individual experiments carried

Significantly different (P $0.05) than the correspond- out in triplicate are given.

ing controls.

maximally-effective concentrations of prostaglandins or prostaglandin-like compounds on adenylate cyclase in human antral mucosal biopsies in the presence or absence of GTP are summarized in Table 1.

Basal enzyme activity averaged 260 pmoles of CAMP per mg protein per I5 min which is lower than that in Billroth I1 resection material. This apparent discrepancy might be attributed to the different tissue specimens used or to differences in the frequency of ulcer disease in both groups [ 191.

Stimulation or inhibition of human gastric adeny- late cyclase by different prostaglandins was not depen- dent on the presence of GTP in the assay mixture.

2000 -

1 5 0 0 .

1000.

500.

1 il \-SC 29333

-

Basa l - 0

Figure 2. (A) Left part. Dose-response curves for PGEz. SC-29333 and the prostaglandin-like compound Ca. Values arc means of six individual experiments carried out with gastric mucosal homogenates from different subjects. (B) Right part. Inhibition of PGE~-stimulated adenylate cyclase activity in human gastric mucosal homogenates by SC-29333 and C83. Values arc means of Seven individual experiments carried out with homogenates from different subjects. The prostaglandin E2-concentration was 0.1 mmol/l.

484 BERND SIMON & HORST KATHER

PGE2, PG12, the orally effective PGE2-analogue. 1 S(S)-methyl-PGEz and PGF2p induced identical in- creases of CAMP-formation in the presence and absence of 0.1 mmol/l GTP. In addition, the expres- sion of the inhibitory actions of SC-29333 and CB) on human gastric mucosal adenylate cyclase was the same in both groups.

Discussion

Various prostaglandin compounds and their synthetic analogues have been shown to possess potent anrisec- refory and cyfoprofecfive actions in gastric mucosa. The actions are independent [8], but may be related to the adenylate cyclase/cAMP system. However, the relationship between the prostaglandins and the pro- duction of cyclic AMP in gastric mucosa seems rather complex.

In fractions enriched with parietal cells, prostaglan- dins of the E- and I-type [20-221 inhibited histamine- stimulated cyclic AMP production. This CAMP-lower- ing effect is probably related to the antisecretory effect of these compounds. In biopsy specimens which are relatively poor in parietal cells, prostaglandins were found to increase markedly adenylate cyclase activity in several species including man [23-261. This suggests that cyclic AMP in cell types other than parietal cells may be the mediator ofcytoprotection [S, 231. Accord- ing with this concept is the observation that PGF2p, a prostaglandin analogue without antisecretory proper- ties, stimulates adenylate cyclase activity in human gastric mucosal homogenates.

However, the data presented in this communication argue against a close relationship between cyclic AMP and cytoprotection, at least in man. Prostaglandin D2, although at least 60 times less effective than PGE2, has been shown to prevent the development of nonsteroi- dal anti-inflammatory drug-induced lesions in rats [27]. This compound, in concentrations up to 300 pmol/l, was unable to stimulate adenylate cyclase activity not only in man but also in the rat (unpub- lished observations).

Prostaglandin D2, however, was active in human colonic tissue [28], indicating that the lack of effect of PGD2 in gastric mucosa is not attributable to biologi- cal inactivity of the prostaglandin preparation used. In addition, this prostaglandin E2 isomer did not in- fluence the response of the enzyme to PGE2, suggesting that PGDz does not compete with E-type prostaglan- dins for receptor sites in human gastric mucosa. Our findings are therefore compatible with the concept of distinct receptor sites for PGD2 and PGEz in the human alimentary tract, as has been previously shown in platelets [29].

The prostaglandin E 1-analogue, SC-29333, may strengthen the integrity of the gastric mucosal bamer against the damaging effects of nonsteroidal anti- inflammatory agents such as aspirin, as determined by measurements of the gastric transmucosal potential difference and transmucosal ionic fluxes (301. In addi-

tion, gastric lesions induced by nonsteroidal anti- inflammatory compounds such as indomethacin, by pyloric ligation or by necrotizing agents such as abso- lute ethanol were effectively and dose-dependently in- hibited by the prostaglandin-like compound C8, [3 I ] .

Neither compound stimulated adenylate cyclase activity in human gastric mucosa. In contrast to PGD2, both analogues inhibited basal enzyme activity and reversed the stimulatory action of PGE2 and PGIl (not shown) on human gastric cyclase in concentrations higher than 10 pmol/l.

The finding that different cytoprotective prostaglan- dins or prostaglandin-like compounds possess inverse effects on human gastric mucosal adenylate cyclase, i.e. stimulation (PGE2, PGI2, lS(S)-me-PGEl) or inhibi- tion (SC-29333, Cs,), lends credence to the concept that cyclic AMP is not primarily involved in this aspect of prostaglandin action. The most convincing evidence against cyclic AMP'S being a mediator of prostaglan- din-induced cytoprotection derives from studies with prostaglandin Dz. This compound is cytoprotective in rats (271 but obviously lacks cyclase-coupled receptor sites in rat and human gastric mucosa.

Most of the studies concerning the cytoprotective effects of prostaglandins in gastric mucosa have been done with rats. We cannot therefore exclude that our findings merely reflect species differences. However, the observation that prostaglandin D2 is also ineffec- tive in the rat where cytoprotection has been described supports the contention that the lack of a close correla- tion between cyclic AMP and cytoprotection is a general phenomenon.

Acknowledgments The authors wish to thank Ms Finkensieper for excel- lent technical assistance.

They also are grateful to Dr Andre Robert, The Upjohn Company, Kalamazoo, Mich., U.S.A., and Professor Dr Losert and Dr Radiichel, Schering AG, Berlin, F.R.G., who supplied prostaglandin E2, pro- staglandin D2. prostacyclin (PGIZ) and 1 5(S)-methyl- prostaglandin Ez.

Thanks are also to T. B. Martinez and D. C. Colton, Searle Co., Chicago, Illinois, U.S.A. for supplying the prostaglandin E 1 -analogue, SC-29333.

The prostaglandin-like compound C8] was a kind gift of Fa. Maddaus Co., Koln, F.R.G.

This study was supported by grants from the Deutsche Forschungs-gemeinschaft, Bad Godesberg, F.R.G.

References I Robert A., Nezamis J.E. & Phillips J.P. (1967) Inhibition of

gastric secretion by prostaglandins. Am J Dig Dis 12. 1073-1076. 2 Robert A. (1968) Antisecretory property of prostaglandins. Pro-

staglandin Symposium of Worcester Foundrrtion for Experimental Biology (ed. by P. W. Ramwcll and J. E. Shaw), pp. 47-54. Interxicnce, New York.

3 Robert A., Nczamis J.E. & Phillips J.P. (1968) Effect of prosta-

HUMAN GASTRIC MUCOSAL ADENYLATE CYCLASE ACTIVITY 485

glandin Ez on gastric acid secretion and ulcer formation in thc rat, Gu.rfroc,nrero/~J~pI. 55. 48 1-487.

4 Wilson D.E., Phillips C. & Levine R.A. (1971) Inhibition of gastric acid secretion in man by prostaglandin A I. Gusfroonrero-

5 ClassenM..Koch H..BickhardtJ..TopfH.&DemlingL.(1971) The effect of prostaglandin E? on the pentagastrin-stimulated gastric secretion in man. Digesrion 4, 333-336.

6 Konturek S.. Kwiecien M.. Swierczek J . . Olesky J.. Sit0 E. & Robert A. ( 1976) Comparison of methylated prostaglandin Ez-analogues given orally in the inhibition ofgastric responses to pentagastrin and peptone meal in man. Gu.yfromrerologF 70.

7 Ippoliti A.F.. Isenberg J.1.. Maxell V.J. & Walsh J.H. (1976)The effect of 16. I6-dimethyl-PGE? on meal-stimulated gastric acid secretion and serum gastrin in duodenal ulcer patients. Guyfroen- rerology 70,48849 I.

8 Robert A,. Nezamis J.E. & Hanchar A J . (1979) Cytoprotection by prostaglandins in rats: prevention of gastric necrosis pro- duced by alcohol. HCI. NaOH. hypertonic NaCl and thermal injury. Gu.rirornrerolo.qv 17,433-443.

9 Simon 8. & Kather H . (1980) Der zytoprotektive Effekt der Prostaglandine. Experimentelle Befunde und klinische Bedeu- lung. Z Gusrroenferol 18, 32G324.

10 Robert A. ( 1979) Cytoprotection by prostaglandins. Gusfroen- rerology 77, 76 1-767.

I I Domschke W.. Domschke S.. Hornig D. & Demling L. (1978) Prostaglandin-stimulated gastric mucus secretion in man. Acfu Hepuro-gusrroenr 25, 292-294.

I2 Salomon Y.. Londos C. & Rodbell M. (1974) A highly sensitive adenylate cyclase assay. Anul Biochum 58. 541-548.

13 Simon B.. Czygan P.. Stiehl A. & Kather H. (1978) Human colonic mucosal adenylate cyclase: effects ofbile acids. Eur J Clin

14 Simon B. & Kather H. (1980) Interaction of laxatives with enzymes of cyclic AMP metabolism from human colonic mucosa. Eur J Clin Inrerr 10, 23 1-234.

15 Lowry0.H.. Rosebrough N.F.. FarrA.L.& Randall R.J.(1951) Protein measurement with the fohn phenol reagent. J BivlClrem 193,265-275.

16 lmmich H. (1974) Medi-inische S~u/i.srik. pp. 318-322. Schat- tauer. Stuttgart.

17 Hanvood I.P., Low H. & Rodbell M. (1973) Stimulatory and inhibitory effects of guanyl nucleotides on Fit cell adenylate cyclase. J Biol Chem 248,6239-6245.

18 Simon 8. & Kather H. (1978) Activation of human adenylate

IO*CJ 61, 201-206.

683-687.

Itirrsf a. 32 1-323.

cyclase in the upper gastrointestinal tract by vasoactive intestinal peptide. Gusrroenrerology 74. 722-725.

19 Kather H. & Simon B. (1979) Biphasic effects of prostaglandin E2 on the human fat cell adenylate cyclase. J Clin Inrest 64,

20 Wollin A,. Sol1 A.H. & Samloff I.M. (1979) Actions of histamine. secretin and PGE: on cyclic AMP production by isolated fundic mucosal cells. A m J Phpiol237(5). E 437-E 433.

21 Major J.S. & Scholes P. (1978) The localization of histamine-Hz- receptor adenylate cyclase system in canine parietal cells and i ts inhibition by prostaglandins. Agents und Actions 8, 3%33 I ,

22 Sol1 A.H. & Whittle B.J.R. (1979) Activity of prostacyclin. a stable analogue. 6P-PGIl and 6-0x0-PG FI, on canine isolated- parietal cells. Br J Phurmucol68.97P-98P.

23 Dousa T.P. & Dozois R.R. (1977) Interrelationships between histamine, prostaglandins and cyclic AMP in gastric secretion: a hypothesis. Gusrroenrerology 73. 904-9 12.

24 Perrier C.V. & Laster L. (1976) Action of Hi and Hz-inhibitors on the response of histamine-sensitive adenylyl cyclase from guinea pig gastric mucosa. Eur J Clin I n i m 6. I 13-1 20.

25 Simon B.. Kather H. & Kommerell B. (1978) Effects of prostag- landins and their methylated analogues upon human adenylate cyclase in the upper gastrointestinal tract. Digesfion 17,547-553.

26 Simon B. & Kather H. (1979) Modulation of human gastric mucosal adenylate cyclase activity by prostacyclin. DiRr.rrion 19,

27 Robert A. ( 1976) Antisecretory. antiulcer. cytoprotective and diarrheogenic properties of prostaglandins. Adruncw m Pro.trug- lundin and Tliroinho.rune Reseurcli. Vol. 2. pp. 507-520. Raven Press. New York.

28 Simon B.. Kather H. & Kommerell B. (1980) Activation of human colonic mucosal adenylate cyclase by prostaglandlns. Acliwnces in Pri~.riugIunditi und T/iroinho.rune Reseurth. Vol. 8. pp. 1617-1620. Raven Press, New York.

29 Miller 0. & Gorman R.R. (1979) Evidence for distinct prostag- landin I: and D: receptors in human platelets. J Phurmurol E.rp Tlrrr210. 134-140.

30 Colton D.G.. Callison D.A. & Dajani E.Z. (1979) Effects of a prostaglandin E i derivative. SC-29333, and aspirin on gastric ionic fluxes and potential difference in dogs. J Pharniacol ESP

3 I Long-Peschlow E. (1980) Prevention of experimental gastric lesions in rats by C83. a new prostaglandin analogue. XI Interna- tional Congress of Gastroenterology. 8-13 June 1980. Hamburg. Hepufo-Gusrroenferology. Suppl. E26.16. p. 210.

609-6 I 2.

I 37- 1 39.

T k 210,283-288.