Gut 1995; 36: 657-663 657 Non-steroidal anti-inflammatory drugs and prostaglandin effects on pepsinogen by dispersed human peptic cells Gut: first published as 10.1136/gut.36.5.657 on 1 May 1995. Downloaded from

A I Lanas, J Nerin, F Esteva, R Saiinz

Abstract It is widely accepted that non-steroidal anti- The effects of and on inflammatory drug (NSAID) use is associated pepsinogen secretion were studied in with a high prevalence of gastroduodenal isolated human peptic cells prepared from injury.1 It is also recognised that the presence endoscopically obtained biopsy specimens of acid and pepsin enhances the potential after collagenase , mechanical damaging effect of these drugsl 2 and, there- disruption, and percoli gradient centrifu- fore, anti-secretory therapy is commonly gation. Pharmacological concentrations prescribed to treat or prevent damage induced of aspirin and ibuprofen (10-8-10-4 M), by NSAIDs.3 The pathogenic mechanisms of potentiated histamine (10-6-10-4 M) and NSAID induced gastroduodenal damage forskolin (10-5 M) stimulated pepsinogen remains unclear,l 2 although different mech- secretion without affecting basal secre- anisms, most of them related to prostaglandin tion, (10-6 M) stimulated inhibition, have been proposed.2 4 pepsinogen secretion or cell vitality. Aspirin and other non-salicylate NSAIDs Augmentation of secretagogue stimulated have been reported to increase or potentiate pepsinogen secretion was dependent on histamine stimulated acid secretion in vivo5 6 extracellular calcium because potentia- and in vitro in different animal models,7-11 tion was abolished by calcium depletion of suggesting that enhanced secretion the medium. inhibited the could play a part in the pathogenesis ofNSAID

potentiation effect on histamine but not induced gastroduodenal mucosal damage. The http://gut.bmj.com/ on forskolin stimulated pepsinogen secre- mechanisms of NSAID stimulated acid secre- tion, thus suggesting that this augmenta- tion have been extensively explored by Levine tion was independent of histamine H2 et al 7-9 in an in vitro rabbit model, receptors. Of interest, potentiation was suggesting that NSAID induced acid secretion also independent of endogenous prosta- is regulated by the presence of calcium and the glandin inhibition because exogenous inhibition of prostaglandins.

addition of and D2 Although available information provides on September 28, 2021 by guest. Protected copyright. increased both basal and acetylcholine sufficient evidence that intramucosal activa- stimulated pepsinogen secretion in a tion of acid proteases, mainly pepsinogens, dose dependent way, but they did not plays an important part in peptic diseases of modify histamine or histamine plus the upper ,'2 there are aspirin or ibuprofen stimulated pepsino- almost no data on the potential effects of gen secretion. In conclusion, aspirin NSAIDs on pepsinogen secretion by the and ibuprofen potentiate secretagogue . In this study we have examined in Service of stimulated pepsinogen secretion by vitro the effects of two widely and more com- Gastroenterology, dispersed human peptic celis and this monly used NSAIDs (aspirin and ibuprofen) Universitario, might be an additional mechanism of on pepsinogen secretion by dispersed human Zaragoza, Spain non-steroidal anti-inflammatory drug peptic cells obtained from endoscopic biopsy A I Lanas J Nerin (NSAID) induced gastric injury.M This specimens. F Esteva potentiation effect is regulated by calcium, R Sainz independent of endogenous prostaglandin Correspondence to: inhibition and seems to act on pepsinogen Methods Dr A I Lanas, Servicio de secretion at a post- site. Aparato Digestivo, Hospital. Clinico Universitario, 50009 (GutGul93657631995; 36: 657-663) Chemicals Zaragoza, Spain. Chemicals were obtained from the following Accepted for publication Keywords: non-steroidal anti-inflammatory drugs, sources: bovine serum albumin (fraction V), 16 August 1994 pepsinogen, prostaglandins. acetylcholine, histamine, cimetidine, aspirin, ibuprofen, forskolin, HEPES, percoll, crys- Effects ofaspirin and ibuprofen on basalpepsinogen secretionfrom dispersed human peptic cells talline porcine pepsinogen, soybean trypsin inhibitor, ethylene glycol tetraacetic acid Agent Number 10-4 M 10-6 M 10-8 M (EGTA), and glucose from Sigma Chemical Aspirin 6 -0.05 (0.56)% +0-66 (0.72)% +0 68 (0 70)% (St Louis, MO). Collagenase (type IV) from Ibuprofen 6 +0-14 (0 20)% +0.01 (0 25)% -0 55 (0.27)% Worthington Biochemical (Freeholdd, NJ); Pepsinogen released into the medium was expressed as percentage (%) of the total pepsinogen bovine haemoglobin from Gibco Diagnostics initially present in cells and secretory responses to stimuli were calculated in the output per unit and and D2 time with basal output subtracted. Basal secretion in these experiments was 5.79 (0.58)% of (Madison, WI), prostaglandin E2 total/30 min. Data expressed as mean (SEM). from Calbiochem (La Jolla, CA). 658 Lanas, Nein, Esteva, Sdinz

69 patients undergoing upper gastrointestinal a HIS endoscopy (45 men and 15 women). The 0E 4.0F HIS 104 + NSAID HIS + ASA mean (SEM) age of this population was 39.11 (0 3.5%,-,W F ii (1-59) (18-61). In 20 of these patients the 1- * HIS+ IBU endoscopy was normal, 18 had an active Gut: first published as 10.1136/gut.36.5.657 on 1 May 1995. Downloaded from Qto 3.0 i duodenal ulcer, 14 a healed duodenal ulcer, i nine a hiatal hernia, six oesophagitis, five 0 2.5 duodenitis, five antritis, and two a gastric 1 ulcer. Most of them (55 of 69) were or had 0 2.0 ez recently been receiving H2 receptor antago- 0 nists, and the rest were receiving antacids or 1.5 were free of any treatment. Helicobacter pylori c status in the corpus of the stomach was not 1.0 CD determined, but it was in the antrum (CLO C test) in 43 of 69 patients (positive in 40). The 0.5 ._ study was approved by the Institutional Review c Board and written informed consent was -44-8 -4 -6 -4- -4 -6 obtained from all patients. Molar concentration Figure 1: Effects ofdifferent concentrations ofboth aspirin (ASA) and ibuprofen (IBU) on two different concentrations of histamine (HIS) stimulated pepsinogen secretion from Cell isolation dispersed human peptic cells. Basal secretion in these experiments was 4-86 (0.57) % of The method used to isolate peptic cells had total/30 min. n=4-7. *p<0 05 with respect to histamine stimulated secretion, tp<0 05 with respect to basal secretion. been previously described elsewhere.'3 In brief, eight to 10 endoscopically obtained gastric biopsy specimens (jumbo forceps) from Unless otherwise stated, the standard Ringer the oxintic area were transported in Ringer's solution contained 82.2 mM NaCl, 4.0 mM solution that had been gassed with 95% 02 KCl, 1.8 mM CaCl2, 0.8 mM MgCl2, 0.8 and 5°/O CO2. The specimens then received NaSO4, 17.8 mM NaHCO3, 0-8 mM a 20 minute digestion in 0.1% collagenase NaH2PO4, 11.5 mM glucose, and 0-2% bovine solution with 0.2% trypsin inhibitor in a serum albumin. The medium was equilibrated shaking water bath at 370C. This initial colla- with 95% 02-5% C02 and pH was adjusted to genase solution was discarded and replaced 7-25. Call low medium contained 0.1 mM and digestion continued for 45 more minutes. CaCl2 and no MgCl2, and Ca free medium The cells and were collected by contained no CaCl2 or MgCl2. centrifugation, resuspended in Ringer's solu- http://gut.bmj.com/ tion without calcium and magnesium, and subjected to repetitive pipetting for several Biopsy specimens minutes. The cells were filtered through 100 Gastric biopsy specimens were obtained from ,um Teflon mesh and resuspended in Ringer's solution containing calcium and magnesium. The cells were then taken up in 1 ml of U Ca' Ringer's solution and layered onto 2 ml 50/O on September 28, 2021 by guest. Protected copyright. D (v/v) percoll containing isotonic HEPES Ca' free medium buffered Ringer's solution, which contained 82.2 mM NaCl, 4 mM KC1, 1.8 mM CaCl2, 0.8 mM MgCl2, 19.4 mM HEPES, 11.5 mM 6 * glucose, and 02% bovine serum albumin. The cells were then centrifuged at 14 000 g for 20 E minutes at 4°C (Eppendorf Centrifuge, Model 0 5412). After this centrifugation the bottom of InCo CoCD, the gradient containing mainly red blood cells cc was discarded, and the remaining cells were 4 washed and layered again onto a 2 ml 33% 0 CCo (v/v) containing isotonic HEPES buffered Ringer's solution. After another centrifugation 3 at 14000 g for 10 minutes, the top of the .._J, 0 gradient containing mainly the parietal cells 0cJ was discarded and a clearly different layer a) 2 at the bottom contained the chief cells. The isolated chief cells were greater than 90/o

._ pure by microscopy (haematoxylin and eosin 0.a) .7n0 staining procedure and periodic acid schiff CL reaction), and >90%/O viable by trypan blue exclusion and >89% by flow cytometry (Epics o Elite, Coulter, Hialeah Fl, USA) using the FK FK + ASA FK + fluorochromes, rhodamine 123 and propidium Figure 2: Effects ofboth aspirin (ASA) (10-6 M) and ibuprofen (IBU) (i0-4 M) on iodide to measure live and dead cells respec- forskolin (FK) (10-5 M) stimulated pepsinogen secretion from dispersed human peptic tively.'4 15 The pepsinogen content of this cells. Removal ofextracellular calcium from the medium suppressed the potentiation effect of was 100 times than the both drugs on forskolin stimulated pepsinogen secretion. Basal secretion in these experiments layer higher pepsinogen was 3.40 (0.29) % oftotalJ30 min. n= 4-5. *p< 0. 05 with respect to forskolin. obtained from the top layer containing mainly Non-steroidal anti-inflammatoty drugs and prostaglandin effects on pepsinogen secretion by dispersed human peptic cells 659

* Ringer's solution and counted (haemato- * cytometer). Aliquots of 105 cells were then 200r c placed in 1.5 ml polypropylene test tubes con- Mi- .2 taining the appropriate agents and incubated 0 in a shaking water bath under an atmosphere Gut: first published as 10.1136/gut.36.5.657 on 1 May 1995. Downloaded from CE150 * o~-0 of 95% 02-5% CO2 at 37°C for various I-Cv times. In general, cells from a single patient 00 were sufficient for 9-11 aliquots (including 001 c duplicates). If the experiment was high in cc number ofaliquots or samples, two sets of cells Ooe from two different subjects were put together. X E 50 In this case n=4 means eight people. The 0a._cc experiments were finished by centrifuging the tubes for 17 seconds at 14 000 g and pipetting off the An of the 0 H~- supernatants. aliquot super- HIS HIS HIS FK FK FK natant was assayed for pepsinogen activity + ASA + ASA + ASA + ASA using a modified Anson-Mirsky method using +CIM + CIM acid denaturated haemoglobin as the substrate Figure 3: Effects ofcimetidine (CIM) (2X 10-4 M) on aspirin (10-6 M) potentiation of as previously described.'6 The mixture was either histamine (HIS) (10-4 M) orforskolin (FK) (l0-5 M) stiimulated pepsinogen incubated at 37.C for five hours and the secretion from dispersed human peptic cells. Results are expressed azs percentage of pepsinogen secretion obtained after stimulation with either histamilne orforskolin minus reaction stopped by adding 1 ml of 70/0 basal secretion and set at 100%. Basal secretionfor both histamin,0eandforskolin was trichloroacetic acid solution followed by similar to those values expressed in previousfigures. n= 3-5, *p< *. centrifugation at 13 000 rpm for five minutes. The optical absorbance of the supernatant was parietal cells. No signtificant pepsinogen read at 280 nm. Pepsinogen released into the secretion was obtained irom the top layer medium was expressed as percentage of the when stimulated with acetylcholine or total pepsinogen initially present in cells. octapeptidle. Secretory responses to stimuli were calculated Eight biopsy specimen.s from one young in the output per unit time with basal output patient yielded about 1 06 cells that were subtracted. The mean (SEM) basal rate of suspended in Ringer's solh,ition under 95% 02 pepsinogen secretion in the experiments was 5°/O CO2 atmosphere and refrigerated 0-18 (0-01)% of total pepsinogen content per overnight. The next morining, the cells were minute. The concentrations of aspirin and

allowed to warm at room temperature before ibuprofen used in these experiments are within http://gut.bmj.com/ they were used in experiments. No experi- the range that can be expected in patients who ments were performed if cell viability was take these drugs when prescribed to treat lower than 89% and experiiments with low total rheumatological conditions,'7 18 and they pepsinogen content were dliscarded. do not modify optical absorbance when measuring pepsinogen activity. In some experiments, carried out to study Experiments and calculation.ts the role of calcium in NSAIDs on pepsinogen on September 28, 2021 by guest. Protected copyright. The warmed cells were r(esuspended in fresh secretion, cells were washed once in calcium free Ringer's solutions, resuspended in calcium free Ringer's solution containing 0.5 mM a Ach10 M EGTA for three minutes to remove extra- cellular washed and 2.5 AchlO4M+ ASA Ca++, again, finally EAch 10-M+ IBU resuspended in calcium free Ringer's solution plus the drugs to be tested. cv E 0 .0 2.0 0 1 Cell labelling and detection ofeicosanoid 0 production Synthesis of prostaglandins from endogenous 1.5 cellular sources was measured by radiolabelling cells with 10 1-14C-arachi- U resuspended ,uM 0 donic acid (New England Nuclear, .0 Daimlertrasse, Germany) for 30 minutes. Cells (0 1.0 were resuspended in standard Ringer's solution 0 cJ with 0-2% essentially free bovine 0 serum albumin. After incubation, the super- natant (1 ml) was acidified with 1 M citric CD 0.5 0. acid to pH 3.0 and injected into C18 cartridges 0. (Sep-Pack, Millipore Iberica, Madrid) con- 0L taining octadecylsilyl silica. Elution and detection of metabolites have 0 1- -6 -4 -6 -6 -4 -6 been described elsewhere.'9 In brief, the Molar concentration cartridges were first washed with 3 ml of acidi- Figure 4: Effects ofboth aspirin (ASA) and ibuprofen (IBU) on acetylcholiine (Ach) fied water (pH 3.0) and 3 ml of water, then stimulated pepsinogen secretion from dispersed human peptic cells. Basal secrretion in these with 1 ml of hexane for partial elution of experiments was 5-24 (0.27) % oftotal/30 min. n=3-4. non-polar , and finally with 2 ml ofethanol 660 Lanas, Nerin, Esteva, Sdinz

Results nedium Effects of aspirin and ibuprofen on pepsinogen .2 E 20- secretion Neither aspirin nor ibuprofen had any Gut: first published as 10.1136/gut.36.5.657 on 1 May 1995. Downloaded from X1.5 significant effect on basal pepsinogen secretion 0i 1.5 * (Table), but both drugs potentiated maximal C0 (10-4 M) histamine stimulated pepsinogen

0 , secretion in a dose dependent way (Fig 1). Maximal stimulation was obtained with 10-4 M aspirin (2-96 (0.38)%/30 min) and 10-6 M 0 ibuprofen (3.08 (0.56)0/o/30 min). Cell via- HIS HIS + ASA HIS + IBU bility was not affected by these concentrations Figure 5: Calcium dependence ofaspirin (ASA) (10-4M) as shown by both trypan blue exclusions and ibuprofen (IBU) (10-6 M) potentiation ofhistamine and the fluorchromes rhodamine 123 and (HIS) (10-4 M) stimulated pepsinogen secreticmfrom propidium iodide (89.9 (2.7)% of viable cells dispersed human peptic cells. Removal ofextrac calcium from the medium suppressed the potentitellilarefect in controls v 90.8 (2.5)% with 10-4 M aspirin ofaspirin and ibuprofen. Basal secretion in thesRe and 90.8 (2.7)% with 10-6 M ibuprofen; experiments was 5.19 (0.6) % oftotal/30 min. n=i4-6, n=5). Potentiation of pepsinogen secretion *p<0005. was also seen when peptic cells were pre- stimulated with lower doses of histamine for elution of arachidonic acid metabc)lites. The (10-5 M) and reinforces the physiological ethanol fraction was evaporated t(o dryness significance of these effects. Like histamine, under N2, then dissolved in 7 5 p.l of pepsinogen secretion stimulated by forskolin chloroform-methanol (2:1) and applied (10-5 M), an intracellular secretagogue that onto silica gel 60 thin layer chromLatography activates the catalytic subunit of adenylate plates (E Merck, Darmstadt, Csermany). cyclase, was also potentiated by both aspirin The plates were developed using tile solvent and ibuprofen (Fig 2). Cimetidine (2x10-4 system - ethylacetate:acetic acid:isooctane: M) blocked aspirin potentiation of histamine water (110:20:50:100) - for separation of stimulated pepsinogen secretion but did not cycloxygenase products. Five ,ug each of inhibit significantly aspirin effects on forskolin authentic standards of prostaglandin ]E2 and D2 stimulated pepsinogen secretion (Fig 3), show- were also applied onto each tihin layer ing that cimetidine inhibited histamine stimu- chromatography plate, and were visiaalised by lation in itself rather than the potentiation of exposure to iodine vapour. Each bandi was then aspirin. http://gut.bmj.com/ scraped offinto scintillation vials, mixted with 10 Aspirin and ibuprofen (10 -i- 0 8 M) failed ml of scintillation cocktail, and couinted in a to potentiate acetylcholine stimulated pep- scintillation counter (Packard) after ione hour. sinogen secretion suggesting that the potentiat- Results are presented as mean (S;EM) and ing effect of these two drugs may be mediated the number (n) in each experiment is equal to by mechanisms that share some intracellular the number of separate cell parations. pathway (for example, Ca++) but do prel not on September 28, 2021 by guest. Protected copyright. Statistical significance was deternnined by change muscarinic related responses (Fig 4). Student's t test for unpaired values aand either equal or unequal variances, and p< o0o05 was taken to represent a significant differ(ence. Role ofcalcium in aspirin and ibuprofen potentiation ofpepsinogen secretion Aspirin and ibuprofen potentiate histamine and forskolin but not acetylcholine stimulated pepsinogen secretion showing that these drugs M PGE2 may activate mechanisms of action different E 40 - PGD2 from those used by histamine or forskolin and c similar to those used by acetylcholine. To a I* 0 3.5 co study whether such mechanisms were calcium -0 * * dependent, cells were deprived of extracellular 3-0 0 calcium and then stimulated. Under these 2.5 1 conditions aspirin and ibuprofen completely 4-0 failed to potentiate either histamine or 2-0 forskolin stimulated pepsinogen secretion (Figs 2 and 5). 0) 0 1-5 ._0 C 1-0 Co Role ofprostaglandins in aspirin and ibuprofen a)0. C0 potentiation ofpepsinogen secretion Q 0-5 0~ NSAIDs inhibit prostaglandin secretion in 0 I I + + gastric cells8 9 and prostaglandins inhibit ~6 -7 -8 -9 -6 -8 _9 gastric acid secretion in vivo and in vitro.2 11 Molar c(oncentration This could be one mechanism to explain the potentiation effects of NSAIDs on Figure 6: Prostaglandins (PGE2 and PGD2) stimiulate pepsinogen secretion from dispersed pepsinogen human peptic cells. Basal secretion in these experimrents was 6-33 (0-60) % oftotal/30 min. secretion. In our system, we have seen that n=4-5, *p<0-05. the addition of 10 -4 M aspirin inhibited Non-steroidal anti-inflammatory drugs and prostaglandin effects on pepsinogen secretion by dispersed human peptic cells 661

humans and different animal models.7- 1 These o 10 effects have suggested that enhanced gastric t secretion by NSAIDs could be a factor in the CD pathogenesis of NSAID induced gastroduo- denal injury.7-10 Species differences have been Gut: first published as 10.1136/gut.36.5.657 on 1 May 1995. Downloaded from .05 reported on the actions of NSAIDs in rabbits, frogs, and humans on parietal cells6 11 20-22 0,, and show that data from one species cannot be fully extrapolated to others (for 0 example, humans). 14 Unlike gastric acid secretion, there are A2~ ~ ~ c+PE HI+P almost no data on the potential effects of 0 NSAIDs on pepsinogen secretion. A report from Ohe et al 23 showed that aspirin increased the ratio of alkali-labile to total pepsinogen in the homogenates of gastric mucosa in rats and considered that activated pepsinogen was an essential step in ulcer formation by the hydro- exeiet0 a -5( 1)% ofttl3 i.n4,*< S r 1 gen ion back diffusion. By using an 'in vitro' model of isolated human peptic cells, we now a. M~~edoeolarproducnrtionofbtraileld provide evidence for the first time that both Ach2an HISpeti cel aspirin and ibuprofen (two widely used PGE2 PGE2=33 NSAIDs) potentiate secretagogue stimulated FigreProtalanin(PGE2)stimlatingopeprogeinserto from disperse human pepsinogen secretion in humans. The poten- peptic~~cel(596)*ito48 t ha btie withacetychoine n=3-5;tt s o*p0o5) tiation of histamine stimulated pepsinogen pepsnognscreionobtInenwthistramn(Hs)stithmublaishedBasal secetioundithese secretion was seen at concentrations of both aspirin and ibuprofen that are commonly endogenous productind of bothraiolabeled reached in plasma of patients who regularly secetiondin human cells used these drugs.17 18 Furthermore, the peptic (FigE2=393 potentiation effect was also seen on different (1235 cp0g(1r1)%30ein)v n15 M26) with histamine concentrations (10-4-1 0-6 M) that a,~ ~ ~ ~~~~apirin PW2=419I (486)3 cpin).gPprioteinv are in the range required for maximum Inecrtontrs withducdblihe thasti acid stimulated in the vascularly perfused rat daota.,landiond stomach.24 http://gut.bmj.com/ ~~~~2~ ~ ~ . tandtDbtinducedbyaelcoi. exitogntousPE We have also explored the mechanisms aHosevder,pendetagadnaldtincreaseinpsnoe by which aspirin and ibuprofen potentiate 0~~~~~~ hsertaione sinmulane pepticogell secretion 0~~~~~~0~~~~~~~~~ stimulated pepsinogen secretion by human C~~~~~~~~CMaximalsoentiation washistamined atimulated peptic cells. Receptor and 0 -7-9 pepsinogen- -7- secretion30-4 6 -7-8C- pathway studies have clearly defined two sets min).rPepsinoge of pepsinogen secretagogues: (a) those that activate adenylyl cyclase (for example, hista- on September 28, 2021 by guest. Protected copyright. aDditcsiontotainuebyaeycln. mine, forskolin, vasoactive intestinal , Howevr., potgandoteNSIns failedtienreasred etc) and stimulate intracellular cAMP and (b) toices rhistamineaa stimulatedpesngnscrion calcium mediated secretagogues (for example, (Freigond did nothinovivoya aaspirinandr ibu acetylcholine, cholecystokinin, , etc); these agents interact with receptors in chief cells and result in the activation of that mediate hydrolysis, release of calcium from intracellular stores, and the Asii aD oHer NSAIDs havMbeen activation of calcium-calmodulin 'dependent repote protein kinases.25.27 Aspirin and ibuprofen did not seem to potentiate pepsinogen secretion by affecting the adenylate cyclase/cAMP system or the H2 receptor. Cimetidine inhibited the potentiation of histamine stimulated pep- sinogen secretion by these drugs but failed to block the potentiation effect on forskolin 6Eo HS+NAD1 10 M PGE stimulated pepsinogen secretion. Because forskolin activates the catalytic subunit of adenylate cyclase intracellularly, these results suggest a regulatory role for aspirin and ibuprofen in conjunction with secretagogues distal to the site of the catalytic subunit of HIS ASA IBU ASA IBU adenylate cyclase activation. On the other -4 -4 -6 -4 -6 hand, aspirin and ibuprofen effects on pep- Molar concentration sinogen secretion seem dependent on extra- Figure 8: Exogenous prostaglandin (PGE2) addition does cellular calcium, as these two drugs failed not affect aspirin (ASA) and ibuprofen (IBU) potentiation ofhistamine (HIS) stimulation ofpepsinogen secretion to potentiate either histamine or forskolin from dispersed human peptic cells. Basal secretion in these stimulated pepsinogen secretion when extra- experiments was 4 80 (0 60) % of total/30 min. n=3. cellular calcium was removed from the media. 662 Lanas, Nerin, Esteva, Sdinz

In rabbit parietal cells aspirin also potentiates (potentiation, inhibition, etc). Moreover, our histamine, forskolin, and dbcAMP stimulated results show that the effects of both NSAIDs aminopyrine uptake by calcium dependent and prostaglandins on pepsinogen secretion mechanisms.7 In such a system aspirin are independent of H2 receptors. Finally, the increases intracellular calcium from intracellu- process of isolating cells includes repeated Gut: first published as 10.1136/gut.36.5.657 on 1 May 1995. Downloaded from lar sources, although an effect of aspirin to washing of cells with fresh Kreb's Ringer refill intemal Ca++ stores from the extracellu- solution (at least 10 times), which eliminates lar space was not excluded.7 Prolonged EGTA the potential presence of any drug taken by the washes deplete both intra and extracellular patient before the endoscopy. Ca++ and a brief EGTA wash only depletes Our results show also some variability in extracellular Ca+ + .7 28We have exposed peptic basal pepsinogen secretion in the different cells only to a brief EGTA wash followed by experiments. Studies on pepsinogen secretion incubation in a calcium free Ringer solution in experimental conditions always show some and this was sufficient to eliminate the potenti- interassay variability, but this has been found ation effect of NSAIDs on histamine and to be greater in isolated human peptic cells.13 forskolin stimulated pepsinogen secretion. The reasons for this are not known, but it may This suggests that extracellular calcium is depend, at least in part, on different factors needed for aspirin and ibuprofen to potentiate such as the different age of the subjects used pepsinogen secretion in human peptic cells in the study, the presence of a peptic ulcer probably by refilling intracellular calcium disease, the presence of Helicobacter pylorn stores. Furthermore, both aspirin and infection, etc. In this way, serum pepsinogen ibuprofen failed to potentiate acetylcholine has been found to decrease in ulcer patients stimulated pepsinogen secretion, a natural after H pylori eradication31 and in vitro calcium dependent secretagogue. It is possible sonicated preparations of H pylori increased that intracellular calcium mobilisation by pepsinogen secretion from isolated rabbit acetylcholine is potent enough to prevent gastric glands.32 However, H pylori had no further calcium increase by NSAIDs. effect in isolated guinea pig gastric glands.33 By finding that exogenous prostaglandins Whether H pylori affects pepsinogen secretion actually stimulated peptic cells, we ruled out a of human peptic cells, or the response of prostaglandin mechanism by which aspirin and these cells to natural agonists (acetylcholine, other NSAIDs might potentiate pepsinogen histamine, prostaglandins), and whether this secretion.7 8 Furthermore, this increase was in infection or any other factor are responsible for addition to that produced by acetylcholine the variability in basal pepsinogen secretion

suggesting that prostaglandins and acetyl- from isolated human peptic cells warrant http://gut.bmj.com/ choline act through independent pathways. further studies. Previous reports had already shown that In conclusion, our data show that pharma- prostaglandin effects on chief cells from dogs cological doses of two of the most widely used and guinea pigs were mediated by cAMP.2930 NSAIDs (aspirin and ibuprofen) potentiate We have confirmed this in our system because secretagogues stimulated pepsinogen secretion the exogenous addition of prostaglandin E2 to from dispersed human peptic cells and this

histamine stimulated pepsinogen secretion was might be an additional mechanism of NSAID on September 28, 2021 by guest. Protected copyright. not different to that obtained with histamine induced gastric mucosal injury in humans. alone. Furthermore, aspirin potentiation of The potentiation effect of these drugs is histamine stimulated pepsinogen secretion was calcium dependent, independent of endogen- neither inhibited nor potentiated by the ous prostaglandin inhibition, and may be addition of exogenous prostaglandin E2. mediated in peptic cells at a post-receptor site. By using patient material it can be ques- This study was supported by a grant from the Asociaci6n para tioned whether longterm drug use may disturb Investigaciones Grastoenterol6gicas de la Provincia de Zaragoza. The authors thank Professor Basil I Hirschowitz for the responsiveness of the mucosa (for example, critical discussion and review of the manuscript. H2 receptor antagonists used in most of our patients). However, we think that such a 1 Soll AH, Weinstein WN, Kurata J, McCarthy D. possibility does not invalidate our results. As in Nonsteroidal antiinflammatory drugs and (UCLA Conference). Ann Intern Med 1991; 114: most studies of isolated peptic cells, in this 307-19. work each single experiment had its own 2 Schoen RT, Vender RJ. Mechanisms of nonsteroidal anti- inflammatory drug-induced gastric damage. Am J Physiol control samples and responses to different 1989; 86: 449-58. stimuli are always expressed as a reference to 3 Walan A, Bader JP, Classen M, Lamers CBH, Piper DW, Rutgersson K, et al. Effect of and the values obtained in such control samples. on ulcer healing and relapse rates in patients with benign Therefore, if the basal state of peptic cells gastric ulcer. NEnglJfMed 1989; 320: 69-75. 4 Leung FW, Robert A, Guth P. Gastric mucosal blood was changed by longterm drug use, such a flow in rats after administration of 16,16-dimethyl difference is eliminated by expressing the prostaglandin E2 at a cytoprotective dose. Gastroenterology 1985; 88: 1948-53. results as percentage of total pepsinogen 5 Levine RA, Schwartzel EH Jr. Effect of indomethacin on content and basal subtracted. Furthermore, basal and histamine-stimulated human gastric acid secre- tion. Gut 1984; 25: 718-22. because most of our patients were receiving H2 6 Hunt JN, Smith JL, Jiang CL, Kessler L. Effect of synthetic blockers, the potential effects of these drugs on prostaglandin El analog on aspirin-induced gastric bleeding and secretion. Dig Dis Sci 1983; 28: 897-902. the responsiveness of the mucosa (increase 7 Levine RA, Nandi J, King RL. Aspirin potentiates prestim- or decrease secretion to histamine, etc) ulated acid secretion and mobilizes intracellular calcium in rabbit parietal cells. J Clin Invest 1990; 86: 400-8. were present in all samples of a particular 8 Levine RA, Nandi J, King RL. Nonsalicylate nonsteroidal experiment and probably did not affect the antiinflammatory drugs augment prestimulated acid secretion in rabbit parietal cells. Investigation of the main effect of the drug tested (NSAID, mechanisms of action. Gastroenterology 1991; 101: prostaglandins, etc) in our experiments 756-65. Non-steroidal anti-inflammatory drugs and prostaglandin effects on pepsinogen secretion by dispersed human peptic cells 663

9 Murthy UK, Levine RA. Aspirin induces morphological 21 Shea-Donohue T, Steel L, Montcalm-Mazzilli E, Dubois transformation to the secretory state in isolated rabbit A. Aspirin-induced changes in gastric function: role of parietal cells. Gastroenterology 1991; 101: 404-9. endogenous prostaglandins and mucosal damage. 10 Reeves JJ, Stables R. Effect of indomethacin, piroxican and Gastroenterology 1990; 98: 284-92. selected on gastric acid secretion by the 22 Rowe PH, Starlinguer MJ, Kasdon E, Marrone G, Silen W. rat isolated gastric mucosa. Br Pharmacol 1985; 86: Effect of simulated systemic administration of aspirin, sal-

J7 Gut: first published as 10.1136/gut.36.5.657 on 1 May 1995. Downloaded from 677-84. icylate and indomethacin on amphibian gastric mucosa. 11 Jaramillo E, Mardh S, Green K, Persson B, Rubio C. The Gastroenterology 1986; 90: 559-69. effect of arachidonic acid and its metabolites on acid pro- 23 Ohe K, Yokoya H, Kitaura T, Kunita T, Miyoshi A. duction in isolated human parietal cells. Scand J Increase in pepsin content in gastric mucosa during the Gastroenterol 1989; 24: 1231-7. course of aspirin- and taurocholate-induced gastric ulcer- 12 Hirschowitz BI. Pepsin in the pathogenesis ofpeptic ulcera- ation in rats. Dig Dis Sci 1980; 25: 849-56. tion. In: Halter F, Garner A, Tytgat GNJ, eds. Falk 24 Sandvik AK, Waldum HL, Kleveland PM, Schulze SB. Symposium 59. Mechanisms of peptic ulcer healing. Gastrin produces an immediate and dose-dependent hist- Dordrecht, The Netherlands: Kluwer Academic, 1991: amine release preceding acid secretion in the totally 183-94. isolated, vascularly perfused rat stomach. Scand Jf 13 Lanas AI, Anderson JW, Uemura N, Hirschowitz BI. Gastroenterol 1987; 22: 803-8. Effects of cholinergic, histaminergic and peptidergic stim- 25 Raufman JP. Gastric chief cells: receptors and signal-trans- ulation on pepsinogen secretion by isolated human peptic duction mechanisms. Gastroenterology 1992; 102: cells. ScandJ3 Gastroenterol 1994; 29: 578-83. 699-710. 14 Johnson LU, Walsh ML, Bokus BJ, Chen LB. Monitoring 26 Defize J, Meuwissen GM. Pepsinogens: an update of bio- of relative mithocondrial membrane potential in living chemical, physiological and clinical aspects. J Pediatr cells by fluorescence microscopy. J Cell Biol 1981; 83: Gastroenterol Nutr 1987; 6: 493-508. 526-35. 27 Basson MD, Modlin IM. Pepsinogen: biological and 15 Crissman HP, Egmond JV, Holdrinet RS, Pennings JV, pathophysiologic significance. J Surg Res 1988; 44: Haanen C. Simplified method for DNA and protein stain- 82-97. ing ofhuman hematopoietic cell samples. Cytometry 1981; 28 Hirschowitz BI, Uemura N, Matsumoto H, Dickinson 2: 59-62. KEJ. The role of Ca'+ in the time-dependent pepsino- 16 Matsumoto K, Dickinson KEJ, Anderson W, Hirschowitz gen secretion of frog oesophageal peptic glands stimu- BI. Pepsinogen secretion from perifused frog peptic lated by bombesin. Acta Physiol Scand 1990; 140: glands: rapid transiets detected with a modified pepsin 401-12. assay. Life Sci 1988; 42: 1237-44. 29 Defize J, Hunt RH. Effect of hydrochloric acid and 17 Mandelli M, Tognoni G. Monitoring plasma concentration prostaglandins on pepsinogen synthesis and secretion in of salicylate. Clin Pharmacokinet 1980; 5: 424-40. canine gastric chief cell monolayer cultures. Gut 1989; 30: 18 Geisslinger G, Schuster 0, Stock KP, Loew D, Bach GL, 774-81. Brune K. Pharmacokinetics of S(+)- and R(-)-ibuprofen 30 Raufman J-P, Cosowsky L. Relation of prostaglandin- in volunteers and first clinical experience of S(+)-ibupro- induced increases in cellular cAMP to stimulation of fen in rheumatoid arthritis. EurJ Clin Pharmacol 1990; 38: pepsinogen secretion from dispersed chief cells. J Cell 493-7. Physiol 1987; 132: 137-42. 19 ElAttar TMA, Lin HS, Platt RD. Effect of effervescent 31 Chittajallu RS, Dorian CA, McColl KEL. Serum pepsino- buffered aspirin on prostaglandin synthesis by human gin- gen I in duodenal ulcer-effect of eradication of H pylori gival fibroblasts. Prostaglandins Leukot Med 1987; 29: and correlation with serum gastrin and antral . Gut 237-47. 1990; 31: Al 199. 20 Seeger AJM, Jager LP, VanNoordwijk. Effects of 32 Cave TR, Cave DR. Helicobacter pylori stimulates pepsin , and caffeine on the erosive activ- secretion from isolated rabbit gastric glands. Scand J ity of acetylsalicylic acid in the rat stomach: dose-response Gastroenterol 1991; 26 (suppl 181): 9-14. relationships, time course of erosion development and 33 Rademaker JW, Hunt RH. Helicobacter pylori and gastric effects on acid secretion. J Pharm Pharmacol 1979; 31: acid secretion: the ulcer link? Scand J Gastroenterol 1991; 840-8. 26 (suppl 187): 71-7. http://gut.bmj.com/ on September 28, 2021 by guest. Protected copyright.