Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

Gut, 1986, 27, 1484-1492

Synthesis of E2, B2 and prostaglandin catabolism in gastritis and gastric ulcer

C J HAWKEY (WITH THE TECHNICAL ASSISTANCE OF N K BHASKAR AND B FILIPOWICZ) From the Department of Therapeutics, University Hospital, Nottingham

SUMMARY Because endogenous may protect the gastric mucosa a study was conducted to determine factors influencing (a) the synthesis of immunoreactive prostaglandin (iPG) E2 and thromboxane (iTx) B2 as measured by radioimmunoassay and (b) prostaglandin catabolism measured radiometrically, in human gastric mucosa. Gastric mucosa was obtained at endoscopy. Synthesis of iPE2 and iTxB2 was inhibited in vitro by indomethacin; iTxB2 synthesis was also selectively inhibited by the thromboxane synthesis inhibitor dazmegrel. Prostaglandin catabolism was inhibited by carbenoxolone. Multivariate analysis showed that synthesis of iPGE2 from endogenous precursor during homogenisation was decreased in patients on non-steroidal anti-inflammatory drugs. Mucosal inflammation was associated with significantly increased synthesis of iPGE2 and decreased prostaglandin catabolism. There were no differences between the mucosa of patients with or without gastric ulcers, nor between the ulcer rim and mucosa 5 cm away. Age, sex, smoking history, and ingestion of antisecretory drugs appeared to exert no influence. In this study gastritis was the major influence on prostaglandin synthesis. It seems unlikely that prostaglandin deficiency is a strong predisposing factor for gastric ulceration. http://gut.bmj.com/

Reports suggesting gastric mucosal protection by ment of thromboxane synthesis as thromboxane endogenous prostaglandin synthesis'-3 raise the synthesis is associated with gastric ulceration in rats possibility that this process might be defective in and inhibitors of thromboxane synthesis appear to patients who develop gastric ulcers. Previous evi- be protective.7 Alternatively reduced levels of dence on this point has been conflicting. Some protective prostaglandins might result from in- studies have found reduced prostaglandin synthesis creased prostaglandin catabolism.'2I3 on September 27, 2021 by guest. Protected copyright. in gastric ulcer patients,4 even in the presence of In this study synthesis of immunoreactive prosta- gastritis: this would be surprising because other glandin (iPG)E, and of thromboxane (iTx)B,, alnd inflammatory lesions lead to enhanced prostaglan- prostaglandin catabolism in humatn gastric mucosa din synthesis. Another study reported increased have been investigated with reference both to prostaglandin synthesis in patients with gastric histological appearances and to the presence or ulcers6 and attributed this to the associated gastritis: absence of gastric ulcerattion. The relationship of however, even where this occurs it remains possible age, sex or smoking or ingestion of non-steroidal that prostaglandin synthesis could be deficient in anti-inflammatory drugs (NSAIDS) or antisecretory relation to the prevailing levels of mucosal in- drugs to these parameters has also been investigated flammation. in view of the known influence of these factors on Factors other than reduced prostaglandin synth- the incidence or natural history of gastric ulceration. esis might render the mucosa liable to ulcerate. These could include an absolute or relative enhance- Methods

Addrcss for corrcspondcicncc: I)r ( J llawikcy. i)cpartnicrit of ThhcralpcLtics. PATI ENTS Univcrsity tIospitl.ll Nottiighitm NG7 2L i-. The following radiolabelled compounds were used: Reccivc(d for puhlication 22 April 1986 (H3)PGE,, 160 Ci/mmol (Amersham International); 1484 Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

HiTll(t711l g(lstric Iro1sftalloi(ds 1485 (H3) TxB1, 139 3 Ci/mmol (New England Nuclear) slight increase in inflarmmatory cell infiltrate) or and (9 H3)PGF1. 13 7 Ci/mmol (Amersham Interna- inflamed (moderate or matrked increatse in in- tional). The following compounds were gifts: in- flammatory cell infiltraite) using staindard histological domethaicin (Merck Sharp & Dohme). dazmegrel criteria of inflammation. No attempt was maide to (Pfizer UK). carbenoxolone (Biorex Laboratories). differentiate between inflammnation caused malinly TxB, antiserum wtas kindlv donated by Dr Lawrence by neutrophils and inflarmmation caused mainly by Levine. PGE, antiserum wias bought from the lymphocytes. As mucosal atrophy and Imlucosail Institut Pasteur. Other chemicals were obtained inflammation were frequently associated no attempt from standard lAboratory suppliers. was made to assess them independently. Patients undergoing upper gastrointestinal endo- scopy with an Olympus T endoscope were studied. HOMOGENATES ANI) INCUBATIONS Where there was a possibility that mucosa might be The rest of the biopsy specimens were homogenised taken for research patients were asked to give immediately on ice in Tris HCI t)()5M. pH 7-4 (20 written informed consent before the endoscopy. The mg wet weight/ml finatl proportions) for 15 seconds study was approved by the South Nottingham using an Ultraturax homogenizer. Ethical Committee. In studies used to develop and validate the SYNTHESIS FROM ENDOGENOUS STORES incubation conditions and radioimmunoassalys and (ENDOGENOUS iPGE2ANI) iTxB,) to assess reproducibility with repeated study. both Ten milligram aliquots in 0 5 ml (final proportions) normail subjects aind paitients with a vairiety of of the homogenaite were extracted immediately into gastroduodenal conditions were studied. During the chloroform,'1 for the determination of mnain study 27 patients with gastric ulcers and 43 synthesised from endogenous stores (henceforth control patients were studied. As gastritis was one of designated Endogenous iPGE, or iTxB ). the factors which the study wats designed to investi- gate. endoscopic or histological evidence of in- INHIBITOR EXPERIMENTS flarmmation was not an exclusion criterion. In all Some samples were pre-incubated in Tris HCI other respects. the control patients had normal containing either indomethacin 50 [tg/ml (4x 10 4M) endoscopic findings. The characteristics of the for 15 minutes before being homogenised in fresh patients studied are detailed in Table 1. Tris HCI also containing indomethacin 50 Rg/ml

(4x10-4 M). These samples were extracted in the http://gut.bmj.com/ HIS (IOGY1 usual way and results compared with those obtained Four to five mucosail biopsy specimens were taIken from biopsies from the same patient but not exposed from the lesser curve between 45 and 50 cm from the to indomethacin. incisor teeth and used immediately. A small piece from two of the specimens, or a separate biopsy SYNTHESIS AFTER INCUBATION WITH ARACHIDONIC specimen from the sarme irea wats exatmined histolo- ACID (TOTAL iPGE2 SYNTHESIS) gically (without knowledge of the biochemical re- Other aliquots (10 mg, 0 5 ml, final proportions) sults) and graded as uninflamed (normal or only were incubated with 2 utg for 30 on September 27, 2021 by guest. Protected copyright. minutes at 37°C in a rocking water bath before extraction into chloroform.'5 Because this proce- Table I (Characteristics of tile 1i)(lit.e.s Istdied dure results in synthesis of prostanoids from both endogenous and exogenous arachidonic acid it is Control GU p)atielIS designated Total iPGE2 synthesis. Total 43 27 Mcn 29 13 INHIBITOR EXPERIMENTS Stilokcrs 16 17 In some experiments aliquots of homogenates were Antisecrctory dtirgs 9 17 incubated with or without indomethacin 1 [ig/ml NSAIDS 5 4 (2*8x10-6M) or dazmegrel 0-3 [tg/ml (10-6M). The Agc <3(0 ysears 7 3(-50 II- effect of the inhibitor was calculated after allowance 50-65 9 11 for synthesis during homogenisation. 65-75 11 14 75+ 4 RADIOIMMUNOASSAYS After extraction into chloroform, samples were Pa'aticnts on rcgular NSAID's. last dosc in cach case approximiatcly 12 hours bcforc cindoscopsy Individual druLgS aIS follows:- resuspended in phosphate buffered saline (pH 7-4 indomcthacin: flivc, ispirin: tsAo. naproxen: onc. flurhuprotcn: 0-15 M), and stored at -50°C for radioimmunoas- onc. say, usually within one month (maximum three Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

1486 Hawkey months). iTxBI3 was measured by radioimmunoassay dilutions gave similar results (value at higher dilu- using anti-serum kindly donated by Dr Lawrence tion 99+3%/,, SEM, of value at lower dilution, n=6). Levine. For thromboxane assays standard TxB,, Authentic PGE, added to samples before extraction TxBI anti-serum and H13] TxB2 tracer (Amersham) was assayed as 103±4°/O (n=48) of that added. were dissolved and samples were diluted in Tris , re-assayed in samples after one to saline pH 7-4, 0-15 M with 1% gelatin. iPGE2 was three months of storage was 105±9'S (n=101) of measured by radioimmunoassay as previously initial value. The PGE, antiserum showed selective described. 16 17 reactivity with material cochromatographing with authentic PGE,. CATABOLISM Aliquots of homogenate (10 mg 0 5 ml, final propor- (ii) iTxB2 tions) were also incubated with (90-H3)PGF2, Under the assay conditions used, the cross reactivity (1 mg, 0-1 iCi) and NAD 1 mm for five minutes at of the TxB2 antiserum with other prostanoids (based 37°C. They were extracted into chloroform at pH on concentrations required to produce 50% dis- 3 5 and re-suspended in 75 Rl; chloroform:ethanol placement of (H3) TxB2) was as follows: PGD2: (1:1). PGF,(L and its metabolites were separated by 1-2%; PGFia: 0(12%; PGE2: 0 1%; PGF2a: 0-01%; thin layer chromatography using Watman LK6D arachidonic acid, PGA2, 6 keto PGFI(, and 13 14 silica gel chromatography plates and ethyl acetate: dihydro 15 keto PGE2: all <0-001%. Assay of acetone:acetic acid (90:10: 1) as solvent. Regions samples at two dilutions gave similar results (value corresponding to authentic PGF,(,, 15 keto PGF,,, at higher dilution=105+5% of value at lower 13 14 dihydro PGF2, and 13 14 dihydro 15 keto dilution, n=15). Authentic TxB2 added to samples PGF2, were scraped and added direct to LKB before or after extraction was assayed as 90+9% Optiphase scintillant for quantitation by liquid (n=20) of that added. TxB2 re-assayed after one to scintillation counting. Radioactivity cochromato- three months of storage was 109+1% (n=53) of graphing with metabolites was expressed as a per initial value. TxB2 anti-serum showed selective cent of total radioactivity scraped. A similar value reactivity with material cochromatographing with derived using boiled mucosa from the same patient authentic TxB2. was then deducted to obtain values for the percen- tage of enzymatic catabolism in samples. QUAILITY CONTROL Preliminary experiments showed that iPGE2 and iTxB, (measured by radioimmunoassay) were re- INHIBITOR EXPERIMENTS http://gut.bmj.com/ In some experiments, aliquots of the homogenate covered to a similar extent (69+1% for PGE2 and were preincubated on ice for 15 minutes with 69±2"/o for TxB,) (n=3 each). In subsequent carbenoxolone 57 sg/ml (10-4 M) or 5-7 [sg/ml experiments (H3)PGE, was used to measure recov- 10-5 M) and then incubated with (9ji-H3)PGF2(, ery: mean recoveries were 72%. Three external (1 [tg, 25 nCi) and NAD 1 mm for 15 minutes at standards of PGE, or of TxB, were included in each 370C. radioimmunoassay and two samples previously assayed satisfactorily were re-assayed to ensure on September 27, 2021 by guest. Protected copyright. STATISTICAI. MLTHOI)S accuracy and consistency from one assay to another. The catabolism data approximated to a normal HOMOGENATES distribution but all synthesis data were transformed EFFECTS OF INHIBITORS IN VITRO logarithmically to obtain a normal distribution for (a) Immediate extraction (Endogenous iPGE2 and analysis. The influence of gastritis, gastric ulcera- TxB,) tion. age, sex, smoking, drug ingestion, and study Synthesis of iPGE2 was inhibited by 60±8% (n=8) number (to allow for any changes in experimental if homogenisation was carried out in the presence of technique with time) upon prostaglandin synthesis indomethacin, 50 sg/ml (1 4x 10-4 M). Under these and catabolism were investigated by multivariate conditions synthesis of iTxB2 was reduced by analysis (Statistical Package for Social Sciences 68±6% (n=8). Programme). Paired data were evaluated using the paired t test. (b) Incubated synthesis Synthesis of iPGE2 was inhibited by 72±9% and of Results iTxB2 by 63±15% (n=4) by indomethacin 1 sg/ml (2.8x 10-6 M). The thromboxane synthesis inhibitor VAl II)ITY Of RADIOIMMUNOASSAYS dazmegrel 0-3 Fg/ml (10-6 M) inhibited iTxB2 (i) iPGE, synthesis by 44+21%; synthesis of iPGE2 was In these experiments assay of samples at two-fold 243±69% control (n=4). Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

Human gastric prostanoids 1487

4^ nnn _ 10 000 r Endogenous iPGE2 Iuuuu Total iPGE2

0

0

* 0 0 0 0 * .. 1000 01000 0 F 0~~~~~~~ ......

0 4e...... 4 w

w 0. a- 1- * o 'S'0 100

* , I I * No GU GU NoGU GU No GU GU No GU GU Uninf lamed Inf lamed Uninflamed Inflamed

10000 http://gut.bmj.com/

iTXB2 Catabol ism 10'°r

*

* on September 27, 2021 by guest. Protected copyright. ^ 1000 *4 a

I-CN *~~~~ 0 * 0 * 0 x @00 100

8 :.:F: :-*:j;.

* So5 * so *0 0

0 0 I I I a 1 I C . Ui No GU GU No GU GU No GU GU NoGU GU Uninframed Inflamed Uninflamed Inf lamed Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

1488 Hanvkev,

(c) Catabolism used. Inflammation wits aI miajor influence and was Carbenoxolone inhibited prostaglandin catabolism associated with increased synthesis of iPGE2 (both by 29±9% at 10-5 M (n=8) and by 79±5% at endogenous and total) as well as reduced prosta- 10-4 M) (n=6). glandin catabolism (Fig. 1). Patients takinig non- steroidal anti-inflammrinatory drugs showed reduced REPRODUCIBILITY synthesis of prostanoids (Fig. 1): this was statisti- Eighteen patients were studied on two occasions cally significant for endogenous iPGE,. The study after receiving in the intervening period, misopros- number was a significant influence on prostagliandin tol (n=7), trimoprostil (n=4), placebo (n=4), cime- catabolism, Total iPGE. and iTxB.: values of alll tidine (n=2) and ranitidine (n= 1). Two way analysis three increatsed gradually as the study progressed. of variance showed that patients differed significantly Finally there were trends towards lower levels of from each other but that there was no significant iPGE, and of iTxB. in patients with gastric ulceria- change after the period of treatment. Based on the tion but these did not reach conventional levels of residual variance, the coefficients of variation for significance. No other factor showed statistically repeated study were: for endogenous iPGE2: 14%; significant influence. for total iPGE2: 10%; for iTxB2: 17% (all logarith- mically transformed data) and for prostaglandin UlCER RIM VERSUS INTAC[r MUCOSA catabolism (untransformed data): 33%. In lX patients mucosa wals obtained from the rim of the gastric ulcer: ulcerated slough and fibrous tissue INFI UENCE OF PATIENT CHARACTERISTICS ON in the base were avoided. Synthesis of iPGE., iTxB, SYNTHESIS AND CATABOI ISM and prostaglandin catabolism in the ulcer rim was Table 2 shows all factors having a statistically compared with that from mucosa on the lesser curve significant or near significant influence on the at least 5 cm away. As seen in Figure 2 there is little dependent variables studied. Where patients were difference between the ulcer rim and the intact studied more than once the first data only were mucosa.

PAIRFI) DATA, GASTRIC ULICER VERSUS Table 2 Factors influencing the synthesis and catabolism NON-GASTRIC Ul.CER PATIENTS of prostanoids in human gastric mucosa Twelve pairs of patients could be identified, matched for levels of inflammation, drug intake, sex, age (to )ependent within seven years) and study date (to within three http://gut.bmj.com/ vari(blec 1i IR F p Influence months), differing only with regard to the presence Ln(Endo- 64 NSAIDS -0-31 8X3( <()0()1 or absence of gastric ulceration. There were no gcnous Inliammation () 2X X 39 <()0()1 significant differences in the synthesis of iPGE,, PGE,) C1u -0-15 3-73* <() 1 iTxB, or of prostaglandin catabolism (Fig. 3). Ln(Total 60 Inflamiaiation 0(37 9X81 <((I1 PGE,) Study numiibcr (032 7-015 <()0()5 Discussion on September 27, 2021 by guest. Protected copyright. Ln(TXB,) 45 Studs uLmtibicr 0(42 8X99 <()0()1 The main conclusions from these data are that GlU -()028 3-85, <01; iPGE, synthesis as meiasured in the study is in- Catabolism 7(0 StuLy nunibcr 032 8X68 <01 creased and that prostaglandin catiabolism is re- Inflaimmi,ation -0-23 5-1(0 <()-()5 duced in the presence of gastritis. After allowing for the influence of coexisting inflammiation, however, Ln= natural logarithm there were no significant differences between gastric R=simplc corrclation cociticicnt F=variancc iatio mucosa from patients with or without gastric ulcera- "Critical F vislucs - 40)) (p<) ()5), 2X80 (p<() 1()) tion. -,-Critical F salLICs - 4-08 (p<()-()5). 2-85 (p<()-I() The Tablc only shows thosc varirlbcs whcrc a p vlaluc Icss than ()0 1 Radioimmunoassays were used as the bases of was obtaincd. measurement for synthesis. In these studies as with

Fig. 1 Svntihesis anid catabolism of prostatloids in gastric mnuco.sa: effect of gastrilis and ga.stric ulceration. (a) iPGE, sPthesised from enidogeioul sub.strate (Endogenoli.s iP(E,) in pgmlng wet weight. (h) Total iP(GE, (after incubation- willi exogenous arachidonic acid) in pgl ing wet ueight. (c) iTvB, (.synilhesisedfromergdogenous .sore.s) iip mg et ucight. (d) Catabolism of (913)H- PGF,(,:t(Y conversion to metabolite.s calculated Jas described ini text. 0 repre.sent fpatietsA taking non-steroidal anti-in flainmatory drugs. * - all otlher patients.. s/iou'meanis. Stipp)le(l area,s shlow SEM's. Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

Human gastric prostanoids1 1489)

Endogenous iPGE2 Total iPGE2 TXB2 Gatabolism 00 r-

1000 80O E 0 cm E 60 VI In .8 .4_

-r a _ 100 40 _-

20

I10 I I Lesser Rim Lesser Rim Lesser Rim Lesser Rim curve curv curve curve Fig. 2 Synthesis and catabolism of prostanoids it gastric ulcer patietnts: comparisotn oJ muoilsa away from the ulcer (lesser curve) and from the ulcer rim. Results from lesser curve andfrom the ulcer rim are joinedl or each individual patient - show means. Stippled areas show SEM's.

Endogenous iPGE2 Total iPGE2 TXB2 Catabolism 100 r- http://gut.bmj.com/

1000 801-

E E 60 -c in

. . _ 1o00- I0 4 on September 27, 2021 by guest. Protected copyright. s£r ....:::: i,,: 20[-

11 I I I I I I I 01 I I No GU GU NoGU GUl NoGU GU No GU GU Fig. 3 Synthesis and catabolism ofprostanoids in gastric mucosa: comparisot ofgastric ulcer patienlts with matchled controls. Each pair ofconnected points shows results from an individual gastric ulcer patient joined to those obtainedf'rom the matched control. - show means. Stipped areas show SEM's. others using radioimmunoassay it is possible that noid and that its synthesis would be inhibited in an crossreacting substances were at least in part mea- appropriate way by indomethacin or dazmegrel. The sured. This is acknowledged by use of the terms rather modest reduction (60-70%) in immuno- immunoreactive PGE2 or TxB2 although it seems reactive prostanoids synthesised during homogenis- unlikely that another s-ibstance would cross react ation achieved by high concentration ofindomethacin with the antiserum, cochromograph with the prosta- could be taken as evidence that substances other Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

149() Hawkey than cyclo-oxygenase products were measured. In both situations, it is difficult to ascribe a central Greater inhibition was seen, however, with lower role to prostaglandins because they do not have concentration of indomethacin in the incubated properties which make them likely to cause recruit- homogenates and other explanations for the results ment of inflammatory cells or to lead to mucosal obtained during homogenisation are more likely: atrophy. Vasodilator prostaglandins may, however, there may have been incomplete diffusion of in- give rise to mucosal redness sometimes seen in domethacin and consequently only partial cyclo- association with acute superficial gastritis. The oxygenase inhibition so that some synthesis occurred mediators responsible for the recruitment of during homogenisation. The present results are also inflammatory cells in gastritis however remain consistent with earlier data and the suggestion unidentified. Obvious candidates which require accompanying them that in gastric mucosa there investigation are B4 and other related may in fact be a small resting content of prostanoids'8 chemotactic lipoxygenase products. though methodological difficulties make it impossible Although there were large differences between to resolve this point with certainty. patients the data show reasonable within patient Prostaglandin catabolism was measured radiomet- consistency on repeated study, despite intervening rically. Prostaglandin F,,, was used as substrate treatment with a variety of therapeutic agents. The because of its resistance to non-enzymatic catabol- purpose of the reproducibility study was to validate ism. The assay is of enzyme activity in a defined the methodology rather than investigate the effects system and should not be taken to imply that PGF),, of individual drugs and the numbers are too small to is normailly the main prostaglandin catabolised by exclude the possibility that some of the treatments this pathway. Inhibition of prostaglandin catabolism influenced subsequent synthesis. Anti- by carbenoxolone has been reported before. In the secretory drug treatment, however, was not a present study different methodology has been used significant influence on any of the variable investi- to confirm that carbenoxolone inhibits prostaglan- gated in the main study in contrast with the earlier din catabolism with a potency similar to that suggestion that cimetidine enhances gastric mucosal previously reported. 9 prostaglandin synthesis. 22 The enhianced synthesis of iPGE, with gastritis The present results are different from those presumably reflects the infiltration of the mucosa by obtained by Wright et a both in showing enhanced inflammatory cells: human neutrophils and mac- synthesis with gastritis and in failing to find signifi- rophages both have the capacity to synthesise cant differences in patients with gastric ulceration. PGE,.-") These differences between inflamed and Trends toward lower levels of endogenous iPGE, http://gut.bmj.com/ uninflamed mucosa are similar to those shown by and of iTxB, in gastric ulcer patients do not reach Schlegel et al." Doubt has been expressed as to the statistical significance and no difference was seen in validity of these earlier data as the amounts of PGE, the comparison of matched pairs of patients. synthesised appeared to be very high. In the present Moreover, there were no local changes in PGE, or study the quantities measured were much lower but TxB. synthesis or prostaglandin catabolism in the the pattern of results was the same. The reduction in ulcer rim. Thus in this study gastric ulceration seems

prostaglandin catabolism in inflamed tissue has not at best to be a much weaker influence than on September 27, 2021 by guest. Protected copyright. been shown before. Two main explanations are suggested by Wright et al who reported reduced possible. One is that a primary defect in prostaglan- levels of PGE, in patients with gastric ulcers even din catabolism leads to increased mucosal levels of when the mucosa was inflamed.4 There was no pro-inflammatory prostaglandins and makes the evidence of an imbalance between PGE2 and TxB2. mucosa more liable to become inflamed. An The reasons for the differences betweeen the alternative explanation is that the reduced prosta- present data and those of Schlegel et al on the one glandin catabolism is a consequence rather than a hand and those reported by Wright et a14 on the cause of the inflammation: gastric mucosa has an other are not clear. The use of homogenates might exceptionally high capacity for prostaglandin be insensitive to subtle changes in cyclooxygenase catabolism'12 13 which might fall because of the activity and it remains possible that changes in mucosal atrophy which accompanies the inflamma- prostanoid synthesis could occur in vivo because of tion in gastritis. phospholipase activation, the influence of neural Whatever the mechanism, increased synthesis and stimulation or changes in endogenous synthesis. reduced catabolism of prostaglandins would both Specific evidence for such changes in gastric ulcer lead to increased tissue levels in gastritis. As with patients is lacking, however, and these possibilities ulcerative colitis, the significance of increased pros- could not account for the differences between the taglandin synthesis in gastritis must be viewed in the three studies as all used homogenates. The controls context of the known properties of prostaglandins.2' in the present study were 43 patients in whom no Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

Humanl gastric prostfanoids 1491

pathology was found whereas Wright et al used 5 Konturek SJ. Actions of nonsteroid ainti-inflatmmatory seven normal volunteers. One point that emerged compounds on gastric mucosal integrity and prostaglan- from the use of the patient study number as a quality din formaition in heialthy subjects and peptic ulcer patients. Adi, IlflJamnl(ltioln Res 1984; 6: 29-37. control measure was that there was a tendency for 6 Schlegel W, Wenk K. Dollinger HC, Raptis S. Concen- all parameters measured to rise slightly in value trations of prostaglandin A-, E- and F-like substances during the course of the study (presumably because in gastric mucosat of normal subjects and of patients of subtle changes in methodology). Because the with various gaistric diseaises. Clin Sci Mol Medl 1977; gastric ulcer patients and their controls were studied 52: 255-8. concurrently these changes do not influence the 7 Whittle BJR. Kauffman GL, Moncada S. Vasoconstric- comparison between them. Comparable data are tion with thromboxane A, induces ulceration of the not provided in the other studies. gatstric mucosa. Nature 1981; 292: 472-4. One stimulus for this study was the observation 8 Konturek SJ, Brzozowski T, Piastucki I, Radecki T, Dembinska-Kiec A. Role of prostaglandin aind throm- that the incidence of gastric and duodenal ulceration boxane biosynthesis in gastric necrosis produced by is static or rising in older women in the United taurocholate and ethanol. Dig Dis Sci 1983; 28: 154-60. Kingdom but falling in other groups.23 24 Although a 9 Whittle BJR. Cellulair mediiators in gastric damage: balanced group of patients with a wide age distribu- actions of thromboxiane A, and its inhibitors. In: Allen tion was studied there was no evidence that the A, Flemstrom G, Giarner A, Silen W. Turnlberg LA, parameters measured were influenced by age, sex, eds. Mechanismns otf sinicosail protectioni in t11e ppler or smoking. Patients on NSAID's, however, showed gastroinitestinal tract. Raven Press: New York, 1984: reduced prostaglandin synthesis even though they 295-302. were studied more than 12 hours after their last 1() Hawkey CJ, Kemp R. Filipowicz B, Bh3askar NK, Walt RW. Gaistric mucosail protection with dazmegrel, an dose. Comparable differences between patients not inhibitor of gastric mucosal thromboxane synthesis. on NSAID's and controls were not seen suggesting (Gastroeniterology 1985; 88: 1415. that reduced prostaglandin dependent processes are 11 Bhaskair NK, Filipowicz B. Hawkey CJ. Inhibition of less important in this group than in patients taking humain gastric mucosal thromboxane synthesis by NSAID's. Wright et al regarded the reduction in WHR 2348A. Br J Pharmacol 1985: 85: 268p. synthesis of PGE, which they observed with gastric 12 Spenney JG. Barton JC. 15 hydroxyprostaglandin ulcer to be a secondary phenomenon. The data dehydrogenase aind 13 reductiasc content of gastrointes- reported in the present paiper give even fewer tinal organs of rabbits and rats. Prostaglandinis 1981; 21 suppl: 15-24. for that a of http://gut.bmj.com/ grounds believing primary deficiency 13 Peskar BM, Seyberth HW, Peskiar BA. Synthesis ind prostaglandin synthesis accounts for the develop- metabolism of endogenous prostaglandins by humian ment of gastric ulceration in patients not on gastric mucosa. In: Sommelsson B. Ramwell PW, NSAID's and thus cast some doubt on the clinical Paoletti R. eds. Advances in prostaglandin and1l tlhroitn- relevance of using prostaglandins for mucosal pro- boxanie resealrch, vol 8, New York: Riaven Press, 1980: tection of these patients. 1511-4. 14 Whitehead R, Truelove SC. Gear MW. The histological diagnosis of chronic gastritis in fibreoptic gastroscope This work was supported by a project grant from the biopsy specimens. J Clin Patliol 1972; 25: 1-1 1. on September 27, 2021 by guest. Protected copyright. Medical Research Council. 15 Hawkey CJ. Truelove SC. Inhibition of prostaglandin synthetase in humatn rectal mucosat. Glut 1983; 24: 213-7. References 16 Hawkey CJ, Truelove SC. Effect of' prednisolone on prostaglandin synthesis by rectal mucosa in ulcerative I Konturek SJi Pialstucki I. Brzozowski T. et til. Roic of colitis: investigation by latminar flow bioiassiay and locally generaited prostaglandins in adaptivc gastric radioimmunoassay. Git 1981; 22: 190-3. cytoprotection. Dig Dis Sci 1982; 27: 967-71. 17 Hawkey CJ. Evidence that prednisolone is inhibitory to 2 Robcrt A, Nezamis JE. Liancaster C. Davis JP, Ficid the cyclo-oxygenase aictivity of humral colonic mucosat. SO, Hanchar AJ. Mild irritiants present gastric necrosis Prostaglandinis 1982; 23: 397-4(09. through 'adiaptive cytoprotection' mediated by prostia- 18 Bennett A. Stamford IF, Stockley HL. Estimation and glandins. Am J Physiol 1983; 245: G113-G21. characterization of prostaglandins in the human gais- 3 Hawkey CJ, Rampton DS. Prostaiglandins iand the trointestinial triact. Br J Phliarmriacol 1977; 61: 573-8. gastrointestinal mucosa; are they importaint in its 19 Peskatr BM, Holland A, Peskair BA. Effect of carben- function, disease or treatment'? Gastroentterology 1985; oxolone on prostagliandin synthesis ind degraidation. 89: 1162-88. J Pharinr Pharinacol 1976; 28: 146-8. 4 Wright JP, Young GO. Klaft LJ, Wcerss LA, Price SK, 20 Morley J, Bray MA. Jones RW, Nugteren DM, Vian Marks IN. Gatstric mucosal prostaiglandin E levels in Dorp DA. Prostaglandin and thromboxane production patients with gastric ulcer disease and carcinoma. by human and guinea pig macrophages aind leucocytes. Gastroenterology 1982; 82: 263-7. Proastglantidinis 1979; 17: 730-6. Gut: first published as 10.1136/gut.27.12.1484 on 1 December 1986. Downloaded from

1492 Hawkey

21 Rampton DS, Hawkey CJ. Prostaglandins and 23 Coggan D, Lambert P, Langman MJS. 20 years of ulcerative colitis. Progress report. Gut 1984; 24: hospital admissions for peptic ulcer in England and 1399-413. Wales. Lancet 1981; i: 1302-5. 22 Branski D, Sharon P, Karmeli F, Rachmilewitz D. 24 Walt RPW, Katschinski B, Logan RFA, Ashley J, Effect of cimetidine on human gastric and duodenal Langman MJS. Rising frequency of ulcer perforation in prostanoid synthesis. Scand J Gastroenterol 1984; 19: elderly people in the United Kingdom. Lancet 1986; i: 457-60. 489-92. http://gut.bmj.com/ on September 27, 2021 by guest. Protected copyright.