Baillie re's Clinical Gastroenterology Vol. 14, No. 1, pp.

147159, 2000

doi:10.1053/bega.1999.0065, available online at http://www.idealibrary.com on

11 How do NSAIDs cause ulcer disease?

John L. Wallace
PhD

Professor Department of Pharmacology and Therapeutics, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada

Gastroduodenal ulceration and bleeding are the major limitations to the use of non-steroidal anti-inammatory drugs (NSAIDs). The development of safer NSAIDs or of eective therapies for the prevention of the adverse eects of existing NSAIDs requires a better understanding of the pathogenesis of NSAID-induced ulcer disease. NSAIDs can cause damage to the gastroduodenal mucosa via several mechanisms, including the topical irritant eect of these drugs on the epithelium, impairment of the barrier properties of the mucosa, suppression of gastric prostaglandin synthesis, reduction of gastric mucosal blood ow and interference with the repair of supercial injury. The presence of acid in the lumen of the stomach also contributes to the pathogenesis of NSAID-induced ulcers and bleeding, by impairing the restitution process, interfering with haemostasis and inactivating several growth factors that are important in mucosal defence and repair. In recent years, a fuller understanding of the pathogenesis of NSAID-induced ulcer disease has facilitated some new, very promising approaches to the development of stomach-sparing NSAIDs. Key words: ulcer; non-steroidal anti-inammatory drug; acid; neutrophils; cyclo-oxygenase; nitric oxide.

The ability of non-steroidal anti-inammatory drugs (NSAIDs) to cause ulceration and bleeding in the upper gastrointestinal tract was rst documented by the endoscopic study of Douthwaite and Lintott in 1938.1 In that study, the investigators demonstrated the ability of aspirin to damage the stomach. In the years that followed, more potent NSAIDs, such as indomethacin, phenylbutazone and the fenamates, were brought to market. Shortly thereafter, case reports of melaena associated with the use of aspirin and the newer NSAIDs began to appear in the literature, and in the 1960s and 70s, case-control studies began to document the gastrointestinal toxicity of this class of drug. In recent years, the upper gastrointestinal tract damage caused by NSAIDs has been referred to as an `epidemic' by a number of investigators.2,3 This is in part attributable to the widespread use of these drugs, particularly by patients with osteoarthritis and rheumatoid arthritis. The world market for NSAIDs is approximately $8 billion. The cost of treating NSAID-related gastrointestinal adverse eects almost certainly exceeds this amount (the annual cost of treating NSAID gastropathy in rheumatoid arthritis patients in the USA, for example, having been reported to exceed $4 billion).4 Developing anti-inammatory drugs that do not produce ulceration or bleeding in the gastrointestinal tract, or developing prophylactic therapies that substantially
1521-6918/00/010147+13 $35.00/00 c 2000 Harcourt Publishers Ltd. *

148 J. L. Wallace

reduce the toxicity of existing NSAIDs, depends upon clearly understanding the pathogenesis of NSAID-induced ulceration. In this chapter, the mechanisms by which NSAIDs damage the mucosa, interfere with repair and promote bleeding will be reviewed. In general, the properties of NSAIDs that contribute to ulcerogenesis can be divided into two categories: (1) topical irritancy, and (2) the suppression of prostaglandin synthase activity. In addition, the presence in the stomach and duodenum of acid and, in some cases, Helicobacter pylori (H. pylori), may contribute to the ability of NSAIDs to damage the mucosa. TOPICAL IRRITATION OF THE MUCOSA Studies in the 1960s by Davenport suggested that aspirin could directly damage the gastric epithelium.5 The breaking of the `barrier' permitted the back-diusion of acid into the mucosa, which eventually led to the rupture of mucosal blood vessels. These topical irritant properties were subsequently found to be predominantly associated with those NSAIDs with a carboxylic acid residue. The unionized forms of these drugs can enter epithelial cells in the stomach and duodenum. Once in the neutral intracellular environment, the drugs are converted to an ionized state and cannot diuse out. This has been referred to as `ion trapping'.6 As the drug accumulates within the epithelial cell, the osmotic movement of water into the cell results in swelling of the epithelial cell, eventually to the point of lysis.6,7 Another mechanism by which NSAIDs could damage the gastroduodenal epithelium is via the uncoupling of oxidative phosphorylation in the epithelial cells.7,8 Various NSAIDs have been shown to uncouple mitochondrial respiration7,8, leading to a depletion of ATP and therefore a reduced ability to regulate normal cellular functions, such as the maintenance of intracellular pH. The ability of NSAIDs to uncouple oxidative phosphorylation also appears to be related to some extent to acidic moieties (such as carboxylic acid residues), since substitution at these sites interferes with the ability of these compounds to act as uncouplers.8 The theory that NSAIDs cause topical injury by virtue of their eects on mitochondrial respiration has been challenged, mainly based on the observation that gastric and duodenal ulcers are observed following the parenteral or rectal administration of NSAIDs9,10. Erosions and ulcers can also be produced in experimental circumstances in which NSAIDs are administered parenterally.11 It is also dicult to comprehend how the uncoupling of oxidative phosphorylation would occur in gastrointestinal epithelial cells but not in the numerous other cells with which an NSAID would have contact subsequent to its absorption. On the other hand, the small intestine would be repeatedly exposed to NSAIDs that are excreted in bile and recycled enterohepatically, so it is conceivable that the uncoupling of oxidative phosphorylation is an important component of the pathogenesis of NSAID-induced enteropathy. A third mechanism that could account for the topical irritant properties of NSAIDs is their ability to decrease the hydrophobicity of the mucus gel layer in the stomach. Lichtenberger and co-workers have proposed that this layer is a primary barrier to acid-induced damage in the stomach.12,13 They demonstrated that the surface of the stomach is hydrophobic and that this hydrophobicity can be reduced by various pharmacological agents. For example, NSAIDs were shown to associate with the surface-active phospholipids within the mucus gel layer, thereby reducing its hydrophobic properties.1315 These investigators further demonstrated that the mucus gel layer in the stomach of rats and mice given NSAIDs was converted from a

Pathogenesis of NSAID ulceration 149

non-wettable to a wettable state. This eect was found to persist for several weeks or months after the cessation of NSAID administration.12,14 Attempts have been made to produce NSAIDs with reduced topical irritant eects. These include formulations in slow-release or enteric coated tablets, as well as the preparation of the drug as a pro-drug that requires hepatic metabolism in order to be active (e.g. sulindac). However, the incidence of gastroduodenal ulceration with prodrugs is comparable to that seen with standard NSAIDs.1618 This is taken as evidence that the topical irritant properties of NSAIDs are not paramount in terms of their ability to induce frank ulceration in the stomach. However, one cannot completely exclude this possibility. Most NSAIDs are excreted in bile. Even with parenteral administration of the NSAID, therefore, a reux of duodenal contents into the stomach would result in the topical exposure of the gastric mucosa to these drugs. Eorts have recently been made to develop gastrointestinally safe NSAIDs on the basis of a reduced ability to interfere with the surface-active phospholipid layer in the gastrointestinal mucus. Lichtenberger et al proposed that pre-associating NSAIDs with zwitterionic phospholipids prior to their administration should reduce the ability of the NSAIDs to associate the phospholipids in the mucus gel, and should therefore reduce their ulcerogenicity.12 They demonstrated this to be the case by pre-associating aspirin and other NSAIDs with dipalmitoyl-phosphatidylcholine (DPPC) and demonstrating that the complex produced signicantly less damage in the gastrointestinal tract than did the parent drug.14 Importantly, the pre-association of aspirin with DPPC did not interfere with the eectiveness of the aspirin to reduce fever or inammation.19 Similar eects could be obtained with other NSAIDs.14,19 SUPPRESSION OF PROSTAGLANDIN SYNTHESIS Vane's discovery in 1971 that NSAIDs inhibit prostaglandin synthesis20 led directly to the discovery of the ability of minute quantities of exogenous prostaglandins to protect the gastrointestinal tract from injury induced by topical irritants and NSAIDs.21 This in turn led to extensive research into the roles of prostaglandins in mucosal defence. There is substantial evidence that the ability of an NSAID to cause gastric damage correlates well with its ability to suppress gastric prostaglandin synthesis2225; agents that are weak inhibitors of gastric prostaglandin synthesis are less ulcerogenic.24,25 There is also a good correlation between the time- and dose-dependency of suppression of gastric prostaglandin synthesis by NSAIDs and their ability to cause gastric ulcers.2224 Why does the suppression of gastric prostaglandin synthesis lead to mucosal injury? As mentioned above, it is now clear that prostaglandins play a central role in modulating mucosal defence. Endogenous prostaglandins are involved in the regulation of mucus and bicarbonate secretion by the gastric and duodenal epithelium, mucosal blood ow, epithelial cell proliferation, epithelial restitution and mucosal immunocyte function.26 This is not to say that prostaglandins are the only endogenous substances regulating these components of mucosal defence; indeed, nitric oxide appears to perform many of the same functions as the prostaglandins in this respect.26 Thus, the inhibition of prostaglandin synthesis alone may not result in the formation of gastric erosions or ulcers. For example, mice in which the gene for cyclo-oxygenase-1 was disrupted exhibited negligible levels of gastric prostaglandin synthesis yet did not spontaneously develop gastric erosions or ulcers.27 Of course, one would, in such a situation, anticipate that adaptation of the stomach to the chronic absence of prostaglandins would occur (i.e. there might be an enhanced production of other

150 J. L. Wallace

gastroprotective substances). On the other hand, the suppression of mucosal prostaglandin synthesis, while not necessarily resulting in ulcer formation, will reduce the ability of the gastric mucosa to defend itself against luminal irritants. This has been demonstrated very clearly in experimental animals. Doses of NSAIDs that substantially reduced mucosal prostaglandin synthesis did not cause overt gastric injury but did greatly increase the susceptibility of the gastric mucosa to damage induced by irritants (e.g. bile salts).28 In terms of the pathogenesis of ulcer disease, is there one component of mucosal defence that is more important than others with respect to its impairment by NSAIDs? This question has not yet been completely answered. The fact that gastric ulcers can be induced by parenterally administered NSAIDs911, and that this damage can be produced independently of the presence of luminal acid29,30, suggests that the impairment of mucus and/or bicarbonate secretion may be of lesser signicance. While prostaglandins are extremely potent inhibitors of mast cell degranulation31, and mast cells are capable of releasing a variety of mediators (e.g. leukotriene C4 and plateletactivating factor) that can contribute to mucosal injury32,33, the eects of NSAIDs on mast cells do not seem to be crucial to the pathogenesis of mucosal injury. This latter conclusion is based on the observation that rats in which mucosal mast cells have been depleted by chemical means, or mice that are genetically decient of mast cells, exhibit similar susceptibility to NSAID-induced gastric injury as do their respective controls.34 The rate of epithelial turnover has been shown to be reduced by NSAIDs and could contribute to the pathogenesis of ulcer disease.35 However, it is unlikely that this eect is the principal one that leads to the development of an ulcer. The component of mucosal defence that appears to be most profoundly altered by NSAIDs is the gastric microcirculatory response to injury. When the mucosa is exposed to an irritant, or when supercial epithelial injury occurs, mucosal blood ow substantially increases. This is probably a response aimed at removing any toxins or bacterial products that enter the lamina propria, neutralizing back-diusing acid and contributing to the formation of a microenvironment at the surface of the mucosa that is conducive to repair.36 As described in greater detail below NSAIDs can reduce gastric mucosal blood ow and profoundly alter the behaviour of neutrophils owing through the gastric microcirculation. EFFECTS ON THE MICROCIRCULATION The ability of NSAIDs to reduce gastric mucosal blood ow has been recognized for several decades.3739 Prostaglandins of the E and I series are potent vasodilators that are continuously produced by the vascular endothelium, so the inhibition of their synthesis by an NSAID leads to a reduction in vascular tone. Several lines of evidence have suggested that damage to the vascular endothelium is an early event following the administration of NSAIDs to experimental animals.40,41 Endothelial injury is also an early event in the pathogenesis of gastrointestinal damage associated with ischaemia reperfusion42, in which neutrophils have been demonstrated to play a critical role as mediators of endothelial injury.43 This observation led us to examine the possible role of the neutrophil in the pathogenesis of experimental NSAID gastropathy (Figure 1). Our studies and those of others demonstrated that NSAID administration to rats resulted in a rapid and signicant increase in the number of neutrophils adhering to the vascular endothelium in both gastric and mesenteric venules.25,4446 This eect was typically seen within 1560 minutes after the administration of an NSAID, consistent

Pathogenesis of NSAID ulceration 151

Figure 1. Role of changes in the gastric microcirculation in the pathogenesis of NSAID-induced ulceration. NSAIDs suppress prostaglandin (PG) synthesis, and cause an increase in the liberation of leukotriene (LT) B4 and tumour necrosis factor (TNF). The net result is an increase in expression of various adhesion molecules, leading to neutrophil adherence to the vascular endothelium.

with the period of time required for the suppression of prostaglandin synthesis by these drugs.47 Subsequent studies demonstrated that this adherence was dependent on the expression of the b2-integrins (CD11/CD18) on the neutrophil and intercellular adhesion molecule-1 (ICAM-1) on the vascular endothelium.46 Such adherence could also be demonstrated to occur in vitro using isolated human neutrophils and endothelial cells from human umbilical vein.48 Furthermore, the upregulation of ICAM-1 on the vascular endothelium in the gastric microcirculation was shown to occur within 1530 minutes of the administration of an NSAID to rats, and this could be prevented by the administration of exogenous prostaglandins.49 Of course, the fact that neutrophils adhere to the vascular endothelium following NSAID administration does not necessarily mean that these cells contribute to the endothelial injury of the mucosal tissue injury that can be induced by NSAIDs. To test this hypothesis, a number of studies were carried out. First, we demonstrated that depleting circulating neutrophils in the rat, either with an anti-neutrophil serum or with methotrexate, greatly reduced the severity of the gastric damage induced by indomethacin or naproxen.50 Importantly, the induction of neutropenia also prevented

152 J. L. Wallace

the damage to the vascular endothelium induced by the NSAIDs.50 Second, we demonstrated that treating rabbits or rats with monoclonal antibodies that blocked NSAID-induced neutrophil adherence to the vascular endothelium also markedly reduced the extent of NSAID-induced gastric damage.41,46 Third, we found that administration of prostaglandins at doses previously shown to prevent gastric injury prevented NSAID-induced leukocyte adherence.44,45 Further evidence for a link between neutrophil adherence and NSAID-induced gastric injury came from studies of arthritic rats.51 These animals, like humans with rheumatoid arthritis, exhibit an increased susceptibility to NSAID-induced gastric damage. Interestingly, these rats also exhibited an increased level of neutrophil adherence to the vascular endothelium. This appeared to be due to an increase in expression of ICAM-1 on the vascular endothelium of arthritic rats, since pre-treatment with an antibody against ICAM-1 reduced leukocyte adherence and the susceptibility to NSAID-induced gastric damage to levels seen in healthy controls.51 These studies suggest that, in response to the suppression of prostaglandin synthesis by NSAIDs, there is a rapid upregulation of expression of ICAM-1 on the vascular endothelium and possibly an upregulation of b2-integrin expression on circulating neutrophils, resulting in an increased adherence of neutrophils to the endothelium. This begs the question of whether the enhanced adhesion molecule expression is purely a consequence of the reduction of prostaglandin synthesis, or whether there is another chemical signal produced in response to diminished prostaglandin synthesis that triggers the events leading to neutrophil adherence. Santucci et al suggested that tumour necrosis factor-a (TNFa) might be the key signal for NSAID-induced neutrophil adherence within the gastric microcirculation.52 The release of TNFa from macrophages and mast cells has been shown to be suppressed by prostaglandins31,53, and TNFa is a well-characterized stimulus for adhesion molecule expression.54 Santucci et al demonstrated that the levels of TNFa in the plasma of rats signicantly increased following the administration of indomethacin, this being accompanied by a parallel accumulation of neutrophils within the gastric microcirculation and the development of gastric injury.52 Furthermore, pre-treatment with a TNFa synthesis inhibitor, pentoxifylline, dose-dependently reduced neutrophil accumulation in the gastric microcirculation and gastric damage.52 These results have been conrmed and extended by Appleyard et al using a number of structurally unrelated inhibitors of TNFa synthesis and an anti-TNFa antibody.55 Another group of mediators that might contribute to the increase in neutrophil adherence following NSAID administration is the leukotrienes. Like prostaglandins, leukotrienes are derived from arachidonic acid. They have been shown to be capable of altering the susceptibility of the gastric mucosa to injury44,5658, at least in part through stimulatory eects on neutrophil adherence to the vascular endothelium.44 Inhibitors of leukotriene synthesis and leukotriene receptor antagonists have been shown to exert some protective eects in experimental models of NSAID-induced gastric damage.5658 There is also evidence of an elevated leukotriene B4 production following NSAID administration to laboratory animals44 and man59, and inhibitors of leukotriene synthesis and a leukotriene B4 receptor antagonist have been shown to prevent NSAID-induced neutrophil adherence to the vascular endothelium both in vivo44 and in vitro.48 Evidence for a role for neutrophils in the pathogenesis of gastric ulceration in humans has recently been provided.60 Patients taking NSAID therapy who had signicant numbers of neutrophils within the gastric mucosa were approximately six times more likely to develop an ulcer over the course of 24 weeks than were patients who

Pathogenesis of NSAID ulceration 153

did not have signicant numbers of neutrophils in their gastric mucosa. It is also noteworthy that, in a clinical trial examining the potential benet of famotidine for the prevention of NSAID-related gastroduodenal ulcers, Taha et al observed that an increased peripheral white cell count was a signicant risk factor for ulcer development.61 They suggested that this observation was consistent with the hypothesis that NSAID-induced gastropathy was mediated, at least partially, by neutrophils. How would the adherence of neutrophils to the vascular endothelium contribute to the formation of gastroduodenal ulcers? First, the adherence of neutrophils to the vascular endothelium is accompanied by an activation of these cells, leading to the release of proteases (e.g. elastase and collagenase) and oxygen-derived free radicals (e.g. superoxide anions). These substances may mediate much of the endothelial and epithelial injury caused by NSAIDs. This is supported by observations that the severity of NSAID-induced mucosal injury can be markedly diminished by compounds that scavenge oxygen-derived free radicals62,63 and by inhibitors of neutrophil-derived proteases.64,65 Second, neutrophil adherence to the endothelium, and the subsequent recruitment of other elements of blood (e.g. platelets), could produce an obstruction of the capillaries, thereby reducing gastric mucosal blood ow. In this respect, it should be noted that the well-characterized ability of NSAIDs to reduce gastric blood ow3739 has been shown to occur subsequent to the appearance of `white thrombi' in the gastric microcirculation.38 INHIBITION OF RESTITUTION Damage to the gastric epithelium probably occurs on a daily basis but does not usually lead to deeper mucosal injury because of the ability of rapid (i.e. within minutes) repair to occur via the process of restitution. This process involves the rapid migration of healthy cells from the gastric pits to re-establish an intact epithelial barrier.66 The cells move along the denuded basement membrane, which acts as a template and has been shown to be crucial to the restitution process.6669 The basement membrane can be damaged by acid, leading to an inhibition of restitution and the progression of necrosis to deeper layers of the mucosa.47,70,71 This does not, however, occur in normal circumstances because of the formation over sites of injury (i.e. exposed basement membrane) of a microenvironment in which the pH is maintained at near neutral, even in the presence of a signicant acid load in the lumen.47,70 A `mucoid cap' consisting of mucus, cellular debris and plasma proteins (particularly brin) forms within seconds of gastric epithelial injury, trapping the plasma that leaks from the underlying microcirculation.70 It is this plasma which accounts for the near-neutral pH within the protective mucoid cap, since even a very brief cessation of mucosal blood ow results in a rapid decrease in the pH within the mucoid cap, which in turn results in the formation of haemorrhagic erosions.41 As discussed above, NSAIDs can decrease mucosal blood ow. Thus, NSAIDs can cause gastric injury by interfering with the appropriate function of the mucoid cap and therefore the process of restitution. Indeed, we observed that, following the systemic administration of an NSAID to an animal in which supercial epithelial injury had been induced, the pH within the mucoid cap that had formed over the sites of epithelial damage began to decline in parallel with the inhibition of prostaglandin synthesis.41 Within minutes thereafter, the formation of haemorrhagic erosions was clearly evident. This eect could be prevented through the luminal delivery of exogenous prostaglandins, as could the formation of haemorrhagic erosions.41

154 J. L. Wallace

REPAIR OF ULCERS In addition to causing ulcer formation, NSAIDs can delay the healing of pre-existing ulcers and promote their bleeding.72,73 The eects on ulcer healing are probably related, once again, to the ability of NSAIDs to suppress prostaglandin synthesis. In normal gastric mucosa, prostaglandin synthesis occurs mainly via the cyclo-oxygenase-1 isoform. However, at a site of ulceration, and particularly around the ulcer margin, cyclo-oxygenase-2 appears to be the primary contributor to prostaglandin synthesis. Studies in mice and rats initially demonstrated the marked upregulation of cyclooxygenase-2 in ulcerated gastric tissue74,75, and this has recently been conrmed in humans.76 Moreover, the treatment of rats or mice with selective inhibitors of cyclooxygenase-2 results in a signicant delay of ulcer healing.74,75 These observations, and reports of selective cyclo-oxygenase-2 inhibitors exacerbating intestinal inammation and ulceration77, suggest that caution should be exercised in regarding the new cyclooxygenase-2 inhibitors as gastrointestinally safe.78,79 The ability of NSAIDs to promote the bleeding of pre-existing ulcers is most probably related to their inhibitory eects on platelet aggregation.80,81 The inhibition of platelet aggregation by NSAIDs occurs as a consequence of the inhibition of thromboxane synthesis. It should be noted that, unlike other NSAIDs, aspirin produces an irreversible inhibition of thromboxane synthesis in the platelet. Thus, even the low doses of aspirin used for the prophylaxis of myocardial infarction and stroke can signicantly increase the risk of gastrointestinal bleeding.8284 ROLE OF ACID The observation that NSAID-induced ulcers can develop in achlorhydric individuals29,30 has contributed to a widely held belief that acid is not involved in the pathogenesis of these lesions. Further reinforcing this hypothesis are several studies demonstrating that treatment with histamine H2-receptor antagonists did not reduce the incidence of NSAID-induced ulceration.8587 However, many of these types of studies have demonstrated that H2-antagonists and proton pump inhibitors can prevent NSAID-induced gastric lesions, but not the formation of the clinically more signicant ulcers, as well as ulcer complications. Recently, however, Taha et al reported that a high dose of famotidine (40 mg twice daily) was eective in preventing NSAID-induced ulcers61, and Hawkey et al88 demonstrated that omeprazole could signicantly reduce the incidence of NSAID-induced ulceration. These studies suggested that a profound suppression of acid secretion, as is produced by omeprazole or by a high dose of famotidine, was necessary in order to have a signicant impact on the incidence of NSAID-induced ulcers. Acid may contribute to NSAID-induced ulcer formation in several ways. First, acid can exacerbate damage to the gastric mucosa induced by other agents. For example, acid can convert regions of ethanol-induced vascular congestion in the mucosa to actively bleeding erosions.66 Second, acid will contribute to ulcer formation by interfering with haemostasis. Platelet aggregation, for example, is inhibited at a pH of less than 4.89 Third, as outlined above, acid can convert supercial injury to deeper mucosal necrosis by interfering with the process of restitution. Fourth, acid can inactivate several growth factors (e.g. broblast growth factor) that are important for the maintenance of mucosal integrity and for the repair of supercial injury, since these growth factors are acid-labile.90

Pathogenesis of NSAID ulceration 155

It is important to note that NSAIDs can increase gastric acid secretion, although it is not clear whether such eects have any impact on ulcer formation or healing. Prostaglandins exert inhibitory eects on parietal cells91, so the inhibition of their synthesis by NSAIDs can result in an increase in gastric acid secretion.92

SUMMARY A great deal has been learned about the pathogenesis of NSAID-induced gastric injury over the past two decades. As a result, several new NSAIDs, which will have greatly reduced gastrointestinal toxicity relative to current drugs, are in the process of being introduced to the marketplace. For example, selective inhibitors of cyclo-oxygenase-2, which will spare the major form of cyclo-oxygenase in the gastrointestinal tract, produce substantially less injury to the stomach than do the current NSAIDs that inhibit both cyclo-oxygenase-1 and -2.93 Whether or not these agents prove to be as eective for the treatment of the full range of inammatory conditions currently treated with NSAIDs remains to be seen, and some experimental studies suggest that this may not be the case.94,95 Nitric oxide-releasing NSAIDs represent another approach to reducing the adverse eects of this class of drug.96 The design of these new NSAIDs would not have been possible had it not been for an increased understanding of the pathogenesis of NSAID-induced ulceration. Acknowledgements
Dr Wallace is a Medical Research Council of Canada Senior Scientist and an Alberta Heritage Foundation for Medical Research Senior Scientist.

REFERENCES
1. Douthwaite AH & Lintott SAM. Gastroscopic observation of the eect of aspirin and certain other substances on the stomach. Lancet 1938; 2: 12221225. 2. Gabriel SE & Bombardier C. NSAID induced ulcers. An emerging epidemic? Journal of Rheumatology 1990; 17: 14. 3. Singh G, Ramey DR, Morfeld D et al. Gastrointestinal tract complications of nonsteroidal antiinammatory drug treatment in rheumatoid arthritis. A prospective observation cohort study. Archives of Internal Medicine 1996; 156: 15301536. 4. Bloom BS. Direct medical costs of disease and gastrointestinal side eects during treatment of arthritis. American Journal of Medicine 1988; 83 (supplement 2A): 2024. 5. Davenport HW. Gastric mucosal hemorrhage in dogs. Eects of acid, aspirin, and alcohol. Gastroenterology 1969; 56: 439449. 6. Fromm D. How do non-steroidal anti-inammatory drugs aects gastric mucosal defenses? Clinical and Investigative Medicine 1987; 10: 251258. 7. Somasundaram S, Hayllar H, Ra S et al. The biochemical basis of non-steroidal anti-inammatory druginduced damage to the gastrointestinal tract: a review and a hypothesis. Scandinavian Journal of Gastroenterology 1995; 30: 289299. 8. Mahmud T, Wrigglesworth JM, Scott DL & Bjarnason I. Mitochondrial function and modication of NSAID carboxyl moiety. Inammopharmacology 1996; 42: 189194. 9. Estes LL, Fuhs DW, Heaton AH & Butwinick CS. Gastric ulcer perforation, associated with the use of injectable ketorolac. Annals of Pharmacotherapy 1993; 27: 4243. 10. Henry D, Dobson A & Turner C. Variability in the risk of major gastrointestinal complications from nonaspirin nonsteroidal anti-inammatory drugs. Gastroenterology 1993; 10: 10781088.

156 J. L. Wallace 11. Wallace JL & McKnight GW. Characterization of a simple animal model for nonsteroidal antiinammatory drug induced antral ulcer. Canadian Journal of Physiology and Pharmacology 1993; 71: 447452. *12. Lichtenberger LM. The hydrophobic barrier properties of gastrointestinal mucus. Annual Review of Physiology 1995; 57: 565583. 13. Goddard PJ, Hills BA & Lichtenberger LM. Does aspirin damage canine gastric mucosa by reducing its hydrophobicity? American Journal of Physiology 1987; 252: G421G430. 14. Lichtenberger LM, Wang ZM, Romero JJ et al. Non-steroidal anti-inammatory drugs (NSAIDs) associate with zwitterionic phospholipids: insight into the mechanism and reversal of NSAID-induced gastrointestinal injury. Nature Medicine 1995; 1: 154158. 15. Lichtenberger LM, Wang ZM, Giraud MN et al. Eect of naproxen on gastric mucosal hydrophobicity. Gastroenterology 1995; 108: A149. 16. Graham DY, Smith JL, Holmes GI & Davies RO. Nonsteroidal anti-inammatory eect of sulindac sulfoxide and sulde on gastric mucosa. Clinical Pharmacology and Therapeutics 1985; 38: 6570. 17. Carson JL, Strom BL, Soper KA et al. The relative gastrointestinal toxicity of the nonsteroidal antiinammatory drugs. Archives of Internal Medicine 1987; 147: 10541059. 18. Hawthorne AB, Mahida YR, Cole AT & Hawkey CJ. Aspirin-induced gastric mucosal damage: prevention by enteric-coating and relation to prostaglandin synthesis. British Journal of Clinical Pharmacology 1991; 32: 7783. 19. Lichtenberger LM, Ulloa C, Vanous AL et al. Zwitterionic phospholipids enhance aspirin's therapeutic activity, as demonstrated in rodent model systems. Journal of Pharmacology and Experimental Therapeutics 1996; 277: 12211227. 20. Vane JR. Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs. Nature 1971; 231: 232235. *21. Robert A, Schultz JR, Nezamis JE & Lancaster C. Gastric antisecretory and antiulcer properties of PGE2 , 15-methyl PGE2 , and 16,16-dimethyl PGE2. Gastroenterology 1976; 70: 359370. 22. Lanza FL. A review of gastric ulcer and gastroduodenal injury in normal volunteers receiving aspirin and other nonsteroidal anti-inammatory drugs. Scandinavian Journal of Gastroenterology 1989; 24 (supplement 163): 2431. 23. Rainsford KD & Willis C. Relationship of gastric mucosal damage induced in pigs by antiinammatory drugs to their eects on prostaglandin production. Digestive Diseases and Sciences 1982; 27: 624635. 24. Whittle BJR. Temporal relationship between cyclooxygenase inhibition, as measured by prostacyclin biosynthesis, and the gastrointestinal damage induced by indomethacin in the rat. Gastroenterology 1981; 80: 9498. 25. Wallace JL, McCaerty D-M, Carter L et al. Tissue-selective inhibition of prostaglandin synthesis in rat by tepoxalin: anti-inammatory without gastropathy. Gastroenterology 1993; 105: 16301636. 26. Wallace JL & Tigley AW. New insights into prostaglandins and mucosal defence. Alimentary Pharmacology and Therapeutics 1995; 9: 227235. 27. Langenbach R, Morham SG, Tiano HF et al. Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acid-induced inammation and indomethacin-induced gastric ulceration. Cell 1995; 83: 483492. 28. Whittle BJR. Mechanisms underlying gastric mucosal damage induced by indomethacin and bile salts, and the actions of prostaglandins. British Journal of Pharmacology 1977; 60: 455460. 29. Rashid F & Toolon EP. Misoprostol healed a benign nonsteroidal antiinammatory drug-induced gastric ulcer in a patient with pentagastrin-fast achlorhydria. Journal of Clinical Gastroenterology 1993; 16: 219221. 30. Janssen M, Dijkmans BAC, Vanderbroucke JP et al. Achlorhydria does not protect against benign upper gastrointestinal ulcers during NSAID use. Digestive Diseases and Sciences 1994; 39: 362365. 31. Hogaboam CM, Bissonnette EY, Chin BC et al. Prostaglandins inhibit inammatory mediator release from rat mast cells. Gastroenterology 1993; 104: 122129. 32. Rioux KP & Wallace JL. Mast cell activation increases gastric mucosal injury through peptido-leukotriene release. American Journal of Physiology 1994; 266: G863G869. 33. Rosam AC, Wallace JL & Whittle BJR. Potent ulcerogenic actions of platelet-activating factor on the stomach. Nature 1986; 319: 5456. 34. Rioux KP & Wallace JL. Mast cells do not contribute to NSAID-induced gastric mucosal injury in rodents. Alimentary Pharmacology and Therapeutics 1996; 10: 173180. 35. Eastwood GL & Quimby GF. Eect of chronic aspirin ingestion on epithelial proliferation in rat fundus, antrum, and duodenum. Gastroenterology 1982; 82: 852856. 36. Wallace JL & Granger DN. The cellular and molecular basis for gastroduodenal mucosal defense. FASEB Journal 1996; 10: 731740.

Pathogenesis of NSAID ulceration 157 37. Ashley SW, Sonnenschein LA & Cheung LY. Focal gastric mucosal blood ow at the site of aspirininduced ulceration. American Journal of Surgery 1985; 149: 5359. 38. Kitahora T & Guth PH. Eect of aspirin plus hydrochloric acid on the gastric mucosal microcirculation. Gastroenterology 1987; 93: 810817. 39. Gana TJ, Huhlewych R & Koo J. Focal gastric mucosal blood ow in aspirin-induced ulceration. Annals of Surgery 1987; 205: 399403. 40. Rainsford KD. Microvascular injury during gastric damage by anti-inammatory drugs in pigs and rats. Agents and Actions 1983; 13: 457460. *41. Wallace JL, Keenan CM & Granger DN. Gastric ulceration induced by nonsteroidal anti-inammatory drugs is a neutrophil-dependent process. American Journal of Physiology 1990; 259: G462G467. 42. Granger DN, Kvietys PR & Perry MA. Leukocyte-endothelial cell adhesion caused by ischemia and reperfusion. Canadian Journal of Physiology and Pharmacology 1993; 71: 6775. 43. Hernandez LA, Grisham MB, Twohig B et al. Granulocytes: the culprit in ischemic damage to the intestine. American Journal of Physiology 1987; 253: H669H703. 44. Asako H, Kubes P, Wallace JL et al. Indomethacin-induced leukocyte adhesion in mesenteric venules: role of lipoxygenase products. American Journal of Physiology 1992; 262: G903G908. 45. Asako H, Kubes P, Wallace JL et al. Modulation of leukocyte adhesion to rat mesenteric venules by aspirin and salicylate. Gastroenterology 1992; 103: 146152. 46. Wallace JL, McKnight W, Miyasaka M et al. Role of endothelial adhesion molecules in NSAID-induced gastric mucosal injury. American Journal of Physiology 1993; 265: G993G998. 47. Wallace JL & McKnight GW. The mucoid cap over supercial gastric damage in the rat. A highpH microenvironment dissipated by nonsteroidal antiinammatory drugs and endothelin. Gastroenterology 1990; 99: 295304. 48. Yoshida N, Takemura T, Granger DN et al. Molecular determinants of aspirin-induced neutrophil adherence to endothelial cells. Gastroenterology 1993; 105: 715724. 49. Andrews FJ, Malcontenti-Wilson C & O'Brien PE. Eect of nonsteroidal anti-inammatory drugs on LFA1 and ICAM-1 expression in gastric mucosa. American Journal of Physiology 1994; 266: G657G664. 50. Wallace JL, Arfors K-E & McKnight GW. A monoclonal antibody against the CD18 leukocyte adhesion molecule prevents indomethacin-induced gastric damage in the rabbit. Gastroenterology 1991; 100: 878883. 51. McCaerty DM, Granger DN & Wallace JL. Indomethacin-induced gastric injury and leukocyte adherence in arthritic versus healthy rats. Gastroenterology 1995; 109: 11731180. *52. Santucci L, Fiorucci S, Giansanti M et al. Pentoxifylline prevents indomethacin induced acute gastric mucosal damage in rats: role of tumour necrosis factor alpha. Gut 1994; 35: 909915. 53. Kunkel SL, Wiggins RC, Chensue SW & Larrick J. Regulation of macrophage tumor necrosis factor production by prostaglandin E2. Biochemical and Biophysics Research Communications 1986; 137: 404410. 54. Rothlein R, Czajkowski M, O'Neill MM et al. Induction of intercellular adhesion molecule 1 on primary and continuous cell lines by pro-inammatory cytokines. Regulation by pharmacologic agents and neutralizing antibodies. Journal of Immunology 1988; 141: 16651669. 55. Appleyard CB, McCaerty DM, Tigley AW et al. Tumour necrosis factor mediation of NSAID-induced gastric damage: role of leukocyte adherence. American Journal of Physiology 1996; 270: G42G48. 56. Rainsford KD. The eects of 5-lipoxygenase inhibitors and leukotriene antagonists on the development of gastric lesions induced by nonsteroidal antiinammatory drugs in mice. Agents and Actions 1987; 21: 316319. 57. Pihan G, Rogers C & Szabo S. Vascular injury in acute gastric mucosal damage: mediatory role of leukotrienes. Digestive Diseases and Sciences 1988; 33: 625632. 58. Vaananen PM, Keenan CM, Grisham MB & Wallace JL. A pharmacological investigation of the role of leukotrienes in the pathogenesis of experimental NSAID-gastropathy. Inammation 1992; 16: 227240. 59. Hudson N, Balsitis M, Everitt S & Hawkey CJ. Enhanced gastric mucosal leukotriene B4 synthesis in patients taking non-steroidal anti-inammatory drugs. Gut 1993; 34: 742747. 60. Taha AS, Dahill S, Morran C et al. Neutrophils, Helicobacter pylori, and nonsteroidal anti-inammatory drug ulcers. Gastroenterology 1999; 116: 254258. 61. Taha AS, Hudson N, Hawkey CJ et al. Famotidine for the prevention of gastric and duodenal ulcers caused by nonsteroidal antiinammatory drugs. New England Journal of Medicine 1996; 334: 14351439. 62. Del Soldato P, Foschi D, Benoni G & Scarpignato C. Oxygen free radicals interact with indomethacin to cause gastrointestinal injury. Agents and Actions 1985; 17: 484488. 63. Vaananen PM, Meddings JB & Wallace JL. Role of oxygen-derived free radicals in indomethacin-induced gastric injury. American Journal of Physiology 1991; 261: G470G475. 64. Yoshida N, Cepinskas G, Granger DN et al. Aspirin-induced, neutrophil-mediated injury to vascular endothelium. Inammation 1995; 19: 297312.

158 J. L. Wallace 65. Murakami K, Okajima K, Harada N et al. Plaunotol prevents indomethacin-induced gastric mucosal injury in rats by inhibiting neutrophil activation. Alimentary Pharmacology and Therapeutics 1999; 13: 521520. 66. Morris GP & Wallace JL. The roles of ethanol and of acid in the production of gastric mucosal erosions in rats. Virchows Archives B (Cell Pathology) 1981; 38: 2338. 67. Svanes K, Ito S, Takeuchi K & Silen W. Restitution of the surface epithelium of the in vitro frog gastric mucosa after damage with hyperosmolar sodium chloride. Gastroenterology 1982; 82: 14091426. *68. Lacy ER & Ito S. Rapid epithelial restitution of the rat gastric mucosa after ethanol injury. Laboratory Investigation 1984; 51: 573583. 69. Moore R, Carlson S & Madara JL. Rapid barrier restitution in an in vitro model of intestinal epithelial injury. Laboratory Investigation 1989; 60: 237244. 70. Wallace JL & Whittle BJR. Role of mucus in repair of gastric epithelial damage in the rat. Inhibition of epithelial recovery by mucolytic agents. Gastroenterology 1986; 91: 603611. 71. Black BA, Morris GP & Wallace JL. Eects of acid on the basal lamina of the rat stomach and duodenum. Virchows Archives B (Cell Pathology) 1985; 50: 109118. 72. Armstrong CP & Blower AL. Non-steroidal anti-inammatory drugs and life threatening complications of peptic ulceration. Gut 1997; 28: 527532. 73. Stadler P, Armstrong D, Margalith D et al. Diclofenac delays healing of gastroduodenal mucosal lesions. Double-blind, placebo-controlled endoscopic study in healthy volunteers. Digestive Diseases and Sciences 1991; 36: 594600. *74. Mizuno H, Sakamoto C & Matsuda K. Induction of cyclooxygenase-2 in gastric mucosal lesions and its inhibition by the specic antagonist delays healing in mice. Gastroenterology 1997; 112: 387397. 75. Schmassmann A, Peskar BM, Stettler C et al. Eects of inhibition of prostaglandin endoperoxide synthase-2 in chronic gastro-intestinal ulcer models in rats. British Journal of Pharmacology 1998; 123: 795804. 76. Jackson LM, Wu KC, Mahida YR et al. Cyclooxygenase (COX)-1 and -2 in normal, inamed and ulcerated human gastric mucosa. Gut 1999; (in press). 77. Reuter BK, Asfaha S, Buret A et al. Exacerbation of inammation-associated colonic injury in rat through inhibition of cyclooxygenase-2. Journal of Clinical Investigation 1996; 98: 20762085. 78. Wallace JL. Selective COX-2 inhibitors: is the water becoming muddy? Trends in Pharmacological Sciences 1999; 20: 46. 79. Wallace JL, McKnight W & Bak A. Selective inhibitors of COX-2: are they really eective? Are they really GI-safe? Journal of Clinical Gastroenterology 1998; 27: S28S34. 80. Prichard PJ, Kitchingman GK, Walt RP et al. Human gastric mucosal bleeding induced by low dose aspirin, but not warfarin. British Medical Journal 1989; 298: 493496. 81. Hawkey CJ, Hawthorne AB, Hudson N et al. Separation of the impairment of haemostasis by aspirin from mucosal injury in the human stomach. Clinical Science 1991; 81: 565573. 82. The SALT Collaborative Group. Swedish Aspirin Low-Dose Trial (SALT) of 75 mg aspirin as secondary prophylaxis after cerebrovascular ischaemic events. Lancet 1991; 338: 13451349. 83. Meade TW, Roderick PJ, Brennan PJ et al. Extracranial bleeding and other symptoms due to low dose aspirin and low intensity oral anticoagulation. Thrombosis and Haemostasis 1992; 68: 16. 84. Cryer B, Luk G & Feldman M. Eects of very low doses of aspirin (ASA) on gastric, duodenal & rectal prostaglandins (PGs) & mucosal injury. Gastroenterology 1995; 108: A77. 85. Agrawal NM. Epidemiology and prevention of non-steroidal anti-inammatory drug eects in the gastrointestinal tract. British Journal of Rheumatology 1995; 34 (supplement 1): 510. 86. Koch M, Capurso L, Dezi A et al. Prevention of NSAID-induced gastroduodenal mucosal injury: metaanalysis of clinical trials with misoprostol and H2-receptor antagonists. Digestive Diseases 1995; 13 (supplement 1): 6274. 87. Simon TJ, Berger ML, Hoover ME et al. A dose-ranging study of famotidine in prevention of gastroduodenal lesions associated with non-steroidal anti-inammatory drugs (NSAIDs): results of a US multicenter trial. American Journal of Gastroenterology 1994; 89: 1644 (abstract). *88. Hawkey CJ, Karrasch JA, Szczepanski L et al. Omeprazole compared with misoprostol for ulcers associated with nonsteroidal antiinammatory drugs. New England Journal of Medicine 1998; 338: 727734. 89. Green FW, Kaplan MM, Curtis LE & Levine PH. Eect of acid and pepsin on blood coagulation and platelet aggregation: a possible contributor to prolonged gastroduodenal mucosal hemorrhage. Gastroenterology 1978; 74: 3843. 90. Szabo S, Folkman J, Vattay P et al. Accelerated healing of duodenal ulcers by oral administration of a basic broblast growth factor in rats. Gastroenterology 1994; 106: 11061111. 91. Soll AH. Mechanisms of action of antisecretory drugs: studies on isolated canine fundic mucosal cells. Scandinavian Journal of Gastroenterology 1986; 21: 16.

Pathogenesis of NSAID ulceration 159 92. Ligumsky M, Goto Y & Yamada T. Prostaglandins mediate inhibition of gastric acid secretion by somatostatin in the rat. Science 1983; 219: 301303. *93. Xie W, Robertson DL & Simmons DL. Mitogen-inducible prostaglandin G/H synthase: a new target for nonsteroidal antiinammatory drugs. Drug Development Research 1992; 25: 249265. *94. Wallace JL, Bak A, McKnight W et al. Cyclooxygenase-1 contributes to inammatory responses in rats and mice: implications for GI toxicity. Gastroenterology 1998; 115: 101109. 95. Wallace JL, Chapman K & McKnight W. Limited anti-inammatory ecacy of cyclooxygenase-2 inhibition in carrageenan-airpouch inammation. British Journal of Pharmacology 1999; 126: 12001204. 96. Wallace JL, Elliott SN, Del Soldato P et al. Gastrointestinal sparing anti-inammatory drugs: the development of nitric oxide-releasing NSAIDs. Drug Development Research 1997; 42: 144149.