Nitric oxide as a mediator of gastrointestinal mucosal injury?—Say it ain't so

Nitric oxide has been suggested as a contributor to tissue injury in various experimental models of gastrointestinal inflammation. However, there is overwhelming evidence that nitric oxide is one of the most important mediators of mucosal defence, influencing such factors as mucus secretion, mucosal blood flow, ulcer repair and the activity of a variety of mucosal immunocytes. Nitric oxide has the capacity to down-regulate inflammatory responses in the gastrointestinal tract, to scavenge various free radical species and to protect the mucosa from injury induced by topical irritants. Moreover, questions can be raised regarding the evidence purported to support a role for nitric oxide in producing tissue injury. In this review, we provide an overview of the evidence supporting a role for nitric oxide in protecting the gastrointestinal tract from injury.


Introduction
It is now widely accepted that nitric oxide (NO) synthesized from -arginine is the substance that accounts for the activity originally described by Furchgott and Zawadzki as 'endothelium-derived relaxing factor'. Nitric oxide is produced by virtually all mammalian cells. In the intestine, nitric oxide is produced by three distinct enzyme systems, referred to as nitric oxide synthases (NOS). A neurally associated, constitutive form of NOS found in neurons of the enteric nervous system is referred to as NOS1 or ncNOS. An inducible form of the enzyme, referred to as NOS2 or iNOS, is found under appropriate circumstances in endothelial cells, the epithelium and in a number of immunocytes. This enzyme can be induced by exposure to various cytokines and endotoxin. NOS3 (or ecNOS) is a constitutively expressed enzyme found primarily in the endothelium lining blood vessels, but may also exist in intestinal epithelial cells and mast cells. One of the important distinguishing features between the constitutive and inducible forms of NOS is the calcium-dependency of the former but not the latter. For the most part, NOS1 and NOS3 can be regarded as the forms that are responsible for producing nitric oxide in a physiological context, whereas NOS2 produces nitric oxide in pathophysiological circumstances. However, this statement should not be construed as meaning nitric oxide derived from NOS2 only acts in a detrimental manner.
In recent years, a plethora of studies have implicated NO as a critical mediator of mucosal defence in the gastrointestinal tract. On the other hand, it has been suggested by many groups that nitric oxide contributes to mucosal injury in several human gastrointestinal disorders and in experimental animal models of gastrointestinal injury. Like oxygen, high concentrations of NO oxide are cytotoxic. While there are considerable data suggesting that NO production is greatly elevated in situations of mucosal inflammation, such as in inflammatory bowel disease, it has yet to be clearly established that NO plays a primarily damaging role in those circumstances, or that it ever accumulates to concentrations that could achieve such results.
In this review, we summarize some of the key evidence suggesting a role for nitric oxide in the maintenance of gastrointestinal mucosal integrity, and we challenge the hypothesis that nitric oxide, when present in 'large' amounts, necessarily results in mucosal injury and dysfunction.
Nitric oxide and mucosal defence The mucosa of the gastrointestinal tract is continuously exposed to potentially damaging substances, such as acid, digestive enzymes, bacterial toxins and bile. The ability of the mucosa of the gastrointestinal tract to resist injury by these sub-(C) 1995 Rapid Science Publishers P. Kubes andJ. L. Wallace stances is attributable to a complex and dynamic network of factors collectively referred to as 'mucosal defence'. This includes extracellular factors, such as mucus, which helps to prevent abrasive injuey to the mucosa and to impair attachment of microbes to the epithelium. The epithelium of the gastrointestinal tract itself forms an important barrier, restricting the movement of large molecules into the mucosa and secreting fluid into the lumen in order to wash away potentially harmful microbes or toxins. The subepithelium is richly inneeeated by both the extrinsic and enteric nervous systems, which are capable of triggering changes in several of the components of the mucosal defence network when damage to the epithelium is detected or if toxins enter the lamina propria in sufficient quantities. One of the most rapid and important mucosal defence responses is an increase in mucosal blood flow, which acts to dilute the toxic substances, neutralize them and remove them from the tissue before they can accumulate to damaging concentrations. This has been particularly well characterized in the stomach, where back-diffusing acid can trigger sensory afferent neurons beneath the epithelium. The subsequent release of calcitonin gene-related peptide in the vicinity of the submucosal arterioles results in a rapid and profound increase in mucosal blood flow, thereby effectively neutralizing the backdiffusing acid that had triggered the response. The various components of mucosal defence are co-ordinated and modulated via the release of a number of soluble mediators. For example, prostaglandins play a very important role in the maintenance of mucosal integrity, stimulating virtually every component of the mucosal defence network. Over the past 10 years, substantial data 398 Mediators of Inflammation Vol 4. 1995 have been generated to suggest that nitric oxide is another key mediator of mucosal defence in the gastrointestinal tract.
Nitric oxide can protect the GI tract MacNaughton et al. 2 demonstrated that nitric oxide itself or an NO donor could protect the gastric mucosa from damage induced by topically applied ethanol. Moreover, administration of the guanylate cyclase inhibitor, methylene blue, or the nitric oxide scavenger, oxyhaemoglobin, prevented the protective effects of exogenous NO and increased the susceptibility of the gastric mucosa to injury. These findings have now been extended by Kitagawa et al., who showed protective effects of nitric oxide donors in an acidinduced model of damage in the rat, by Lopez-Belmonte et al. 4 who demonstrated that nitric oxide donors reduced the severity of ethanolinduced damage, and by Masuda et al. 5 who demonstrated that blockade of nitric oxide synthesis greatly increased the susceptibility of the stomach to damage induced by ethanol, a summary of the studies implicating nitric oxide in gastrointestinal mucosal resistance to injury is provided in Table 1.
The ability of nitric oxide to reduce the severity of damage induced bynonsteroidal antiinflammatory drugs (NSAIDs) has recently been exploited in an attempt to produce NSAIDs which are not ulcerogenic. By attaching a nitric oxide releasing moiety to standard NSAIDs, the ulcerogenic effects of these compounds was greatly reduced without adversely affecting their 68 anti-inflammatory properties. These compounds appear to release nitric oxide very slowly and in small amounts. Thus, the compounds do not affect systemic arterial blood pressure, but mucosa is exposed to an irritant. Blockade of the release of NO is sufficient to maintain gastric this response, by ablation of sensory afferent mucosal blood flow and prevent leukocyte neurons, renders the mucosa unable to withstand adherence to post-capillary venules. 7'u Decreased even brief exposure to topical irritants such as mucosal blood flow and leukocyte adherence are 15% ethanol.4The reactive hyperaemic response obseeeed with standard NSAIDs and are believed is a nitric oxide-dependent response. Calcitonin to contribute to the pathogenesis of NSAID-gene-related peptide accounts for the largest induced gastropathy. 9 component of the gastric reactive hyperaemic There is also recent evidence to suggest that response, 15 and its effects on vascular smooth endogenous nitric oxide derived from the induci-muscle tone are mediated through the generable isoform of nitric oxide synthase (NOS2) is tion, presumably by the vascular endothelium, of capable of protecting the stomach, which flies in nitric oxide. Prior treatment of rats with inhibithe face of claims that the large amounts of nitric tors of nitric oxide synthase abolishes the reacoxide produced via this enzyme will cause tissue rive hyperaemic response to topical irritants and injury. Tepperman and Soper demonstrated greatly increases the susceptibility of the stomach that administration of endotoxin to rats led to a to damage. significant increase in NOS2 activity within the Nitric oxide also plays a key role in modulating gastric mucosa. When the stomach was chal-basal gastric blood flow, since it can be reduced lenged by intragastric administration of ethanol, by administration of NOS inhibitors. 7 Moreover, the rats previously given endotoxin were found this reduction of mucosal blood flow by NOS to be significantly more resistant to damage. That inhibitors probably underlies their ability to this protection was mediated via nitric oxide was increase the susceptibility of the gastric mucosa supported by their observation that administrato iniury induced by various topical irritants, such tion of L-NAME (a non-selective NOS inhibitor) as ethanol and NSAIDs. 4'18'19 or dexamethasone inhibited the protective effects associated with endotoxin treatment. In the case of dexamethasone, they demonstrated a selective inhibition of the inducible nitric oxide synthase within the gastric mucosa.

Nitric oxide modulates mucus secretion
Nitric oxide is important in repair of mucosal injury Repair of gastrointestinal injury can be subdivided into two major types. When damage is limited to the most superficial layers of cells, rapid repair of the epithelium is possible through Mucus secretion by the epithelium is believed the process of restitution. This involves migration to form one of the primary levels of mucosal of healthy cells from the uninjured margins, over defence against acid in the stomach and duode-the exposed basement membrane, g The restitunum. Mucus also plays a key role in preventing tion process in vivo is dependent upon the abrasive injury to the gut epithelium, in limiting maintenance, in the region of injury, of a microthe access of luminal bacteria to the epithelial environment of near neutral pH (the basement surface and as an anti-oxidant. Bicarbonate is membrane is very susceptible to damage by acid produced in small amounts relative to the and restitution will not occur if the basement amount of acid produced by the stomach, but membrane is damaged). 21 Mucosal hyperaemia when trapped within mucus on the surface of and the leakage of plasma at the site of injury is the epithelium can form a pH gradient. Nitric responsible for maintaining the neutral pH oxide appears to be one of the most important microenvironment. Takeuchi et a/. 22 recently signals for mucus secretion. Brown et al. 2 demonstrated that blockade of nitric oxide prodemonstrated that nitric oxide donors and a duction results in an impairment of the vascular cGMP analogue could stimulate mucus release, response and the consequent alkaline flux into Furthermore, Price et al. reported that nitric the lumen and in turn, to an impairment of the oxide was the mediator responsible for cholinerrestitution process. gic-stimulated mucus release.
The second major type of repair that occurs in the gastrointestinal tract is that associated with injury that involves the full thickness of the Nitric oxide mediates mucosal mucosa, and perhaps even the submucosa and muscularis (i.e. true ulcers). This process hyperaemic responses involves the formation of granulation tissue at the As outlined above, a very important compo-ulcer base, the process of angiogenesis and the nent of mucosal defence is the elevation of subsequent re-establishment of glandular strucmucosal blood flow that occurs when the ture (this occurs primarily from the margins of the ulcer). While the mechanisms of action are the capacity to produce nitric oxide, and could yet to be determined, there is now clear evidence therefore modulate epithelial permeability. There that nitric oxide can accelerate this healing is now strong evidence for mast cells exerting process. Konturek et a/. 23 first demonstrated that such control through the release of nitric oxide administration of a nitric oxide donor (glyceryl (see below). trinitrate) significantly accelerated the healing of gastric ulcers in the rat. They also showed that Nitric oxide modulates the mucosal administration of an inhibitor of nitric oxide immune system synthesis impaired ulcer healing. These results were recently confirmed by Elliott et al. 24 who The mucosal immune system plays an essential showed that oral but not intraperitoneal glyceryl role in protecting us from the enormous quantrinitrate accelerated ulcer healing in the rat. tities of microbes, microbial products and other Moreover, administration of a nitric oxide-releasantigens present in the lumen of the gut. The ing NSAID derivative, but not the parent NSAID, normal gut could be said to be in a state of significantly accelerated ulcer healing. 24 'controlled inflammation'. While there is a constant exposure of the gut to substances that Nitric oxide regulates intestinal mucosal could stimulate an immune response, which barrier function could lead to inappropriate injury to the host tissue, a number of mediators act to keep the Inhibition of nitric oxide synthase results, inflammatory response in check. Included among within minutes, in a substantial increase in small these down-regulatory molecules are interleukinintestinal epithelial permeability. For example, 10, prostaglandins and nitric oxide. the transepithelial movement of a marker such as The ability of nitric Oxide to down-regulate 5Cr-EDTA is greatly enhanced following adminisimmunocyte function has been demonstrated in tration of the NO synthase inhibitor L-NAME. 25 a number of cell systems. For example, macro-This increase in epithelial permeability can also phage-derived NO has been suggested to supbe induced by other inhibitors of nitric oxide press the proliferation of T lymphocytes and to synthesis, and can be reversed by nitric oxide mediate several aspects of macrophage funcdonors, -arginine or cGMP analogues. On the tion. As outlined below, nitric oxide is a potent other hand, administration of biologically inactive suppressor of neutrophil function, inhibiting enantiomers of the NO synthase inhibitors had both adherence and activation. With respect to no effect on epithelial permeability. This role for the gastrointestinal tract, perhaps the best charnitric oxide in the intestine has now been acterized immunoregulatory effect of nitric oxide demonstrated in the cat, 25 guinea-pig 2 and rat. 2v is its ability to inhibit mediator release from mast It is important to note that in the studies cells. Salvemini et al. first reported the ability of described above, inhibition of nitric oxide synthexogenous nitric oxide to inhibit release of histaesis did not result in widespread disruption of mine from peritoneal mast cells. These studies the epithelium. Indeed, the permeability of the were later extended by Masini et al., 32 who epithelium to molecules such as albumin (MW demonstrated that mast cells could produce 60 000), was not affected by the inhibitors of NO nitric oxide, and by Hogaboam et aL, who synthase. Moreover, the effect of NO synthase reported that the nitric oxide produced by the inhibition on permeability to EDTA was very mast cell autoregulated the release of other rapidly reversed by administration of an NO inflammatory mediators, such as platelet-activatdonor, demonstrating that the permeability ing factor (PAF). The ability of the mast cell to change was not due to damage of the epithelium produce nitric oxide could be greatly enhanced or to the tight junctions. As epithelial cells in the by very brief exposure to interleukin-1. Nitric gastrointestinal tract have the capacity to produce oxide release from mast cells also contributes to nitric oxide, 28 it is possible that these cells can the ability of these cells to kill target tumour exert autoregulation of their own permeability, cells. 4 The factors that increase or decrease nitric oxide These in vitro studies of mast cells have now production by the epithelium have not yet been been extended to in vivo models. Kanwar et well characterized, although it is interesting to a/. 27 demonstrated that administration of a nitric note that the reduction of epithelial permeability oxide synthase inhibitor resulted in a marked that can be induced by interferon-gamma was increase in serum levels of a protease specific to recently shown to be mediated via the produc-mucosal mast cells. Using histochemical tion of nitric oxide. 29 Of course, numerous other methods, degranulation of connective tissue-type cells in the lamina propria, including enteric mast cells in the mesentery was also demonneurons, mast cells and macrophages, also have strated. 5 Associated with the degranulation of mast cells was a rapid increase in permeability of Nitric oxide inhibits neutrophil adherence the intestinal epithelium to the small molecular and activation weight marker, EDTA. 2v The permeability changes observed could be prevented by pre-Neutrophils are a major contributor to gastrotreating the rats with various mast cell stabilizers intestinal injury, including that associated with or with receptor antagonists to PAF or the histainflammatory bowel disease 42 and Helicobacter mine H-1 receptor. Taken together, these studies pylori-associated ulcer disease. 43'44 Moreover, a demonstrate that inhibition of nitric oxide synth-key role for neutrophils as effectors of tissue esis leads to mast cell degranulation. The release damage has been demonstrated in various of mediators such as histamine and PAF results experimental models of intestinal injury and in increases in intestinal epithelial permeability, inflammation, including that induced by ischae-In addition, inhibition of nitric oxide synthesis mia-reperfusion, 45'4 PAF, 47 indomethacin  led to increased numbers of leukocvtes adhering and ethanol. 51 '52 The evidence for a role for neu- 36 to mesenteric post-capillary venules. This effect, trophils in these models includes the observawhich is discussed in more detail below, also tions that: (1)neutrophils infiltrate the inflamed appeared to be mast cell-dependent. 5 tissue; (2) rendering animals neutropenic Similar findings have been reported by Kurose reduces tissue injury; and (3) antibodies against et al., r who studied the effects of ischaemianeutrophil and/or endothelial adhesion molereperfusion on intestinal mast cells. They visually cules reduce the extent of injury. assessed mast cell integrity after 30 min of reper-Nitric oxide has been suggested as an imporfusion of ischaemic mesentery and noted a sig-rant modulator of neutrophil adherence to the nificant increase in degranulated mast cells that vascular endothelium. Inhibition of nitric oxide was entirely inhibited by nitric oxide donors, synthesis leads to elevated numbers of neu-They also demonstrated that endogenous nitric trophils adhering to walls of vessels. 5 As outoxide was reduced by 90% following the period lined above, this may be in part a mast cellof ischaemia, while supplementation of the tissue dependent phenomenon. Based on these obserwith nitric oxide donors significantly reduced vations, it has been suggested that one of the leukocyte infiltration and microvascular dysfuncmechanisms through which nitric oxide can tion. The authors concluded that the reduction in reduce injury in the gastrointestinal tract is nitric oxide production associated with ischae-through inhibition of neutrophil recruitment to mia-reperfusion led to destabilization of mast the tissue. Indeed, in numerous studies, adminiscells, as a consequence of reduced nitric oxide tration of nitric oxide donors reduced gastroproduction, and the subsequent release of prointestinal myeloperoxidase activity (an index of adhesive molecules such as histamine, PAF and tissue granulocyte numbers) as well as reducing leukotrienes. It is these mediators which ultitissue injury. Andrews eta/. 54 demonstrated that mately elicited the inflammatory response, fewer neutrophils infiltrated the post-ischaemic Although the platelet is generally viewed in the gastric mucosa following nitroprusside or acetcontext of its important role in thrombosis and ylcholine administration and this regimen of coagulation, there is considerable evidence that nitric oxide generators protected the reperfused platelets also contribute significantly to inflamtissue. Kurose et aL 55 demonstrated that adminismatory processes. 8 Nitric oxide is a very potent tration of nitric oxide donors resulted in a inhibitor of platelet aggregation and adherence, 9 marked attenuation of leukocyte adhesion and and may therefore exert anti-inflammatory effects emigration, as well as vascular dysfunction in the by suppressing platelet activation. Nishida et al. 4 post-ischaemic mesentery. Mackendrick et al. 56 recently reported that suppression of nitric demonstrated that inhibition of nitric oxide prooxide synthesis during administration of endoduction resulted in an exacerbation of PAFtoxin profoundly increased platelet and fibrin induced neutrophil influx into the mucosa as deposition in liver sinusoids. They suggested that well as an increase in tissue injury. As mentioned inhibition of endogenous nitric oxide results in above, the addition of a nitric oxide-releasing platelet aggregation which contributes to plugmoiety to standard NSAIDs resulted in prevention ging of the sinusoids and tissue injury. Har-of the neutrophil adherence to mesenteric postbrecht et al. 4 also suggested an important role capillar venules normally seen with the parent for inducible NOS in preventing thrombus for-NSAID.
mation in situations of endotoxaemia. They It should be noted that prevention of infiltrafound that suppression of NO synthesis in endotion of circulating neutrophils into the intestine toxaemic animals led to intrahepatic thrombus may not fully explain the protective effects of formation and oxygen radical-mediated liver nitric oxide. For example, in experimental damage, ischaemia-reperfusion-induced injury, circulating neutrophils appear to contribute significantly to influx) and both superoxide dismutase and NO the microvascular but not the mucosal dysfunc-donors directly inhibited this response, further tion. 5 This is based on the fact that pretreatment suggesting a lotential anti-oxidant ability of these with an anti-CD18 antibody, which prevents leu-NO donors. Finally, Wink et al. 69 recently prokocyte adhesion to post-capillary venules, sig-posed a protective role for NO donors in cellular nificantly attenuated the increase in microvascular injury induced by hydrogen peroxide. Since catapermeability but did not affect the rise in lase, which is a hydrogen peroxide-detoxiTin mucosal permeability. 5v On the other hand, nitric agent, can also reduce intestinal inflammation, oxide donors have been shown to reduce both the possibility that NO donors counter both microvascular and mucosal injury in the intestine, superoxide and hydrogen peroxide-induced Since adhering neutrophils do not mediate the tissue injury cannot be excluded. mucosal permeability alterations in reperfused intestine, then nitric oxide must have other bene-Large amounts of nitric oxide do not ficial effects in addition to its anti-adhesive propcause intestinal injury erties. One possibility is that nitric oxide can inhibit functions of neutrophils other than their It has repeatedly been suggested that producability to adhere to the vascular endothelium. For tion of 'large' amounts of nitric oxide, such as example, nitric oxide may directly inactivate the may occur when the inducible form of NOS is enzyme responsible for the oxidative burst that is present, will result in damage within the gastrogenerated by activated neutrophils. Clancy et al. 58 intestinal tract or, indeed, in other tissues. In an have demonstrated that nitric oxide can inhibit attempt to test this hypothesis, in vivo and in superoxide production from neutrophils by vitro studies on the feline intestine and human directly inhibiting NADPH oxidase. Therefore, in epithelial cells, respectively, were performed addition to scavenging superoxide (discussed which involved local infusion of high concentrabelow), nitric oxide can prevent the synthesis of tions of nitric oxide donors (CAS 754 or SIN-l) superoxide and associated oxidants including into autoperfused segments of cat ileum. 7 Rather hydrogen peroxide. Thus, in addition to prevent-than looking for overt tissue damage, very subtle ing leukocyte adhesion and infiltration into the functional alterations were used as indices of reperfused intestine, nitric oxide may also intestinal dysfunction. These included measuredirectly affect the cytotoxicity of neutrophils by ments of both microvascular and mucosal peraltering the ability of this cell to produce oxi-meability. These studies clearly demonstrated that dants. Other inhibitory effects of nitric oxide local infusion of compounds that would release donors on neutrophil function have also been 'large' amounts of NO failed to cause changes in reported. 59-microvascular permeability, changes in the integrity of the mucosal barrier or impairment in the Nitric oxide is an anti-oxidant functional aspects (absorption or secretion) of the small bowel. Furthermore, when these NO In addition to being anti-adhesive, nitric oxide donors were added to endothelial or epithelial has the capacity to inhibit reactive oxygen meta-cells in vitro, they failed to produce detectable bolites, including superoxide anion, and can damage. This is in sharp contrast to the proprevent the cellular damage attributable to found injury (i.e. lysis) observed when a hypoxhydrogen peroxide. 2' Both superoxide and anthine/xanthine oxidase system, which hydrogen peroxide have been implicated in the generates superoxide anions, was added to the mucosal injury associated with ischaemia-reler-cultured cells. Similar observations have been fusion 64 and the administration of ethanol 5-v or reported by others. For example, nitric oxide NSAIDs. 68 Nitric oxide may play a role in redu-released from DEA/NO or spermine-NO themcing the injury in these models through its ability selves did not cause cytotoxicity and in fact preto inactivate oxidants. Indeed, nitric oxide has vented cellular injury to Chinese hamster lung been shown to rapidly react with sup.eroxide to fibroblasts or rat mesencephalic dopaminergic 62 abolish its biological activity in vitro. There is cells induced by either hydrogen peroxide or also evidence suggesting that this occurs in vivo, superoxide. 9 These data raise some questions since the anti-oxidant capacity of plasma was about the contention that nitric oxide per se doubled by administration of nitric oxide donors directly injures cells of the intestine or other at concentrations that lrevented reperfusion-organs. 63 r induced mucosal injury. Othe in v,vo expert-It might be argued that large amounts of nitric ments demonstrated that administration of a oxide may only produce tissue injury if released superoxide-generating system could induce at in a setting of active inflammation. In an attempt least one feature of inflammation (neutrophil to test this hypothesis, further experiments using an autoperfused segment of cat intestine were derived from a reaction dependent on nitric performed, but in addition to administering nitric oxide synthesis. 2 This evidence supports the oxide donors, the pro-inflammatory mediator concept that peroxynitrite is formed under con-PAF was infused, v By itself, PAF caused a proditions of intestinal inflammation, but does not found increase in both mucosal and micro-address whether or not it contributes to the vascular permeability. Concomitant administration tissue injury. Blockade of nitric oxide synthase of the nitric oxide donor, CAS 754, did not with L-NAME has been shown to reduce the augment these permeability changes; in fact, the severity of experimental colitis v5 and experinitric oxide donor inhibited the permeability mental ileitis. 2< Administration of aminoguanienhancing effects of PAF.
dine, which is a more selective inhibitor for inducible NOS, produced an even more impress-What about peroxynitrite? ive reduction of injury and inflammation. 7 While these data support a role for nitric oxide in pro-Nitric oxide is often described as being highly ducing tissue injury and inflammation, they do reactive and a gas. In biological settings, not directly address the role of peroxynitrite.
however, it is neither, v Nitric oxide reacts very There is also evidence that intracolonic adminisslowly with most biological molecules and theretration of peroxynitrite will cause extensive fore its proposed cytotoxicity is dependent upon damage and inflammation in the rat. v8 Again, conversion to much more reactive oxidants. In however, this does not address the issue of fact, much of the apparent cytotoxicity of nitric whether or not endogenous peroxynitrite prooxide has been attributed to its reaction with duction contributes to tissue injury in colitis. It is superoxide anion to form peroxynitrite] 2 Perox-clear that real progress in answering this quesynitrite is itself cytotoxic, and it can rapidly tion will be dependent upon the development of decompose to the highly reactive and toxic highly selective inhibitors of the inducible hydroxyl radical (OH.) and nitrogen dioxide isoform of NOS, and improved methods for (NO2 o).-71 The tissue damage that can be caused detecting the production of peroxynitrite in viva by peroxynitrite is attributable to its ability to Finally, some caution should be exercised oxidize sulfhydryl groups within cells, thereby when interpreting the results from studies in depleting an important scavenging mechanism, which >NAME is chronically administered and This leads to increased susceptibility of the cell found to reduce the severity of experimental to oxidant-induced damage to membrane lipids intestinal inflammation. While it is possible that and proteins, various enzymes and DNA. 73 this effect of L-NAME is attributable to suppres-There are a couple of points that should be sion of nitric oxide synthesis, and indicative of considered with respect to the putative role of NO playing a role in the production of tissue peroxynitrite in intestinal injury. Rubbo et a/. 74 injury, there are other possible explanations. recently reported that NO actually inhibits perox-First, t-NAME is capable of reducing blood flow ynitrite-induced lipid peroxidation. While NO to tissues, and in doing so, may reduce the influx could participate in the generation of peroxyni-of granulocytes. Secondly, t-NAME may cause trite, by reacting with superoxide anion, and had leukocyte adhesion in tissues other than the one the capacity to cause lipid peroxidation, this only being studied, therefore resulting in an apparent occurred when the concentrations of NO and decrease in infiltration of these inflammatory superoxide anion present were equal. When NO cells into the intestine. Thirdly, in light of the concentrations were increased, as may well be observed ability of L-NAME to cause degranula- 27 the case in situations of inflammation, NO inhibtion of mast cells in the gastrointestinal tract, ited peroxynitrite-dependent lipid peroxidation, one must consider the possibility that with What evidence is there supporting a role for chronic administration of this compound there peroxynitrite in the production of gastrointestinal could be a progressive depletion of inflammatory injury? A problem with attempting to provide mediators from these cells. It is therefore possisuch evidence is the lack of a selective inhibitor ble that such a depletion could contribute to the of the formation or actions of this substance. It beneficial effects of L-NAME administration over a is possible to stain tissues for the presence of prolonged period of time. Finally, virtually all the nitrotyrosine, a product produced when peroxexperimental models of intestinal inflammation ynitrite reacts with tyrosine residues in a tissue. 71 are dependent upon the presence within the gut Miller et a/. 26 demonstrated positive staining for lumen of microbes. It is entirely possible that nitrotyrosine in inflamed ileum of guinea-pig. The chronic L-NAME administration alters the nature fact that inhibition of nitric oxide synthase resul-and/or numbers of microbes within the gut ted in diminished staining for these products lumen and in doing so, alters the magnitude of supports the concept that the nitrotyrosine was the mucosal inflammatory response. These alter-native explanations for the beneficial effects of chronic t-NAME administration in experimental models of gastrointestinal inflammation have yet to be systematically evaluated.

Conclusions
There can be little doubt that nitric oxide is among the most important mediators of gastrointestinal rnucosal defence, influencing virtually every component of the mucosal defence network. There is emerging evidence that NO is an important endogenous scavenger of various free radical species, and that suppression of NO synthesis can lead to oxidant-mediated tissue injury. More controversial is the putative role of NO, when produced in 'large' amounts via the inducible isoform of NOS, in the production of gastrointestinal injury. While there is evidence that chronic suppression of NO synthesis can reduce some indices of tissue injury and inflammation in experimental models, these effects may not be due to inhibition of the production of cytotoxic concentrations of NO. Indeed, there is convincing evidence that administration of large amounts of NO does not cause detectable damage to the mucosa or vasculature of the intestine. There is also recent in vitro data which raise questions about the role of peroxynitfite in the production of tissue injury; specifically, whether conditions ever exist, physiologically or pathophysiologically, in which this substance would be produced in sufficient quantities to cause significant injury. In our view, the vast majority of available data point to nitric oxide serving a critical role in protecting the gastrointestinal mucosa from injury induced by reactive oxygen metabolites and other cytotoxic substances.