Nitric oxide as a mediator of inflammation?—You had better believe it

Nitric oxide has enigmatic qualities in inflammation. In order to appreciate the precise contributions of nitric oxide to a pathophysiological process, one must account for enzyme source, coproduction of oxidants and antioxidant defences, time, rate of nitric oxide production, cellular source, peroxynitrite formation and effects on DNA (mutagenesis/apoptosis). We contend that there is ample evidence to consider nitric oxide as a molecular aggressor in inflammation, particularly chronic inflammation. Therapeutic benefit can be achieved by inhibition of inducible nitric oxide synthase and not the donation of additional nitric oxide. Furthermore, there is growing appreciation that nitric oxide and products derived thereof, are critical components linking the increased incidence of cancer in states of chronic inflammation.

Nmuc oxide has enigmatic qualities in inflammation. In order to appreciate the precise contributions of nitric oxide to a pathophysiological process, one must account for enzyme source, coproduction of oxidants and antioxidant defences, time, rate of nitric oxide production, cellular source, peroxynitrite formation and effects on DNA (mutagenesis/apoptosis). We contend that there is ample evidence to consider nitric oxide as a molecular aggressor in inflammation, particularly chronic inflammation. Therapeutic benefit can be achieved by inhibition of inducible nitric oxide synthase and not the donation of additional nitric oxide. Furthermore, there is growing appreciation that nitric oxide and products derived thereof, are critical components linking the increased incidence of cancer in states of chronic inflammation.
Key words: Apoptosis, inflammation, immune, mutagenesis, peroxynitrite Nitric oxide as a mediator of inflammation?--You had better believe it Mark J. S. Miller Table 1). Many of these same nitrogen oxides, collectively called NOx and which are ultimately derived from NO, are produced by mammalian cells.
NO is a colourless gas that contains an odd number of electrons and is, by definition, a free radical. The unpaired electron is delocalized over the nitrogen and oxygen atoms. NO is relatively unstable in the presence of molecular oxygen with an apparent half-life of approximately 3-5 s .I t will rapidly decompose in the presence of O2 to yield a variety of nitrogen oxides in a 12 complex series of interactions:' 2NO + 02 2NO2 2NO + 2NO2 > 2N203 2N203 + 2H20 > 4NO2-+ 4H + where NO2, N203 and NO2represent nitrogen dioxide, dinitrogen trioxide and nitrite, respectively.
Although this reaction pathway has been proposed to occur in aqueous (i.e. physiological) solutions, there is some controversy as to whether N203 is formed as an intermediate. '4 There is no doubt that the only stable product formed by the spontaneous auto-oxidation of (C) 1995 Rapid Science Publishers NO in oxygenated solutions is NO2-, 2 However, when urine or plasma are analysed, predominantly NOis found in these extracellular fluids. Although the mechanisms by which NO is converted to NOin vivo are not entirely clear, there are at least two possibilities. Ignarro et al. 2 have suggested one mechanism in which the NO2-derived from NO autooxidation is rapidly converted to NOvia its oxidation by certain oxyhaemoproteins (P-Fe 2+ O2) such as oxyhae moglobin or oxymyoglobin: 2P-Fe2+O2 + 3NO2-+ 3NO2-+ 2H + > 2P-Fe 3+ + 3NO3-+ H20 or 4P-Fe2+O 2 + 4NO2-+ 4H+ + 4H + : 4P-Fe 3+ + 4NO3-+ 02 + 2H20 It should be noted, however, that these investigators used large concentrations of NO (300 t.tM) which will rapidly auto-oxidize to NO2-. Although the authors suggested that the NO2-would in Thus, it is within this background that the complexity of NO chemistry can be appreciated. A family of oxidative products is generated under physiological conditions. These oxidative products may have similar or contrasting effects to the parent nitric oxide or alternatively no biological consequence ( Table 2).
Mechanisms of NO-dependent tissue injury and dysfunction Metabolic inhibition: Several groups of investigators have demonstrated that the large amounts of NO produced from activated macrophages are capable of injuring hepatocytes, pancreatic islet cells, and lymphocytes. The mechanisms by which NO injures cells appears to include intracellular iron release, inhibition of mitochondrial function and inhibition of DNA synthesis. It is known for example, that NO will inhibit three mitochondrial enzymes including acotinase (tricarboxylic acid cycle), NADPH-ubiquinone oxidoreductase and succinate-ubiquinone oxidoreductase (Complex I and Complex II or the mitochondrial respiratory chain). -s This inhibition is a result of the NO-mediated degradation of the iron-sulfur clusters associated with these three enzymes. This interaction most probably accounts for the NO-mediated release of nonhaem iron from certain target cells. 9 In addition, NO has been shown to inactivate ribonucleotide reductase, which may account for its ability to inhibit DNA synthesis.
Influence of time: There is a wealth of literature indicating either that NO donors protect or NOS inhibitors exacerbate tissue injury to an acute insult. Virtually all these studies have all been acute in nature, i.e. experimental protocols were only a few hours in duration. One may pose the question, "How can time influence the role of NO in a state of inflammation?" The answer lies primarily in expression of iNOS. Nitric oxide promotes tissue injury when it is produced in excess, but the isoform which produces large quantities of NO (iNOS) is regulated at the transcriptional level. In other words, to proceed from stimulus and activation of iNOS expression via transcription of the gene and its translation into new iNOS protein and the formation of excess NO, takes a defined period of time. Under ideal circumstances this process takes several hours but in a disease where the stimuli may vary in amount and relative importance, then the expression of iNOS may be delayed even further. Until this point is reached only limited production of NO can be achieved.
An excellent example of this phenomenon has come from the laboratory of Drs Brendan Whittle and Salvador Moncada. In 1990 they reported that the NO donor S-nitroso-N-acetylpenicillamine attenuated endotoxin-induced acute intestinal damage, 1 suggesting that NO was protective over the 3 h of the experimental protocol. In a follow-up study reported 4 years later, 1 they confirmed the initial report; the NOS inhibitors t-NAME or t-NMMA, exacerbated endotoxininduced vascular injury when administered at the same time as endotoxin and evaluated 2h later. If, however, administration of the NOS inhibitors was delayed until 3 h after endotoxin administration and the protocol extended for an additional 2h (total of 5h) then the NOS inhibitors were protective. Thus, in the matter of hours NO had transformed from an agent which was antiinflammatory to one which was now pro-inflammatory. Although not confirmed by this group, the likely reason for this transformation is the time required for the expression of iNOS. Endotoxin is a classic stimulant for the expression of iNOS and expression of iNOS has been noted after ischaemia/reperfusion injury. 12 Few studies have addressed the effects of NO donors on a chronic inflammatory process, although there is ample evidence of NOS inhibitors being therapeutic in chronic inflamma- 13 16 tion.
Our colleagues in this debate have shown in models of gastric ulceration that NSAIDs that contain a NO-donating moiety are protective when compared to the parent com- 17 18 pound without the NO donor structure.
The mechanisms for this phenomenon are not clear. While a vascular effect and a role of cGMP-based mechanisms has been proposed, an effect on microbial flora remains a possibility. Nitric oxide is a mediator of host defence and any limitation on bacterial translocation at the site of the ulcer been highlighted .by the application of stroke may potentially promote healing, models to nNOS knock-out mice. 22 In animals Are there conditions where cNOS acts like lacking nNOS the administration of an NOS inhi-iNOS; and if so, under these conditions should bitor resulted in an exacerbation of the injury, NOS inhibition be protective even if the proto-when studied acutely. Furthermore, there is evicols were acute? The answers to these questions dence that the redox state of NO may account are yes and yes! Stroke is an acute condition for 23 neuroprotective or neurodestructive roles. associated with an immediate and sustained If nNOS acts like iNOS in acute stroke is there release of NO, and neuronal injury can be ame-a role for iNOS itself in stroke? It appears that if liorated by attenuating the activity of the neuro-the protocol length is extended an iNOS-only nal isoform of cNOS (nNOS). Neuronal NOS, like contribution to the overall condition 4 can be the endothelial constitutive isoform, eNOS, is seen. Inducible NOS can be expressed in regulated by the intracellular levels of calcium neurons and glia. [25][26][27] In this respect there is an and activation of calmodulin. In contrast, the inflammatory component to cell death in stroke, inducible isoform, iNOS, has calcium-activated this component involves iNOS and is apparent calmodulin as an integral subunit, and therefore hours to days after the initiation of neuronal iNOS activity is independent of intracellular injury. Administration of aminoguanidine, an NOS calcium fluxes. 19 In stroke NMDA receptor activa-inhibitor with selectivity for iNOS, was beneficial tion maintains intracellular calcium at a high in experimental stroke when administered 6h level. Consequently nNOS activation is sustained, after the initiation of stroke. 28 as opposed to the normal pulsitile manner, and as a result, large, sustained release of NO results Influence of enzyme source---iNOS vs. cNOS: in exacerbation of the initial neuronal injury.
With the exception of acute neuronal injury asso-There are reports that NOS inhibitors are not ciated with stroke, a reasonable generalization is 20 21 neuroprotective.
This can be reconciled as that one can predict the role nitric oxide will the result of impairing endothelial NOS activity play in a disease process by determining the preand exacerbating the vascular complications and sence or absence of iNOS. If iNOS is absent then granulocyte infiltration response associated with NO will primarily exhibit anti-inflammatory stroke. These contrasting roles of the neuronal effects. Conversely, the presence of iNOS is an and endothelial NOS isoforms in stroke have excellent indicator that NO is directly contribut-ing to tissue injury. This generalization is blurred by pharmacological probes which were not in states of infection where NO generated from selective for iNOS (either L-NAME or t-NMMA) in iNOS is necessary for killing the invading micro-TNBS ileitis 3 streptococcal cell wall arthritis or organisms and inhibition of NO formation may colitis 4'5 or adjuvant arthritis, 2''5 or spontapromote the growth and extent of infiltration of neous glomerulonephritis in mice. 36 With the this infection and enhance the resultant tissue availability of inhibitors with greater selectivity for injury. Nevertheless, there is evidence that in inducible NOS, e.g. aminoguanidine or -NIL (Nexperimental pneumococcal meningitis excess iminoethyl-lAysine) therapeutic benefit has also NO formation contributes to the brain oedema been observed in a range of chronic inflammaand meningeal inflammation. 29 For the purposes tory states, e.g. peptidoglycan-polysaccharide of this debate we will exclude the complications induced chronic granulomatous colitis, 5 adjuvant of infection, focusing rather on autoimmune dis-arthritis, 6 TNBS and adjuvant ileitis 26' and autoeases or non-infectious inflammation, immune encephalomyelitis in SJL mice. 37 There are relatively few studies which have We have also described a mild ileitis induce6 directly assessed iNOS gene expression in disease by chronic administration of the NOS inhibitor, states, particularly over the time-course of the L-NAME, in otherwise normal guinea-pigs. 8'39 inflammatory condition. In models of gut inflam-While superficially this is consistent with a loss of mation we have observed iNOS gene expression the physiological anti-inflammatory actions of at the onset of symptoms in the TNBS (trininitric oxide, when we further addressed indices trobenzene sulfonic acid) model of ileitis and of nitric oxide formation there was no evidence colitis 3 and adjuvant ileitis. In the TNBS model of attenuated NO formation. Rather, all indices iNOS gene expression is apparent 1 day after the suggested that nitric oxide formation was administration of TNBS and maintained for the increased in conjunction with increased levels of first week of inflammation, as determined by RT-cGMP. In order to explain this perplexing phe-PCR and Western blotting. In contrast, in the nomenon we addressed iNOS gene expression adjuvant model of ileitis, in an analogous manner and noted that chronic administration of t-NAME to adjuvant arthritis, the onset of symptoms is results in the expression of iNOS. This response delayed until about 10 days after administration of was not isolated to the gut, being apparent in the adjuvant. In keeping with the role of iNOS in lung, vasculature and uterus, in rats and guineathe initiation of inflammation, the expression of pigs. It appears that chronic reductions in NO iNOS is apparent in this model at day 14 but not release can result in the expression of iNOS in at day 7. Similarly, increased production in HLAorder to compensate for inadequate levels of B27 transgenic rats is only noted at the onset of NO. This effective mechanism for restoring NO symptoms and persists for the duration of the production is not without complications as iNOS, spontaneous colitis that they exhibit. 3 once formed, is poorly regulated and can result The expression of iNOS has also been reporin excess NO formation which may account for ted in models of arthritis initiated by adjuvant 6 the leukocytosis, ileitis and increased endothelin and streptococcal cell wall fragments. 4 The levels associated with chronic L-NAME administraexpression of iNOS coincides with increased tion. 38'4 This phenomenon is probably only production of nitric oxide and the initiation of achievable because t-NAME is slightly selective inflammatory symptoms. While the loss of matrix for cNOS; iNOS selective inhibitors do not proteoglycans from articular cartilage is a halldisplay this response. Thus it is important to mark of destructive joint disease like osteoar-veriB7 if the administration of an NOS inhibitor thritis or rheumatoid arthritis and may well has reduced NO formation, particularly if admiinvolve NO-related mechanisms, it is now apprenistered chronically, otherwise the data may be ciated that nitric oxide can inhibit glycosamiinterpreted incorrectly. noglycan synthesis. [32][33][34] In other words natural If NO is primarily anti-inflammatory, states of repair mechanisms may also be compromised by tissue injury and inflammation should be charnitric oxide. While this is compelling circum-acterized by a reduction in NO levels. This is not stantial evidence suggesting that NO is involved the case; states of inflammation are virtually in the disease process, these associations do not always characterized by increased NO levels. This reveal the role of NO in that process. Neveris more consistent with the pro-inflammatory role theless, with chronic administration of NOS inhi-of NO, otherwise the interpretation is that NO is bitors it is readily demonstrated that the anti-inflammatory but its efficacy is quite weak. expression of iNOS and increased synthesis of As many of the disease states demonstrating a NO in inflammation is a major contributor to pro-inflammatory role of NO via expression of tissue injury in these states of inflammation, iNOS are autoimmune in character, it is logical to Initially, therapeutic benefit was demonstrated expect that organ transplant rejection responses may also involve iNOS. Recently Worrall et al.
arthritis. 62 Biochemical or immunohistochemical described that iNOS is expressed in allogenic evidence suggests that iNOS is expressed in these cardiac transplants in rats during the rejection disease states and probably accounts for the response, in association with increased systemic increased production of nitric oxide. We postulevels of nitrate/nitrite. 41 Further, administration late that the role of this exaggerated production of the iNOS inhibitor, aminoguanidine, pro-of nitric oxide in these human diseases will parlonged graft survival and significantly reduced allel animal models, i.e. nitric oxide will be prohistological indices of rejection, inflammatory. Them is an expanding list of induced and spontaneous disease states which are character-Cellular sources of iNOS in inflammation: Tradiized by exaggerated release of nitric oxide and tionally, iNOS is associated with macrophages inhibitors of NOS display therapeutic efficacy, and to a lesser extent granulocytes, because These include HLA B27 transgenic rats with these are the cell types in which iNOS activity spontaneous intestinal inflammation, 31 colonic was discovered. It is now apparent that many cell inflammation induced by the sulfhydryl blocker, types can express iNOS but in contrast, the neuiodoacetamide, 42 insulin-dependent diabetes mel-ronal and endothelial forms of NOS remain localitus, 43 carrageenan-induced pleurisy and footpad lized in their original classification with the 44 45 46 oedema, endotoxin-induced uveitis, air exception of epithelial cells and skeletal muscle 47 pouch models of granuloma formation, TNBS which express cNOS. 63 Thus, in an inflammatory and acetic acid colitis in rats 48'49 and sponta-disease state iNOS is expressed in a variety of neous, idiopathic colitis in rhesus macaques. 5 cells and not just in infiltrating leukocytes. In Nitric oxide displays anti-inflammatory chararthritis, chondocytes, osteoblasts and synovium acteristics in ischaemia-reperfusion injury. [51][52][53] can be a source of NO. 16'32'64'65 In gut inflamma-These protocols are acute and therefore are tion the lredominant sources of iNOS are epi-iNOS-independent models, although with exten-thelia 26''6'6 and the enteric nervous sion of the study period iNOS can be detected in system, 26'3 in addition to resident and infiltrating previously injured tissue. 2 Recently, a report has leukocytes in the lamina propria. Tepperman and addressed the role of nitric oxide in ischaemia-colleagues first described the ability of endotoxin mperfusion injury in a condition where iNOS is to increase NO synthesis in cultured epithelial present, i.e. after administration of endotoxin. Ma cells, a response which led to epithelial cell and colleagues noted that reperfusion injury was death via NO-dependent mechanisms. 66'67 Furattenuated by the NOS inhibitor t-NMMA under thermore, it appears that the epithelium these conditions, as opposed to the usual responds to this NOx-induced stress and resulexacerbation seen in the absence of endotoxin. 54 tant DNA damage by undergoing apoptosis. 68 They proposed, based on the evaluation of che-Epithelial barrier function, conferred by tight miluminescence signals, that excess NO reacted junctions between epithelia, is influenced by with superoxide induced by ischaemia-reperfunitric oxide. Inhibition of nitric oxide synthesis sion, to form the highly toxic peroxynitrite, compromises epithelial barrier dysfunction under which in turn contributed to hepatocellular basal conditions and in ischaemia reperfusion injury. This is another example of NO promoting injury. 52'53'69 Following an acute injury where the tissue injury when conditions are favourable, viz. epithelial barrier becomes leaky there is a rapid during states in which iNOS is expressed, repair/restitution response. This response is In human disease there remains a lack of solid impaired by NOS inhibitors, 7 although a conevidence that inhibition of NOS is protective, trasting result in the Caco-2 cell line has been largely because compounds with sufficient selec-reported. 7 Because of the importance of the tivity for iNOS are not available for testing. The epithelial barrier in limiting the translocation on exception is the preliminary work in septic shock luminal flora and contents any impairment of this with -NMMA. 55 In shock a complete withdrawal barrier would promote the inflammatory of nitric oxide synthesis may be deleterious, response. In the acute (iNOS absent) models NO leading to suggestions that scavengers that moppromotes the integrity of this barrier and thereup excess NO may be viable alternatives. One fore is anti-inflammatory. In chronic gut inflaminteresting NO scavenger is hydroxocobalamin, a mation there is also epithelial barrier dysfunction form of vitamin B12 lacking cyanide, which preconcomitant with exaggerated nitric oxide vents and reverses endotoxic shock in rats and release, 1'2' but in these states NOS inhibitors mice. 56 are protective.
Nevertheless, there is ample evidence of What could be the explanation for this apparexcess nitric oxide formation in human diseases, ent discrepancy? We propose that these differincluding ulcerative colitis, 57-59 gastritis 6'61 and ences are mediated by contrasting mechanisms.
Under acute conditions the beneficial effects are mediated by cGMP, indeed cGMP mimics decreased epithelial permeability. In contrast, in iNOS-dependent chronic inflammation the increase in epithelial permeability is the result of increased epithelial cell injury and/or death. It does not matter how tight the tight junctions are if there is rampant epithelial cell death, leukocyte emigration and substantial morphological derangementscrypt abscesses, ulceration and bursting of villus tips. These morphological derangements, characteristic of mucosal inflammation, will promote epithelial barrier dysfunction.
Pathways of NO metabolism--is injury due to NO or NOx? Nitric oxide itself is a poorly reactive free radical, and in order for it to be a major component of cellular injury it is likely to be metabolized to alternative species. In an analogous manner superoxide (O2-) is a weakly reactive reactive oxygen species but it is metabolized in inflammation to a cascade of more reactive products, e.g. HOCl and OH'. Nitric oxide is inactivated by oxidation to nitrite (NO2-) and subsequently nitrate (NO-) but other oxidized forms may be formed which are not innocuous. Some products require two molecules of nitric oxide as substrate for the generation of the product, e.g. N203 and N204. In other words, these reactions display second-order kinetics and large amounts of the substrate are required to drive the reaction. Thus, it is likely that only in conditions in which iNOS is expressed will these products be formed. One of the consequences of these reactive nitrogen species results from their ability to act as nitrosating and nitrating species. If the target is DNA then this will lead to deamination and point mutations, as indicated in Table 3. 74,75 Interaction between Of and NO.. Recently biochemical studies have demonstrated that 0 2 and NO rapidly interact via a radical-radical reaction at a diffusion-limited rate (k 6.7 x 10 9 m-l.s-1) to generate the potent oxidant peroxynitrite 77 78 (ONOO-). 76 Beckman and co-workers have suggested that the interaction between these two Although ONOO-is relatively stable, especially at alkaline pH, it has a pK of 6.6 which dictates that substantial amounts of this compound will be protonated at physiological pH to yield peroxynitrous acid (ONOOH). This compound is very unstable and has been suggested to rapidly decompose to yield a hydroxyl radical (OH') and NOi-like compound.
This hydroxyl radical-like oxidant will react with virtually all biomolecules at diffusion limited rates of reactions ( 10v-10 M-1. s-). Nitrogen dioxide is also a very reactive radical with an ability to react with alkanes and alkenes via free radical-mediated mechanisms. 8'8 Furthermore, Pryor and coworkers have demonstrated that NO2 = will initiate lipid peroxidation in vitra 85'86 They also demonstrated that abstraction of H atoms from unsaturated lipid by NO2" resulted in the formation of nitrous acid HONO which was shown to N-nitrosate certain amines to yield nitrosamines. 85 '86 In addition, NO2" is known to react with haemoproteins such as haemoglobin and oxidize thiols and thioethers (methionine). 85'86 Taken together, these data suggest that if the decomposition of ONOOH proceeds via the formation of OH-like species and NO2" then it represents a potentially important pathway in which toxic oxygen and nitrogen-derived free radicals may be formed from the interaction between two relatively unreactive radicals, O2and NO. In fact, this mechanism has been proposed to account for the 02-dependent microvascular injury produced by ischaemia and reperfusion of various organ systems.
Because of the ephemeral nature of peroxynitrite it is difficult to measure it in vitro, but it can be tracked in situ by its ability to promote nitration reactions, in particular the formation of nitrotyrosine. We have demonstrated that nitrotyrosine formation is negligible under basal conditions but in inflammation it is readily demonstrable by immunohistochemistry. Importantly, nitrotyrosine co-localizes with the expres-sion of iNOS, and therapy with inhibitors of NOS but yet is accompanied by cellular/tissue growth prevents the formation of nitrotyrosine (and pre-and immunosuppression. The explanation for vents the inflammation). Thus, although nitric this dichotomy may lie in the production of oxide itself cannot form nitrotyrosine (it would other oxidants. Unlike inflammatory conditions form nitrosylate not nitrate tyrosine residues), the where nitrotyrosine and iNOS co-localize 26'3 formation of NO from iNOS is linked to nitrotyr-nitrotyrosine is not present in the pregnant osine formation, uterus or placenta. This suggests that interactions Nitrotyrosine has been demonstrated in several between NO and other reactive oxygen species human-disease states--atherosclerosis and amyare critical determinants of the cytotoxic effects trophic lateral sclerosis. [87][88][89] Using HPLC techni-of nitric oxide. If pregnant rats are challenged ques nitrotyrosine was readily detectable in joints with endotoxin, nitrotyrosine formation and celof patients with rheumatoid or osteoarthritis 9 lular injury can be demonstrated in the uteroand absent in healthy individuals. Furthermore, placental unit (unpublished data, MJSM). As the we have found that nitrotyrosine co-localizes with synthesis of oxygen free radicals is elevated by iNOS and DNA fragmentation in Helicobacter endotoxin and inflammation in general, interacpylori gastritis and that eradication of the infections with nitric oxide may be a critical determition or treatment with the antioxidants, ascorbic nant of the role reversal of nitric oxide. acid and/or beta-carotene, reduces the staining for nitrotyrosine. 1 Mutagenesis, DNA damage and Studies from Radi's laboratory have clearly apoptosis demonstrated some of the potential toxic effects As previously mentioned, NO will spontaneously of peroxynitrite, including inhibition of cell auto-oxidize to yield potent N-nitrosating agents proliferation, DNA synthesis, succinate de-such as N203. This nitrosating agent may then hydrogenase, fumarate reductase, thereby interact with secondary amines to yield potensuppressing mitochondrial electron transport tially carcinogenic nitrosamines. It is known that chain in the parasite Trypanosoma. 91 '92 In addinitrosamines must be activated into mutagenic tion, inhibition of alpha-proteinase inhibitor has and carcinogenic species via the action of the been reported. 9 Direct luminal administration of cytochrome P450 system. In addition, it has been peroxynitrite can cause colitis in rats. 94 demonstrated that electrophilic or mutagenic While it is unlikely that all the deleterious metabolites may be generated from nitrosamines effects of nitric oxide are due to peroxynitrite, via a cytochrome P-450 independent pathway the kinetics for the formation of peroxynitrite involving the interaction of nitrosamines with exceed that for dismutation of superoxide by highly reactive oxy radicals, such as OH', or with superoxide dismutase. 77 '78 In other words, co-ultraviolet light. Both the P450-dependent and production of superoxide and nitric oxide will independent pathways appear to involve the forlead to the formation of peroxynitrite unless the mation of the a-hydroxynitrosamine intermediate. NO is sequestered or oxidized. Thus the dis-This unstable intermediate, will spontaneously crepancies which outline the current debate, i.e. decompose to yield the monoalkylnitrosamine is nitric oxide inherantly anti-or proinflammatory derivative. Through a series of spontaneous remay be moot because the benefit/injury may be arrangement and decomposition reactions this mediated by entirely different chemical species, intermediate will produce the alkylcarbonium Low doses of nitric oxide are anti-inflammatory cation which is the ultimate carcinogen. This whereas high concentrations of NO in conjunc-electrophilic agent will rapidly alkylate a variety of tion with the release of other reactive oxygen different nucleophilic sites in cellular compospecies, will lead to the formation of NO-derived nents including protein, DNA and RNA. It has reactive nitrogen species that are pro-inflamma-been determined that alkylation of DNA may tory.
occur at as many as twelve different sites includ-Further evidence that this maybe an important ing the ring nitrogen positions in adenine, explanation is found in pregnancy. Our postulate guanine, cytosine and thymine, the oxygen atoms is that nitric oxide production only has proassociated with the hydroxyl or carbonyl groups inflammatory effects when it is produced in of guanine, thymine and cytosine, and the phosexcess, i.e. when iNOS is expressed. If this pos-phate groups. Experimental data suggests that tulate is valid then exceptions should not be Cfi-alkylation of guanine represents one of the observed. In other words, are there conditions in most important reactions because there is a which iNOS is expressed but not associated with better correlation of the extent of O6-alkylation tissue injury? Pregnancy is the exception. Pregwith carcinogenicity and mutagenicity than with nancy is a state in which iNOS is expressed in a any other type of alkylation reaction. Alkylation sustained manner in the uterus and placenta 95 of guanine for example would alter the base pairing such that during DNA replication thymine and not cytosine may be incorporated into the newly synthesized strand of the nucleic acid. Mutational effects of this type may then activate certain oncogenes, resulting in malignant transformation.
More recent studies have demonstrated that NO may promote mutagenesis via nitrosative deamination of DNA. 96 '97 It has been known for many years that NO-derived nitrosating agents (e.g. N203) will nitrosatively deaminate a variety of primary aromatic amines (AR-NH2) to yield their hydroxy derivatives. The mechanism appears to involve the N-nitrosation of AR-NH2 to yield a nitrosamine (Ar-NH-NO) intermediate.
This unstable intermediate will rapidl+y rearrange to produce the diazonium ion (ar-N2) which in turn decomposes to yield the hydroxyl derivative (Ar-OH) of the aromatic amine: From these reactions it has been suggested that deamination of cytosine (C), methyl cytosine (mC), guanine (G) or adenine (A) would result in the formation of uracil (U), thymine (T), xanthine (X), and hypoxanthine (HX), respectively. 97 If these types of reactions occur in intact DNA then base pair substitution mutations may occur. There are two major types of base pair substitution mutations: one type involves the substitution of a purine for purine or a pyrimidine for a pyrimidine and is termed a transition mutation. The other type is called a transversion mutation and is characterized by the substitution of pyrimidine for a purine or vice versa.
Deamination of cytosine to yield uracil for example, will result in the removal of uracil by the enzyme uracil glycosylase leaving an abasic site. This type of lesion is commonly misrepaired by insertion of adenine opposite the site during replication resulting in the observed G:C to A:T base pair substitution mutation (Table 3). 97 If uracil is not repaired it will pair with an incoming adenine to give the same mutation upon replication. Deamination of methyl cytosine to yield thymidine would also produce a G:C to A:T transition (Table 3). This type of mutation is especially significant in mammalian cells because the pattern of methylation of DNA is very important in regulating gene expression and differentiation. In addition to the misrepair of deaminated bases, the instability of hypoxanthine and xanthine in DNA formed from the deamination of adenine and guanine, respectively, leads to rapid depurination forming apurinic sites. It is thought that depurination leads to DNA breakage and subsequent cytotoxicity as well as mutageni-city. Base pair substitution mutations arising from the nitrosative deamination of adenine and guanine would lead to A:T to G:C and G:C to T:A transversion mutations, respectively (Table 3). 97 Based upon the relative rates of NO-dependent deamination of the different bases in DNA, one would expect the G:C to A:T transition mutations to predominate. Indeed, this type of mutation has been shown to predominate in mutants (revertants) of Salmonella yphimurium produced by exposure of these cells to NO. 96 Nitric oxide has been shown to promote apoptosis in cancer cell lines, macrophages and colonic epithelial cells. 68'98'99 The mechanisms by which NO may initiate cell death (apoptosis or necrosis) vs uncontrolled cell replication (cancer) may not be as divergent as they superficially appear. Both involve DNA damage. A normal cell whose DNA is damaged by NO, oxidants or the combination, will arrest its cell cycle, undergo DNA repair, or alternatively undergo apoptosis, if the damage is perceived as too extensive. Both the repair and apoptosis processes protect the host from the duplication of damaged DNA, with injured cells repaired or eliminated. Survival of transformed cells may occur if either the repair of apoptosis mechanisms are not effectively evoked. It is commonly considered that cancer results in failures of multiple fail-safes (mmour suppressor and protooncogene) and NO may play a role at all levels. In those 'acute' protocols where no deleterious effects are noted, and where high doses of nitric oxide are administered, or in conditions favourable for peroxynitrite formation, an evaluation of DNA status is invariably absent. Unless mutagenesis or induction of apoptosis has been ruled out, there is no assurance that serious complications have not been set into motion. Complications that may remain cloaked to standard physiology-based tests and functional assays.
Concluding remarks There is ample evidence that nitric oxide is a molecular aggressor in inflammation. The reported anti-inflammatory actions are either very acute in duration and/or pertain to conditions which are not classically defined as inflammation, but rather states which are primarily driven by oxygen free radicals, e.g. ischaemia/reperfusion injury. Furthermore, NO formation is greatly elevated in inflammation suggesting that if its primary role was anti-inflammatory its actions are too weak to negate tissue damage. Rather, it appears that NO is primarily pro-inflammatory in inflammatory disease states. For this reason new therapeutic approaches for the treatment of chronic inflammatory diseases such as arthritis and inflammatory bowel disease, focus on the inhibition of inducible nitric oxide synthase, and not the use of NO donors.
In the last few years we have come to appreciate that the pro-inflammatory actions of nitric oxide involve a multitude of mechanisms. It is likely that the critical events may involve reactive nitrogen species derived from nitric oxide, but the parent radical is not, in itself, the mediator of tissue destruction. Peroxynitrite, at this time, appears to be the best candidate but other reactive nitrogen species may assume the mantle currently bestowed on peroxynitrite. Further, nitric oxide and related products appear to be chemical links between inflammation and cancer by virtue of their ability to react with DNA bases, causing point mutations.
In summary, nitric oxide is clearly a critical pro-inflammatory mediator. The dilemma established by its anti-inflammatory actions under basal conditions can be reconciled by evaluating the influence of time on the observations, the enzyme source of nitric oxide production, the cellular sources of NO in the condition being evaluated, and the generation of other reactive nitrogen species from NO, particularly those involved in nitrosation and nitration reactions. When these influences are accounted for one can predict if nitric oxide will play an anti-or proinflammatory role.