Peroxidative metabolism of carcinogenic N-arylhydroxamic acids: implications for tumorigenesis.

Peroxidative oxidations of chemical carcinogens including N-substituted aryl compounds could result in their metabolic activation because the products react with cellular molecules and lead to cytotoxicity, mutagenicity, and carcinogenicity. In vivo, peroxidative activities are chiefly of neutrophilic leukocyte origin. Neutrophils may be attracted to the site(s) of exposure to carcinogen and, via phagocytosis and respiratory burst, release oxidants that catalyze carcinogen activation and/or cause DNA damage. Our studies, presented herein, concern oxidations of carcinogenic N-arylhydroxamic acids, N-hydroxy-N-2-fluorenylacetamide (N-OH-2-FAA), and N-hydroxy-N-2-fluorenylbenzamide (N-OH-2-FBA), by enzymatic and chemical systems simulating those of neutrophils, myeloperoxidase and hydrogen peroxide (H2O2) +/- halide, and hypohalous acid and halide at the physiologic concentrations (0.1 M Cl- and/or 0.1 mM Br-) and the pH (4-6.5) of phagocytosis. Studies also concern oxidations of the hydroxamic acids by rat peritoneal neutrophils stimulated to undergo respiratory burst and release myeloperoxidase in medium-containing 0.14 M Cl- +/- 0.1 mM Br-. The metabolites formed in the presence of exogenous H2O2 are consistent with two peroxidative mechanisms: one electron-oxidation to a radical that dismutates to equimolar 2-nitrosofluorene (2-NOF) and the ester of the respective hydroxamic acid and halide-dependent oxidative cleavage, especially efficient in the presence of Br-, to equimolar 2-NOF and the respective acyl moiety. 2-NOF and the esters undergo further enzymatic and nonenzymatic conversions to unreactive products and/or may bind to cellular macromolecules. The results suggest that peroxidative metabolism of N-arylhydroxamic acids by neutrophils, yielding the potent direct mutagen 2-NOF and the electrophilic esters, occurs in vivo and is involved in the activation and thus local tumorigenicities of the hydroxamic acids at the site(s) of application.

during oxidation of arachidonic acid to prostaglandins. Oxidations of these compounds, termed cosubstrates, were catalyzed by the peroxidase activity of PHS. Investigations from several laboratories on activation of carcinogenic N-arylamines via PHS-mediated oxidations have recently been reviewed (2,3). PHS catalyzed one electron (le-)-oxidation of N-2-fluorenamine (2-FA) to nitrogen-centered free radicals, which yielded dimeric and polymeric products and 2-nitrofluorene (2-NO2F). However, the pathway to 2-NO2F still remains obscure because its expected precursors, N-hydroxy-2-FA and 2-nitrosofluorene (2-NOF), were not detected among the products (4). The oxidation of 2-FA by PHS yielded two, as yet uncharacterized, adducts with DNA (5) that were different from N-(deoxyguanosin-8-yl)-2-FA, which presumably originates from the nitrenium ion. The ease of oxidation of Narylamines is altered by substituents that affect the electron density on the nitrogen. For example, benzidine (BZ) is a much better cosubstrate for PHS peroxidase than 2-FA. PHS oxidation of BZ to BZ diimine, which may undergo deprotonation to nitrenium ion, appears to be the source of a major DNA adduct, N-(deoxyguanosin-8-yl)-BZ (6,7). This adduct may be oxidized in vivo to N,3-(deoxyguanosin-7,8-yl)-BZ. Adducts consistent with PHS-mediated oxidation of BZ, 2-naphthylamine, and 2-FA were found in dog bladder epithelia, and may be involved in bladder cancer induction (8,9).
In addition to ubiquitous PHS, peroxidative activities in human or animal tissues also may be derived from endogenous peroxidase in various cells such as glandular acinar cells, epithelial and endothelial cells, and leukocytes (10). Neutrophils, which constitute 95% of polymorphonuclear leukocytes (PMNL), contain myeloperoxidase (MPO) (11). MPO is also found in monocytes or mononuclear phagocytes (MNL). Eosinophils (3.8% of total PMNL) and, presumably, mast cells contain eosinophil peroxidase. In various tissue eosinophilias, eosinophils may be the predominant infiltrating leukocytes (12). Their peroxidase activity per cell is approximately 2.5-fold greater than that of neutrophils. During phagocytosis, the leukocytes undergo degranulation, releasing peroxidase, and a respiratory burst generating superoxide anion (O°) from the reduction of°b y NADPH oxidase (11). 0°is, in turn, converted to hydrogen peroxide Environmental Health Perspectives (H202) through spontaneous or superoxide dismutase-catalyzed dismutation. Using H202, MPO and eosinophil peroxidase catalyze oxidation of halides, preferentially chloride (CF) and bromide (Br-), respectively, to generate the powerful oxidants hypohalous and hypobromous acids (HOCI and HOBr) (13,14). The hypohalous acids are capable of halogenation of endogenous amines such as taurine to N-chloroand Nbromotaurine, which are more stable oxidants.
MPO-dependent oxidations were implicated in activation of N-arylamines to DNA-bound products by PMNL (Table 1). In these experiments, incubations of 14Clabeled BZ, methylaminoazobenzene, 2-FA, or acetaminophen with phorbol myristate acetate (PMA)-treated guinea pig or human PMNL (15)(16)(17), or neutrophils of human leukemic cell line origin (18,19) resulted in covalent binding of radioactivity to leukocytic DNA. The release of MPO activity apparently was accomplished by treatment of leukocytes with PMA in buffers containing Ca++ or solubilization with a cationic detergent. The leukocyte oxidizing system was simulated by MPO and H202/CF-, which yielded covalently bound radioactivity to exogenous DNA (15)(16)(17). When 2-FA was reacted with human neutrophil lysates supplemented with H202, a single DNA adduct, Nsdeoxyguanosin-8-yl)-2-FA, was detected by 2P-postlabeling (20). Thus, even though the mechanism(s) of oxidation leading to the electrophile reacting with DNA is unknown, MPO-dependent metabolism of carcinogenic N-arylamines leads to their activation and a DNA adduct associated with initiation of carcinogenesis.
The leukocytes, both PMNL and MNL, also were used in studies of metabolism of drugs containing free -NH2 in para position to other substitutents (Table 1) (21)(22)(23). The -NH2 group was oxidized to the hydroxylamino-(detected only in the presence of ascorbic acid) and then to the nitro-compound. However, the presumed intermediary nitroso compound was not detected from any of the drugs investigated. Simulations of these oxidations with MPO and H202/Cl-yielded the same metabolites as above and also N-chloramino derivatives that rearranged to o-chlorocompounds (24). The MPO-derived oxidation products from drugs containing the free -NH2 group have been implicated in adverse effects such as agranulocytosis and generalized hypersensitivity reactions during drug therapy (25). The detection of the hydroxylamino derivatives from these drugs may be the result of peroxygenation, the mechanism recently elucidated for chloroperoxidase-catalyzed N-oxidation of p-substituted N-arylamines (26). It is thus becoming increasingly evident that the structure of N-arylamine and the type of the substituent on the aromatic ring are factors in determining the oxidation products. The latter may also depend on the type of peroxidase present and the environment in a particular tissue.
The above evidence linking leukocytic peroxidative systems to activation of N-arylamines suggested that these systems also might be involved in the activation of the carcinogenic N-arylhydroxamic acids. Oxidations of Carcinogenic N-Arylhydroxamic Acids by Chemical and Enzymatic Systems Simulating Oxidants of Leukocyte Origin A particular characteristic of N-arylhydroxamic acids is their ability to induce tumors at the sites of application (27). Thus, multiple ip injections of an aqueous suspension of N-OH-2-FAA (cumulative dose of -0.5 mmole/rat) yielded approximately 62% peritoneal tumor incidence in male or female rats. The structural analogue, N-OH-2-FBA, was more potent since approximately 0.2 mmole/rat yielded pleomorphic sarcomas in 75 to 100% rats. The hydroxamic acids and the oxidation products, 2-NOF and N-acetoxy-2-FAA (N-AcG-2-FAA), also were carcinogenic upon sc or im injections and upon direct application of the solid to the mammary gland. The sites of carcinogen application may be infiltrated by leukocytes in response to physical and chemical stimuli. It is thus possible that oxidants released by leukocytes at the sites of carcinogen deposits may be involved in their oxidations and yield products relevant to their local carcinogenicities.
To investigate this possibility, we first examined oxidations of the hydroxamic acids by chemical and enzymatic oxidants simulating those generated by leukocytes. Two peroxidative pathways in oxidation of the hydroxamic acids were considered ( Figure 1): le--oxidation to nitroxyl free radicals, which dismutate to equimolar Cnitroso compound and the ester of the hydroxamic acid, the reaction initially reported by Bartsch et al. (28)(29)(30) for chemical and enzymatic oxidations of Narylacetohydroxamic acids; and hypohalous acid-catalyzed oxidative cleavage of the hydroxamic acid to the C-nitroso compound and the respective acid, the reaction that we reported for halide-dependent enzymatic oxidations of N-OH-2-FAA (31). Our earlier investigation also showed that enzymatically or chemically generated HOCI, HOBr, or N-bromotaurine, but not N-chlorotaurine, were capable of oxidation of N-OH-2-FAA via both le and oxidative cleavage (32). Accordingly, N-OH-2-FBA would undergo le--oxidation to a nitroxyl free radical that would dismutate to 2-NOF and the ester N-benzoyloxy-2-FBA (N-BzO-2-FBA) or halide-dependent oxidative cleavage or both to 2-NOF and benzoic acid (BA). Although evidence for le-oxidation of N-OH-2-FBA by alkaline K3Fe(CN)6 in benzene was reported by Bartsch et al. (28), the presumed N-BzO-2-FBA was not quantified, and 2-NOF accounted for only 14%, rather than the theoretical 50%, of the substrate. Hence, we reexamined the oxidation of N-OH-2-FBA by K3Fe(CN)6 and confirmed the presence of a nitroxyl free radical by electron spin resonance (ESR) spectroscopy ( Figure 2). However, the area and maximum amplitude of the signal produced within 1 min from N-OH-2-FBA were approximately 50 and 60%, respectively, of those from N-OH-2-FAA, indicating that N-OH-2-FBA was more slowly oxidized. Separation of the radical dismutation products by HPLC and estimation of their amounts in the oxidation mixtures based on the synthesized standards (33) showed that 2-NOF and the ester accounted for 44 and 40% of the substrate, respectively. This result confirmed the originally postulated mechanism of le--oxidation of N-OH-2-FBA by a chemical oxidant in nonaqueous media (28).
In our attempts to examine enzymatic oxidations of N-OH-2-FBA, we encountered problems related to the poor solubility of the hydroxamic acid and instability of N-BzO-2-FBA, which, in part, underwent ortho rearrangement (Figure 1 (Figure 3), it appears that the lower concentration was near saturating for the enzyme.
We also compared oxidations of the two hydroxamic acids by MPO and H202 + halide under conditions optimal for product formation from N-OH-2-FAA ( Figure  4). At pH 6.5, the presence of halide did not affect the amounts of 2-NOF formed from N-OH-2-FAA. However, in the absence of halide, the amounts of N-AcO-FAA were greater, indicating increased level of le--oxidation. Under these conditions, the presence of halide increased the amounts of 2-NOF formed from N-OH-2-FBA but had no effect on the small amounts of its ester. A similar product profile was determined from N-OH-2-FBA oxidized by MPO and H202/halide at pH 4 and 5 and suggested that the major source of 2-NOF was halide-dependent oxidation of N-OH-2-FBA. However, N-OH-2-FBA yielded approximately 10 times less 2-NOF than did N-OH-2-FAA, which was nearly completely converted to 2-NOF. In the presence of the less easily oxidized N-OH-2-FBA, H202 may have led to inactivation of MPO or destroyed HOBr, the product of oxidation of Br-by MPO and H202. As with HRP and H202, oxidations of N-OH-2-FBA at 30 or 10 pM by MPO and H202 ± halide yielded similar amounts of the products, suggesting saturation or inhibition of MPO or both.
Oxidations of the hydroxamic acids by hypohalous acids also were compared ( Table 2). More 2-NOF, the major product, was produced from both compounds by equimolar HOBr at pH 4.0 than by excess HOCI at pH 5.0, reflecting the greater oxidizing potential of HOBr. The determination of small amounts of the esters from oxidations of the hydroxamic acids by hypohalous acids indicated relatively low levels of le--oxidation. The ESR signal of nitroxyl-free radical, indicative of le--oxidation, was detected only during oxidation of N-OH-2-FAA by HOCI (Figure 2). With the controlled rate of oxidation of N-OH-2-FAA by HRP and H202, a detectable  Br or both, and determination of the metabolites were described previously (33). Values are the means from two to six incubations ± SD. aHOBr (30 pM) + 0.1 mM Br, pH 4.0, or HOCI (3 x 60 pM) + 0.1 M Cl, pH 5.0. bDetermined after 1 0-min incubation at 25°C as described previously (33). Values are the means from two to four incubations ± SD. cNot applicable.
level of radical was maintained while HOCl-catalyzed radical generation was rapid and brief. With the more powerful oxidant HOBr, no radical signal from N-OH-2-FAA and only a faint signal from N-OH-2-FBA were detected. Based on the relative amplitudes of the radical signals and the amounts of the esters, excessive amounts of 2-NOF were produced from hypohalous acid-catalyzed oxidations ( Table 2). This supported our original proposal that the major source of 2-NOF from the oxidations of the hydroxamic acid by HOCI or HOBr is via oxidative cleavage (32). Likewise, the determinations of acetate (31) and BA (Table 2) from these oxidations supported oxidative cleavage of the hydroxamic acid. BA also was found from MPO and H202 and Br--catalyzed oxidation of N-OH-2-FBA (33), which supported oxidative cleavage as the major reaction for enzymatically generated HOBr. We also found that the esters did not con-tribute to 2-NOF formation. 2-NOF occasionally yielded small amounts of 2-NO2F. The elucidation of the conditions for enzymatic or chemical oxidations of the Narylhydroxamic acids, by oxidants simulating those of leukocyte origin, prompted our investigation of metabolism of N-OH-2-FAA and N-OH-2-FBA by rat PMNL in vitro.

Peroxidative Metabolism of Cardnogenic N-Arylhydroxamic Acids by Rat Peritoneal Neutrophils in Vitro
The carcinogenicities for rat peritoneum of N-OH-2-FAA, and especially N-OH-2-FBA, administered ip in aqueous suspensions (27), suggested influx of leukocytes in response to the foreign compound deposits. Hence, neutrophils, the predominant leukocytes, were elicited into rat peritoneum through ip injections of proteose peptone and harvested within 4 hr for Environmental Health Perspectives In the absence of Ca++, azurophil granules containing MPO remained intact when the cells were treated with PMA to effect the respiratory burst. Addition of a cationic detergent, cetyltrimethylammonium chloride (Cetac) at 0.002% or 0.02%, prompted loss of granule structure. These microscopic changes, observed with the stained cell preparations, were reflected in the spectrum of the oxidants produced by neutrophils ( Figure 5). Thus, treatment with PMA in the absence of Ca++ effected the respiratory burst as shown by O-and H2G2 production. Some hypohalous acid but no MPO activity was detected. Addition of Cetac caused immediate release of MPO activity, and the level of the activity was higher at the higher concentration of Cetac.
Metabolism of N-OH-2-FAA by neutrophils was determined before and after release of MPO ( Figure 6). Because C1or Bror both had a significant effect on the formation of 2-NOF by MPO and H202 (Figure 4), we examined metabolism of N-OH-2-FAA by neutrophils in HBSS ( Figure 6A) and HBSS + 0.1 mM Br-( Figure 6B). Both the substrate depletion and metabolite generation were monitored. In these experiments, N-OH-2-FAA was added 30 min after initiation of respiratory burst by PMA. However, no substrate  30 pM, was added followed 15 min later by 0.002% Cetac and 50 pM H202. After 6 min the second aliquot of H202 was added and the mixture incubated for an additional 18 min. Aliquots (0.9 ml) of the incubation mixtures were removed for determination of metabolites and unreacted N-OH-2-FAA (insert panels) as described previously (34). Values in panels A and B are the means ± SD from three and five experiments, respectively, and were each carried out in duplicate. The effect of 0.1 mM Br on the amounts N-OH-2-FAA metabolized was significant at p<0.05 at 12, 18, and 24 min after addition of Cetac and H202. Among the metabolites, the effect of Br was significant at p<0.01 for amounts of 2-NOF and 2-NO2F at 6, 12, 18, and 24 min after addition of Cetac and H202. depletion or metabolite formation occurred in the presence of 0or H22 or both, the products of the respiratory burst. Release of MPO activity by 0.002% Cetac and addition of 50 pM H202 initiated metabolism which was augmented by the addition of the second aliquot of H202 (no metabolism was detectable in the absence of H202).
The metabolites included N-AcG-2-FAA, 2-NO2F, 2-FAA, and 2-NOF. The amounts of 2-NO2F, and especially 2-NOF, were significantly greater in the presence of Br-. These data were consistent with our previous findings that N-OH-2-FAA was oxidized chiefly to 2-NOF by MPO and H202 in the presence of Bror CI-+Br- (32,33). The increased level of oxidation of N-OH-2-FAA was likely because of the formation of Br--derived oxidants. However, Br-had no apparent effect on the level of le--oxidation because the amounts of N-AcG-2-FAA and 2-FAA were similar.
We also have investigated the fate of 2-NOF and N-AcG-2-FAA because their reactivities in cellular systems might affect the amounts detected in metabolism studies. The recovery of 2-NOF was only 22% and decreased with time, especially in active cells (34). Formation of small amounts of 2-NO2F that was H202-dependent and both enzymatic and nonenzymatic, could not account for all the losses of 2-NOF. The low recoveries of the latter were likely because of reactivity with cellular components such as glutathione or protein binding or both (27). The recovery of N-AcG-2-FAA decreased from 75 to 8% in 24 min (34). In the absence of H202, 73% of the ester was recovered as 4-OH-2-FAA ( Figure 1). Similarly, large amounts of 4-OH-2-FAA were formed from N-AcG-2-FAA incubated with heat-inactivated cell lysates or in HBSS. The data were consistent with the nonenzymatic conversion of N-AcG-2-FAA to 4-OH-2-FAA in buffers containing 0.14 M CF, reported by Scribner (35). Addition of H202 disrupted the time-dependent increase in the formation of 4-OH-2-FAA, which probably was oxidized to a phenoxy radical. However, no (1) or N-OH-2-FAA (A) at 10 pM was added (0 point on the graph) and incubation continued for 24 min. Aliquots (0.9 ml) of the incubation mixtures were analyzed for unreacted compounds and metabolites as described previously (34). Values are the means ± SD from two experiments, each carried out in duplicate. A denotes N-OH-2-FAA formed from N-AcO-2-FAA.
OH-2-FAA was detected in the peroxidative metabolism of N-OH-2-FAA ( Figure 6), probably because the levels of N-AcO-2-FAA were too low for its conversion to detectable 4-OH-2-FAA. Small amounts of N-AcO-2-FAA were found to be O-deacetylated to N-OH-2-FAA, especially in the absence of H202 (Figure 7). O-Deacetylation was inhibited by paraoxon. The data were indicative of O-esterase activity in the neutrophils, which might lead to regeneration of the substrate from the product N-AcO-2-FAA ( Figure 1). On the other hand, we found no evidence for Ndeacetylase activity (Figure 7) that would catalyze formation of N-hydroxy-2-FA from N-OH-2-FAA and thus lead to 2-NOF via oxidation of the N-hydroxy-2-FA.
Metabolism of N-OH-2-FBA by rat peritoneal neutrophils was examined under somewhat modified conditions (Figure 8). Because of its low solubility, N-OH-2-FBA  (Figure 8) that was more efficiently extracted in the presence of the detergent. Additions of H202 significantly increased consumption of the substrate and formation of metabolites, especially of 2-FBA and 2-NO2F. The latter was formed, in part, from 2-NOF because in a separate experiment, approximately 15% of the added 2-NOF was found to be metabolized to 2-NO2F in the presence of H2*2' Comparison of the metabolite ratios (2-NOF + 2-NO2F:ester) and the amounts of the amides obtained from peroxidative metabolism of N-arylhydroxamic acids by neutrophils (Figures 6, 8) suggested a higher level of le--oxidation for N-OH-2-FBA than N-OH-2-FAA relative to the oxidative cleavage. Thus, as with the Narylamines, the extent of oxidation and the product profile are determined by the structure and, hence, physico-chemical properties of the N-arylhydroxamic acids. It is possible that less soluble compounds such as N-OH-2-FBA remain longer at the site of application and cause prolonged infiltrations and oxidant release by leukocytes. Under such circumstances, in addition to DNA damage from reactive metabolites of carcinogens, leukocytic oxidants may cause direct damage to DNA. Demonstrating that these oxidative processes occur in vivo is a complex undertaking. Based on our studies, several approaches appear viable: a) demonstration of the presence of the more stable derivatives (e.g., 4-OH-2-FAA, 2-NO2F, o-BzO-2-FBA) ( Figure 1); b) characterization of both carcinogen-and oxidant-induced DNA damage in target tissues (e.g., peritoneal serosa); c) determination of the effect of increasing physiologic concentration of Br-, a significant environmental pollutant (36), on the tumorigenicity of the hydroxamic acids; and d) demonstration of the interaction products of the major metabolite 2-NOF with glutathione, protein, and unsaturated lipids. Because of the exceptional direct mutagenicity of 2-NOF in the bacterial systems (27), it especially merits investigation. Recent evidence that interaction of C-nitroso compounds with unsaturated lipids is a source of 0°- (37) or genotoxic 2 products via lipid peroxidation (38) may be pertinent to the biologic significance of 2-NOF.