Phagocytes Produce 5-Chlorouracil and 5-Bromouracil, Two Mutagenic Products of Myeloperoxidase, in Human Inflammatory Tissue

Oxidative damage to DNA has been implicated in carcinogenesis during chronic inflammation. Epidemiological and biochemical studies suggest that one potential mechanism involves myeloperoxidase, a heme protein secreted by human phagocytes. In this study, we demonstrate that human neutrophils use myeloperoxidase to oxidize uracil to 5-chlorouracil in vitro . Uracil chlorination by myeloperoxidase or reagent HOCl exhibited an unusual pH dependence, being minimal around pH 5 but increasing markedly under either acidic or mildly basic conditions. This bimodal curve suggests that myeloperoxidase initially produces HOCl, which subsequently chlorinates uracil by acid- or base-catalyzed reactions. Human neutrophils use myeloperoxidase and H 2 O 2 to chlorinate uracil, suggesting that nucleobase halogenation reactions may be physiologically relevant. Using a sensitive and specific mass spectrometric method, we detected two products of myeloperoxidase − 5-chlorouracil and 5-bromouracil − in neutrophil-rich human inflammatory tissue. Myeloperoxidase is the most likely source of 5-chlorouracil in vivo because the halogenated uracil is a specific product of the myeloperoxidase system in vitro . In contrast, previous studies have demonstrated that 5-bromouracil could be generated by either eosinophil peroxidase or myeloperoxidase, which preferentially brominates uracil at plasma concentrations of halide and under moderately acidic conditions. These observations indicate that the myeloperoxidase system promotes nucleobase halogenation in vivo . Because 5-chlorouracil and 5-bromouracil can be incorporated into nuclear DNA, and these thymine analogs are well-known mutagens, our observations raise the possibility that halogenation reactions initiated by phagocytes provide one pathway for mutagenesis and cytotoxicity at sites of inflammation.


Introduction
Chronic inflammation is implicated in the pathogenesis of human cancer, and many lines of evidence point to oxidative modification of DNA as one potential mechanism (1,2).
Phagocytic white blood cells, the cellular hallmark of inflammation, possess an NADPH oxidase that is a well-characterized source of reactive intermediates that damage DNA (3,4). The enzyme's initial product is superoxide, which dismutates to hydrogen peroxide (H 2 O 2 ). The oxidizing equivalents in H 2 O 2 are used by myeloperoxidase (5) (6,7), a heme protein expressed by neutrophils, monocytes, and macrophages (8) (9). At plasma halide concentrations, the major activity of myeloperoxidase is oxidation of chloride to hypochlorous acid (HOCl) (10,11).
HOCl + H + + Br -BrCl + H 2 O HOBr + H + + Cl -(Equation 3) which was then back-extracted with 0.2 M ammonium hydroxide and concentrated to dryness under vacuum. The extract was derivatized for electron capture gas chromatography mass spectrometry as described below.
Reverse-phase HPLC. Uracil oxidation products were analyzed and purified for NMR analysis as described (34). For mass spectrometric analysis, HPLC fractions were collected and concentrated under nitrogen. Uracils extracted from human tissue were eluted from an HPLC reverse-phase column (34)  interfaced with a Hewlett-Packard 5973 mass spectrometer. Ions were monitored in the negative ion electron capture ionization mode with methane as the reagent gas. Injector and interface temperatures were 280°C and 290°C, respectively. The initial GC oven temperature was 170°C for 1 min, followed by a 10°C/min increasing ramp to 250°C. Toluene was analyzed between

Results
The myeloperoxidase-H 2 O 2 -Clsystem chlorinates uracil at plasma concentrations of halide ions. We exposed uracil to the myeloperoxidase-H 2 O 2 -Clsystem in buffer A (100 mM sodium chloride, 100 µM DTPA, 50 mM sodium phosphate, pH 7) and used HPLC to characterize the product. The reaction mixture yielded a new peak of material with a retention time and UV/visible absorption spectrum identical to that of authentic 5-chlorouracil ( Fig.   1A,B). The structure of the uracil oxidation product was investigated by GC/MS (Fig. 1C). The Myeloperoxidase chlorinates uracil at C-5 of the pyrimidine ring. To determine the position of chlorine on the pyrimidine ring of uracil, we used HPLC to isolate the product from the myeloperoxidase reaction mixture and analyzed it with high-resolution 1 H NMR. The spectrum of the product was essentially identical to that of authentic 5-chlorouracil (data not shown), but it was distinct from that of 6-chlorouracil. Significant features included loss of the C-5 proton resonance, a downfield shift in the C-6 proton, and conversion of the C-6 proton resonance from a doublet to a singlet. These findings are consistent with substitution of a chlorine atom at the C-5 position of the pyrimidine ring.

5-Chlorouracil is a specific product of myeloperoxidase.
We used reverse-phase HPLC to characterize uracil chlorination by various peroxidases. 5-Chlorouracil was generated by myeloperoxidase but not by eosinophil peroxidase, lactoperoxidase, or horseradish peroxidase. Ascorbate and thiols potently inhibit uracil chlorination by myeloperoxidase. We used HPLC to examine the effects of antioxidants on the yield of 5-chlorouracil produced by the myeloperoxidase system ( Table 2). The water-soluble antioxidants ascorbate, N-acetyl cysteine, and glutathione, which react with HOCl to respectively form dehydroascorbate and oxygenated thiols, inhibited uracil chlorination almost completely. Taurine, which reacts with HOCl to produce chloramines, significantly reduced the yield of 5-chlorouracil. Nitrite, a nitric oxide breakdown product, also inhibited uracil chlorination, perhaps by acting as an alternative substrate for myeloperoxidase or by reacting with HOCl to produce nitrosyl chloride and nitrate.
Less effective were trolox, a water-soluble vitamin E analog, and phenol, which would be expected to form p-chlorophenol. Mannitol, frequently used as a hydroxyl radical scavenger, had virtually no effect on 5-chlorouracil production. These observations indicate that water-soluble antioxidants that scavenge HOCl are effective inhibitors of uracil chlorination by myeloperoxidase.
Uracil chlorination by myeloperoxidase has both acidic and neutral pH optima. Uracil chlorination by myeloperoxidase at neutral pH increased with increasing substrate concentration and reaction time ( Fig. 2) and was optimal at plasma concentrations of Cl -(100 mM). The product yield of 5-chlorouracil was ~20% relative to peroxide. The pH dependence for chlorination by myeloperoxidase showed an unusual bimodality, with optima at pH 4 and pH 7 ( Fig. 2). These observations indicate that myeloperoxidase produces high yields of 5-chlorouracil under both neutral and acidic conditions.
Two distinct reaction pathways chlorinate uracil. To better understand the mechanism for 5-chlorouracil production, we examined the pH dependence of 5-chlorouracil formation by HOCl, the chlorinating intermediate generated by myeloperoxidase (Fig. 3). Reagent HOCl produced quantitative yields of 5-chlorouracil under both strongly acidic and mildly basic conditions, with the product yield being lowest at pH 5. Because Cl 2 is in equilibrium with HOCl by a reaction requiring Cland H + (38), we explored its role in the reaction using chloride-free HOCl. Omitting NaCl ( Fig. 3; -NaCl) had little effect on 5-chlorouracil yield when the reaction mixture was > pH 5. In contrast, omission of Clmarkedly diminished product yield when the pH was < 5. Under strongly acidic conditions, chlorination of uracil was independent of Cl -, possibly due to generation of Cl 2 from trace amounts of Clcontaminant.
At plasma concentrations of Cl -, chlorination of uracil was directly proportional to the concentration of HOCl in the reaction mixture at both pH 4 and pH 8 (Fig. 3). However, the yield of 5-chlorouracil was increased markedly by Clat pH 4 but was independent of Clat pH 8. The rate of uracil chlorination was also significantly increased by 100 mM Clat pH 4 but not at pH 8 (Fig. 3). It is noteworthy that the reaction rate was highest at pH 8. In contrast, ring chlorination of deoxycytidine by HOCl is optimal under acidic conditions (18). These  (Table 3). In contrast, superoxide dismutase modestly enhanced the product yield, perhaps by increasing the availability of H 2 O 2 or by preventing superoxide from inactivating myeloperoxidase (39). uracil. Three of these samples also contained 5-bromouracil; levels of this halogenated nucleobase ranged from 3.9 to 59 µmol/mol uracil. When 5-bromouracil was detected, it was more abundant than 5-chlorouracil, suggesting that eosinophil peroxidase contributed to uracil halogenation or that Brwas present at high enough levels to favor bromination by myeloperoxidase (13,40).
To determine whether the 5-chlorouracil or 5-bromouracil detected in the samples might have been generated artifactually during workup, we performed three sets of experiments. First, the analytic procedure was applied to samples of plasma or urine from healthy donors. Under these conditions, neither halogenated base was detectable. Second, we quantified the levels of halogenated nucleobases in liver harvested from apparently healthy mice. We, again, failed to detect either 5-chlorouracil or 5-bromouracil (limit of detection, ~ 0.9 nmol/mol uracil) providing strong evidence that artifactual halogenation of uracil was not taking place during tissue work-up or analysis. Third, uracil labeled with two 13 C atoms and one 15  Previous studies have demonstrated that the artifactual generation of halogenated and nitrated oxidation products is a major problem during the analysis of biological material (42)(43)(44)(45)(46)(47).
Indeed, some studies have found that virtually all of the oxidation products detected in tissue were generated artifactually (46,47). This is particularly a problem when acidic conditions are used for sample hydrolysis or derivatization. However, we were unable to detect halogenated nucleobases in human plasma or urine, or in liver harvested from apparently healthy mice.
Moreover, we directly monitored the ex vivo formation of halogenated nucleobases by adding [ 13 C 2 , 15 N 1 ]uracil to tissue samples and then monitoring the appearance of halogenated [ 13 C 2 , 15 N 1 ]uracil. Artifactual halogenation accounted for <1% of 5-chlorouracil and <5% of 5bromouracil that we detected in inflammatory tissue. Collectively, these observations provide strong evidence that 5-chlorouracil and 5-bromouracil were present in inflamed human tissue and that the halogenated pyrimidines were not generated ex vivo during extraction or analysis.
While many studies have focused on direct oxidation of chromosomal DNA as a mechanism for mutagenesis (1,2), we recently showed that myeloperoxidase and eosinophil peroxidase brominate the free nucleobase uracil and the free nucleoside deoxycytidine in vitro at plasma concentrations of chloride and bromide (13,19,34). Mass spectrometry, flow cytometric analysis, and immunohistochemical studies revealed that 5-bromouracil was incorporated into the DNA of dividing mammalian cells exposed to purified 5-bromodeoxycytidine produced by eosinophil peroxidase (19). Previous studies have also shown that 5-chlorouracil and 5bromouracil can be taken up by mammalian cells and tissues and subsequently converted to their corresponding deoxynucleosides by thymidine phosphorylase (48,49). DNA polymerase then incorporates the resulting chlorodeoxyuridine and bromodeoxyuridine into DNA (50,51). Thus our detection of halogenated uracil may be significant, as 5-chlorodeoxyuridine and 5-bromodeoxyuridine are well-established thymidine analog mutagens that mispair with guanine, causing GC to AT and AT to GC transitions (52). Based on our own observations and previous findings, we propose that nucleobases are halogenated by phagocytes and incorporated into host DNA during replication or repair in a process we term nucleotide precursor mutagenesis. In this way DNA may be damaged both indirectly by incorporation of halogenated nucleosides and directly by oxidation of DNA in situ.
We have demonstrated that 5-chlorouracil is a specific product of myeloperoxidase, and have detected two products of myeloperoxidase−5-chlorouracil and 5-bromouracil−in neutrophilrich inflammatory tissue from humans. These observations indicate that human neutrophils use myeloperoxidase to halogenate uracil in vivo. Transition mutations occurring in genes containing these halogenated nucleobases could then alter cell cycle regulation, DNA replication, or DNA repair, creating a link between myeloperoxidase activity and increased cancer risk.