5-Chlorouracil, a Marker of DNA Damage from Hypochlorous Acid during Inflammation

Hypochlorous acid (HOCl), generated from H2O2 and Cl– by myeloperoxidase in activated neutrophils, causes tissue damage during inflammation. We have developed a simple, sensitive (∼0.2 fmol on column) and specific GC-MS assay for the detection of 5-chlorouracil (5-ClUra), a signature product of HOCl-mediated damage to nucleobases. In this assay, 5-ClUra is released from isolated DNA by a digestion with nuclease P1, alkaline phosphatase, and thymidine phosphorylase (TP), or from chlorinated nucleosides in biological fluids by TP. The freed 5-ClUra is derivatized with 3, 5-bis-(trifluoromethyl)-benzyl bromide, which is detected by negative chemical ionization mass spectrometry. The assay can be used to simultaneously detect other halogenated uracils including bromouracil. Using this assay, we showed that 5-ClUra is generated by the reaction of low micromolar HOCl with (deoxy)cytidine, (deoxy)uridine, and DNA. In cell cultures, an increase of 5-ClUra was detected in DNA when cells were treated with sublethal doses of HOCl and allowed to proliferate. The elevation of 5-ClUra was markedly accentuated when physiologically relevant concentrations of (deoxy)uridine, (deoxy) cytidine, uracil, or cytosine were present in the medium during HOCl treatment. In the carrageenan-induced inflammation model in rats, chlorinated nucleosides was significantly increased, compared with controls, in the exudate fluid isolated from the inflammation site. Our study provides the direct evidence that chlorinated nucleosides are found in the inflammation site and can be incorporated in DNA during cell/tissue proliferation. These findings may be relevant to the carcinogensis associated with chronic inflammation.

Chronic inflammation contributes to the development of cancer (1,2). Oxidative modification of DNA bases may cause mutation and therefore could contribute to carcinogenesis. Hypochlorous acid (HOCl), generated from H 2 O 2 and Cl Ϫ by myeloperoxidase (MPO) 1 in activated neutrophils, plays a key role in the host defense against bacteria and other invaders (3). Consistent with this notion, the amount of intraphagosomal production of HOCl during phagocytosis is estimated to be sufficient to exert potent bactericidal activity (4), and MPO knock-out mice are susceptible to infection (5). However, despite its protective role against infection, excessive HOCl released during inflammation can contribute to tissue damage. Several biomarkers specific for HOCl-caused damage have been identified. Chlorination of cholesterol gives rise to cholesterol chlorohydrin (6). 3-Chlorotyrosine, a protein oxidation marker from HOCl or molecular chorine (7,8), was found in inflammatory tissues and the lesions of inflammation-associated diseases, such as atherosclerosis (9).
In addition to proteins and lipids, HOCl also oxidizes nucleobases, which are electron-rich. It has long been recognized that bleach, composed mainly of sodium hypochlorite, oxidizes uracil to 5-chlorouracil (5-ClUra) (10). Whiteman et al. (11) recently demonstrated that 5-ClUra can be produced by the reaction of HOCl with DNA, although relatively high concentrations of HOCl were used in the study. Henderson et al. (12) showed that under relatively acidic conditions (pH 4.5-6), HOCl, or MPO-H 2 O 2 -Cl Ϫ oxidizes cytosine mainly to 5-chlorocytosine. 5-Chlorocytosine appears to be relatively unstable and converts to 5-ClUra by spontaneous (13) or enzymatic deamination (14). Therefore, 5-ClUra seems to be a specific marker of HOCl-caused damage to nucleobases in DNA.
Although these in vitro studies suggest that excessive generation of HOCl may lead to an increase in 5-ClUra during inflammation, due to the lack of a suitable assay, no evidence shows that this adduct is present in DNA of living cells after an insult of sublethal doses of HOCl. Herein we describe a simple, sensitive, and reliable GC-MS assay for the measurement of 5-ClUra. Using this assay, we found that 5-ClUra can be detected in the DNA isolated from cells that were treated with sublethal doses of HOCl, or in the inflammation fluid in a rat model.
DNA Isolation-DNA was isolated from cells and tissues using a Qiagen kit (69504, Valencia, CA) or phenol-chloroform extraction. In the latter method, nuclei were isolated from tissue homogenate by a published protocol (15), then digested with RNase A to remove RNAs, incubated with proteinase K to degrade proteins, and finally extracted with phenol and chloroform. DNA was then precipitated by ethanol in the presence of 0.3 M sodium acetate at pH 5.1. The amount of DNA was quantified by the picogreen assay (Molecular Probes, Eugene, OR).
Enzymatic Hydrolysis of DNA-20 -150 g of DNA, dissolved in 200 l of 20 mM sodium acetate at pH 5.1, was digested to nucleotides by 3 l, 3.3 mg/ml nuclease P1 under 65°C for 10 min. The pH of the resulting solution was adjusted to ϳ8.0 using 40 l of 0.1 M sodium borate at pH 8.7. Nucleosides were obtained on incubation with 2 l of 4 units/l alkaline phosphatase at 37°C for 1 h. Thymine and 5-chlorouracil were released on incubation with 1 unit of purified thymidine phosphorylase (see "Experimental Procedures") at 37°C for 1 h, with the addition of 50 l of 0.2 M potassium phosphate (pH 7.2).
Extraction of Thymine and 5-ClUra-After enzymatic digestion, 128 pg of 5-[ 15 N 2 , 13 C]chlorouracil and 95 ng of [D 4 ]thymine were added as internal standards. 1 ml of ethyl acetate was added into the resulting mixture, and free bases were extracted by vigorously shaking on a vortex mixer for 30 s. After a brief centrifugation, 0.85 ml of ethyl acetate layer was collected, and the solvent was evaporated under N 2 . Complete dehydration of samples was achieved by further evaporation for 1 h in a speed vaccum (Savant, Farmingdale, NY).
Derivatization-After extraction and drying, the residue was resuspended in 50 l of acetonitrile, 5 l of triethylamine, and 1 l of 3,5-bis(trifluoromethyl)benzyl bromide (BTFMBzBr). The resulting mixture was incubated by vigorously shaking in a G24 shaker (New Brunswick Scientific Co, Inc., New Brunswick, NJ) at 37°C for 25 min. After derivatization, 50 l of water and 110 l of isooctane were added. BTFMBz-5-chlorouracil and thymine derivative were extracted into isooctane by shaking vigorously on a vortex mixer for 30 s followed by a brief centrifugation. The top isooctane layer was collected and injected in a gas chromatograph-mass spectrometer.
Gas Chromatography-Mass Spectrometry-One microliter of sample in isooctane was injected onto a 5890 II gas chromatograph (GC) with a 7673 autosampler (Hewlett-Packard, Palo Alto, CA). GC was set up in the splitless mode with a purge activation time of 1 min and split vent flow of 50 ml/min. Ultra-high purity helium (Praxair, CA) was used as the carrier gas at a linear velocity of 55 cm/s. The injection port was maintained at 280°C. An adequate separation of thymine and 5-chlorouracil and a good peak shape were obtained on a capillary column (30 m ϫ 0.25 mm internal diameter ϫ 0.25 mm film thickness (HP-5MS), Hewlett Packard) by holding the oven at 100°C for 1 min and then ramping to 220°C at 20°C/min and staying for 3 min, to 250°C at 30°C/min and staying for 2 min, and to 300°C at 40°C/min and staying for 3 min. A 5989A mass spectrometer (Hewlett Packard) was used for acquiring both electron impact (EI) and negative chemical ionization (NCI) mass spectra. EI data was obtained using an electron energy of 70 eV. NCI-MS was obtained using the following parameters: electron energy of 230 eV, ion source temperature of 250°C, quadrapole temperature of 100°C, high energy dynode on, and ultra high purity methane reagent gas (Matheson, Newark, CA) maintained at ion source pressure of 2.0 -2.2 torr. The instrument was controlled using Hewlett Packard Chemstation 1034C version C.01.05 software.
The maximum sensitivity was achieved by using the selective ion monitoring (SIM) in the NCI. In the SIM mode, m/z of 371 and 374 was used to detect and quantify BTFMBz-derivatized 5-ClUra and 5-[ 15 N 2 , 13 C]chlorouracil. In order to measure thymine, a 2500-time dilution was made before injection and m/z at 351 and 355 was used to quantify BTFMBz-thymine and BTFMBz-[D 4 ]thymine. The quantification was based on the ratio of the peak area of 5-ClUra and internal standard and the calibration curve. The level of 5-ClUra in DNA was expressed as the ratio of 5-ClUra and Thy (mol/10 6 mol).
Cell Culture Experiments-Both human Hela cells and murine embryo C3H10T1/2 cells were obtained from the American Type Culture Collection (Manassas, VA). Cells were routinely kept in Dulbecco's modified Eagle's medium (DMEM) with 10% of fetal bovine serum (FBS), split, and seeded with ϳ60 -70% confluence at 24 h before treated with HOCl. In some experiments, cells were incubated with 10 -40 M of HOCl in the Hank's Balanced Saline solution (HBSS) at 37°C for 10 -30 min and then immediately harvested. In other experiments, after HOCl treatment, cells were replenished by DMEM with 5% FBS, or cells were treated with HOCl in DMEM with 5% FBS, and then further incubated for 24 h before harvesting.
Detection of 5-ClUra in the Air Pouch Inflammation Model in Rats-The animal use protocol was approved by the animal care committee at Children's Hospital Oakland Research Institute and was strictly followed. The details of the animal study was described elsewhere (16). Briefly, male Wistar rats (250 -330 g) (Charles River, CA) were caged singly and routinely fed ad libitum with Purina Chow with free access to the tap water. An air pouch was created by a subcutaneous injection of 12 ml of sterile air into the intrascapular area of the rat's back. Thirty hours later, 2 ml of 0.5% carrageenan (Sigma) or PBS was injected into the air pouch. Twenty hours after the induction of inflammation, the air pouch was lavaged with PBS without Ca 2ϩ /Mg 2ϩ and the exudate fluid was collected. After a brief centrifugation, the supernatant was collected and stored under Ϫ80°C until use.
To detect 5-ClUra in the isolated fluid, 400 l of exudate was vigorously mixed with 2:5 ml of methanol/hexane, in the presence of 1 mM methionine, to remove most proteins or lipids. After a brief centrifugation, the methanol layer was collected and evaporated until the volume was reduced to ϳ0.25 ml. 5-ClUra was extracted from the resulting aqueous phase into 1 ml of ethyl acetate after the addition of 128 pg of 5-[ 15 N 2 , 13 C]chlorouracil. To release 5-ClUra from nucleosides, the resulting aqueous phase (ϳ0.25 ml) was incubated at 37°C for 1 h in the presence of 1 unit of TP and 50 l, 0.2 M potassium phosphate (pH 7.2). 5-ClUra was then extracted using 1 ml of ethyl acetate after the addition of internal standard. To detect 5-ClUra in DNA, tissues along the liner of the pouch was collected and DNA was isolated using Qiagen kit. Samples were then processed, derivatized, and analyzed by GC-MS as described above.
Statistics-A non-paired Student's t test was used in all conditions. A p value less than 0.05 was considered statistically significant.

Development of GC-MS Assay for the Detection of 5-ClUra-
We previously developed a sensitive GC-MS method for the detection of uracil where uracil is derivatized with BTFMBzBr (17). Here we directly adapted this technique for the assay of 5-ClUra. Derivatization of 5-ClUra with BTFMBzBr yields a volatile, electron-capture derivative that is readily detected by gas chromatography coupled with EI-or NCI-MS. Fig. 1A illustrates the EI spectrum and structure of the derivatized compound, 5-ClUra-diBEFMBz. A mass to charge ratio (m/z) at 598 corresponds to the molecular ion In the NCI mass spectrum (Fig. 1B), the predominate peak was found at m/z 371, which resulted from a cleavage of one BT-FMBz from the parental ion. It is noticeable that the two major isotopes of choride atoms, i.e. 35 and 37 with natural isotopic abundance of 3:1, give rise to the characteristic peaks, as indicated, for instance, by the abundance ratio of 3:1 for the m/z at 371 to 373 [M-BTFMBz] (Fig. 1). The EI and NCI fragmentations of the internal standard, [ 15 N 2 , 13 C]5ClUra-BTFMBz, were shifted up three mass units as a result of three heavier atoms (not shown).
In order to effectively release 5ClUra from DNA, we have tried several methods. Although DNA can be completely hydrolyzed by formic acid at high temperature (11) (13), in order to achieve good sensitivity and reproducibility, multiple complicated purification and enrichment steps had to be subsequently performed (13). We also attempted to digest DNA using HFpyridine at room temperature for 1 h. However, contaminates

5-Chlorouracil
from this reagent showed a significant signal at 371, which co-migrates with 5-ClUra on the chromatograph, and thus interfered with 5-ClUra detection. We then turned to the enzymatic hydrolysis. Uracil glycosylase, which removes uracil from DNA, failed to recognize 5-ClUra. TP, an enzyme documented to excise thymine and its analog, bromouracil, from the corresponding nucleosides (18), was then tried. We found that TP is able to effectively remove 5-ClUra from nucleosides (5-CldU or 5-ClUridine), but the commercially available TP contains significant amounts of contaminate which interferes with 5-ClUra detection. To eliminate contaminates, we purified the commercial TP by filtration (see "Experimental Procedures"). The purified enzyme proved to be satisfactory for the current use. Because 5-bromouracil can also be removed by TP from the corresponding brominated nucleoside (18), another halogenated adduct and inflammation-associated biomarker from eosinophile peroxidase-H 2 O 2 -Br Ϫ (19), and the BTFMBz-derivatized 5-bromouracil and 5-ClUra were separated well on the GC chromatography (Fig. 2), these two adducts can be readily measured simultaneously using the current assay.
Scheme I summarizes the assay for 5-ClUra from DNA or chlorinated nucleosides in biological fluids. For DNA, the procedure includes a digestion of isolated DNA to nucleotides using nuclease P1, a conversion of nucleotides to nucleosides by dephosphorylation using AP, and a release of 5-ClUra and thymine by TP. In biological fluids, 5-ClUra is freed from chlorinated nucleosides by an incubation of TP after a removal of proteins and lipids by methanol/hexane extraction (see "Experimental Procedures"). Because the reaction of TP is reversible, the complete excision of 5-ClUra from 5-CldU or 5-chlorouridine was achieved in the presence of excess amount of phosphate (with a final concentration of 50 mM). The complete release of 5-ClUra was confirmed by the quantitative excision of 5-ClUra from known amounts of 5-CldU (Table I) (Table I). The quantitative removal of thymine is evident from the fact that the amount of DNA measured by picogreen assay is equal to that calculated from the GC-MS assay where thymine is assumed to represent 29% of the total base (Fig. 3A).
The detection limit of the current assay is ϳ10 -20 fg of 5-ClUra on column, with a signal to noise ratio of more than 5. The minimum reliable detected value from DNA is 0.1-0.3 ClUra/10 6 Thy. A linear response curve was achieved from low picogram (Fig. 3B) to nanogram ranges (not shown).
5-ClUra Was Detected When Low Micromolar (1-10 M) HOCl Reacts with Nucleosides and DNA-To test whether low micromolar HOCl leads to the production of detectable 5-ClUra, deoxycytidine (dC), deoxyuridine (dU), and calf thymus DNA were incubated with 1-10 M of HOCl in the presence of physiological concentrations of chloride ion. We observed a dose-dependent increase of 5-ClUra under the conditions examined (Fig. 4). At neutral pH, the reaction of HOCl with dU produced higher yields of 5-ClUra than with dC (Fig. 4A). This is likely due to the fact that dC may be primarily oxidized to the chloramine, in contrast to dU which can be directly converted to 5-ClUra (20). The yield from chlorination of dC is higher under slightly acidic conditions (pH 5), compared with pH 7, which is likely due to the formation of molecular chlorine from HOCl under acidic conditions, which then chlorinates dC to 5-chlorocytidine (12). Chen et al. (13) recently reported that 5-chlorocytidine is unstable and quickly deaminates to 5-chlorouridine. Besides dU and dC, uridine, and cytidine were oxidized by HOCl to generate 5-ClUra with a similar product yield (data not shown). A dose-dependent increase of 5-ClUra was detected from calf thymus DNA when treated with 1-10 M of HOCl (Fig. 4B), presumably as a result of the chlorination of deoxycytidine.

Elevated Levels of 5-ClUra Were Detected in the DNA of Proliferating Cells that Were Treated with Sublethal Doses of HOCl and Allowed to Proliferate, but not after an Immediate
Insult-To test whether 5-ClUra can be generated by the direct insult of HOCl in live cells, both human Hela cells and murine embryo C3H10T1/2 cells were treated with non-lethal doses of HOCl. Immediately after HOCl (0 -40 M) treatment, no significant increase in 5-ClUra was detected from isolated DNA (Fig. 5A). At doses lower than 40 M, no apparent cell death was observed as assayed by the trypan blue exclusion (data not shown). Even at higher dose of HOCl, e.g. 100 M for 10 min, where substantial amounts of cells became detached or stained in the trypan blue exclusion assay, no significant increase of 5-ClUra could be found in isolated DNA (data not shown). On the other hand, after cells were treated with HOCl for 20 min, replenished with complete medium with FBS and further incubated for 20 -24 h, a mild but significant increase in 5-ClUra was detected in the isolated DNA (Fig. 5A, similar results were also obtained with Hela cells, not shown). The elevation of 5-ClUra was markedly accentuated by the addition of 1 M dU and dC during HOCl treatment (Fig. 5B). This intensified elevation of 5-ClUra in DNA was blocked up to ϳ98% with the addition of 0.5 mM of methionine, a potent scavenger of HOCl. The presence of complete medium plus 5% FBS and dU or dC during HOCl treatment resulted in a reduction of 5-ClUra (Fig.  5C) compared with HBSS as the medium. Besides dU and dC, the addition of low micromolar uracil, cytosine, uridine, and cytidine to the culture medium during HOCl treatment also led to significant elevations of 5ClUra in the DNA (Fig. 5D). These observations suggest that although HOCl may not penetrate to the nuclei to directly oxidize DNA, it could modify dC, dU, uracil, cytosine, uridine, and cytosine in the surrounding tissues to generate chlorinated adducts, which might be incorpo-

5-Chlorouracil
rated into the DNA during cell growth.
To test the incorporating efficiency of chlorinated adducts into the DNA, 10 pM to 10 M of 5-CldU or 5-ClUra were added to the medium when cells were in their log growth phase. As low as 10 pM of 5-CldU was effectively incorporated into DNA with an overall efficiency of 30 -40% (Table II). On the other hand, the direct incorporation yield for 5-ClUra was more than 10 4 -10 5 -fold lower than that of 5-CldU (Table II).

Elevated Levels of Chlorinated Uracil Adduct(s) Were Detected in the Exudate Fluid at the Inflammation Site in a Rat
Inflammation Model-To test whether 5-ClUra is elevated during inflammatory response in vivo, we measured this adduct in the exudate fluid and the DNA from tissues isolated at the inflammation site in the carrageenan-induced inflammation model in rats, which is believed to mimic joint diseases (16,21). In this model, a single injection of carrageenan in the precreated air pouch in the intrascapular area of rat backs led to localized inflammation as indicated by a marked increase in neutrophil infiltration, pro-inflammatory eicosanoids and oxidative damage (16,21). A markedly enhanced neutrophil infiltration to the inflammation site has been previously reported (21) and was also confirmed by us (16). At the time when

5-Chlorouracil
samples were collected, neutrophils accounted for Ͼ90% of the total cells in the exudate. Using the current GC-MS assay, we detected a significant increase in 5-ClUra in the isolated exudate fluid in carrageenan-treated rats, compared with PBS-injected controls (Fig. 6). It is important to note that the significant increase in 5-ClUra was observed only after the fluid was treated with TP, which releases 5-ClUra from 5-CldU or 5-Cl-uridine, suggesting that the predominate chlorinated adducts may be chlorinated nuclosides, but not 5-ClUra. We did not observe a statistically significant elevation of chlorination in the DNA that was isolated from the inflammation site tissues (data not shown). DISCUSSION We have developed a simple, specific, and sensitive assay for the detection of 5-ClUra. The assay employs thymidine phosphorylase-catalyzed de-pyrimidination to achieve a complete release of derivatizable 5-ClUra and thymine from DNA under physiological condition. The specificity of the enzymatic cleavage and the subsequent separation of freed bases from the nucleosides by a one-step ethyl acetate extraction (Scheme I), result in a simple and clean product profile, which is critical for obtaining an efficient and reproducible derivatizating outcome and high sensitivity. Our assay is thus simpler and yet highly sensitive, compared with a recently published GC-MS assay (13), which uses acid hydrolysis to completely digest DNA, and involves multiple clean-up and enriching steps in order to achieve a similar sensitivity. Our assay is also capable of measuring chlorinated nucleosides in biological fluids (Fig. 6), which are difficult to directly analyze by acid hydrolysis. In addition, the simultaneous removal of both 5-ClUra and thymine by TP makes it convenient to use thymine to estimate the amount of DNA. The assay can simultaneously measure chlorinated and brominated uracil, while the specificity and sensitivity for 5-ClUra is particularly valuable because an equally sensitive and specific ELISA (enzyme-linked immunosorbent assay) is not likely to be available due to the poor antigenicity of the chloride atom. In contrast, ELISA assays are commonly available for detecting 5-bromouracil, due to the strong antigenicity of bromide atom. Although TP indistinguishably releases 5-ClUra from 5-CldU or 5-Cl-Uridine, in the current study, RNA was completely removed by the digestion with RNase A during DNA isolation (data not shown). On the other hand, the method can be adapted to measure the chlorinated and brominated nucleobases in RNA, the feasibility of which is currently under investigation.
Previous studies together with our current investigation indicate that 5-ClUra appears to be a sensitive, specific and stable marker of HOCl-caused damage. It has been shown that the reaction of HOCl with nucleobases primarily leads to extensive damage to pyrimidines such as cytosine, uracil and thymine, but minor damage to purines (11,12,22). HOCl oxidizes uracil to 5-ClUra (20), and primarily oxidizes cytosine to chloramines under neutral pH or 5-chlorocytosine at acidic pH (12). Chloramine-cytosine slowly gives rise to other base lesions, including 5-ClUra (23-25). 5-Chlorocytosine is also relatively unstable and spontaneously deaminates (13) or during acid hydrolysis to generate 5-ClUra (11,13). In biological systems, cytidine deaminase readily converts 5-chloro-dC to 5-chloro-dU and the apparent half-life of 5-chloro-dC is about 10 min in vivo (14). On the other hand, 5-ClUra seems to be quite stable because it is resistant to further modification by acid hydrolysis or strong oxidants including peroxynitrite (26). Therefore, 5-ClUra appears to be a stable and specific product of HOCl-mediated modification of cytosine and uracil. In the current study, we find that low micromolar HOCl oxidizes (deoxy)uridine or (deoxy)cytidine and DNA to generate detectable 5-ClUra at both pH 7 and pH 5. Under slightly acidic conditions (pH 5), HOCl-mediated chlorination of dC yields similar amounts of 5-ClUra compared with that of dU (Fig. 4), suggesting that both (deoxy)uridine and (deoxy)cytidine contribute significantly to the generation of 5-ClUra. Because our assay does not involve acidic hydrolysis, 5-ClUra may be directly derived from chlorinated cytosine by way of chloramine conversion and/or deamination of 5-chlorocytosine. Although high micro-to millimolar HOCl has been estimated at the site of inflammation or inside phagosomes (4), the scavenging by antioxidants and other biomolecules makes low micromolar HOCl more likely to be relevant to the levels that could react with nucleobases in complex biological environments. Thus, in addition to being stable and specific, 5-ClUra is a sensitive marker for HOCl-mediated damage to nucleotides, nucleosides, free nucleobases, DNA, and RNA.
Although HOCl chlorinates nucleobases in DNA in vitro, considering its high reactivity and cytotoxicity (27), the question regarding the physiological importance of HOCl-mediated DNA damage is whether the chlorinated adduct can be found in the nuclear DNA of live cells after being treated with nonlethal doses of HOCl. In the earlier stage of the study, we did not observe a significant increase of chlorinated adduct in the DNA immediately after cells were treated with HOCl, possibly because HOCl is scavenged by defense biomolecules such as glutathione (28) before reaching nuclei and directly oxidizing DNA. On the other hand, we observed an elevation of 5-ClUra after a prolonged incubation subsequent to the HOCl treatment, suggesting that chlorinated adducts, which may be derived from HOCl-oxidized cytosolic nucleotide/nucleoside pools, are incorporated into the DNA during cell proliferation. This mechanism is confirmed by the observation that the addition of physiologically relevant concentrations of nucleosides and nucleobases to the media during HOCl treatment causes a markedly intensified increase of 5-ClUra in the isolated DNA (Fig. 5). The presence of methionine, a HOCl scavenger, almost completely abolishes this increase, although a significant elevation of 5-ClUra is found even in the presence of complete growth medium during HOCl treatment. We further compared the relative efficiency for 5-CldU and 5-ClUra to be incorporated into DNA. In this regard, incorporation of 5-CldU is much more efficient than that of 5-ClUra, suggesting that at similar concentrations, chlorinated deoxynucleosides, but not nuclobases, are likely to be the major precursors to be incorporated. Our cell culture experiments can mimic the situation at the site of inflammation. Under physiological conditions, the plasma concentrations of uracil, cytosine, uridine, and cytidine are in the low micromolar ranges (29). During inflammation, nucleosides and nucleobases in the plasma exude to the inflammatory site and become the targets of HOCl, which is generated by the MPO-H 2 O 2 -Cl Ϫ system of infiltrated neutrophils. The chlorinated nuclosides or bases may then be incorporated into proliferative tissues.
We have detected a significant increase of 5-ClUra in the exudate fluid isolated from the inflammation site in which infiltrated neutrophils are particularly abundant (Fig. 6). Interestingly, in the current study, the enhanced 5-ClUra at the site of inflammation is likely derived from chlorinated nucleosides instead of free bases because a significant increase of 5-ClUra is detected only after samples were de-pyrimidinated by TP. Despite an increase of chlorinated adducts in the inflammatory fluid, we did not detect significant enhancement of 5-ClUra in the DNA isolated from the inflammatory tissues. This may be explained by the possibility that the tissues, when isolated, are not undergoing proliferation. In the future, tissues from a chronic inflammation site or undergoing proliferation subsequent to the inflammation should be tested for the chlorinated products in DNA.
It is conceivable that during chronic inflammation, chlorinated nucleosides may be frequently generated and incorporated into the DNA of proliferating tissues. This may lead to the accumulation of chlorinated bases in the DNA especially because halogenated uracils, the thymine analog, are not efficiently repaired (30). The accumulation of 5-ClUra in DNA has been shown to induce sister chromatid exchange, and mutations (30). Therefore, chronic inflammation-mediated accumulation of modified bases may be relevant to the development of cancer. II Picomolar concentrations of 5-CldU, but not 5-ClUra, were efficiently incorporated into C3H10T1/2 cells Murine embryo C 3 H 10 T 1/2 cells were grown to reach 50 -60% confluency in T-75 flask. Cells were then re-fed with 20 ml of fresh medium containing 5-CldU or 5-ClUra at indicated concentrations and incubated for 20 h. DNA was then isolated from the harvested cells using Qiagen kit. FIG. 6. Elevated 5-ClUra was detected in the exudate fluid at the inflammation site. In the carrageenan-induced inflammation model, exudate was collected after a brief centrifugation of the lavaged fluid isolated from the inflammation site. 400 l of exudate fluid from carrageenan-treated rats (CARR, n ϭ 5) or PBS-injected controls (PBS, n ϭ 3) was treated with TP at 37°C for 1 h in the presence of 1 mM methionine. 5-ClUra was extracted by ethyl acetate, dried under nitrogen gas, derivatized, and then analyzed by GC-MS (see "Experimental Procedures"). diate. They also found enhanced levels of 5-ClUra and 5-bromouracil in human inflammatory tissues. Their study complements our current observations that physiologically relevant concentrations of HOCl oxidize uracil, cytosine, and their nucleosides to generate chlorinated adducts, which are found in the inflammation sites in a rat inflammation model and can be incorporated into nuclear DNA during cell/tissue proliferation.