Neutrophils amplify the formation of DNA adducts by benzo[a]pyrene in lung target cells.

Inflammatory cells and their reactive oxygen metabolites can cause mutagenic effects in lung cells. The purpose of this study was to investigate the ability of activated neutrophils to modulate DNA binding of benzo[a]pyrene (B[a]P), a known carcinogen, in lung target cells. Equivalent numbers of rat lung epithelial cells (RLE-6TN cell line) and freshly isolated human blood neutrophils (PMN) were coincubated in vitro for 2 hr after addition of benzo[a]pyrene (0.5 microM) or two of its trans-diol metabolites, with or without stimulation with phorbol myristate acetate (PMA). DNA adducts of B[a]P-metabolites were determined in target cells using 32P-postlabeling; oxidative DNA damage (7-hydro-8-oxo-2'-deoxyguanosine [8-oxodG]) was evaluated by high performance liquid chromatography with electrochemical detection. Increased DNA adducts were observed in lung cells coincubated with polymorphonuclear leukocytes (PMN). Activation of PMN with PMA, or addition of more activated PMN in relation to the number of lung cells, further increased the number of adducts, the latter in a dose-response manner. Incubation with B[a]P-4,5-diol did not result in any adduct formation, while B[a]P-7,8-diol led to a significant number of adducts. Moreover, PMA-activated PMN strongly enhanced adduct formation by B[a]P-7,8-diol, but not 8-oxodG, in lung cells. The addition of antioxidants to the coincubations significantly reduced the number of adducts. Results suggest that an inflammatory response in the lung may increase the biologically effective dose of polycyclic aromatic hydrocarbons (PAHs), and may be relevant to data interpretation and risk assessment of PAH-containing particulates.


Introduction
An association between inflammationrelated and neoplastic events has been recognized for a long time (1,2). Although a number of mediators released by inflammatory cells may play a role in this process, several studies have demonstrated that reactive oxygen species (ROS) play an important role in this interaction (3)(4)(5). First, indirect mutagenic effects of inflammatory cells were demonstrated in the HPRT gene of rat lung epithelial (RLE) cells isolated after in vivo exposure to low solubility particles such as toxic silica and nontoxic carbon black (6). The increase in inflammatory response paralleled the number of HPRTmutations in RLE cells, which suggested that inflammatory cells, and more specifically their ROS, may be responsible for the mutagenic effects of particle exposure (7). Second, precarcinogens can be activated by the attack of ROS, or cells producing these species (5,8). Increased generation of ROS such as superoxide anion, nitric oxide, hydrogen peroxide, and hydroxyl radical (9) is a characteristic of activated inflammatory cells such as macrophages and polymorphonuclear leukocytes (PMN). In vitro studies have shown that activated PMN can bioactivate polycyclic aromatic hydrocarbons (PAHs), benzene, and drugs such as acetaminophen, chlorpromazine, p-aminobenzoic acid, and some imidazoles (3). An in vivo model that induced topical skin inflammation by initial application of the phorbol ester 12-0tetradecanoylphorbol-13-acetate (TPA), and a second application (16 hr later) of TPA and benzo[a]pyrene (B[a]P)-7,8-diol, was used to study the effect of inflammation on metabolism. Results showed a 50% enhancement of adducts to epidermal DNA (10). Since many environmental particles contain PAHs and induce lung inflammation (11,12), we investigated whether activated neutrophils biotransform B[a]P as a model carcinogen, and whether they are able to modulate DNA binding of one of its major metabolites. To investigate the direct action of ROS on DNA, we evaluated the formation of 7-hydro-8-oxo-2'-deoxyguanosine (8-oxodG) at the same time. Studies were done in an in vitro model coincubating PMN and lung epithelial cells ( Figure 1). genotype (13) was used to measure DNA adducts. RLE cells were grown in Ham's F12 medium supplemented with 1% IM HEPES buffer and 5% heat-inactivated FCS; fresh medium was added to the cells every second day. At 95% confluency, culture dishes were rinsed with phosphatebuffered saline (PBS), cells were detached (trypsin-EDTA, Gibco), and seeded 1:10 in fresh culture medium. All experiments with RLE cells were performed between passages 35 and 50. Duplicate incubations were performed in separate experiments using different cell passages. Coincubations were performed at confluency of RLE cells. PMN were freshly isolated from the blood of a human male volunteer by gradient centrifugation (800 x g, 20 min, 4°C) using lymphoprep (Nycomed, Oslo, Norway). To the lowest layer, 3 volumes of cold (4°C) lysis buffer (155 mM NH4Cl, 10 mM KHCO3, 10 mM EDTA, pH=7.4, Merck) were added to lyse the erythrocytes; this procedure was repeated and PMN were suspended in Ham's F12 medium (2 x 107 cells/ml). Cell number was determined by hemacytometer chamber (Biirker) using 0.4% trypan blue to determine viability of PMN. During all steps of isolation PMN were kept on ice to avoid activation of the cells. At the start of the incubation, PMN (2 x 107) were added'to a T164 flask containing an equal amount of RLE cells (about 1.25 x 105/cm2). All experiments except dose-response experiments (see below) used this 1:1 ratio of PMN and RLE cells. B[a]P and its diols (final concentration 0.5 pM) were added in DMSO without exceeding a final concentration of 0.05% v/v. Stock solutions were kept out of the light during preparation and storage. PMN were activated with 100 ng/ml PMA. Control incubations were T164 flasks without PMN, without RLE cells, or without substrate/PMA. To determine the significance of radical-related mechanisms in DNA adduct formation in coincubations, several scavengers-were evaluated. Coincubations were run as described above in the presence of catalase (5000 U/ml), SOD (2500 U/ml), or mannitol (100 mM). In these experiments, Hanks balanced salt solution (HBSS, Gibco) was used instead of Ham's F12/FCS culture medium. Cytotoxicity Cytotoxicity in RLE cells was determined by the MTT colometric assay developed by Mosmann (14). The test is based on the activity of mitochondrial dehydrogenase in viable cells to convert MTT into a blue formazan dye. Cells were plated into 96well flat-bottom culture plates (Costar, Cambridge, MA) at 1 x 104 cells/100 pl medium and cultured at 37°C and 5% CO2. After 2 days, fresh medium was added to each well. At confluency, samples were added and incubated at 37°C, 5% C02; subsequently, 25 pl of MTT dissolved in sterile PBS (2 mg/ml) was added and followed by a 3-hr incubation (37°C, 5% C02). The medium was then replaced by 200 pl DMSO, and formazan crystals were dissolved by agitation of the plate for 1 min. Absorption was measured with a microplate reader at 540 nm (Biorad, Veenendaal, The Netherlands) and cytotoxicity was calculated as described by Dong et al. (15).

DNA Isolation
After incubation for 2 hr, cells of the coincubations were gently scraped from the culture flasks with a disposable policeman (Greiner, Germany) and spun at 40C (5 min/400 xg). Cell pellets were lysed overnight at 37°C in 5 ml NEP/SDS (75 mM NaCl, 25 mM EDTA, 50 pg/ml proteinase K, 1% SDS), and cellular DNA was isolated as described previously (16). After precipitation of the DNA, sample concentrations were assayed spectrophotometrically. DNA was dissolved in 2 mM Tris (pH = 7.4) and stored at -20°C until analysis. Of each DNA sample, 12 pg was used for 32P-postlabeling; the remaining DNA was used for high performance liquid chromatography/ electrochemical detection (HPLC/ECD) analysis ofoxidative DNA damage.

HPLC/ECD Analysis of 8-oxodG
Oxidative damage in DNA was evaluated as described previously (17), according to Floyd et al. (18), and was expressed as the ratio of 8-oxodG to deoxyguanosine (dG). The lower limit of detection was 40 fmol absolute for 8-oxodG, or 1.5 residue/106 dG, which required a minimum yield of 35 Figure 2). This putative BPDE-DNA  adduct has been quantitated and will be referred to as the BPDE-DNA adduct. BPDE-DNA adduct formation is shown in Table 1. No dose-dependent increase in adducts was observed in RLE cells incubated with B[a]P for 2 hr. Incubation with B[a]P-7,8-diol for 2 hr resulted in higher adduct levels compared to equimolar concentrations of B[a]P, and the increase in BPDE-DNA adduct levels was also dose dependent (Spearman r= 0.96, p < 0.05). No adducts were detected by 32P-postlabeling in RLE cells incubated with B[a]P-4,5diol. PMA had no effect on BPDE-DNA adduct formation in RLE cells exposed to 0.5 pM B[a]P, i.e., 2.6 ± 1.6 adducts/108 nt (without PMA) versus 2.7 ± 1.7 adducts/108 nt (with PMA).
Background oxidative DNA damage, as measured by 8-oxodG/dG, was readily detectable in RLE cells. A 2-hr exposure to 10 pM H202 resulted in a slight but not significant increase of the 8-oxodG/dG ratio (142% compared to the control incubations). Figure 3 shows the calculated BPDE-DNA adduct levels in the coincubations exposed to B[a]P at different ratios of PMN to RLE

cells. RLE cells and PMN
were not separated before DNA was isolated. The adduct levels shown in Figure 3 were corrected for the different number of PMN; previous experiments demonstrated that in PMN exposed to B[a]P, adduct levels are negligible. Figure 3 shows that using unstimulated PMN, no enhanced adduct formation was observed in RLE cells with an increasing number of PMN. However, using activated PMN, adduct formation increased with an increasing number of PMN. As illustrated in Figure 3, PMN activation resulted in a marked enhancement of adducts. An even stronger enhancement of adducts in coincubations was observed by PMN activation of B[a]P-7,8-diol (Table 2). Typical chromatographs of these experiments are shown in Figure 2. No BPDE-DNA adducts were detected in PMN or in coincubations exposed to B[a]P-4,5-diol. Activation of the neutrophils' oxidative burst with PMA also resulted in significantly higher 8-oxodG/dG ratios Within their own Environmental Health Perspectives -Vol 105, Supplement 5 * September 1997 DNA (i.e., 188.5%, n=3 experiments). In unstimulated PMN, B[a]P was able to enhance oxidative DNA damage (118.5%, n = 3). However, in coincubations, no significant effects of PMA or B[a]P or its diols on 8-oxodG/dG were observed (data not shown).
In Figure 4, the effect of various radical scavengers on BPDE-DNA adduct formation in coincubations exposed for 2 hr to B[a]P-7,8-diol is shown. In comparison to the control incubation in HBSS, incubation in culture medium supplemented with FCS reduced BPDE-DNA adducts similar to catalase-supplemented HBSS. In the presence of both catalase and SOD, BPDE-DNA adduct levels further decreased, while mannitol or SOD alone had no scavenging effect. The MTT assay showed no cytotoxicity of SOD and/or catalase at the concentrations used (< 5%). However, 100 mM mannitol reduced the viability of RLE cells by 27%. Discussion In this study we showed that activated neutrophils are able to modulate biotransformation of B[a]P and DNA binding of its crucial genotoxic metabolite BPDE in an in vitro coincubation model of neutrophils and lung epithelial cells. In coincubations exposed to B[a]P and stimulated with PMA, BPDE-DNA adduct levels appeared to be correlated with the number of PMN (Figure 2). These results suggest that during in vivo exposure to B[a]P, the formation of BPDE-DNA adducts in the lung might be related to the extent of inflammation i.e., the percentage of PMN. Since ultrafine particles induce rapid PMN influx after inhalation, these data might be of significance to the interpretation of health effects in environmental exposure to particulates (7). Our results are probably not caused by a direct effect of PMA on RLE cells, as PMA concentration was equal in the three groups and further experiments showed that stimulation of RLE cells (without PMN) with PMA did not increase the level of BPDE-DNA adducts. Our data also indicate that PMN are unable to metabolize B[a]P to B[a]P-7,8diol, probably as a consequence of their negligible CYP450 activity. No adducts were detected by 32P-postlabeling on incubation with B[a]P-4,5-diol in the coincubations, or in either RLE cells or PMN alone, which agrees with the observation that this diol cannot form DNA adducts (19). Exposure to B[a]P-7,8-diol, however, resulted in high adduct levels in PMN, and stimulation of PMN with PMA further increased these levels both in PMN and in coincubations. These observations might be explained by the fact that B[a]P-7,8-diol can be metabolized to BPDE by several metabolic routes other than CYP450, including myeloperoxidase (MPO), lipidperoxidation products (20,21), and probably also by ROS. During activation of PMN (e.g., by PMA), excessive formation of these products occurs. However, our results do not elucidate whether the metabolic conversion of B[a]P-7,8-diol occurs extracellularly by ROS or MPO or intracellularly by lipid peroxidation. In coincubations adduct levels were even higher, both in the absence and presence of PMA, in comparison to PMN (Table 2). This is most likely the consequence of a higher metabolic capacity (including CYP450 activity) in RLE cells. The increase in adduct levels by PMA in coincubations may also result from the extracellular actions of MPO and/or ROS released by PMN. Furthermore, ROS released by activated PMN may cause lipid peroxidation of RLE membranes, which leads to intracellular peroxyradicals, which are known to activate B[a]P-7,8-diol in vitro as well as in vivo (22). Addition of scavengers demonstrated the role of ROS in DNA adduct formation of BPDE, because catalase, as well catalase plus SOD, but not SOD alone, were able to reduce BPDE-DNA adduct formation. Comparison of incubations in HBSS versus culture medium showed that the effects observed in coincubation models occur even in the presence of the scavenging effects of fetal serum. Thus, the effects observed in coincubations would probably be much stronger if HBSS was used, but they indicate that the effects we observed in vitro are also more likely to occur in vivo despite scavenging action of antioxidants. Our results are in agreement with data of Kensler and co-workers, who reported increased DNA adduct formation in mouse keratinocytes exposed to [3H]-B[a]P-7,8-diol in the presence of TPA-activated PMN (10). This DNA binding may be MPO dependent, as supported by the ability of azide to reduce this binding (5,8).
Parallel analysis of oxidative DNA damage showed that stimulation by phorbol ester caused enhanced ratios of 8-oxodG/dG in PMN, which is in agreement with previous observations (18,23) but in contrast to a recent study (24). In all experiments 8-oxodG/dG ratios were higher in PMN than in coincubations, which indicates that PMN are more susceptible to oxidative damage than RLE cells. This can be explained by the high resistance of RLE cells to H202, and the high local concentrations of ROS in or around PMN. Exposure to B [a]P in the absence of PMA resulted in increased 8-oxodG/dG ratios in PMN but not coincubations. This oxidative effect of B[a]P has been reported by Mauthe et al. (25), and might be explained by one-electron oxidation of B[a]P (21). However, this process is negligible compared to the generation of ROS by activated PMN.
In conclusion, this coincubation model is a suitable in vitro system to study mechanisms of PMN-mediated DNA damage in lung target cells. The results of our experiments clearly show that the genotoxic action of B[a]P toward lung epithelial cells is not caused solely by classic biotransformation processes, but also involves ROS produced by PMN. Moreover, if applicable to particle-associated PAHs our data might promote testing and evaluation of many environmental airborne particulates with regard to genotoxic outcomes in the lung. Experiments are currently in progress to elucidate the responsible ROS involved in this model. However, experiments are needed to investigate whether the dose-response observations regarding number of PMN can be confirmed in an in vivo model of lung inflammation and related genotoxicity.