Nitric oxide–mediated inhibition of DNA repair potentiates oxidative DNA damage in cholangiocytes

Nitric oxide–mediated inhibition of DNA repair potentiates oxidative DNA damage in cholangiocytes

GASTROENTEROLOGY 2001;120:190 –199 Nitric Oxide–Mediated Inhibition of DNA Repair Potentiates Oxidative DNA Damage in Cholangiocytes MEETA JAISWAL,* ...

1MB Sizes 3 Downloads 20 Views

GASTROENTEROLOGY 2001;120:190 –199

Nitric Oxide–Mediated Inhibition of DNA Repair Potentiates Oxidative DNA Damage in Cholangiocytes MEETA JAISWAL,* NICHOLAS F. LARUSSO,* RICHARD A. SHAPIRO,‡ TIMOTHY R. BILLIAR,‡ and GREGORY J. GORES* *Center for Basic Research in Digestive Diseases, Division of Gastroenterology and Hepatology, Mayo Medical School, Clinic, and Foundation, Rochester, Minnesota; and ‡Department of Surgery, University of Pittsburgh, Pittsburgh, Pennsylvania

Background & Aims: Chronic inflammation, a risk factor for the development of bile duct cancer, induces inducible nitric oxide synthase (iNOS) with nitric oxide (NO) generation, which promotes oxidative damage of DNA, a process that probably is important in the initiation and progression of malignancies. Because inhibition of DNA repair is required for accumulation of oxidative DNA lesions, our aim was to determine if NO also inhibits repair of oxidative DNA damage. Methods: A cholangiocarcinoma cell line and a cholangiocyte cell line were transfected with iNOS. Results: Extracts from transfected but not untransfected cells were unable to repair 8-oxodeoxyguanine (8-oxodG); this effect was irreversible because addition of dithiothreitol to cell extracts had no effect. NO inhibition of 8-oxodG repair was blocked by NO scavengers but not by peroxynitrite scavengers or inhibitors of the soluble guanylyl cyclase/ protein kinase G pathway. NO also potentiated hydrogen peroxide–induced DNA damage. Finally, immunohistochemistry in human liver samples uniformly demonstrated de novo expression of iNOS and the presence of 3-nitrotyrosine and 8-oxodG formation in the biliary epithelia of 30 patients with primary sclerosing cholangitis (a premalignant disease of the biliary tract) compared with controls. Conclusions: Collectively, these data implicate NO-mediated inhibition of 8-oxodG base excision DNA repair processes as a mechanism potentiating DNA damage in human inflammatory diseases involving the biliary tract.

hronic inflammation and the associated oxidative stress are carcinogenic.1 This is especially true in the biliary tract of the liver, where a wide variety of chronic inflammatory liver diseases are known to place patients at high risk for development of cholangiocarcinoma, an adenocarcinoma arising from the epithelial cells lining the bile ducts known (i.e., cholangiocytes).2 For example, cholangiocarcinoma develops in up to 30% of patients with primary sclerosing cholangitis (PSC), an inflammatory disease of the intrahepatic and extrahepatic bile ducts.3 Despite this clinical association, the cellular


mechanisms linking chronic inflammation to malignant transformation remain obscure. Nitric oxide (NO) is generated from L-arginine in inflamed tissue by inducible nitric oxide synthase (iNOS). Increased concentrations of NO have been implicated in cancer development and progression as an endogenous mutagen, an angiogenesis factor, and/or an inhibitor of apoptosis.4,5 Indeed, several human gastrointestinal neoplasms, including cholangiocarcinoma, express iNOS.6,7 These observations suggest that iNOS with NO production plays a fundamental role in the initiation and promotion/progression of cancers arising from inflamed tissues such as cholangiocarcinoma. The mutagenic aspects of NO may be especially important in cancer development and promotion.8,9 Indeed, NO has been shown to react directly with DNA, causing base deamination, nitration, and oxidation.10 However, the pathogenic significance of these biochemical observations is dependent in part on the efficiency of the DNA repair processes of the cell. If these premutagenic base lesions are repaired before mutations occur, they probably will be biologically inconsequential. We have recently demonstrated that cytokine-mediated induction of iNOS with NO production is associated with a diminished global DNA repair capacity in cholangiocarcinoma cell lines.11 These data suggest that NO may potentiate the accumulation of premutagenic lesions not only by damaging DNA, but also by inhibiting DNA repair enzymes. Thus, the inhibitory effect of NO on DNA repair deserves further attention as a mechanism linking chronic inflammation to cancer. Abbreviations used in this paper: BER, base excision repair; DTT, dithiothreitol; iNOS, inducible nitric oxide synthase; L-NMMA, NG-nitroD-glucamine; NER, nucleotide excision repair; NO, nitric oxide; 8-oxodG, 7-hydro-8-oxodeoxyguanosine; PKG, protein kinase G; PSC, primary sclerosing cholangitis. © 2001 by the American Gastroenterological Association 0016-5085/01/$10.00 doi:10.1053/gast.2001.20875

January 2001

Detection, recognition, and repair of premutagenic DNA lesions is performed by several distinct, if not completely independent, major repair pathways: (1) base excision repair (BER), (2) nucleotide excision repair (NER), and (3) mismatch repair.12,13 Although DNA oxidative lesions, the predominant mechanism of DNA damage in inflammation, can be excised from DNA by multiple pathways, quantitatively BER is the most important.14 For example, the most abundant oxidative DNA lesion, 8-oxo-deoxyguanine (8-oxodG), is recognized and excised by BER glycosylases.14 A human enzyme repairing 8-oxodG, 8-oxo-deoxyguanine DNA glycosylase 1 (hOGG1), has been identified and cloned.15 This glycosylase contains critical thiol moieties and therefore is potentially susceptible to inactivation by nitrosylation reactions (formation of S–NO bonds).15,16 NO may affect BER and other DNA repair processes not only by nitrosylation reactions but also potentially via kinase signaling cascades. NO activates soluble guanylyl cyclase, generating guanosine 3⬘,5⬘-cyclic monophosphate (cGMP), which in turn stimulates protein kinase G (PKG) activity. Indeed, kinase cascades have been shown to modulate DNA repair processes by phosphorylating DNA repair enzymes.17,18 The potential mechanisms by which NO inhibits DNA repair (nitrosylation vs. kinasemediated activities) remain unknown. The overall objective of this study was to determine directly the effect of NO on 8-oxodG DNA repair processes. We used a human cholangiocyte cell line and a cholangiocarcinoma cell line for the studies because biliary epithelia are the targets of inflammation in a wide variety of human diseases. Our results show that cell extracts from iNOS-transfected cells were unable to repair 8-oxodG, and NO also potentiated hydrogen peroxide–induced DNA damage in cells. These data suggest that NO-mediated inhibition of 8-oxodG DNA BER processes is a mechanism potentiating DNA damage in human inflammatory diseases involving the biliary tract.

Materials and Methods Cell Culture The human cholangiocyte cell lines H69, an SV40 T-antigen–transformed cell line derived from normal bile duct epithelia, and KMBC, a cell line derived from human cholangiocarcinoma, were used for these studies.19 The cell lines were cultured in Dulbecco modified Eagle medium (Life Technologies, Gaithersburg, MD) supplemented with 2 mmol/L Lglutamine, 100 U/mL penicillin, 100 mg/mL streptomycin, and 5% fetal bovine serum and maintained at 37°C with 5% CO2 and 95% humidity.



Comet Assay The assay was performed as described in a protocol from Trevigen (Gaithersburg, MD), with slight modification. H69 and KMBC cells were suspended in 0.5% (wt/vol) solution of low melt agarose in phosphate-buffered saline (PBS; pH 7.4) at 37°C and fixed onto a frosted microscope slide. The agarose-embedded cells were covered with Fpg DNA glycosylase (0.1 U/gel; Trevigen), a bacterial enzyme that recognizes and cleaves 8-oxodG lesions on DNA, and incubated at 37°C for 1 hour. Cells were electrophoresed and stained with SYBR green according to the manufacturer’s directions. Statistical evaluation was performed using National Institutes of Health image (Netscape Navigator) and Komet 3.0 Macro software.

Transfection of Cell Lines With iNOS The human iNOS construct was prepared by ligating iNOS cDNA (4013 base pairs [bp]), cloned from cytokinestimulated human hepatocytes, into the expression vector pCIS 5370. Transfection of the cell lines was carried out using a 9:1 (wt/wt) ratio of lipid 2 (Invitrogen Corp., Carlsbad, CA) to plasmid DNA and Lipofectamine (Life Technologies) at a 4:1 (wt/wt) ratio of lipid to plasmid, preincubated for 10 and 30 minutes at room temperature, respectively. The lipid–DNA complexes were overlaid on 70% confluent H69 and KMBC cells and incubated at 37°C with 5% CO2 and 95% humidity in serum-free media for 12 and 24 hours, respectively. After transfection, the cells were incubated for an additional 24 hours with complete growth media. 1400W (50 ␮mol/L), an iNOS inhibitor, was added to the media to inhibit NO synthesis until initiation of the desired experiment. The transfection efficiency was determined by transfecting cells with pEGFP-CI (Clonetech, Palo Alto, CA) and assessing the percent of green fluorescent protein (GFP)-expressing cells by fluorescence microscopy; the transfection efficiency was approximately 40%.

Western Blot Analysis Cells were harvested by trypsinization and lysed by sonication in ice-cold lysis buffer containing100 mmol/L TrisHCl (pH7.5), 0.5 mmol/L EDTA, 0.5 mmol/L EGTA, 2 mmol/L dithiothreitol (DTT) protease inhibitors (5 mg/mL leupeptin, pepstatin, and chymostatin and 87 mg/mL phenylmethylsulfonyl fluoride), 5 mmol/L H4B, 10 mmol/L flavin mononucleotide and 10 mmol/L flavin adenine dinucleotide (Sigma, St. Louis, MO). Whole-cell lysates were boiled in Laemmli buffer (200 mmol/L Tris-HCl [pH, 8], 8% sodium dodecyl sulfate, 1% bromophenol blue, and 400 mmol/L DTT). Protein samples (40 mg/lane) were loaded on a 7.5% sodium dodecyl sulfate–polyacrylamide gel and separated electrophoretically. The proteins were transferred to a nitrocellulose membrane (Schleicher & Schuell, Keene, NH) overnight at 90 mA in a Bio-Rad TransBlot cell. The membrane was blocked with 5% nonfat dried milk in TTBS (Tris 20 mmol/L, 0.05% Tween, and 0.5 mol/L NaCl, pH 7) for 1 hour. The primary antibody for iNOS (Transduction Laboratories, Kensington, KY) was applied at a 1:2500 dilution for 2 hours. The



membrane was washed 3 times in TTBS for 10 minutes each before application of the secondary antibody (Transduction Laboratories) at a 1:5000 dilution for 1 hour. The blot was washed in TTBS 4 times for 10 minutes each. It was then incubated in commercial enhanced chemiluminescence reagent (Amersham, Buckinghamshire, England) and exposed to photographic film.

Chemiluminescence Assay for NO Production NO was measured in the culture media using an NO analyzer (NOA; Sievers, Boulder, CO). Nitrite and nitrate present in the culture medium (100 mL) was converted to NO by a saturated solution of VCl3 in 0.8 mol/L HCl and the NO detected by a gas-phase chemiluminescent reaction between NO and ozone. Nitrite and nitrate concentrations were determined by interpolation from known standards.

Global DNA Repair Incorporation Assay DNA repair was assessed by determining the ability of cell extracts to incorporate radiolabeled nucleotide into oxidatively damaged plasmid DNA, as previously described in detail.20 Briefly, the pBluescript II KS(⫹) 2961-bp [pKS(⫹)] and the pKB-CMV 4518-bp plasmid substrates were prepared by lysozyme–Triton X-100 method.20 Oxidatively damaged pKS II(⫹) plasmid was prepared by treating 0.1 mg/mL DNA with 10 mmol/L methylene blue in 0.01 mol/L sodium phosphate buffer (pH 7.4) and exposed to visible light at a fluence of 117 W/m2 from a 100-W tungsten bulb for 3 minutes at 4°C. Oxidatively damaged pKS II(⫹) and pKB-CMV (control) plasmids were further purified on cesium chloride– ethidium bromide gradient centrifugation steps and one sucrose gradient step. Wholecell extracts were prepared and incubated with the plasmid substrates plus 50 mmol/L each of deoxyadenosine triphosphate, deoxycytidine triphosphate, and deoxythymidine triphosphate and 5 mmol/L deoxyguanosine triphosphate (Sigma) and 1 mCi of [␣-32P]deoxyguanosine triphosphate (NEN Life Science Products, Boston, MA), as described previously in detail.11,20 Plasmid DNA was recovered by phenolchloroform extraction and ethanol precipitation. Plasmids were linearized with EcoRI endonuclease and separated by 1% agarose gel electrophoresis containing 0.5 mg/mL ethidium bromide. The gel was scanned to measure the intensity of ethidium bromide fluorescence (MultiImage light Cabinet; Ionnotech Corp., San Leandro, CA). Fluorescence intensities of linearized plasmid bands in comparison with known amounts of DNA were used to determine DNA loading. The gel was dried under vacuum and exposed to a storage screen (PhosphoImager; Molecular Dynamics, Sunnyvale, CA) for 12 hours. Autoradiograms were analyzed using ImageQuant software (Molecular Dynamics). Background corrected values of the radioactivity incorporated into the plasmids was normalized for the amount of DNA.


Specific DNA Glycosylase Repair Assay Glycosylase activity was assayed by determining the ability of whole-cell extracts to recognize and excise 8-oxodG. A 24-base DNA oligonucleotide, 5⬘-GAACTAGTG8oxoGATCCCCCGGGCTGC–3⬘ containing an 8-oxodG oxidative lesion at position 10 from the 5⬘ end, and its complement were used as substrates. Fifty picomoles of the 24-bp oligonucleotide substrate with the 8-oxodG lesion was radiolabeled with 4 mmol/L [␣-32P]ATP (6000 mCi/mol; NEN Life Science Products) using polynucleotide kinase and incubated for 45 minutes at 37°C. The specific repair glycosylase activity was assayed by incubating 0.5 pmol of radiolabeled olignucleotide with 1.5 pmol of its complement in 10 mmol/L Tris (pH 7.5), 1 mmol/L EDTA, and 50 mmol/L NaCl at 37°C for 2 hours. The reaction was loaded onto a 15% acrylamide/7 mol/L urea/1⫻ TBE gel. The gel was exposed to a storage screen (PhosphoImager; Molecular Dynamics) for 10 minutes. Autoradiograms were analyzed using ImageQuant software (Molecular Dynamics).

Immunohistochemistry Paraffin-embedded normal human liver tissue samples and liver tissue samples obtained from patients with PSC and alcohol-induced cirrhosis were used for iNOS (15 mg/mL, 1:500 dilution; Transduction Laboratories), 8-oxodG (10 mg/ mL, 1:200 dilution; Trevigen), and 3-nitrotyrosine (5 mg/mL, 1:450 dilution; Upstate Biotechnology, Lake Placid, NY) immunohistochemistry. Tissue sections were processed and stained as previously described in detail.11

Reagents Reagents were obtained from the following sources: NG-nitro-D-glucamine (L-NMMA) L-NIL, and 1400W from Cayman Chemicals (Ann Arbor, MI); ODQ, KT5823, dibutyryl cGMP, MnTBAP, heme, PTIO, and C-PTIO from Calbiochem (San Diego, CA); and trolox and ebselen from Sigma. To reduce the residual methemoglobin in commercial hemoglobin to oxyhemoglobin, 20 mg of bovine erythrocyte hemoglobin was dissolved in 1 mL of phosphate-buffered saline containing 2.2 mg of sodium dithionate and purified on a Sephadex G-25 column.

Results Does In Vitro Gene Transfection Produce Functional Expression of iNOS? iNOS protein and NO generation were assessed 48 hours after transfection of H69 and KMBC cell lines with the iNOS expression vector. Transfected H69 and KMBC cells both expressed the 130-kilodalton iNOS protein (Figure 1A). In subsequent studies (see below), several pharmacologic reagents were used. To assess the potential effects of these reagents on iNOS transfection and protein expression, we performed immunoblots for

January 2001



mmol/L after iNOS transfections and 30 ⫾ 4.3 mmol/L after cytokine stimulation).11 To confirm if iNOS was the source of NO production, the cells were incubated with 2 specific iNOS inhibitors, 1400W and L-NIL, and a nonspecific NOS inhibitor, L-NMMA (Figure 1B). All 3 inhibitors blocked NO production in the transfected H69 and KMBC cells, showing that iNOS was the source of the NO released. Thus, the expressed iNOS was catalytically active. Does iNOS-Generated NO Inhibit DNA Repair? To test the direct effect of NO on the global DNA repair capacity of transfected H69 and KMBC cells, we used an in vitro DNA repair assay. Repair activity in

Figure 1. Expression of iNOS and NO generation after transfection are not affected by inhibitors of guanylyl cyclase or PKG. (A ) Cells were transfected with an expression vector for iNOS or the empty plasmid. Twenty-four hours after transfection, iNOS protein expression was assessed by immunoblot analysis. iNOS expression was not affected by the presence of 40 mmol/L ODQ, 0.5 mmol/L KT5823, 50 mmol/L 1400W, 30 mmol/L L-NIL, or 30 mmol/L L-NMMA. (B) Nitrite and nitrate (NO3⫺/NO2⫺) levels in the media measured by a chemiluminescence assay 24 hours after transfection. Each column represents the mean ⫾ SEM of 3 different experiments.

iNOS protein after transfection and cell culture in the presence of these reagents (Figure 1A). iNOS protein expression after transfection was not inhibited in the presence of ODQ, the soluble guanylyl cyclase inhibitor, KT5823, the protein kinase G inhibitor, MnTBAP, peroxynitrite scavenger, or DTT (Figure 1A). iNOS protein expression was similarly unaffected by the presence of dibutyryl cGMP and the other peroxynitrite and NO scavengers used in the study (data not shown). To determine if the expressed iNOS protein was functionally active, we measured metabolites of NO, nitrites, and nitrates in the media of H69 and KMBC cells transfected with iNOS. There was a 9-fold increase in NO3⫺ and NO2⫺ in the media of transfected cells expressing iNOS protein (Figure 1B). The presence of ODQ or KT5823 in the media did not inhibit the generation of NO (Figure 1B). The magnitude of NO generation by KMBC cells after transfection was comparable with that obtained with interleukin 1␤, interferon ␥, and tumor necrosis factor ␣ for 24 hours (50 ⫾ 1.3

Figure 2. iNOS transfection inhibits global and specific 8-oxodG DNA repair in an NO-dependent manner. Cells were transfected with an expression vector for iNOS or an empty plasmid and incubated in the presence or absence of 50 ␮mol/L 1400W. Twenty-four hours later, whole-cell extracts were prepared as described in Materials and Methods. (A ) Global DNA repair was calculated as a percentage of relative repair incorporation of radiolabeled dGMP into photoactivated methylene blue– damaged plasmid DNA by the transfected H69 and KMBC whole-cell extracts. (B) 8-OxodG specific repair. Specific repair activity is indicated by the recognition and excision of the oligonucleotide at the 8-oxodG lesion giving the 15-bp excised product. The bacterial glycosylase Fpg was added as a positive control.




Figure 3. Inhibition of DNA repair is independent of NO-mediated activation of cGMP/PKG pathway. Cells were transfected with an expression vector for iNOS or an empty plasmid and incubated for 24 hours in the presence of 50 mmol/L 1400W to prevent generation of NO before initiation of experimental treatment. (A ) Global repair efficiency was analyzed by in vitro repair incorporation of radiolabeled dGMP into damaged plasmid DNA substrate as described in Figure 1. (B) The cell-permeable cyclic G analogue, 500 mmol/L dibutyryl cGMP, an activator of the PKG pathway, does not result in inhibition of global DNA repair activity. (C ) Specific 8-oxodG repair activity in cell extracts was measured as described in Figure 1. NO-mediated inhibition of 8-oxodG specific DNA repair activity also is not prevented by guanylyl cyclase or PKG inhibitors.

whole-cell extracts from empty vector–transfected cells demonstrated significant DNA repair (Figure 2A). In contrast, there was an 80% decrease in global DNA repair in cells transfected with iNOS for 48 hours. To establish a direct link between NO production and the decrease in global repair efficiency, 1400W, a highly specific inhibitor of iNOS, was added to the culture media of the iNOS-transfected cells. Repair efficiency returned to control values in the presence of this iNOS inhibitor (Figure 2A). Thus, iNOS expression with NO generation is associated with inhibition of global DNA repair. Because BER is a critical mechanism for repair of oxidative DNA lesions, we next determined the effect of NO in this specific repair process. Efficient BER activity for 8-oxodG was observed in nontransfected cells. In contrast, whole-cell extracts from both H69 and KMBC cells transfected with iNOS for 48 hours had no detectable BER activity. Addition of purified Fpg (a BER glycosylase) to the extracts resulted in predictable excision of the substrate, showing that the cell extracts did not contain an inhibitor of BER (Figure 2B). Furthermore, recognition and excision activity were normal in transfected cells incubated with iNOS inhibitor 1400W (Figure 2B). Collectively, these data indicate that BER

for 8-oxodG is markedly diminished in cells expressing functional iNOS. Is the Inhibition of DNA Repair Processes a Response to NO Activation of Soluble Guanylyl Cyclase or NO-Mediated Oxidative Reactions? To determine if NO inhibits DNA repair processes via a guanylyl cyclase–mediated signaling cascade, experiments were carried out in the presence and absence of inhibitors of soluble guanylyl cyclase (ODQ)21 and its downstream target, PKG (KT 5823).22 Neither ODQ nor KT5823 prevented loss of global excision repair or BER after iNOS transfection (Figure 3A and C). Likewise, global excision repair and BER were also unaffected in H69 and KMBC cells incubated for 24 hours in the presence and absence of dibutyryl cGMP, an analogue of cGMP, and an activator of PKG (Figure 3B). These data suggest that NO-activated cGMP-PKG–mediated phosphorylation processes are not involved in the inhibition of DNA repair activity. We next determined whether NO and/or reactive nitrogen oxide species (RNOS; viz. peroxynitrite) generated by iNOS can inhibit repair activity. The compounds carboxy-PTIO,23 PTIO,24 and heme25 were used as NO

January 2001



activity when added to cell extracts from iNOS-transfected cells (Figure 5). Global repair efficiency in wholecell extracts treated with 10 mmol/L DTT increased from 20% to approximately 50%, indicating partial reversal of the nitrosylation process (Figure 5A). However, specific 8-oxodG repair activity was not improved by DTT treatment (Figure 5B). NO-mediated inhibition of 8-oxodG DNA repair appears to be largely an irreversible process. Are 8-oxodG Lesions on DNA Repaired in the Presence of NO?

Figure 4. DNA repair is normal during treatment of iNOS-transfected cells incubated with NO scavengers but not peroxynitrite scavengers. Cells were transfected with an iNOS expression vector or an empty plasmid. Twenty-four hours after transfection, whole-cell extracts were obtained, and (A ) global DNA repair and (B) 8-oxodG–specific DNA repair were measured as described in Figure 1. The NO scavengers (100 mmol/L carboxy-PTIO, 25 mmol/L heme, and 50 mmol/L PTIO) and the peroxynitrite scavengers (40 mmol/L MnTBAP, 10 mmol/L trolox, and 25 mmol/L ebselen) were included in the media throughout the 24-hour incubation.

The alkaline comet assay combined with specific repair enzymes is an optimized tool for the detection of double- and single-stranded breaks and specific DNA lesions at the single cell level. Formamidipyrimidine DNA glycosylases (Fpg), an Escherichia coli–specific repair enzyme that recognizes and cleaves 8-oxodG lesions on DNA, was used to maximize identification of 8-oxodG lesions. KMBC cells were treated with SNAP, a pharmacologic NO donor that releases NO in a stable manner (Figure 6E). 8-oxodG lesions in KMBC cell line (control)

scavengers, and MnTBAP,26 trolox,27 and ebselen28 were used as peroxynitrite scavengers. Treatment of iNOSexpressing cells with the NO scavengers carboxy-PTIO, PTIO, and heme completely prevented loss of global excision repair and BER after iNOS transfection (Figure 4A and B). In contrast, the peroxynitrite scavengers were only able to increase global repair modestly and were completely ineffective in protecting against loss of BER (Figure 4A and B). These data suggest that NO but not peroxynitrite is the principal species responsible for inhibition of DNA repair processes. Is NO-Mediated Inhibition of DNA Repair Reversible? NO is known to nitrosylate cysteine residues, a reaction reversed by the strong reducing agent DTT.29 When DTT was added to the media of iNOS-transfected cells, there was no impairment in either global or specific repair efficiency (Figure 5A and B). However, this effect could be explained by DTT functioning as an NO scavenger rather than as a reductant of –S–NO bonds. Therefore, we determined if DTT restored DNA repair enzyme

Figure 5. DTT prevents loss of DNA repair as a media adjunct but does not restore global or 8-oxodG–specific DNA repair in whole-cell extracts from iNOS-transfected cells. Cells were transfected with an expression vector for iNOS or an empty plasmid. Twenty-four hours later, whole-cell extracts were prepared, and (A ) global and (B) 8oxodG DNA repair were quantitated as described in Figure 1. When included in the media after transfection, DTT prevented loss of DNA repair capability, presumably by acting as an NO scavenger. However, when added only to cell extracts, DTT did not restore either global or specific DNA repair activity.

Figure 6. 8-OxodG DNA lesions are not repaired in the presence of NO. 8-OxodG lesions were quantitated by the comet assay. The percentage of DNA in 50 comet tails was analyzed using computer analysis of cells from the following experimental groups: (1) untreated cells (A and control); (2) cells treated with 150 mmol/L H2O2 for 1 hour (B and H2O2) and subjected immediately to the comet assay; (3) cells treated with 150 mmol/L H2O2 for 1 hour followed by incubation in H2O2-free media for 24 hours (C and H2O2 after 24 hours); (4) cells treated with 150 mmol/L H2O2 for 1 hour followed by incubation with 0.3 mmol/L SNAP for 24 hours (D and H2O2 ⫹ SNAP after 24 hours); (5) cells treated with 0.3 mmol/L SNAP for 24 hours (E ); and (6) cells treated with 0.3 mmol/L SNAP plus the NO scavenger 100 mmol/L c-PTIO for 24 hours (F ). *P ⬍ 0.01 compared with all other treatment as determined using ANOVA.

Figure 7. Immunohistochemistry for iNOS, 3-nitrotyrosine, and 8-oxodG in human liver specimens. Top panel: The biliary epithelia (arrows) did not contain immunoreactivity for (A ) iNOS, (D) 3-nitrotyrosine, or (G) 8-oxodG in normal liver tissue. Middle panels: Intense immunoreactivity was observed in biliary epithelia (arrows) in liver specimens from patients with PSC for (B) iNOS, (E ) 3-nitrotyrosine, and (H ) 8-oxodG. Bottom panel: Immunoreactivity for (C ) iNOS, (F ) 3-nitrotyrosine, and (I ) 8-oxodG was minimal in the biliary epithelia of patients with alcohol-induced cirrhosis.

January 2001

were detected at basal values (1.5%), as indicated by tight nucleoids (Figure 6A). Cells treated with hydrogen peroxide showed significant oxidative DNA damage (Figure 6B); however, this damage was completely repaired 24 hours after hydrogen peroxide treatment (Figure 6C). Cells with hydrogen peroxide followed by exposure to SNAP were unable to repair the oxidative DNA damage (Figure 6D). Incubation of the cells with SNAP and carboxy-PTIO, an NO scavenger, permitted repair of the DNA damage, showing that SNAP was inhibiting repair by an NO-mediated mechanism (Figure 6F). These data show that NO-mediated inhibition of DNA repair is of sufficient magnitude to inhibit repair of oxidatively damaged DNA in cells. Is iNOS Expression Associated With Oxidative DNA Damage in the Biliary Epithelia of Patients With an Inflammatory Cholangiopathy? Our data with isolated cell systems predicted that iNOS-expressing cholangiocytes in inflammatory human liver diseases should contain oxidatively damaged DNA. Therefore, we performed immunohistochemistry for iNOS in tissue specimens from 30 patients with PSC, a chronic inflammatory premalignant human liver disease targeting the biliary tract. Intense immunoreactivity for iNOS was identified in the bile ducts in all 30 specimens (Figure 7B). All of the liver specimens also contained immunoreactivity for 3-nitrotyrosine, a reaction product of NO-derived ONOO⫺ with tyrosine residues (Figure 7E). Immunostaining of 8-oxodG, a marker for oxidative DNA damage, was identified in the biliary epithelia of all 10 liver specimens studied from patients with PSC (Figure 7H ). In contrast, biliary epithelia from normal liver biopsy specimens did not contain immunoreactivity for iNOS nor 3-nitrotyrosine or 8-oxodG (Figure 7A, D, and G). Liver specimens from patients with alcoholinduced cirrhosis were used as a disease control. We identified no immunoreactivity for iNOS or 3-nitrotyrosine in these human liver specimens and only faint immunoreactivity for 8-oxodG (Figure 7C, F, and I ). These data directly demonstrate an association between iNOS expression and oxidative DNA damage in a human liver disease.

Discussion The results of our study relate to NO-mediated genotoxicity. Our data directly demonstrate that (1) iNOS expression with NO generation inhibits both global and 8-oxodG base excision DNA repair processes; (2) NO scavengers, but not guanylyl cyclase or PKG



inhibitors, prevent NO-mediated inhibition of DNA repair; (3) NO-mediated inhibition of DNA repair enhances oxidative DNA damage in cells; and (4) iNOS, 3-nitrotyrosine, and 8-oxodG are present in the bile ducts in liver specimens from patients with PSC but not in normal liver specimens. The last observation shows the relevance of our in vitro observations to a chronic inflammatory, premalignant disease of the biliary tract. Collectively, these mechanistic studies suggest an integral role for NO in promoting DNA damage in inflammatory diseases. We have previously demonstrated that proinflammatory cytokines inhibited global DNA repair processes by an NO-dependent process. The current study significantly extends these observations. Although inflammatory cytokines may alter the redox status of the cell (e.g., tumor necrosis factor ␣–induced oxidative stress), rendering the DNA repair machinery susceptible to oxidation by NO species,30 we were able to directly demonstrate NO-mediated inhibition of DNA repair in the absence of cytokines. Moreover, we were able to show that specific repair of 8-oxodG lesions was impaired by NO. The biological relevance of the observations was highlighted by demonstrating an increase in hydrogen peroxide–induced formation of 8-oxodG lesions in NOtreated cultured cells using the comet assay, an assay with a sensitivity for identifying 8-oxodG lesions equal to that of high-performance liquid chromatography techniques.31 Using an immunohistochemical approach, we also observed an association between iNOS expression and 8-oxodG formation in a human inflammatory biliary tract disease, PSC. These complementary results suggest NO potentiates DNA damage by inhibiting repair of 8-oxodG lesions. If 8-oxodG lesions are not repaired, they promote incorporation of adenine opposite the lesion, resulting in G:C3 T:A transversions.32 Therefore, the failure to repair 8-oxodG is mutagenic33 and would be expected to promote cancer initiation and promotion/progression. Indeed, other studies have implicated loss of oxidative DNA repair glycosylase– dependent BER in human cancers.34,35 Our observations may therefore provide a potential key mechanism linking inflammation to carcinogenesis: NO-dependent inhibition of the BER pathway. We observed that NO inhibition of BER for 8-oxodG was blocked by NO scavengers but not by inhibitors of the guanylyl cyclase/PKG pathway. These data indicate that NO-mediated transduction of soluble guanylyl cyclase does not inhibit DNA repair proteins. Rather, NO-mediated oxidative processes appear to inhibit DNA repair. We anticipated that NO inhibited these repair



proteins by a reversible nitrosylation of their active site cysteine thiol residues. However, addition of the potent reducing agent DTT to the cell extracts did not restore the BER of 8-oxodG substrate. The key repair glycosylases involved in oxidative BER (hOGG1) and NER (e.g., human replication protein A, xeroderma pigmentosum A) have a cysteine 4 –type zinc finger motif, which is necessary for their catalytic function.15,36,37 NO nitrosylation of the cysteine residues results in the irreversible loss/ejection of the zinc ion.38 The loss of tertiary structural integrity of these proteins after removal of the zinc ion is a potential mechanism to explain the irreversible loss of enzyme function. Indeed, a similar mechanism has been shown for the yeast transcription factor Ace1, a copper-binding protein.39 NO-mediated demetallation of Ace1 also results in its rapid proteolytic degradation. A decrease in protein half-life may also be a mechanism for the irreversible loss of 8-oxodG BER by NO. In summary, our observations are germane to the mechanisms linking chronic inflammation to cancer initiation, promotion, and progression. In inflammatory diseases, cytokines stimulate induction of iNOS with high levels of NO generation. Inhibition of DNA repair processes and potentiation of oxidative DNA damage by NO would be expected to increase the rate of mutagenic DNA lesions. The increased rate of mutagenic DNA lesions could result in loss of tumor-suppressor genes and inappropriate activation of growth-promoting oncogenes. These concepts suggest that iNOS inhibition or NO scavengers would have merit as chemoprevention approaches to decrease cancer development in chronic inflammatory diseases of the biliary tract and perhaps also in other organs. Animal models of carcinogenesis and cancer progression in wild-type and iNOS knockout animals will be useful to further test these concepts.

References 1. Ohshima H, Bartsch H. Chronic infections and inflammatory processes as cancer risk factors: possible role of nitric oxide in carcinogenesis. Mutat Res 1994;305:253–264. 2. Roberts SK, Ludwig J, LaRusso NF. The pathobiology of biliary epithelia. Gastroenterology 1997;112:269 –279. 3. Farges O, Malassagne B, Sebagh M, Bismuth H. Primary sclerosing cholangitis: liver transplantation or biliary surgery. Surgery 1995;117:146 –155. 4. Facchetti F, Vermi W, Fiorentini S, Chilosi M, Caruso A, Duse M, Notarangelo LD, Badolato R. Expression of inducible nitric oxide synthase in human granulomas and histiocytic reactions. Am J Pathol 1999;154:145–152. 5. Li J, Billiar TR, Talanian RV, Kim YM. Nitric oxide reversibly inhibits seven members of the caspase family via S-nitrosylation. Biochem Biophys Res Commun 1997;240:419 – 424. 6. Singer II, Kawka DW, Scott S, Weidner JR, Mumford RA, Riehl TE, Stenson WF. Expression of inducible nitric oxide synthase and



8. 9.

10. 11.


13. 14.












nitrotyrosine in colonic epithelium in inflammatory bowel disease. Gastroenterology 1996;111:871– 885. Mirvish SS. Role of N-nitroso compounds (NOC) and N-nitrosation in etiology of gastric, esophageal, nasopharyngeal and bladder cancer and contribution to cancer of known exposures to NOC [erratum in Cancer Lett 1995;97:271]. Cancer Lett 1995;93:17– 48. Liu RH, Hotchkiss JH. Potential genotoxicity of chronically elevated nitric oxide: a review. Mutat Res 1995;339:73– 89. Thomsen LL, Lawton FG, Knowles RG, Beesley JE, RiverosMoreno V, Moncada S. Nitric oxide synthase activity in human gynecological cancer. Cancer Res 1994;54:1352–1354. Tamir S, Burney S, Tannenbaum SR. DNA damage by nitric oxide. Chem Res Toxicol 1996;9:821– 827. Jaiswal M, LaRusso NF, Burgart LJ, Gores GJ. Inflammatory cytokines induce DNA damage and inhibit DNA repair in cholangiocarcinoma cells by a nitric oxide–dependent mechanism. Cancer Res 2000;60:184 –190. Jaiswal M, Lipinski LJ, Bohr VA, Mazur SJ. Efficient in vitro repair of 7-hydro-8-oxodeoxyguanosine by human cell extracts: involvement of multiple pathways. Nucleic Acids Res 1998;26:21842191. Wood RD. DNA repair in eukaryotes. Annu Rev Biochem 1996; 65:135–167. Laval J, Jurado J, Saparbaev M, Sidorkina O. Antimutagenic role of base-excision repair enzymes upon free radical–induced DNA damage. Mutat Res 1998;402:93–102. Rosenquist TA, Zharkov DO, Grollman AP. Cloning and characterization of a mammalian 8-oxoguanine DNA glycosylase. Proc Natl Acad Sci U S A 1997;94:7429 –7434. Tani M, Shinmura K, Kohno T, Shiroishi T, Wakana S, Kim SR, Nohmi T, Kasai H, Takenoshita S, Nagamachi Y, Yokota J. Genomic structure and chromosomal localization of the mouse Ogg1 gene that is involved in the repair of 8-hydroxyguanine in DNA damage. Mamm Genome 1998;9:32–37. Fritz G, Kaina B. Phosphorylation of the DNA repair protein APE/ REF-1 by CKII affects redox regulation of AP-1. Oncogene 1999; 18:1033–1040. Ariza RR, Keyse SM, Moggs JG, Wood RD. Reversible protein phosphorylation modulates nucleotide excision repair of damaged DNA by human cell extracts. Nucleic Acids Res 1996;24: 433– 440. Vroman B, LaRusso NF. Development and characterization of polarized primary cultures of rat intrahepatic bile duct epithelial cells. Lab Invest 1996;74:303–313. Jaiswal M: Repair of oxidative damage in DNA induced by photoactivated methylene blue in human lymphoblastoid whole cell extracts. Thesis, UMI Co., Ann Arbor, MI, 1997. Brown C, Pan X, Hassid A. Nitric oxide and C-type atrial natriuretic peptide stimulate primary aortic smooth muscle cell migration via a cGMP-dependent mechanism: relationship to microfilament dissociation and altered cell morphology. Circ Res 1999;84:655– 667. Murthy KS, Makhlouf GM. Differential regulation of phospholipase A2 (PLA2)-dependent Ca2⫹ signaling in smooth muscle by cAMP- and cGMP-dependent protein kinases. Inhibitory phosphorylation of PLA2 by cyclic nucleotide-dependent protein kinases. J Biol Chem 1998;273:34519 –34526. Yoshida M, Akaike T, Goto S, Takahashi W, Inadome A, Yono M, Seshita H, Maeda H, Ueda S. Effect of the NO scavenger carboxyptio on endothelium-dependent vasorelaxation of various blood vessels from rabbits. Life Sci 1998;62:203–211. Guidarelli A, Sestili P, Cantoni O. Opposite effects of nitric oxide donors on DNA single strand breakage and cytotoxicity caused by tert-butylhydroperoxide. Br J Pharmacol 1998;123:1311–1316. Wang T, Xie Z, Lu B. Nitric oxide mediates activity-dependent

January 2001








33. 34.


synaptic suppression at developing neuromuscular synapses. Nature 1995;374:262–266. Faulkner KM, Liochev SI, Fridovich I. Stable Mn(III) porphyrins mimic superoxide dismutase in vitro and substitute for it in vivo. J Biol Chem 1994;269:23471. Salgo MG, Pryor WA. Trolox inhibits peroxynitrite-mediated oxidative stress and apoptosis in rat thymocytes. Arch Biochem Biophys 1996;333:482– 488. Arteel GE, Briviba K, Sies H. Function of thioredoxin reductase as a peroxynitrite reductase using selenocystine or ebselen. Chem Res Toxicol 1999;12:264 –269. Rohn TT, Quinn MT. Inhibition of peroxynitrite-mediated tyrosine nitration by a novel pyrrolopyrimidine antioxidant. Eur J Pharmacol 1998;353:329 –336. Laskin DL, Heck DE, Laskin JD. Role of inflammatory cytokines and nitric oxide in hepatic and pulmonary toxicity. Toxicol Lett 1998;102-103:289 –293. Gedik CM, Wood SG, Collins AR. Measuring oxidative damage to DNA; HPLC and the comet assay compared. Free Radic Res 1998;29:609 – 615. Cheng KC, Cahill DS, Kasai H, Nishimura S, Loeb LA. 8-Hydroxyguanine, an abundant form of oxidative DNA damage, causes G3 T and A3 C substitutions. J Biol Chem 1992;267:166 –172. Floyd RA. The role of 8-hydroxyguanine in carcinogenesis. Carcinogenesis 1990;11:1447–1450. Chevillard S, Radicella JP, Levalois C, Lebeau J, Poupon MF, Oudard S, Dutrillaux B, Boiteux S. Mutations in OGG1, a gene involved in the repair of oxidative DNA damage, are found in human lung and kidney tumours. Oncogene 1998;16: 3083–3086. Shinmura K, Kohno T, Kasai H, Koda K, Sugimura H, Yokota J. Infrequent mutations of the hOGG1 gene, that is involved in the


36. 37.




excision of 8-hydroxyguanine in damaged DNA, in human gastric cancer. Jpn J Cancer Res 1998;89:825– 828. Dong J, Park JS, Lee SH. In vitro analysis of the zinc-finger motif in human replication protein A. Biochem J 1999;337:311–317. Morikawa K, Okuyama H, Kikuchi Y. [Action mechanism of adjuvants—a study by fluorescent antibody method]. [Japanese]. Nippon Saikingaku Zasshi 1970;25:427– 428. Laval F, Wink DA, Laval J. A discussion of mechanisms of NO genotoxicity: implication of inhibition of DNA repair proteins. Rev Physiol Biochem Pharmacol 1997;131:175–191. Shinyashiki M, Chiang KT, Switzer CH, Gralla EB, Fukuto M. The interaction of nitric oxide (NO) with the yeast transcription factor Ace1: a model system for NO-protein thiol interactions with implications to metal metabolism. Proc Natl Acad Sci USA 2000; 97:2491–2496.

Received June 1, 2000. Accepted August 2, 2000. Address requests for reprints to: Gregory J. Gores, M.D., Department of Medicine, Mayo Medical School, Clinic, and Foundation, Rochester, Minnesota 55905. e-mail: [email protected]; fax: (507) 2840762. Supported by National Institutes of Health grants 24031 (to N.F.L.), DK41876 (to G.J.G.), and GM44100 (to T.R.B.); by a grant from the American Liver Foundation, Cedar Grove, New Jersey (to M.J.); and by the Mayo Comprehensive Cancer Center, Rochester, Minnesota. The authors gratefully acknowledge the intellectual and technical assistance of Dr. Virginia Miller and Kevin Rud in measuring NO and of Dr. Lawrence Burgart and Randi Carlson in procuring the liver biopsies from the Mayo Tissue Archives; and the secretarial assistance of Sara Erickson.