Abstract

In an effort to define the prevalent DNA damage chemistry-associated chronic inflammation, we have quantified 12 DNA damage products in tissues from the SJL mouse model of nitric oxide (NO) overproduction. Using liquid chromatography–mass spectrometry/MS and immunoblot techniques, we analyzed spleen, liver and kidney from RcsX-stimulated and control mice for the level of the following adducts: the DNA oxidation products 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG), guanidinohydantoin (Gh), oxazolone (Ox); 5-guanidino-4-nitroimidazole (NitroIm); spiroiminodihydantoin (Sp) and M 1 dG; the nitrosative deamination products 2′-deoxyxanthosine, 2′-deoxyoxanosine (dO), 2′-deoxyinosine and 2′-deoxyuridine and the lipid peroxidation-derived adducts 1,N 6 -etheno-deoxyadenosine and 1,N 2 -etheno-deoxyguanosine. The levels of dO, Gh, Ox, NitroIm and Sp were all below a detection limit of ∼1 lesion per 10 7 bases. Whereas there were only modest increases in the spleens of RcsX-treated compared with control mice for the nucleobase deamination products (10–30%) and the DNA oxidation products 8-oxodG (10%) and M 1 dG (50%), there were large (3- to 4-fold) increases in the levels of 1,N 6 -etheno-deoxyadenosine and 1,N 2 -etheno-deoxyguanosine. Similar results were obtained with the liver and with an organ not considered to be a target for inflammation in the SJL mouse, the kidney. This latter observation suggests that oxidative and nitrosative stresses associated with inflammation can affect tissues at a distance from the activated macrophages responsible for NO overproduction during chronic inflammation. These results reveal the complexity of NO chemistry in vivo and support an important role for lipids in the pathophysiology of inflammation.

Introduction

In the past 20 years, a growing body of evidence has emerged that supports a causative role for inflammation in the development of cancer ( 1–4 ). One possible mechanistic link to cancer and other diseases involves infiltration of macrophages and neutrophils into tissues at sites of inflammation with subsequent generation of reactive oxygen and nitrogen species ( 5 ). These species, which include hypohalous acids, peroxynitrite (ONOO ) and nitrous anhydride, have been proposed to cause cytotoxic and mutagenic DNA damage either by direct reaction with DNA or indirectly by generation of DNA adduct-forming electrophiles from reaction with lipids, proteins and carbohydrates, with the resulting DNA lesions presumably representing candidate biomarkers of exposure or disease progression ( 5 ). With the goal of identifying biomarkers of the inflammatory process, we have surveyed a battery of DNA lesions in tissues from the SJL mouse model of nitric oxide (NO) overproduction.

Macrophage-derived NO represents one of the fundamental mediators of the small molecule chemistry of inflammation, with strong evidence for a role in the cytotoxic and mutagenic mechanisms of the carcinogenic processes of chronic inflammation ( 1 , 5 , 6 ). Whereas NO at low levels is essential as an endogenous regulator of the cardiovascular, nervous and immune systems, long-term overproduction of NO and its derivatives by stimulated macrophages may underlie the adverse effects of chronic inflammation. The chemical reactivity of NO in vivo leads to the generation of reactive species capable of both oxidation (and nitration) and nitrosation. The latter is illustrated by the auto-oxidation of NO to form nitrous anhydride (N 2 O 3 ), a potent nitrosating agent that reacts with primary amines and sulfhydryl groups ( 5 ). Alternatively, diffusion-controlled reactions of NO with superoxide lead to the formation of ONOO that subsequently reacts with carbon dioxide to form nitrosoperoxycarbonate (ONOOCO 2 ). Both ONOO and ONOOCO 2 undergo homolytic bond scission to generate NO 2· , a weakly oxidizing species capable of nitration reactions, and either hydroxyl radical or carbonate radical anion, respectively, both of which are potent oxidants capable of reactions with proteins, lipids, carbohydrates and nucleic acids ( 5 ). Whereas these two competing chemistries, oxidation and nitrosation, arise at sites of inflammation, the short half-lives of the reactive nitrogen species lead to difficulties in studying their behavior in vivo . To this end, we were motivated to develop biomarkers as surrogates for reactive nitrogen species to define the roles of these chemistries in inflamed tissues, with a focus on DNA lesions given their potential role in the carcinogenic processes associated with chronic inflammation.

DNA is a potential target for all of the chemistries of inflammation. As shown in Figure 1 , nitrosation of DNA nucleobases by N 2 O 3 leads to the conversion of cytosine to uracil [2′-deoxyuridine (dU)], guanine to either xanthine (2′-deoxyxanthosine [dX]) or oxanine [2′-deoxyoxanosine (dO)] and adenine to hypoxanthine [2′-deoxyinosine (dI)] ( 5 , 7 ). DNA is also subject to oxidation and nitration by reactive nitrogen species, mainly as a consequence of reactions with ONOO and ONOOCO 2 . Whereas ONOO causes mainly deoxyribose oxidation in DNA ( 8 ), the presence of millimolar concentrations of carbon dioxide in tissues leads to the formation of ONOOCO 2 that gives rise to carbonate radical anion that preferentially oxidizes guanine in DNA ( 5 ). The resulting oxidation products include primary lesions such as 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) and 5-guanidino-4-nitroimidazole (NitroIm), as well as secondary products such as the pyrimidopurinone adduct of dG, M 1 dG, that arises from reactions with base propenals derived from deoxyribose oxidation ( 5 ). These DNA lesions that arise from direct attack of oxidizing species on DNA stand in contrast to adducts arising indirectly from reactions of DNA bases with electrophiles derived from primary oxidations of polyunsaturated fatty acids and other cellular molecules ( 9 ). For example, lipid peroxidation induces the formation of a variety of α,β-unsaturated aldehydes such as trans -4-hydroxy-2-nonenal, acrolein and 4-oxo-2-nonenal ( 9 , 10 ). These electrophiles can react with DNA bases to form, among other adducts, substituted and unsubstituted etheno adducts, such as the 1,N 6 -etheno-2′-deoxyadenosine (ϵdA), 3,N 4 -etheno-2′-deoxycytidine (ϵdC) and the two etheno adducts of dG, all shown in Figure 1 ( 11–13 ).

Fig. 1.

Chemistry of DNA damage associated with the reactive oxygen and nitrogen species arising from NO overproduction during inflammation.

Fig. 1.

Chemistry of DNA damage associated with the reactive oxygen and nitrogen species arising from NO overproduction during inflammation.

With implications for biomarker development, one of the key questions involving inflammation-related chemical biology is whether there is a predominant chemistry affecting DNA at sites of inflammation, such as direct nitrosative or oxidative DNA damage by chemical mediators of inflammation or indirect damage in the form of adducts derived from inflammation-induced electrophiles. We approached this problem by quantifying a battery of potential DNA biomarkers of inflammation in the SJL mouse model of inflammation ( 14–16 ). In this model, injection of superantigen-bearing, reticulum cell sarcoma-derived RcsX cells leads to widespread activation of macrophages with subsequent generation of large quantities of NO in spleen, lymph nodes and liver, as reflected by a 30- to 40-fold increase in urinary nitrate excretion at the peak of the reaction (12–14 days) ( 14–17 ). The increased production of NO is associated with an increased mutation frequency ( 15 ) and increases in the tissue levels of nitrotyrosine ( 17 ) and the etheno adducts ϵdA and ϵdC ( 18 ). We have now used this model to quantify the in vivo formation of endogenous DNA adducts that reflect the different chemistries proposed to arise at sites of inflammation: the nucleobase deamination products dU, dX, dO and dI; the oxidation products M 1 dG, 8-oxodG, spiroiminodihydantoin (Sp), guanidinohydantoin (Gh), NitroIm and oxazolone (Ox) and the etheno adducts ϵdA, ϵdC and 1,N 2 -etheno-2′-deoxyguanosine (ϵdG), secondary DNA adducts arising from electrophiles derived from lipid peroxidation.

Materials and methods

Materials

All chemicals and reagents were of the highest purity available and were used without further purification unless noted otherwise. N -Methyl- L -arginine acetate, sodium pyruvate and insulin (bovine) were purchased from Bhem Biochem Research (Salt Lake City, UT). Nuclease P1 and the Genomic DNA Isolation kit were obtained from Roche Diagnostic Corporation (Indianapolis, IN). Phosphodiesterase I was purchased from USB (Cleveland, OH). White potato acid phosphatase, alkaline phosphatase, and desferrioxamine were purchased from Sigma Chemical Company (St Louis, MO). Iscove's modified Dulbecco's medium, fetal calf serum, glutamine and penicillin/streptomycin were purchased from Calbiochem (San Diego, CA). Acetonitrile and high-performance liquid chromatography (HPLC)-grade water were purchased from Mallinckrodt Baker (Phillipsburg, NJ). Water purified through a Milli-Q system (Millipore Corp., Bedford, MA) was used in all of the studies. The anti-M 1 dG antibody was purchased from Oxford Biomedical Research (Oxford, MI).

Synthesis of isotopically labeled internal standards

The syntheses of uniformly 15 N-labeled dX, dU, dI and dO were performed as described elsewhere ( 7 ). Uniformly 13 C- and 15 N-labeled 8-oxodG, 13 C- and 15 N-labeled Ox, Sp, NitoIm, and 1,2,7- 15 N 3 -8-oxodG and 3,7,8- 15 N 3 -Gh were synthesized according to the protocols described elsewhere ( 19 , 20 ). 7- 15 N-dG was a generous gift from Dr T.M.Harris of Vanderbilt University. 15 N-labeled ϵdA, ϵdC and ϵdG were synthesized from the reactions of 15 N-labeled dA, dC and dG, respectively, with chloroacetaldehyde according to published procedures ( 21 , 22 ).

Experiments with SJL mice

RcsX cells (kindly supplied by Dr N.Ponzio, University of New Jersey Medical Center, Newark, NJ) were passaged through SJL mice (Jackson Laboratory, Bar Harbor, ME) and harvested from lymph nodes 14 days after inoculation according to published procedures ( 17 ). Cells were manually dissociated from lymph nodes followed by washing in phosphate-buffered saline (PBS; 140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , pH 7.4) and freezing in aliquots of 5 × 10 7 cells in 10% dimethylsulfoxide/fetal bovine serum. To initiate NO overproduction, eight SJL mice (5–6 weeks old) were each injected intra-peritoneally with 10 7 RcsX cells in 200 μl of PBS, whereas 10 mice were injected with 200 μl of PBS as unstimulated controls. Twelve days after injection, five treated and three control mice were killed and their spleens, livers and kidneys removed, weighed and snap frozen in liquid nitrogen for later DNA isolation. Urinary nitrate excretion was quantified on a daily basis by methods described elsewhere ( 16 ).

DNA isolation from tissues

DNA from spleens, livers and kidneys was isolated using a Roche Genomic DNA isolation kit according to the manufacturer's instructions with additional precautions taken to reduce artifactual oxidative damage during the isolation. Briefly, 400 mg samples of frozen tissue were homogenized in 10 ml of cellular lysis buffer containing a combination of deaminase inhibitors [5 μg/ml coformycin, 50 μg/ml tetrahydrouridine; ( 23 )] and an antioxidant (0.1 mM desferrioxamine) for 20–30 s using a Brinkman Polytron homogenizer on a medium setting. The homogenate was digested with proteinase K (33 μl, 20 mg/ml) at 65°C for 1 h, followed by the addition of DNase-free RNase A (400 μl, 25 mg/ml) for an additional 30 min incubation at 37°C. Proteins were removed by adding 4.2 ml of protein precipitation solution followed by 5 min incubation on ice and centrifugation at 26 900 g for 20 min at 4°C. The supernatant was carefully transferred to a fresh tube and DNA was recovered by precipitation in 200 mM NaCl and 2.5 volumes of 100% ethanol. The floating DNA filament was recovered with a micropipette tip, washed twice with 70% cold ethanol, air-dried at ambient temperature and re-suspended in Milli-Q water. DNA concentration was determined by UV spectroscopy and samples were stored at −80°C until analyzed.

Quantification of DNA deamination products and etheno adducts

Deoxynucleoside forms of deamination products and etheno adducts were quantified using a liquid chromatography–tandem mass spectrometry (LC–MS/MS) variation of a published LC–MS method ( 7 , 23 ). DNA (100 μg) was dissolved in 200 μl of sodium acetate buffer (30 mM, pH 5.6) containing 0.2 mM zinc chloride; 10 pmol labeled 15 N-dX, dI and dU internal standards; 33.3 fmol of 15 N-ϵdA, -ϵdC and -ϵdG standards; 5 μg/ml coformycin; 50 μg/ml tetrahydrouridine and 0.1 mM desferrioxamine. The DNA was digested to deoxynucleoside monophosphates by incubation with 4 U of nuclease P1 at 37°C for 3 h. Following addition of 200 μl of sodium acetate buffer (30 mM, pH 8.1), phosphate groups were removed with 20 U alkaline phosphatase and 1 U of phosphodiesterase I by incubation at 37°C for 6 h. The enzymes were subsequently removed by passing the reaction mixture over a Microcon YM-10 column and were concentrated under vacuum. Deoxynucleoside forms of the various lesions were isolated following HPLC resolution on a Hewlett-Packard model 1100 HPLC equipped with a Phenomenex Synergi C18 reversed-phase column (250 × 4.6 mm, 4 μm particle size, 80 Å pore size; Torrance, CA) and a 1100 A diode array detector eluted with the following gradient of acetonitrile in 8 mM sodium acetate buffer (pH 6.9) at a flow rate of 0.5 ml/min: 0–30 min, 2–10%; 30–43 min, 10–17%; 43–44 min, 100% and 44–53 min, 100%. Individual lesions were isolated by collection of HPLC fractions bracketing empirically determined elution times for each product: dX, 12 min; dU, 17 min; dI, 21 min; dO, 28 min; ϵdG, 34 min and ϵdA, 39 min.

Deoxynucleosides in the HPLC fractions were then analyzed by LC–MS/MS. The deoxynucleoside-containing fractions were concentrated and resolved by reversed-phase HPLC using a Phenomenex C18 column (150 × 1.0 mm, 5 μm; 20 μl injection volume) eluted isocratically at a flow rate of 100 μl/min with a mobile phase consisting of permanganate-treated H 2 O (0.1% acetic acid): acetonitrile (0.1% acetic acid) at 97:3 for dX/dO, dI and dU or at 98:2 for ϵdG, ϵdC and ϵdA. The HPLC column was coupled to an API 3000 tandem quadrupole MS (Applied Biosystems, Foster City, CA) with turbo ion spray used as the ion source and the temperature set at 380°C. The MS was operated in positive ion mode with the First and third quadrupoles (Q1 and Q3, respectively) fixed to unit resolution. The voltages and source gas were optimized for maximal sensitivity, and the parameters were as follows: ion spray source: 4.0 kV; nebulizer gas: 8; curtain gas: 8; collision gas (nitrogen): 4; declustering potential: 20; focusing potential: 100; entrance potential: 5; collision energy: 10 and collision cell exit potential: 10. Multiple reaction monitoring mode was used for detection of deamination samples, with a dwell time to 200 ms. The first quadrupole (Q1) was set to transmit the precursor ions MH + at m/z 273, 257, 231, 297, 255 and 281 for the internal standards U- 15 N-dX/dO, -dI, -dU ϵdG, -ϵdC and -ϵdA, respectively, and at m/z 269, 253, 229, 292, 252 and 276 for unlabeled dX/dO, dI, dU, ϵdG, ϵdC and ϵdA, respectively. The product ions (deglycosylated X/O, I, U, ϵG, ϵC and ϵA) were monitored in the third quadrupole (Q3) at m/z 157, 141, 115, 181, 139 and 165 for U- 15 N internal standards, respectively, and at m/z 153, 137, 113, 176, 136 and 160 for unlabeled bases, respectively.

Quantification of 8-oxodG and its secondary oxidation products

Quantification of dG oxidation products was performed by LC–MS/MS as described elsewhere ( 19 ), with minor changes. Briefly, DNA (117 μg) was digested to the level of 2′-deoxynucleosides by overnight incubation (37°C) with 10 U of nuclease P1 and 1 U of white potato acid phosphatase in a mixture of 150 μl of H 2 O, 65 μl of sodium acetate (1 M, pH 5.1), 50 μl of ZnCl 2 (2 mM) and labeled internal standards. The enzymes were then removed by passing the solution over Microcon YM-10 centrifugal columns and the filtrate was dried under vacuum. Purification of the oxidized nucleosides was carried out by a two-step HPLC method. The first step involved resolution of the mixture on a Zorbax XDB-C18 column [4.6 × 250 mm, 5 μm; mobile phase: water (A) and acetonitrile (B); flow and gradient: 0.8 ml/min, 0–9% B over 36 min]. Sp, Gh and Ox were collected in fractions eluting between 0 and 6 min, whereas NitroIm and 8-oxodG were collected at 9.5–11 min and 23.5–24.5 min, respectively. Following drying under vacuum, samples containing Sp, Gh and Ox were dissolved in water and further purified by HPLC on a Thermo-Keystone Hypercarb column [3.0 × 150 mm, 5 μm; mobile phase: ammonium acetate (10 mM, pH 7.0) (A) and acetonitrile (B); flow and gradient: 0.2 ml/min, 0–18.7% B over 28 min]. Sp was collected at 14–16 min of elution, Gh at 18 min and Ox at 21 min, with the collected samples dried under vacuum.

The dried samples containing 8-oxodG, NitroIm, Ox, Sp and Gh and their corresponding isotopomers were dissolved in water. An aliquot of each sample was analyzed by HPLC–ESI–MS/MS essentially as described earlier with the following changes. Samples were resolved on a 15 cm PicoFrit capillary column (360 μm outer diameter × 75 μm internal diameter) (New Objective, Woburn, MA) packed with Alltech C18 solid phase (5 μm) and eluted with a mobile phase consisting of a 97:3 mixture of 0.4% acetic acid in water (A) and 0.4% acetic acid in acetonitrile (B) delivered at 90 nl/min. A nanoelectrospray ionization source was employed and the MS was operated in the positive ion mode, with all instrument paramenters optimized for maximal sensitivity. Samples were analyzed in multiple reaction monitoring mode, with the following transitions: m/z 284 → 168, 285 → 169 and 287 → 171 for 8-oxodG, 7- 15 N-8-oxodG, 1,2,7- 15 N 3 -8oxo-dG, respectively; m/z 287 → 171 and 291 → 174 for NitroIm and labeled NitroIm; m/z 247 → 131 and 259 → 138 for Ox and uniformly 13 C- and 15 N-labeled Ox; m/z 300 → 184 and 315 → 194 for Sp and uniformly 13 C and 15 N-labeled Sp; m/z 274 → 158 and 277 → 161 for Gh and 3,7,8- 15 N 3 -5-Gh. Calibration curves were performed daily and were linear in all cases [ r2 > 0.99; data not shown; the correction factor for detection of 8-oxodG ( 19 ) was 0.98].

Immunoblot analysis of M 1 dG

Immunoblot assays were carried out as described previously ( 24 ). Briefly, DNA samples were immobilized on nitrocellulose membranes, and M 1 dG content was measured by enhanced chemiluminescence using an anti-M 1 G monoclonal antibody. The chemiluminescence signals were normalized for possible variation in the amount of immobilized DNA by subsequently staining membranes with propidium iodide.

Results

Adaptation and development of analytical methods

The key feature of these studies was the application of sensitive analytical methods to quantify a spectrum of DNA lesions proposed to arise as a result of exposure of DNA to the various chemical mediators of inflammation. Whereas the M 1 dG immunoblot assay and the LC–MS/MS method for the guanine oxidation products have been described in detail elsewhere ( 19 , 24 ), the new method for LC–MS/MS quantification of the nucleobase deamination products and etheno adducts warrants consideration of control studies to establish the rigor of the method. An important element of the analytical method is our ability to achieve baseline resolution of all of the normal and damaged nucleosides by reversed-phase HPLC, as illustrated in Figure 2 . This increases sample specificity in detection and reduces potential interfering signals arising from the more abundant normal nucleosides. For example, there is a 1 mass unit difference in the molecular weights of dI and dA, dC and dU, and dG and dX/dO. Hence, isotope natural abundance could create an M + 1 signal in positive ion mode that could interfere with detection of the higher molecular weight species. HPLC pre-purification followed by another, different chromatographic step essentially eliminated this problem (data not shown). In terms of recovery of the damage products, we were able to achieve an 80% recovery as judged by the use of isotopomeric standards (data not shown). Finally, the tandem MS detection was performed in the multiple reaction monitoring mode, with parameters optimized for each species using standards. Calibration curves were all linear ( r2 > 0.99, data not shown) and the detection limits for 100 μg DNA varied from 6 per 10 9 nt for ϵdA to 1 per 10 7 nt for dU.

Fig. 2.

Example of HPLC resolution of normal deoxynucleosides and various DNA damage products prior to LC–MS/MS analysis. This chromatogram was prepared using chemical standards.

Fig. 2.

Example of HPLC resolution of normal deoxynucleosides and various DNA damage products prior to LC–MS/MS analysis. This chromatogram was prepared using chemical standards.

SJL mice and DNA isolation

Following injection of the RcsX cells (or PBS), urinary nitrate was quantified to assess the level of NO overproduction by macrophages. RcsX-injected mice showed a time-dependent increase in urinary nitrate excretion averaging 0.4 μmol/g/day compared with 0.02 μmol/g/day for PBS-injected controls, with nitrate excretion reaching a maximum at 12 days. This is similar to previous studies employing the SJL model ( 16 ). Animals were killed 12 days after injection of tumor cells and genomic DNA was isolated from the tissues in the presence of coformycin, tetrahydrouridine and desferrioxamine to minimize DNA damage artifacts.

Quantification of DNA lesions in SJL mouse tissues

In as much as inducible NO synthase in macrophages is elevated in spleen and liver, but not kidney in SJL mice ( 14 , 17 ), we chose the former as NO-exposed tissues and the latter as a putative negative control. Following isolation of DNA from snap-frozen tissues, LC–MS/MS analyses were performed to quantify the battery of DNA lesions suspected of arising during inflammation in humans. The results of the analyses are shown in Tables I–III .

Table I.

DNA adducts derived in SJL mouse spleen

 DNA oxidation DNA nitrosation Lipid peroxidation 
  M 1 dG per 10 7 nt   8-oxodG per 10 6 nt   dX per 10 7 nt   dI per 10 6 nt   dU per 10 6 nt   ϵdA per 10 9 nt   ϵdG per 10 9 nt  
Control a 1.5 ± 0.1 1.2 ± 0.2 5.8 ± 0.6 1.4 ± 0.04 48 ± 7 12 ± 4 24 ± 9 
RcsX 2.1 ± 1.4 1.3 ± 0.4 7.7 ± 1.4 1.6 ± 0.2 57 ± 7  45 ± 10 b  65 ± 17 b 
Ratio c 1.5 1.1 1.3 1.1 1.2  3.8 b  2.7 b 
 DNA oxidation DNA nitrosation Lipid peroxidation 
  M 1 dG per 10 7 nt   8-oxodG per 10 6 nt   dX per 10 7 nt   dI per 10 6 nt   dU per 10 6 nt   ϵdA per 10 9 nt   ϵdG per 10 9 nt  
Control a 1.5 ± 0.1 1.2 ± 0.2 5.8 ± 0.6 1.4 ± 0.04 48 ± 7 12 ± 4 24 ± 9 
RcsX 2.1 ± 1.4 1.3 ± 0.4 7.7 ± 1.4 1.6 ± 0.2 57 ± 7  45 ± 10 b  65 ± 17 b 
Ratio c 1.5 1.1 1.3 1.1 1.2  3.8 b  2.7 b 

Values represent mean ± SD for N = 3–6.

a

Control samples contained spleens pooled from three mice; three separate pooled sets were analyzed.

b

RcsX-treated samples significantly different from controls by Student's T -test; P < 0.05.

c

RcsX to control.

Table II.

DNA adducts derived in SJL mouse liver

 DNA oxidation DNA nitrosation Lipid peroxidation 
  M 1 dG per 10 7 nt   8-oxodG per 10 6 nt   dX per 10 7 nt   dI per 10 6 nt   dU per 10 6 nt   ϵdA per 10 9 nt   ϵdG per 10 9 nt  
Control a 1.7 ± 0.5 1.3 ± 0.4 4.8 ± 0.6 1.2 ± 0.07 55 ± 2 20 ± 9 39 ± 38 
RcsX 2.2 ± 0.6 1.2 ± 0.6 5.9 ± 0.2 1.4 ± 0.2 60 ± 6 33 ± 9 102 ± 22 
Ratio b 1.3 0.9 1.2 1.2 1.1 1.7  2.6 a 
 DNA oxidation DNA nitrosation Lipid peroxidation 
  M 1 dG per 10 7 nt   8-oxodG per 10 6 nt   dX per 10 7 nt   dI per 10 6 nt   dU per 10 6 nt   ϵdA per 10 9 nt   ϵdG per 10 9 nt  
Control a 1.7 ± 0.5 1.3 ± 0.4 4.8 ± 0.6 1.2 ± 0.07 55 ± 2 20 ± 9 39 ± 38 
RcsX 2.2 ± 0.6 1.2 ± 0.6 5.9 ± 0.2 1.4 ± 0.2 60 ± 6 33 ± 9 102 ± 22 
Ratio b 1.3 0.9 1.2 1.2 1.1 1.7  2.6 a 

Values represent mean ± SD for N = 3–6.

a

RcsX-treated samples significantly different from controls by Student's T -test; P < 0.05.

b

RcsX to control.

Table III.

DNA adducts derived in SJL mouse kidney

  DNA oxidation
 
DNA nitrosation Lipid peroxidation 
  M 1 dG per 10 7 nt   dX per 10 7 nt   dI per 10 6 nt   dU per 10 6 nt   ϵdA per 10 9 nt   ϵdG per 10 9 nt  
Control 1.6 ± 0.2 5.2 ± 0 1.2 ± 0.2 52 ± 0.3 15 ± 5 20 ± 7 
RcsX 1.9 ± 0.5 4.9 ± 0.1 1.1 ± 0.1 55 ± 2 42 ± 13 77 ± 30 
Ratio b 1.2 0.9 0.9 1.1  2.8 a  3.9 a 
  DNA oxidation
 
DNA nitrosation Lipid peroxidation 
  M 1 dG per 10 7 nt   dX per 10 7 nt   dI per 10 6 nt   dU per 10 6 nt   ϵdA per 10 9 nt   ϵdG per 10 9 nt  
Control 1.6 ± 0.2 5.2 ± 0 1.2 ± 0.2 52 ± 0.3 15 ± 5 20 ± 7 
RcsX 1.9 ± 0.5 4.9 ± 0.1 1.1 ± 0.1 55 ± 2 42 ± 13 77 ± 30 
Ratio b 1.2 0.9 0.9 1.1  2.8 a  3.9 a 

Values represent mean ± SD for N = 3–6.

a

RcsX-treated samples significantly different from controls by Student's T -test; P < 0.05.

b

RcsX to control.

For the nucleobase deamination products, only slight increases in dX, dI and dU were observed in the spleen ( Table I ) and liver ( Table II ) of the RcsX-treated mice compared with controls, whereas in kidney, dU was slightly elevated (<10%) and dX and dI decreased to a small extent ( Table III ). dO was not detected in any of the samples (<1 per 10 8 nt), which is consistent with our previous studies in cultured human cells ( 23 ) and in Escherichia coli and mouse genomic DNA (Pang B., Dong M. and Dedon P.C., unpublished observations). This suggests that dO is not a major deamination product under the physiological conditions, as expected from the model recently proposed by Glaser et al. ( 25 ) in which the G:C base pairing limits the formation of dO in the double-stranded DNA.

Of the guanine oxidation products, only 8-oxodG was observed above its limit of detection of the LC–MS/MS assay ( Tables I–III ). However, the increase in 8-oxodG in RcsX-treated mice was insignificant at 14% in spleen ( Table I ) and liver ( Table II ). The secondary oxidation products Sp, Gh and Ox were not detectable at levels above 1 per 10 7 nt. Similarly, M 1 dG was modestly increased in kidney (18%), liver (30%) and spleen (67%), as shown in Tables I , II and III , respectively.

The only significant changes in DNA damage in the SJL mice occurred with the etheno adducts. As indicated in Tables I , II and III for spleen, liver and kidney, respectively, the levels of ϵdA and ϵdG increased 2- to 3-fold relative to controls. Although kidney was considered to be a negative control organ for NO overproduction due to the apparent absence of elevated inducible NO synthase activity ( 14 , 17 ), we observed an increase in both ϵdA and ϵdG in DNA from the kidneys of RcsX-injected mice ( Table III ). This change is similar to that observed in liver ( Table II ) and spleen ( Table I ). These results suggest that the oxidative stresses associated with NO overproduction are not restricted to the SJL target organs that are sites of NO overproduction.

Discussion

In an attempt to define the predominant chemistry, if any, at sites of inflammation, we have quantified a battery of DNA damage products representative of the NO-associated chemical mediators of inflammation in tissues from a mouse model of NO overproduction. This approach not only provides insights into the fundamental chemistry of inflammation and the connection between inflammation and cancer but also has the potential to reveal novel biomarkers of risk for inflammation-associated diseases.

Our studies entailed the development and application of sensitive analytical methods to quantify DNA lesions presumably associated with inflammation. A critical problem faced by any bioanalytical method is that of adventitious formation of damage during isolation, hydrolysis and other steps in the processing of DNA. Previous studies performed by our group revealed the adventitious formation of dI and dU during DNA isolation and hydrolysis as a result of endogenous and contaminating nucleobase deaminase activity ( 7 , 23 ). This problem was completely controlled by liberal use of deaminase inhibitors such as coformycin and tetrahydrouridine ( 7 , 23 ). In terms of oxidative damage artifacts, a problem that has received considerable attention ( 26 , 27 ), we included desferrioxamine in the DNA isolation and hydrolysis steps to minimize the oxidation of dG and its oxidation products. As illustrated in detail in our published work ( 19 ), oxidative artifacts could not be completely suppressed in spite of the presence of antioxidants and we therefore applied a correction factor (0.98) to account for adventitious oxidation of an isotopically labeled internal standard ( 19 ).

As a result of careful attention to artifacts, the values we obtained in the present studies compare well with the most rigorously performed analyses in the literature. Lim et al. ( 28 ) performed one of the few well-controlled studies, in terms of deamination artifacts, of hypoxanthine formation in purified DNA and in DNA isolated from NO-exposed tissues. They observed background levels of dI in DNA from Jurkat cells and from rat liver and kidney ranging from 1 to 3 per 10 6 nt ( 28 ), which compares favorably with our observation of 1–2 dI per 10 6 nt in SJL spleen, liver and kidney. Of equal importance was their observation that NO overproduction in endotoxin-treated rats did not increase the levels of dI in various tissues ( 28 ). This is again consistent with our observations in the RcsX-treated SJL mice.

Whereas the background levels of M 1 dG we observed using the immunoblot method are several-fold higher than the values determined by Jeong et al. ( 29 ) in rat tissue DNA using a recently developed oxime derivatization-based LC–MS/MS method (15 per 10 8 nt versus 1–4 per 10 8 nt), our values for background levels of 8-oxodG (1–2 per 10 6 nt) compare favorably with the most widely accepted background levels of 8-oxodG of 1–4 per 10 6 nt, as established in the recent ESCOD studies summarized by Collins et al. ( 30 ). As discussed earlier, our values for background levels of ϵdA determined by LC–MS/MS (12 per 10 9 nt) are similar to those determined by Nair et al. ( 18 ) in SJL mice using a 32 P post-labeling method (9 per 10 9 nt).

The changes in DNA adduct levels as a result of the increased oxidative and nitrosative stresses in the RcsX-treated SJL mice have important parallels with other studies. While 8-oxodG has been claimed to be a biomarker of oxidative stress and to correlate with cancer risk in several studies ( 31 , 32 ), we found that the level of 8-oxodG remains unchanged from background in DNA from liver and spleen in the inflamed SJL mice ( Table I ). This is consistent with the observations of Gal et al. ( 17 ), who used an immunohistochemical approach to quantify 8-oxodG, and those of Kadlubar et al. ( 33 ), who studied DNA adduct levels associated with oxidative stress in human pancreas and found no correlation between M 1 dG and either ϵdC or ϵdA, whereas there was a significant correlation between 8-oxodG and M 1 dG. This apparent lack of change in the steady state level of 8-oxodG in SJL mice and other oxidative stress models could be explained by a balance between formation and the consumption expected in light of the lower redox potential of 8-oxodG compared with dG (0.74 V/NHE versus 1.29 V/NHE; reviewed in 5 ). This was the rationale for attempting to quantify both primary dG oxidation products (8-oxodG and NitroIm) and the oxidation products of 8-oxodG (Sp, Gh and Ox). Unfortunately, we were unable to detect the background levels of the Sp, Gh, Ox and NitroIm lesions above the limit of detection of the LC–MS/MS assays (∼1 lesion per 10 7 nt in 100 μg of DNA) in liver or spleen tissues from the SJL mice.

There are several possible explanations for the minimal increases in the levels of the oxidative and nitrosative DNA lesions. One involves DNA repair. The steady state levels of DNA damage products measured here reflect an equilibrium between formation and repair. So it is possible that there are substantial amounts of DNA damage arising as a result of NO overproduction, with formation balanced by efficient repair of the lesions. Arguing against this model, however, is the observation of limited repair of nucleobase deamination products in TK6 cells exposed to NO and O 2in vitro ( 23 ), coupled with evidence for NO-induced inhibition of DNA repair enzymes ( 34–38 ).

Another possible explanation for the small changes in DNA adducts burden is a lesion dilution effect. The basis for this effect lies in the analysis of DNA extracted from the variety of cell types that comprise a whole organ. For the RcsX-inflamed spleen in the SJL mouse, this includes the lymphocytes of the white and red pulp, epithelial cells of the stroma, the endothelial cells and contents of the vascular compartment and fibroblasts and other connective tissue components, as well as the large influx of macrophages and neutrophils associated with the inflammation. Not all of these cells will be targets for or in proximity to the reactive oxygen and nitrogen species arising from the generator cells responsible for the NO overproduction. Thus, only a small fraction of the cells present in the spleen or other organs may be sustaining the bulk of the DNA damage and the adduct load is diluted by undamaged DNA from unaffected cells.

A third possible explanation is that reactive nitrogen and oxygen species arising from the macrophage-derived NO are unable to access the nucleus or are consumed before reacting with nuclear DNA, thus obviating direct reactions of N 2 O 3 , NO 2· or ONOOCO 2 with DNA. This is consistent with the observation of minimal nucleobase deamination in TK6 cells exposed to cytotoxic doses of NO and O 2in vitro ( 23 ).

A lack of direct DNA damage caused by NO-derived species does not preclude an indirect mechanism of DNA adducts formation, as demonstrated by the 3- to 4-fold increases in the levels of etheno adducts arising from lipid peroxidation in the NO overproducing SJL mice ( Tables I–III ). These results are quantitatively similar to the 6-fold increases in etheno adducts observed by Nair et al. ( 18 ) using a completely different analytical method, 32 P post-labeling, and it suggests that electrophiles derived from lipid peroxidation play a major role in the pathophysiology of inflammation. That cellular lipids should be involved with NO chemistry is not surprising given the high solubility NO, O 2 and N 2 O 3 in lipid or hydrophobic environments ( 39 , 40 ) and the high reactivity of the oxidants derived from NO (OH · , NO 2· and CO 3−· ) with polyunsaturated fatty acids [e.g. ( 24 )]. Indeed, studies in E.coli suggest that polyunsaturated fatty acids may preferentially react with these oxidants at the expense of reactions with DNA ( 24 ). The importance of lipid chemistry in inflammation is substantiated by numerous observations of significantly elevated etheno adducts in DNA in a variety of clinical disorders ( 9 ), including Wilson's disease ( 41 , 42 ) and colonic polyps of familial adenomatous polyposis patients ( 43 ).

Of particular interest is the observation that the levels of etheno adducts were increased in parallel in the three organs analyzed. This includes the kidney, which is not supposed to be a target organ for NO overproduction as judged by the absence of changes in apoptosis, mutation frequency and inducible NO synthase expression in the RcsX-treated mice ( 14 , 15 , 17 ). Nonetheless, our recent studies revealed up-regulation of glutathione synthesis enzymes in the kidney and spleen of the RcsX-treated SJL mice ( 44 ). These results suggest that, while kidney does not experience the massive invasion of macrophages as does the spleen, these organs may experience similar or secondary oxidative stresses as a result of NO overproduction, with the kidney affected ‘at a distance’. Possible mechanisms for oxidative stress in the kidneys of the SJL mice include the effects of circulating cytokines arising from the splenic infiltration of macrophages or oxidation of nitrite in blood or kidney to yield lipid-peroxidizing NO 2· , among many other possibilities. The role of local oxidative stresses in the generation of renal etheno adducts could be assessed by administration of antioxidants, whereas the role of NO in the phenomenon could be determined by inhibition of NO synthase by administration of N -methyl- L -arginine acetate.

In terms of biomarker development, a consideration of the results of the present studies in the context of other published observations ( 9 , 41–43 ) suggests that etheno adducts, and possibly other adducts derived from lipid peroxidation, could be suitable as biomarkers of the pathophysiology of inflammation. The utility of etheno adducts in tissue-derived DNA as biomarkers of disease risk in humans remains to be established. However, other lesions in DNA isolated from inflamed tissues may not serve well as biomarkers of inflammation due to the lack of correlation with the inflammatory process. Further, the limited dynamic range even for etheno adduct formation observed in the present studies (2- to 3-fold elevation) suggests that other sources of etheno adducts in DNA and RNA, such as DNA repair products and cell degradation products in blood, or other types of adducts arising from lipid peroxidation, such as protein carbonyls and other protein adducts, could serve as better biomarkers of inflammation and disease risk.

In summary, we have developed and applied analytical methods to quantify a battery of DNA damage products that reflect the different chemistries associated with chronic inflammation. These methods were applied to tissues from the NO overproducing SJL mouse model of inflammation, with the discovery that the only significant increases in DNA lesions occurred with the lipid peroxidation-derived etheno adducts. These observations warrant further studies in other animal models of inflammation and motivate us to examine more thoroughly the role of the lipid environment in the pathological chemistry of chronic inflammation.

Abbreviations

    Abbreviations
  • dI

    2′-deoxyinosine

  • dO

    2′-deoxyoxanosine

  • dU

    2′-deoxyuridine

  • dX

    2′-deoxyxanthosine

  • ϵdA

    1,N 6 -etheno-2′-deoxyadenosine

  • ϵdC

    3,N 4 -etheno-2′-deoxycytidine

  • ϵdG

    1,N 2 -etheno-2′-deoxyguanosine

  • Gh

    guanidinohydantoin or N(1)-(β- D -erythro-pentofuranosyl)-5-guanidinohydantoin

  • HPLC

    high performance liquid chromatography

  • LC

    liquid chromatography

  • MS

    mass spectrometry/meter

  • MS/MS

    tandem or triple quadrupole mass spectrometry

  • NitroIm

    5-guanidino-4-nitroimidazole

  • NO

    nitric oxide

  • 8-oxodG

    8-oxo-7,8-dihydro-2′-deoxyguanosine

  • Ox

    oxazolone

  • ONOOCO 2

    nitrosoperoxycarbonate

  • ONOO

    peroxynitrite

  • PBS

    phosphate-buffered saline

  • Sp

    spiroiminodihydantoin

The authors wish to thank Prof. David Schauer and Ms Laura Trudel for their expert assistance with the SJL mice. This work was supported by National Cancer Institute grants PO1 CA026731 and RO1 GM59790. The MS and HPLC analyses were performed in the Bioanalytical Facilities Core of the Massachusetts Institute of Technology Center for Environmental Health Sciences, which is supported by National Institute of Environmental Health Sciences grant P30 ES002109.

Conflict of Interest Statement: None declared.

References

1.
Ohshima
H
Genetic and epigenetic damage induced by reactive nitrogen species: implications in carcinogenesis
Toxicol. Lett.
 , 
2003
, vol. 
140–141
 (pg. 
99
-
104
)
2.
Shacter
E
, et al.  . 
Chronic inflammation and cancer
Oncology (Huntingt)
 , 
2002
, vol. 
16
 (pg. 
217
-
226
)
3.
Klaunig
JE
, et al.  . 
The role of oxidative stress in carcinogenesis
Annu. Rev. Pharmacol. Toxicol.
 , 
2004
, vol. 
44
 (pg. 
239
-
267
)
4.
Ohshima
H
, et al.  . 
Chronic infections and inflammatory processes as cancer risk factors: possible role of nitric oxide in carcinogenesis
Mutat. Res.
 , 
1994
, vol. 
305
 (pg. 
253
-
264
)
5.
Dedon
PC
, et al.  . 
Reactive nitrogen species in the chemical biology of inflammation
Arch. Biochem. Biophys.
 , 
2004
, vol. 
423
 (pg. 
12
-
22
)
6.
Ohshima
H
, et al.  . 
Chemical basis of inflammation-induced carcinogenesis
Arch. Biochem. Biophys.
 , 
2003
, vol. 
417
 (pg. 
3
-
11
)
7.
Dong
M
, et al.  . 
Absence of 2′-deoxyoxanosine and presence of abasic sites in DNA exposed to nitric oxide at controlled physiological concentrations
Chem. Res. Toxicol.
 , 
2003
, vol. 
16
 (pg. 
1044
-
1055
)
8.
Tretyakova
NY
, et al.  . 
Peroxynitrite-induced DNA damage in the supF gene: correlation with the mutational spectrum
Mutat. Res.
 , 
2000
, vol. 
447
 (pg. 
287
-
303
)
9.
Bartsch
H
, et al.  . 
Oxidative stress and lipid peroxidation-derived DNA-lesions in inflammation driven carcinogenesis
Cancer Detect. Prev.
 , 
2004
, vol. 
28
 (pg. 
385
-
391
)
10.
Marnett
LJ
Oxy radicals, lipid peroxidation and DNA damage
Toxicology
 , 
2002
, vol. 
181-182
 (pg. 
219
-
222
)
11.
Lee
SH
, et al.  . 
4-hydroperoxy-2-nonenal-induced formation of 1,N2-etheno-2′-deoxyguanosine adducts
Chem. Res. Toxicol.
 , 
2005
, vol. 
18
 (pg. 
780
-
786
)
12.
Williams
MV
, et al.  . 
Endogenous lipid hydroperoxide-mediated DNA-adduct formation in min mice
J. Biol. Chem.
 , 
2006
, vol. 
281
 (pg. 
10127
-
10133
)
13.
Nair
J
, et al.  . 
Etheno DNA-base adducts from endogenous reactive species
Mutat. Res.
 , 
1999
, vol. 
424
 (pg. 
59
-
69
)
14.
Gal
A
, et al.  . 
Nitric oxide production in SJL mice bearing the RcsX lymphoma: a model for in vivo toxicological evaluation of NO
Proc. Natl Acad. Sci. USA
 , 
1996
, vol. 
93
 (pg. 
11499
-
11503
)
15.
Gal
A
, et al.  . 
Mutagenesis associated with nitric oxide production in transgenic SJL mice
Proc. Natl Acad. Sci. USA
 , 
1996
, vol. 
93
 (pg. 
15102
-
15107
)
16.
Tamir
S
, et al.  . 
Nitric oxide production in relation to spontaneous B-cell lymphoma and myositis in SJL mice
Cancer Res.
 , 
1995
, vol. 
55
 (pg. 
4391
-
4397
)
17.
Gal
A
, et al.  . 
Nitrotyrosine formation, apoptosis, and oxidative damage: relationships to nitric oxide production in SJL mice bearing the RcsX tumor
Cancer Res.
 , 
1997
, vol. 
57
 (pg. 
1823
-
1828
)
18.
Nair
J
, et al.  . 
Etheno adducts in spleen DNA of SJL mice stimulated to overproduce nitric oxide
Carcinogenesis
 , 
1998
, vol. 
19
 (pg. 
2081
-
2084
)
19.
Yu
H
, et al.  . 
Quantitation of four guanine oxidation products from reaction of DNA with varying doses of peroxynitrite
Chem. Res. Toxicol.
 , 
2005
, vol. 
18
 (pg. 
1849
-
1857
)
20.
Gu
F
, et al.  . 
Peroxynitrite-induced reactions of synthetic oligo 2′-deoxynucleotides and DNA containing guanine: formation and stability of a 5-guanidino-4-nitroimidazole lesion
Biochemistry
 , 
2002
, vol. 
41
 (pg. 
7508
-
7518
)
21.
Bobalova
J
, et al.  . 
High-performance liquid chromatographic technique for detection of a fluorescent analogue of ADP-ribose in isolated blood vessel preparations
Anal. Biochem.
 , 
2002
, vol. 
305
 (pg. 
269
-
276
)
22.
Chen
HJ
, et al.  . 
Analysis of 3,N(4)-ethenocytosine in DNA and in human urine by isotope dilution gas chromatography/negative ion chemical ionization/mass spectrometry
Chem. Res. Toxicol.
 , 
2001
, vol. 
14
 (pg. 
1612
-
1619
)
23.
Dong
M
, et al.  . 
Relatively small increases in the steady-state levels of nucleobase deamination products in DNA from human TK6 cells exposed to toxic levels of nitric oxide
Chem. Res. Toxicol.
 , 
2006
, vol. 
19
 (pg. 
50
-
57
)
24.
Zhou
X
, et al.  . 
Chemical and biological evidence for base propenals as the major source of the endogenous M 1 dG adduct in cellular DNA
J. Biol. Chem.
 , 
2005
, vol. 
280
 (pg. 
25377
-
25382
)
25.
Glaser
R
, et al.  . 
Cytosine catalysis of nitrosative guanine deamination and interstrand cross-link formation
J. Am. Chem. Soc.
 , 
2005
, vol. 
127
 (pg. 
7346
-
7358
)
26.
Cadet
J
, et al.  . 
Facts and artifacts in the measurement of oxidative base damage to DNA
Free Radic. Res.
 , 
1998
, vol. 
29
 (pg. 
541
-
550
)
27.
Cadet
J
, et al.  . 
Artifacts associated with the measurement of oxidized DNA bases
Environ. Health Perspect.
 , 
1997
, vol. 
105
 (pg. 
1034
-
1039
)
28.
Lim
KS
, et al.  . 
Potential artifacts in the measurement of DNA deamination
Free Radic. Biol. Med.
 , 
2006
, vol. 
40
 (pg. 
1939
-
1948
)
29.
Jeong
YC
, et al.  . 
Pyrimido[1,2-a]-purin-10(3H)-one, M1G, is less prone to artifact than base oxidation
Nucleic Acids Res.
 , 
2005
, vol. 
33
 (pg. 
6426
-
6434
)
30.
Collins
AR
, et al.  . 
Are we sure we know how to measure 8-oxo-7,8-dihydroguanine in DNA from human cells?
Arch. Biochem. Biophys.
 , 
2004
, vol. 
423
 (pg. 
57
-
65
)
31.
Feig
DI
, et al.  . 
Reactive oxygen species in tumorigenesis
Cancer Res.
 , 
1994
, vol. 
54
 (pg. 
1890s
-
1894s
)
32.
Frelon
S
, et al.  . 
High-performance liquid chromatography–tandem mass spectrometry measurement of radiation-induced base damage to isolated and cellular DNA
Chem. Res. Toxicol.
 , 
2000
, vol. 
13
 (pg. 
1002
-
1010
)
33.
Kadlubar
FF
, et al.  . 
Comparison of DNA adduct levels associated with oxidative stress in human pancreas
Mutat. Res.
 , 
1998
, vol. 
405
 (pg. 
125
-
133
)
34.
Laval
F
, et al.  . 
Inhibition by nitric oxide of the repair protein, O6-methylguanine-DNA-methyltransferase
Carcinogenesis
 , 
1994
, vol. 
15
 (pg. 
443
-
447
)
35.
Laval
F
, et al.  . 
A discussion of mechanisms of NO genotoxicity: implication of inhibition of DNA repair proteins
Rev. Physiol. Biochem. Pharmacol.
 , 
1997
, vol. 
131
 (pg. 
175
-
191
)
36.
Liu
L
, et al.  . 
Inactivation and degradation of O(6)-alkylguanine-DNA alkyltransferase after reaction with nitric oxide
Cancer Res.
 , 
2002
, vol. 
62
 (pg. 
3037
-
3043
)
37.
Jaiswal
M
, et al.  . 
Human ogg1, a protein involved in the repair of 8-oxoguanine, is inhibited by nitric oxide
Cancer Res.
 , 
2001
, vol. 
61
 (pg. 
6388
-
6393
)
38.
Sidorkina
O
, et al.  . 
Inhibition of poly(ADP-ribose) polymerase (PARP) by nitric oxide and reactive nitrogen oxide species
Free Radic. Biol. Med.
 , 
2003
, vol. 
35
 (pg. 
1431
-
1438
)
39.
Dix
TA
, et al.  . 
Mechanisms and biological relevance of lipid peroxidation initiation
Chem. Res. Toxicol.
 , 
1993
, vol. 
6
 (pg. 
2
-
18
)
40.
Liu
X
, et al.  . 
Accelerated reaction of nitric oxide with O 2 within the hydrophobic interior of biological membranes
Proc. Natl Acad. Sci. USA
 , 
1998
, vol. 
95
 (pg. 
2175
-
2179
)
41.
Nair
J
, et al.  . 
Lipid peroxidation-induced etheno-DNA adducts in the liver of patients with the genetic metal storage disorders Wilson's disease and primary hemochromatosis
Cancer Epidemiol. Biomarkers Prev.
 , 
1998
, vol. 
7
 (pg. 
435
-
440
)
42.
Nair
J
, et al.  . 
Copper-dependent formation of miscoding etheno-DNA adducts in the liver of Long Evans cinnamon (LEC) rats developing hereditary hepatitis and hepatocellular carcinoma
Cancer Res.
 , 
1996
, vol. 
56
 (pg. 
1267
-
1271
)
43.
Schmid
K
, et al.  . 
Increased levels of promutagenic etheno-DNA adducts in colonic polyps of fap patients
Int. J. Cancer
 , 
2000
, vol. 
87
 (pg. 
1
-
4
)
44.
Li
CQ
, et al.  . 
Biological role of glutathione in nitric oxide-induced toxicity in cell culture and animal models
Free Radic. Biol. Med.
 , 
2005
, vol. 
39
 (pg. 
1489
-
1498
)