Original Contribution

In vivo detection of a novel endogenous etheno–DNA adduct derivedfrom arachidonic acid and the effects of antioxidants on its formation

Ying Fu a,1, Raghu G. Nath a,1, Marcin Dyba a,1, Idalia M. Cruz a, Sharanya R. Pondicherry a,Aileen Fernandez a, Casey L. Schultz b, Peiying Yang b, Jishen Pan a, Dhimant Desai c,Jacek Krzeminski c, Shantu Amin c, Plamen P. Christov d, Yukihiko Hara e,Fung-Lung Chung a,nQ1

a Department of Oncology, Lombardi Comprehensive Cancer Center, Georgetown University, Washington, DC 20057, USAb Department of Experimental Therapeutics, University of Texas M.D. Anderson Cancer Center, Houston, TX 77054, USAc Department of Pharmacology, Pennsylvania State College of Medicine, Hershey, PA 17033, USAd Department of Chemistry, Vanderbilt University, Nashville, TN 37235, USAe Tea Solution, Hara Office, Inc., Tokyo, Japan

a r t i c l e i n f o

Article history:Received 19 February 2014Received in revised form28 April 2014Accepted 29 April 2014

Keywords:ω-6 polyunsaturated fatty acidsIsotope dilution liquid chromatography–electrospray ionization–tandem massspectrometryDNA adductAntioxidantsFree radicals

a b s t r a c t

Previous studies showed that 7-(10,20-dihydroxyheptyl)-substituted etheno DNA adducts are products ofreactions with the epoxide of (E)-4-hydroxy-2-nonenal, an oxidation product of ω-6 polyunsaturatedfatty acids (PUFAs). In this work, we report the detection of 7-(10 ,20-dihydroxyheptyl)-1,N6-ethenodeox-yadenosine (DHHedA) in rodent and human tissues by two independent methods: a 32P-postlabeling/HPLC method and an isotope dilution liquid chromatography–electrospray ionization–tandem massspectrometry method, demonstrating for the first time that DHHedA is a background DNA lesion in vivo.We showed that DHHedA can be formed upon incubation of arachidonic acid with deoxyadenosine,supporting the notion that ω-6 PUFAs are the endogenous source of DHHedA formation. Because cyclicadducts are derived from the oxidation of PUFAs, we subsequently examined the effects of antioxidants,α-lipoic acid, Polyphenon E, and vitamin E, on the formation of DHHedA and γ-hydroxy-1,N2-propanodeoxyguanosine (γ-OHPdG), a widely studied acrolein-derived adduct arising from oxidizedPUFAs, in the livers of Long Evans Cinnamon (LEC) rats. LEC rats are afflicted with elevated lipidperoxidation and prone to the development of hepatocellular carcinomas. The results showed thatalthough the survival of LEC rats was increased significantly by α-lipoic acid, none of the antioxidantsinhibited the formation of DHHedA, and only Polyphenon E decreased the formation of γ-OHPdG. Incontrast, vitamin E caused a significant increase in the formation of both γ-OHPdG and DHHedA in thelivers of LEC rats.

& 2014 Elsevier Inc. All rights reserved.

Lipid peroxidation (LPO), a consequence of oxidative stressfrom exposure to environmental pollutants or chronic inflamma-tion, is an endogenous cause of protein and DNA damage [1,2]. LPOhas been implicated in the mechanisms of aging, cancers, neuraldisorders, and other degenerative diseases. Upon oxidation byreactive oxygen species (ROS) or cyclooxygenases (COXs)/lipoxy-genases (LOXs), ω-3 and ω-6 polyunsaturated fatty acids (PUFAs)generate highly reactive α,β-unsaturated aldehydes (enals), whichcan modify DNA bases by forming a variety of exocyclic DNAadducts [3–5].

Dietary ω-6 PUFAs have been strongly implicated in cancerpromotion in experimental animals [6], although the evidencefrom epidemiological studies is less clear [7]. The mechanismsunderlying their tumor-promoting effects are not yet fully under-stood. Arachidonic acid (AA) is one of the most abundant ω-6PUFAs present in cellular phospholipid membranes. During oxida-tive stress and inflammation, AA is released from the membraneby phospholipases [8]. AA can be oxidized by COXs and LOXs toyield eicosanoids that are believed to play a role in cancerdevelopment because of their growth-stimulating activities [9].In addition, enzymatic and nonenzymatic oxidation of AA pro-duces acrolein (Acr) and (E)-4-hydroxy-2-nonenal (HNE) as majorend products, which are capable of modifying DNA bases andcausing mutations and, therefore, may contribute to carcinogen-esis. 1,N2-Propanodeoxyguanosine adducts of Acr and HNE (Acr-dGor OHPdG and HNE-dG) have been detected as background lesions

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

6768697071727374757677787980818283848586

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/freeradbiomed

Free Radical Biology and Medicine

http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.0320891-5849/& 2014 Elsevier Inc. All rights reserved.

n Corresponding author. Fax: þ202 687 1068.E-mail address: flc6@georgetown.edu (F.-L. Chung).1 These authors contributed equally to this work.

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎

in tissue DNA of rodents and humans [10,11]. However, the in vivolevels of HNE-dG are often too low to be quantitatively detected bycurrent methods [12]. This limitation precludes its use as a specificbiomarker of DNA damage caused by ω-6 PUFAs.

HNE is known to be epoxidized by various agents and enzymes,including tert-butylhydroperoxide, hydrogen peroxide, fatty acidhydroperoxide, and lipoxygenase, to 2,3-epoxy-4-hydroxynonanal(EH). EH is considerably more reactive than HNE toward DNAbases [13–15]. Earlier chemical studies showed that EH reacts withdeoxyadenosine (dA) and deoxyguanosine (dG) to form both theunsubstituted and the substituted etheno adducts, such as 1,N6-ethenodeoxyadenosine (edA) and 7-(10,20-dihydroxyheptyl)-1,N6-ethenodeoxyadenosine (DHHedA), respectively [13,14]. AlthoughedA is a widely studied in vivo adduct, there has been no report onthe detection in vivo of DHHedA as an endogenous DNA lesion.In this study, we demonstrate the detection of DHHedA in tissuesof rodents and humans and show that its levels are considerablyhigher than those of HNE-dG, making it a potential DNA damagebiomarker specific for ω-6 PUFAs. We also show that its formationincreases significantly in the presence of AA, suggesting AA is apossible endogenous source (Fig. 1).

Because DHHedA and OHPdG arise from oxidation products ofPUFAs, we further examined the effects of antioxidants, includingα-lipoic acid, Polyphenon E, and vitamin E, on their level in thelivers of Long Evans Cinnamon (LEC) rats, that are afflicted withincreased LPO in the liver owing to abnormal copper accumula-tion, mimicking the human Wilson disease. As a result, LEC ratsdevelop acute hepatitis, followed by chronic hepatitis, and even-tually hepatocellular carcinoma [16]. α-Lipoic acid is an organo-sulfur antioxidant that accumulates in tissues upon dietaryadministration and is converted to dihydrolipoic acid [17].α-Lipoic acid and dihydrolipoic acid exhibit direct free radical-scavenging properties [18,19]. They also show antioxidant effectsby chelating metal ions (e.g., Fe2þ and Fe3þ) and regeneratingendogenous and exogenous antioxidants such as ubiquinone,glutathione (GSH), and ascorbic acid [20,21]. Furthermore, variousstudies have shown that α-lipoic acid decreases LPO ex vivo [22]and in vivo [23]. α-Lipoic acid has been used in treating severaldiseases including hepatic disorders (e.g., mushroom poisoning

and alcoholic liver disease) and diabetes [18]. α-Lipoic acid wasrecommended as one of the promising antioxidants for chemopreven-tion studies in a National Cancer Institute-sponsored screening report[24]. Green tea has been shown to inhibit chemical-induced hepato-carcinogenesis in vivo [25,26]. A cohort study showed that theconsumption of green teawas associated with a reduced risk of humanliver cancer incidence [27], although the collective evidence fromepidemiological studies on the protective effect of green tea consump-tion against liver cancer is not conclusive [28]. Polyphenon E is a well-defined mixture of decaffeinated green tea polyphenols includingepigallocatechin gallate (EGCG), the most abundant and potent anti-oxidative catechin. Because of its highly controlled and reproducibleformulation, it is a suitable form for clinical prevention trials. It hasbeen shown that Polyphenon E can scavenge the stable free radicalsgenerated from lipid peroxidation, and it can inhibit tobaccocarcinogen-induced lung tumor formation in mice [29] and chemicallyinduced bladder cancer [30]. However, it is not known whether it caninhibit spontaneous hepatocarcinogenesis. Vitamin E is a fat-solubleantioxidant and it inhibits carcinogenesis in vivo [31]. It protects livertissues against oxidative stress-induced chromosomal damage intransgenic mice with overexpressed c-myc gene and transforminggrowth factor-α [32]. Vitamin E can inhibit liver dysplasia andadenomas and prevent malignant formation in these mice. However,it did not inhibit LPO in healthy humans [33]. In fact, dietarysupplementation of vitamin E has been shown to be associated withan increased risk of prostate cancer in healthy men, raising publichealth concerns about vitamin E supplementation [34–36]. Never-theless, a more recent work suggested that the various forms andmixture of tocopherols may play a key role in the cancer-preventiveeffect of vitamin E [37].

Materials and methods

Chemicals and enzymes

Calf intestinal alkaline phosphatase grade I was bought fromRoche Diagnostic (Indianapolis, IN, USA). Phosphodiesterase Ifrom Crotalus adamanteus venom, deoxyribonuclease I type II

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 1. Proposed mechanisms of the formation of the cyclic adducts studied in this work induced by PUFAs through hydroperoxy fatty acids (FAOOH): α-OH-1,N2-propano-20-deoxyguanosine (α-OHPdG), γ-OH-1,N2-propano-20-deoxyguanosine (γ-OHPdG), 1,N6-ethenodeoxyadenosine (edA), 7-(10 ,20-dihydroxyheptyl)-1,N6-ethenodeoxyadenosine(DHHedA), and trans-4-hydroxy-2-nonenal–deoxyguanosine (HNE-dG).

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎2

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

from bovine pancreas, dA, dA 30-monophosphate (dA-30-P), dA 50-monophosphate (dA-50-P), dG, dG 30-monophosphate, micrococcalnuclease (MN), RNase A, RNase T1, protease, human placental DNA,calf thymus DNA, glycidaldehyde diethyl acetal, Amberlyst 15 ion-exchange resin, dietary vitamin E (α-tocopherol), and α-lipoic acidwere obtained from Sigma–Aldrich Co. (St. Louis, MO, USA).Polyphenon E was a generous gift from Dr. Hara (Mitsui Norin,Japan). [15N5]dA was from Spectra Stable Isotopes (Columbia, MD,USA). Spleen phosphodiesterase (SPD) was obtained from BoehringerMannheim (Indianapolis, IN, USA), T4 polynucleotide kinase (T4 PNK)was from U.S. Biochemicals (Cleveland, OH, USA). [γ-32P]ATP (sp act3000 Ci/mmol) was from Amersham (Arlington Heights, IL, USA) andmung bean nuclease (MBN) was from Thermo Fisher Scientific (FairLawn, NJ, USA). All other chemicals, solvents, and reagents were fromSigma–Aldrich or Thermo Fisher Scientific.

Animals

LEC rats at the age of 4 weeks were obtained from Charles RiverJapan (Yokohama, Japan.). They were housed in a temperature-controlled light-regulated space with 12-h light and dark cyclesand were given unrestricted access to food and water throughoutthe experiments. The protocol used in this study was approved bythe Georgetown University Animal Care and Use Committee. Theanimals were fed with AIN-76 A powder diets obtained from Dyets(Bethlehem, PA, USA) at the age of 4 weeks; four types of diets(control, α-lipoic acid, Polyphenon E, and vitamin E) were used inthis study. The only difference between the diets was the sucrosecontent to be replaced by the antioxidant (Supplementary TableS1: the doses of the antioxidants, α-lipoic acid (2 g/kg), Polyphe-non E (20 g/kg), and vitamin E (1.8 g/kg), were selected accordingto recent studies [25,26,33,38]). The nutrition for all four diets wascalculated within the range of 3694 to 3766 kcal/kg. The bodyweights of the LEC rats were measured weekly. LEC rats (n¼3–5)were sacrificed after 12 and 20 weeks. The liver tissues weredissected and kept at �80 1C until used for DNA isolation.

HPLC systems

Ten HPLC systems were used and the details are listed in thesupplementary material.

Preparation of DHHedA 30- and 50-monophosphate for 32Ppostlabeling

HNE and EH were synthesized by previously described meth-ods [39,40]. For the preparation of DHHedA 30-monophosphate(DHHedA-30-P) and DHHedA 50-monophosphate (DHHedA-50-P),10 mg of EH was dissolved in 100 ml of dimethyl sulfoxide andmixed with 10 mg of the nucleotides (dA-30-P or dA-50-P) in100 mM sodium phosphate buffer (pH 7.2, total volume of1.5 ml). The mixture was shaken at 37 1C for 72 h and extractedwith 3 ml of chloroform to remove unreacted EH. Portions of theaqueous phase were analyzed by HPLC system 1 and adduct peakswere identified by retention times and UV spectra of standards.The peaks were collected, dried in a SpeedVac, and repurifiedusing HPLC system 2. The purified DHHedA-30-P standards weredesalted using the same solid-phase extraction (SPE) system asunder Detection and quantitation of DHHedA by an HPLC-based 32P-postlabeling assay in tissue DNA. Alternatively, the reactions can beperformed with EH generated in situ from HNE and hydrogenperoxide (see Synthesis of DHHedA and [15N5]DHHedA for LC-MS/MS-multiple reaction monitoring (MRM)), yielding the same twopairs of DHHedA isomers, which were separated by HPLC system 1.Under the conditions used, the yields of the second pair of isomerswere four to five times higher than the first pair. By mixing

approximately equal amounts of each pair of isomers collected, theDHHedA-30-P and DHHedA-50-P were then used as standards in32P postlabeling for labeling and UV reference, respectively.

Synthesis of the 50-monophosphate of 7-hydroxymethyl-substituted 1,N6-edA (HMedA)

Glycidaldehyde diethyl acetal (90% purity, 375 ml) was dissolvedin tetrahydrofuran (THF):water (3:1, 2.25 ml) and mixed withAmberlyst 15 ion-exchange resin (400 mg). The suspension wasgently mixed overnight and filtered and the filtrate was added todA-50-P (15 mg) in sodium phosphate buffer (100 mM, pH 7.2,3 ml). The resulting solution was shaken at 37 1C for 48 h. Thereaction mixture was extracted with chloroform (3�3 ml) toremove the excess of glyceraldehyde and the aqueous phase wasconcentrated on a SpeedVac and purified using HPLC system4 yielding 20% of the product. The identity of the product wasconfirmed by characteristic UV spectrum and coelution with apreviously synthesized compound [39,41] using HPLC system 4.HMedA was then used as a standard for the detection of DHHedAin tissue DNA by the 32P-postlabeling assay (see below).

Detection and quantitation of DHHedA by an HPLC-based32P-postlabeling assay in tissue DNA

DNA was isolated from tissues by a modified Marmur procedureusing chloroform–isoamyl alcohol extraction and RNase A and T1 andprotease treatment as previously described [42]. DNAwas dissolved inwater, its purity was monitored by the 260/280 nm absorbance ratio(1.8–2.0), and it was stored at �80 1C until analysis. DNA samples (50to 100 mg) were digested by MN and SPD as previously described[10,11]. DHHedA 30-monophosphate (20 fmol) together with 100 mgcalf thymus DNA (in duplicate) and one blank (100 ml water) samplewere included in the assay of each set of DNA samples. The SPEmethodwas used to remove the majority of unmodified nucleotides inthe MN/SPD digest before postlabeling. The Bond Elut C18, 100-mg,1-ml volume SPE column (Agilent Technologies, Santa Clara, CA, USA,formerly Varian, Harbor City, CA, USA) was preconditioned with 3 mlmethanol followed by 3ml deionized water before the sample wasloaded. Samples were loaded and the cartridges were washed with3 ml 15% methanol/50 mM ammonium formate (pH 7.5). The adductfraction eluted with 3 ml 1/1 water/methanol was collected and driedin a SpeedVac. The samples were treated with 10 units MBN in 10 mldilution buffer provided by the manufacturer and incubated for60 min at 37 1C. The reaction was terminated by the addition of 3 mlTris base (0.5 M). The samples were 32P postlabeled in the presence of30 units T4 PNK and 20 mCi [γ-32P]ATP (60 min at 37 1C) as previouslyreported [10,11]. To convert the labeled adduct bisphosphate to 50-P,the labeled digest was mixed with 20 ml sodium acetate (0.5 M, pH 5)and 150 units T4 PNK and incubated for 60 min at 37 1C. The majorityof the radioactivity was removed by preliminary purification using theSPE method as outlined above. The fractions were dried, mixed with50-P UV standard, and further purified sequentially by HPLC systems1 and 2. A portion of the purified adduct was analyzed as DHHedA-50-P on HPLC system 3. To confirm its identity, the purified adduct wasconverted to HMedA following a previously reported procedure[39,41]. Briefly, purified 32P-labeled DHHedA-50-P from tissue DNAwas dried, reconstituted in 0.5 ml water, mixed with 100 ml sodiummetaperiodate (40mg/ml), and stirred for 16–18 h at room tempera-ture. To the reaction mixture, 100 ml sodium borohydride (60 mg/ml)was added and stirred at room temperature for 60 min. The resultingmixture was neutralized by treating with 10 ml phosphoric acid, mixedwith authentic UV standard, and analyzed using HPLC system 4.To determine the levels of modification, the radioactivity wasadjusted for decay and procedural loss (recovery) and convertedto fmol (104 dpm¼1 fmol) as previously described [10,11]. To

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

avoid contamination from radioactivity carried over to subsequentanalysis, the system was washed and tested with a blank and/oradduct UV standard before the next analysis.

Synthesis of DHHedA and [15N5]DHHedA for LC-MS/MS-multiplereaction monitoring (MRM)

HNE (24.4 mg) dissolved in 400 ml of THF was mixed with 25 mlof 30% hydrogen peroxide and stirred at room temperature for 1 h.The reaction mixture was then poured into 10 mg of [15N5]dA ordA solubilized in a 1:1 mixture of THF and 50 mM phosphatebuffer, pH 7.3. The resulting mixture was incubated with shakingat 50 1C for 24 h and the solvents were evaporated to drynessusing a SpeedVac. The crude material was redissolved in 50%aqueous methanol and analyzed on HPLC system 6. Adducts (twopeaks) were identified based on retention times and characteristicUV spectra. DHHedA was purified on HPLC system 5 to yield twoHPLC-separable pairs of DHHedA stereoisomers (DHHedA 1,2 and3,4). Standards for mass spectrometry were dissolved in water andquantified by UV spectroscopy using molar extinction coefficientε278¼5200 M�1� cm�1 [43]. The identity of the synthetic stan-dard and internal standard was confirmed by high-resolution massspectrometry using a QSTAR Elite time-of-flight mass spectro-meter (Applied Biosystems, Foster City, CA, USA). Expected massesfor DHHedA: [MþH]þ¼406.20902 m/z, [M-dAþ2H]þ¼290.1617m/z. Masses found: 406.2079 and 290.1605 m/z. Expected massesfor [15N5]DHHedA: [MþH]þ¼411.1942 m/z, [M-dAþ2H]þ¼295.1469 m/z. Masses found: 411.1930 and 295.1456 m/z.

Synthesis of α- and γ-OHPdG, α-[13C10,15N5]OHPdG, andγ-[13C10,15N5]OHPdG for LC-MS/MS-MRM

The synthesis of both OHPdG standards and the stable-isotope-labeled internal standards was described in our earlier publication [44].

DNA isolation and hydrolysis for LC-MS/MS-MRM

DNA samples from human and rat livers were isolated by aQiagen Blood and Cell Culture DNA Maxi Kit (Qiagen, Hilden,Germany) using the tissue protocol as recommended by themanufacturer. For hydrolysis dry DNA (0.4 to 1 mg) was dissolvedin 5 mM magnesium chloride and 0.5 mM GSH solution (1 ml/mgof DNA) and then 100 fmol of α-[13C10,15N5]OHPdG, 50 fmol ofγ-[13C10,15N5]OHPdG, 50 fmol of [15N5]DHHedA 1,2 and 50 fmol of[15N5]DHHedA 3,4 were added as internal standards. DNA washydrolyzed by incubation with DNase I (1300 units/mg of DNA) for30 min at 37 1C followed by a second addition of DNase I(1300 units/mg of DNA) and incubation for an additional 10 minat 37 1C. Finally phosphodiesterase I (0.06 units/mg of DNA), alka-line phosphatase (380 units/mg of DNA), and adenosine deami-nase (0.5 units) were added and the sample was incubated for60 min at 37 1C. After hydrolysis a small portion of hydrolysate wassaved for further dG quantification and the remaining sample waspurified using Phenomenex Strata-X 33 m 30 mg/ml polymericreversed-phase solid-phase extraction columns (Phenomenex,Torrance, CA, USA). Before samples were loaded, the columnswere washed with ACNQ2 (3�1 ml) and stabilized with 25 mMammonium formate, pH 4.00 (3�1 ml). After the DNA hydrolysatewas loaded the columns were washed with 2.5% ACN in 25 mMammonium formate, pH 4.00 (1�1 ml) and then OHPdG wascollected by 5% ACN in 25 mM ammonium formate, pH 4.00(1�1 ml) followed by DHHedA collection by 30% ACN in 25 mMammonium formate, pH 4.00 (1�1 ml). OHPdG and DHHedAfractions were dried using a SpeedVac rotary concentrator, redis-solved in 400 ml 1:1 water:ACN, transferred to HPLC vials, dried,

and kept at �20 1C. Before quantification samples were dissolvedin 60 ml of water and 37 ml was injected on LC-MS.

Quantification of OHPdG in DNA by LC-MS/MS-MRM

Quantification was carried out on an Applied Biosystems/MDSSCIEX 4000 QTRAP triple-quadrupole mass spectrometer (Life Tech-nologies, Carlsbad, CA, USA) interfaced with a Waters Acquity UPLCliquid chromatography system equipped with Waters Acquity UPLCBEH C18 50�2.1-mm, 1.7-mm particle size column (Waters Corp.,Milford, MA, USA). The separation of adducts was performedisocratically by eluting with 3% ACN, 1 mM ammonium formatebuffer over 3.5 min using a 0.5 ml/min flow rate at 40 1C, followed bya 100% ACN wash. The electrospray ionization (ESI) source operatedin positive mode. The MRM experiment was performed usingion transitions of 324.2-208.1 (OHPdG) and 339.2-218.1 m/z([13C10,15N5]OHPdG) with a collision energy (CE) of 20 eV for quanti-fication, and those of 324.2-190.1 (OHPdG) and 339.2-200.1m/z([13C10,15N5]OHPdG) with a CE of 47 eV were used for structuralconfirmation. All other parameters were optimized to achieve max-imum signal intensity. Calibration curves were constructed for allthree HPLC-resolved isomers before each analysis using standardsolutions of α- and γ-OHPdG and α- and γ-[13C10,15N5]OHPdG. Aconstant concentration of [13C10,15N5]OHPdG (1 fmol/ml) was usedwith various concentrations of OHPdG (1.68 amol/ml–220 fmol/ml)and analyzed using 37-ml injections by LC-MS/MS-MRM. The stan-dard curves were linear in the range from 0.41 to 900 fmol of OHPdGon-column for all three HPLC-resolved isomers (1/x weighting;r2¼0.9987, r2¼0.9988, and r2¼0.9994 for OHPdG 1, 2, and 3,respectively). The measured limit of quantification (LOQ) was0.41 fmol/column and limit of detection (LOD) was 0.14 fmol/column.The overall method detection limit (MDL) for DNA samples wascalculated to be 5–10 fmol of each HPLC-resolved isomer/sample.

Quantification of DHHedA in DNA by LC-MS/MS-MRM

Quantification was carried out using the same instrument as forOHPdG adducts. The separation of adducts was performed iso-cratically by eluting with 13.5% ACN, 1 mM ammonium formatebuffer over 6.5 min using a 0.5 ml/min flow rate at 40 1C, followedby 100% ACN wash. The ESI source operated in positive mode.The MRM experiment was performed using ion transitions of406.2-290.2 (DHHedA) and 411.2-295.1 m/z ([15N5]DHHedA)with a CE of 28 eV, for quantification, and those of 406.2-160.1(DHHedA) and 411.2-165.0 m/z ([15N5]DHHedA) with a CE of80 eV were used for structural confirmation. All other parameterswere optimized to achieve maximum signal intensity. Calibrationcurves were constructed for two HPLC-resolved peaks before eachanalysis using standard solutions of DHHedA and [15N5]DHHedA. Aconstant concentration of [15N5]DHHedA (1 fmol/ml) was usedwith various concentrations of DHHedA (3.3 amol/ml–65 fmol/ml)and analyzed using 37-ml injections by LC-MS/MS-MRM. Thestandard curves were linear in the range from 0.37 to 800 fmolof DHHedA on-column for both peaks (1/x weighting; r2¼0.9997and r2¼0.9981 for DHHedA 1,2 and 3,4, respectively). MeasuredLOQ was 0.37 fmol/column and LOD was 0.1 fmol/column for bothpeaks. The overall MDL for DNA samples was calculated to be2–5 fmol of each HPLC-resolved isomer/sample.

Quantification of dG in DNA hydrolysate

dG was quantified using HPLC system 10 with detection at254 nm. A standard curve (from 5 nmol to 5 pmol of dG oncolumn) was constructed using a UV-quantified dG standard(ε254¼13700 M�1� cm�1 in water [45]).

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎4

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

Incubation of dA with AA

dA (10 mM) was incubated with AA (10 mM) and without AA(control) in 500 ml of Tris buffer (100 mM, pH 7.4) at 37 1C. After16 h, the mixtures were extracted with 2 ml of chloroform and theaqueous phase was purified using SPE as described above. AfterSPE, the samples were dried on a SpeedVac and finally DHHedAwas quantified by LC-MS/MS-MRM as described above.

Statistical analysis

The results obtained in rats are expressed as means7standarddeviation throughout the article. The differences in adduct levels wereanalyzed for statistical significance using Student’s t test; differenceswere considered significant when two-tailed tests indicated values of

po0.05. The p values in the survival data were calculated usingMedCalc software (MedCalc Software bvba, Ostend, Belgium).

Results

Detection of DHHedA in rodent and human tissues

HPLC-based 32P-postlabeling methodAlthough DHHedA is a known product of the reaction of EH

with dA, its formation in tissue DNA from rodents and humans hasnot yet been reported. For the detection of DHHedA, DNA samplesafter enzymatic digestion and enriched by SPE were subjected toan HPLC-based 32P-postlabeling assay. Although the recovery ofthe assay based on the adduct standard was only �5%, this assaycan detect as low as 1 fmol of DHHedA (Fig. 2, spectra a–d).

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 2. Detection of DHHedA in Sprague–Dawley rat liver DNA by the HPLC-based 32P-postlabeling method: spectra (a) UV standards, (b) 32P-postlabeled standards, (c) blank,and (d) rat liver DNA. Confirmation of DHHedA detected in vivo by converting DHHedA to HMedA by sodium metaperiodate/sodium borohydride: spectra (e) UV standards,(f) 32P-postlabled standards, (g) blank, and (h) rat liver DNA. Similar results were obtained with DNA from Fischer rat colonic mucosa, human placenta, and calf thymus.

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

Two pairs of stereoisomers of DHHedA were formed owing to twochiral carbons in the side chain. However, unlike DHHedG [46], thestereochemistry of DHHedA isomers has yet to be fully assigned.The HPLC chromatogram (Fig. 2, spectrum d) shows two majorradioactive peaks from the colonic mucosa DNA of a Fischer rat.Although not completely resolved, they have the same retentiontimes as the UV and 32P-labeled standards of DHHedA 3 and 4(Fig. 2, spectra a and b, respectively). A small unresolved, butdiscernible, radioactive peak was also detected in the DNA thatcomigrated with DHHedA 1 and 2. No cross-contamination wasnoted, as the blank sample (Fig. 2c) showed no radioactive peaks.These results indicate that DHHedA adducts are present asendogenous lesions. Interestingly, there is stereoselective forma-tion of DHHedA 3 and 4 in vivo as the recoveries of all four isomersbased on the standards are comparable in this assay. To furtherconfirm the identity of DHHedA detected in vivo, the adductfractions were collected by HPLC and chemically converted tothe corresponding short-chain HMedA with sodium metaperio-date followed by sodium borohydride addition. The HPLC chro-matograms showed a single peak, due to the loss of the chiralcarbons, having the same retention time as the UV standard ofHMedA (Fig. 2, spectra e, f, h). Using the recovery rate ofsimultaneously labeled DHHedA standards, the levels of modifica-tion by DHHedA in Sprague–Dawley rat liver, Fischer rat colonicmucosa, human placenta, and calf thymus DNA are estimated to bein the range of 3 to 960 adducts per 109 dG (Table 1). These resultsindicate that DHHedA is present at a relatively high level in tissueDNA as a background modification [47].

LC-MS/MS methodTo confirm the results from the 32P-postlabeling assay, a

quantitative LC-MS/MS-MRM method was developed for detectingDHHedA level. Calibration curves for both chromatographicallyresolved pairs of DHHedA stereoisomers were constructed usingsynthetic standards and [15N5]DHHedA internal standards. Stan-dard curves were linear between 0.37 and 800 fmol per injection(r2¼0.9997 and r2¼0.9981 for DHHedA 1,2 and 3,4, respectively).The method LOD was as low as 0.1 fmol/column with good LOQaround 0.37 fmol/column. DHHedA was detected as a set of twopeaks with retention time around 3.6 and 4.3 min, respectively(Fig. 3). The detection of the latter peak as the predominant adductisomer is in agreement with its stereoselective formation observedusing the 32P-postlabeling method. Five human liver samples wereanalyzed. The levels of two resolved pairs of isomers, DHHedA 1,2and DHHedA 3,4, were combined and reported as the total amountof DHHedA; the levels in human livers and LEC rat livers were27–63 and 20–75 adducts per 109 dG, respectively. Although thesensitivity of the LC-MS/MS method is not as high as that of the

32P-postlabeling method, the high levels of DHHedA modificationin tissue DNA permit its detection. Together, these results unequi-vocally demonstrated the presence of DHHedA in vivo as endo-genous DNA lesions.

AA as a source of DHHedA

Previously, we reported that HNE-dG is formed by incubatingdG with ω-6 PUFAs under oxidative conditions [3]. In this work, amilder condition via autoxidation (no oxidizing agent was used)was chosen to investigate whether ω-6 PUFAs are a source ofDHHedA. We incubated dA with and without AA; the incubationmixtures were purified on SPE and analyzed by the LC-MS/MSmethod. DHHedA was detected at 5077124 fmol in samplesincubated with AA, whereas no DHHedA was found in samplesincubated without AA. These results suggest that AA is a likelysource of DHHedA in vitro.

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Table 1DHHedA adduct levels detected in tissue DNA by 32P-postlabeling HPLC and LC-ESI-MS/MS-MRM.

Tissue (n¼3–6)a DHHedA/dG�109

Sprague–Dawley rat liverb 3–150Fischer rat colonic mucosab 10–960Human placentab 20–30Calf thymusb 10–370Human liverc 27–63LEC rat liverc 20–75

In general, a wider range of DHHedA was observed using 32P-postlabeling HPLCcompared to LC-ESI-MS/MS-MRM, possibly owing to multiple steps and poorrecovery in the former method compared to the latter, which is consistently morequantitative with internal standards.

a n, sample number.b 32P-postlabeling HPLC.c LC-ESI-MS/MS-MRM.

Fig. 3. Detection of DHHedA in LEC rat liver and human liver DNA by the LC-MS/MSmethod. Mass chromatograms from three experiments are shown. (Top) Masschromatograms for the synthetic DHHedA standard (406.2-290.2 m/z) and [15N5]DHHedA internal standard (411.2-295.1m/z); 29.7 fmol each for DHHedA 1,2 and3,4 standards and 37 fmol each for [15N5]DHHedA 1,2 and 3,4 internal standards.(Middle) Mass chromatograms obtained from DNA of normal human liver (4.8 fmolof DHHedA 1,2 and 12.1 fmol of DHHedA 3,4 detected). (Bottom) Mass chromato-grams for DNA from LEC rat liver (12.1 fmol of DHHedA 1,2 and 21.1 fmol ofDHHedA 3,4 detected).

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎6

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

Effects of antioxidants on DHHedA and γ-OHPdG in liver DNAof LEC rats

Both DHHedA and γ-OHPdG are formed from oxidizedPUFAs. To examine the effects of antioxidants on the formationof DHHedA and γ-OHPdG in vivo, LEC rats were fed with threetypes of antioxidant (α-lipoic acid, Polyphenon E, and vitaminE) diet. The body weights were monitored weekly for 20 weeksthrough the entire experiment (Supplementary Fig. S2). Therewas a 10–20% body weight loss after 10 weeks on α-lipoic acid,although no adverse effects were noted. At the end of theexperiment, the weights were 295.9724.0 (control), 274.3721.1 (α-lipoic acid), 278.1713.4 (Polyphenon E), and 280.3732.2 g (vitamin E) (Supplementary Table S3). The survivalcurves of the four groups are shown in Fig. 4. At termination(24 weeks of age or 20 weeks on antioxidant diet), thepercentages of rats that survived were as follows: 47, 83, 56,and 39% for control, α-lipoic acid, Polyphenon E, and vitamin Ediet, respectively. There was a significant protective effect of α-lipoic acid against the mortality caused by acute hepatitis inLEC rats. The increased survival rate is highly significantcompared to the control group (po0.001).

Three to five LEC rats fed with diets containing antioxidantswere harvested at two time points for DNA isolation: one at theonset (16 weeks of age or 12 weeks on antioxidant diet) andthe other during the phase of acute hepatitis (24 weeks of ageor 20 weeks on antioxidant diet). Digested DNA samples wereanalyzed by LC-MS/MS for DHHedA and OHPdG content. Asexpected, γ-OHPdG was the only isomer detected in all sam-ples [44]. In the rats fed with control diet, there was nosignificant difference in the levels of both adducts between12 and 20 weeks. Previous studies also showed that there waslittle difference between 12- and 20-week-old LEC rats in thelevels of 8-oxo-dG and edA in the liver DNA [48,49]. Among thegroups fed with the different antioxidant diets, a significantdecrease in γ-OHPdG only in rats on the Polyphenon E diet wasobserved. At week 12, the levels of γ-OHPdG were 67727compared to 3077133 per 109 dG in the control group(p¼0.03). After 20 weeks, the levels were 106737 comparedto 291762 per 109 dG in controls (p¼0.04). However, thelevels of γ-OHPdG were not affected by α-lipoic acid. Surpris-ingly, feeding vitamin E for 20 weeks caused a more thantwofold increase in γ-OHPdG from 2917162 to 6717158 per109 dG (p¼0.02) (Fig. 5A). None of the antioxidants showedsignificant inhibitory effects on DHHedA formation. Again,dietary vitamin E dramatically increased the levels of DHHedAby almost fourfold at week 20 (p¼0.02), from 2374 per 109 dG

in the control group to 90725 per 109 dG in the treatmentgroup (Fig. 5B).

Discussion

The detection, formation, repair, mutagenicity, and potentialroles in carcinogenesis of edG and edA as endogenous DNA lesionshave been extensively studied since the 1990 s. Earlier chemicalstudies have demonstrated that both unsubstituted and DHH-substituted etheno adducts are formed in the reactions of EH, theepoxide of HNE, with dA and dG [50,51]. These findings have led tothe assumption that the formation of EH from epoxidation of HNEmay be responsible for the etheno adducts (edA and edG) detectedin vivo [50,51]. In this study, we detected for the first time theDHH-substituted etheno dA in rodent and human tissues andfound that they are present in DNA at relatively high levels.Furthermore, we demonstrated that AA is a potential source ofits formation. These results provide support that epoxidation ofHNE, a unique product of ω-6 PUFAs, may indeed constitute apathway leading to the formation of the etheno adducts in vivo.

The alkyl-substituted etheno adducts have emerged in recentyears as novel endogenous DNA lesions. The heptanone-substitutedetheno adducts, presumably derived from (E)-4-oxo-2-nonenal, weredetected in the small intestine of Min mice [52]. DHHedG andDHHedA were found in cultured human monocytes treated withHNE under peroxidation conditions [53]. In that study, their levelswere reported to be even lower than those of the propano adductsformed directly from HNE. It was suggested that one reason for thepoor yields of DHH adducts was the inefficient conversion of HNE toEH in cells. In this study, we demonstrated unequivocally thatDHHedA is present in rodent and human tissues and found that

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 4. The survival curves of LEC rats on various antioxidant diets (α-lipoic acid,Polyphenon E, and vitamin E) and control diet throughout 20 weeks of bioassay.

Fig. 5. (A) γ-OHPdG and (B) DHHedA levels detected by the LC-MS/MS method inLEC rat liver after feeding various antioxidant diets (α-lipoic acid, Polyphenon E,and vitamin E) and control diet for 12 and 20 weeks; *po0.05.

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

DHHedA levels in tissues are in the range of one modification per107–108 bases, comparable to that of γ-OHPdG and edA. Therelatively high levels of DHHedA in vivo may be attributed to theconversion of HNE to highly reactive EH and/or its slow repair. Theincreased formation of DHHedA upon incubation with AA suggeststhat AA is a potential source for DHH adducts in vivo uponautoxidation.

Chronic inflammation, a risk factor for cancer, is known to beclosely associated with LPO [54]. 8-Oxo-dG is regarded as a generalmarker for oxidative stress, but it is not specifically formed by oxidizedPUFAs; in fact, the mechanism of its formation has been disassociatedfrom LPO-induced DNA damage [55]. The malondialdehyde-derivedcyclic adduct is probably formed primarily from the base propenal, amajor enal generated by oxidation of the ribose moiety of DNA, notfrom lipids [56]. Although acrolein and crotonaldehyde are knownoxidation products of PUFAs, both are ubiquitous environmentalpollutants (cigarette smoke, automobile exhaust, etc.). The ethenomodifications, such as edA and edG, can also be formed from exposureto environmental carcinogens, such as vinyl chloride and ethylcarbamate. Therefore, these adducts are not suitable markers oflipid-derived DNA damage. It has been suggested that the cyclic 1,N2-propano-dG adduct of HNE, a unique oxidation product of ω-6PUFAs, would serve as a specific biomarker of LPO-induced DNAdamage. However, its application is compromised by the low in vivolevel. Little is known about the biological role of the DHH adducts;however, the other structurally related etheno adducts, e.g., theheptanone–etheno adduct of deoxycytidine, have been shown to behighly mutagenic in both Escherichia coli and human cells [57,58]. Thefact that DHHedA is a unique DNA lesion from ω-6 PUFAs thatpresents in tissue at relatively high levels renders it useful as a specificbiomarker of lipid peroxidation, and its role in cancer promotionassociated with ω-6 PUFAs may be further exploited.

Excessive LPO induced by inflammatory processes can causeDNA damage, which may eventually cause cancer [4]. It is knownthat antioxidants can inhibit LPO [59] and carcinogenesis [60–62].We previously demonstrated that EGCG and α-tocopherol haveopposite effects toward γ-OHPdG formation from ω-3 PUFAsunder oxidative conditions in vitro; the former is an effectiveblocker, whereas the latter enhances its formation [63]. Of thethree antioxidants used in this study, only Polyphenon E signifi-cantly inhibited the formation of γ-OHPdG in the livers of LEC ratsat the onset of and during acute hepatitis. These findings agreewith previous in vitro studies, consistent with its potential activityin scavenging ROS and aldehydes and chelating metal ions,including Cu2þ [64,65]. The increased survival rate by α-lipoicacid, yet its failure to reduce the adduct levels, is intriguing,suggesting that a mechanism other than LPO-induced adducts isprobably involved in this survival. These results did confirm aprevious report that α-lipoic acid can ameliorate the mortality ofcopper-induced acute hepatitis in LEC rats [66]. Surprisingly, noneof the antioxidants showed any inhibitory effects against DHHedAformation in the livers of LEC rats. Similar to the previous in vitrostudies, vitamin E caused a significant increase in γ-OHPdG andDHHedA in the livers of LEC rats [63]. It is known that vitamin Epossesses both antioxidant and pro-oxidant properties, dependingon its concentrations and the reaction conditions [67]. Becausecyclic DNA adducts may play a role in carcinogenesis [48], theincrease in both γ-OHPdG and DHHedA by feeding a diet contain-ing vitamin E raises concerns that highlight the importance ofunderstanding their biological roles in liver cancer.

Acknowledgments

We thank the Proteomics & Metabolomics Shared Resource ofLombardi Comprehensive Cancer Center (supported partially by

NIH/NCI Grant P30-CA051008) for providing resources for the DNAadducts analysis. We thank Sanchita Sarangi, Ankeet Naik, andSimrandeep S. Sidhu for their help in DNA isolation and discus-sions; we also thank the Tissue Culture and Animal Core Facilitiesof Lombardi Cancer for the help in the in vivo bioassay. Y.F. thanksthe Prevent Cancer Foundation (USA) for his postdoctoral fellow-ship (Carlucci Family Research Award). This work was supportedby the National Institutes of Health (RO1-CA-134892 to F.-L.Chung).

Appendix A. Supplementary Information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032.

References

[1] Marnett, L. J. Oxy radicals, lipid peroxidation and DNA damage. Toxicology 181-182:219–222; 2002.

[2] Liebler, D. C. Protein damage by reactive electrophiles: targets and conse-quences. Chem. Res. Toxicol. 21:117–128; 2007.

[3] Pan, J.; et al. Formation of cyclic deoxyguanosine adducts from ω-3 and ω-6polyunsaturated fatty acids under oxidative conditions. Chem. Res. Toxicol.15:367–372; 2002.

[4] Nair, U.; et al. Lipid peroxidation-induced DNA damage in cancer-proneinflammatory diseases: a review of published adduct types and levels inhumans. Free Radic. Biol. Med. 43:1109–1120; 2007.

[5] Blair, I. A. DNA adducts with lipid peroxidation products. J. Biol. Chem.283:15545–15549; 2008.

[6] Chen, Y.; et al. Dietary fat-gene interactions in cancer. Cancer Metastasis Rev.26:535–551; 2007.

[7] Astorg, P. Dietary n–6 and n–3 polyunsaturated fatty acids and prostate cancerrisk: a review of epidemiological and experimental evidence. Cancer CausesControl 15:367–386; 2004.

[8] Heller, A.; et al. Lipid mediators in inflammatory disorders. Drugs 55:487–-496; 1998.

[9] Sinicrope, F. A. Targeting cyclooxygenase-2 for prevention and therapy ofcolorectal cancer. Mol. Carcinog. 45:447–454; 2006.

[10] Chung, F. -L.; et al. Deoxyguanosine adducts of t-4-hydroxy-2-nonenal areendogenous DNA lesions in rodents and humans: detection and potentialsources. Cancer Res 60:1507–1511; 2000.

[11] Nath, R. G.; et al. Detection of exocyclic 1,N2-propanodeoxyguanosine adductsas common DNA lesions in rodents and humans. Proc. Natl. Acad. Sci. USA91:7491–7495; 1994.

[12] Chung, F. -L.; et al. Deoxyguanosine adducts of t-4-hydroxy-2-nonenal areendogenous DNA lesions in rodents and humans: detection and potentialsources. Cancer Res. 60:1507–1511; 2000.

[13] Chen, H.; et al. Formation of etheno adducts in reactions of enals viaautoxidation. Chem. Res. Toxicol. 7:857–860; 1994.

[14] Chen, H.; et al. Epoxidation of trans-4-hydroxy-2-nonenal by fatty acidhydroperoxides and hydrogen peroxide. Chem. Res. Toxicol. 9:306–312; 1996.

[15] Lee, S. H.; et al. Oxidative DNA damage and cardiovascular disease. TrendsCardiovasc. Med 11:148–155; 2001.

[16] Masuda, R.; et al. High susceptibility to hepatocellular carcinoma develop-ment in LEC rats with hereditary hepatitis. Cancer Sci. 79:828–835; 1988.

[17] Moini, H.; et al. Antioxidant and prooxidant activities of α-lipoic acid anddihydrolipoic acid. Toxicol. Appl. Pharmacol. 182:84–90; 2002.

[18] Packer, L.; et al. Alpha-lipoic acid as a biological antioxidant. Free Radic. Biol.Med. 19:227–250; 1995.

[19] Podda, M.; et al. α-Lipoic acid supplementation prevents symptoms ofvitamin E deficiency. Biochem. Biophys. Res. Commun. 204:98–104; 1994.

[20] Busse, E.; et al. Influence of alpha-lipoic acid on intracellular glutathionein vitro and in vivo. Arzneim.-Forsch 42:829–831; 1992.

[21] Kozlov, A. V.; et al. Dihydrolipoic acid maintains ubiquinone in the antiox-idant active form by two-electron reduction of ubiquinone and one-electronreduction of ubisemiquinone. Arch. Biochem. Biophys. 363:148–154; 1999.

[22] Marangon, K.; et al. Comparison of the effect of α-lipoic acid andα-tocopherol supplementation on measures of oxidative stress. Free Radic.Biol. Med. 27:1114–1121; 1999.

[23] Khanna, S.; et al. α-Lipoic acid supplementation: tissue glutathione home-ostasis at rest and after exercise. J. Appl. Physiol. 86:1191–1196; 1999.

[24] White, E. L.; et al. Screening of potential cancer preventing chemicals asantioxidants in an in vitro assay. Anticancer Res. 18:769–773; 1998.

[25] Qin, G.; et al. Chemoprevention of aflatoxin B1-initiated and carbontetrachloride-promoted hepatocarcinogenesis in the rat by green tea. Nutr.Cancer 38:215–222; 2000.

[26] Yan, Y.; et al. Efficacy of Polyphenon E, red ginseng, and rapamycin on benzo(a)pyrene-induced lung tumorigenesis in A/J mice. Neoplasia 8:52–58; 2006.

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎8

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i

[27] Ui, A.; et al. Green tea consumption and the risk of liver cancer in Japan: theOhsaki Cohort study. Cancer Causes Control 20:1939–1945; 2009.

[28] Yang, C.; et al. Antioxidative and anti-carcinogenic activities of tea poly-phenols. Arch. Toxicol. 83:11–21; 2009.

[29] Lu, G.; et al. Inhibition of adenoma progression to adenocarcinoma in a4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone–induced lung tumorigen-esis model in A/J mice by tea polyphenols and caffeine. Cancer Res. 66:11494–11501; 2006.

[30] Lubet, R. A.; et al. Preventive effects of Polyphenon E on urinary bladder andmammary cancers in rats and correlations with serum and urine levels of teapolyphenols. Mol. Cancer Ther 6:2022–2028; 2007.

[31] Cook, M. G.; et al. Effect of dietary vitamin E on dimethylhydrazine-inducedcolonic tumors in mice. Cancer Res. 40:1329–1331; 1980.

[32] Factor, V. M.; et al. Vitamin E reduces chromosomal damage and inhibitshepatic tumor formation in a transgenic mouse model. Proc. Natl. Acad. Sci.USA (97):2196–2201; 2000.

[33] Meagher, E. A.; et al. Effects of vitamin E on lipid peroxidation in healthypersons. JAMA 285:1178–1182; 2001.

[34] Lonn, E.; et al. Effects of long-term vitamin E supplementation on cardiovas-cular events and cancer: a randomized controlled trial. JAMA 293:1338–1347;2005.

[35] Klein, E. A.; et al. Vitamin E and the risk of prostate cancer: the Selenium andVitamin E Cancer Prevention Trial (SELECT). JAMA 306:1549–1556; 2011.

[36] Gaziano, J.; et al. Vitamins E and C in the prevention of prostate and totalcancer in men: the Physicians' Health Study II randomized controlled trial.JAMA 301:52–62; 2009.

[37] Smolarek, A. K.; et al. Dietary administration of δ- and γ-tocopherol inhibitstumorigenesis in the animal model of estrogen receptor-positive, but not HER-2 breast cancer. Cancer Prev. Res 5:1310–1320; 2012.

[38] Ho, Y. -S.; et al. Dihydrolipoic acid inhibits skin tumor promotion throughanti-inflammation and anti-oxidation. Biochem. Pharmacol. 73:1786–1795;2007.

[39] Sodum, R. S.; et al. Structural characterization of adducts formed in thereaction of 2,3-epoxy-4-hydroxynonanal with deoxyguanosine. Chem. Res.Toxicol. 2:23–28; 1989.

[40] Bravais, F.; et al. Synthesis of 4-hydroxy[4-3 H]-2(E)-nonen-1-al-diethylacetal.J. Labelled Comp. Radiopharm 36:471–477; 1995.

[41] Sodum, R. S.; et al. Stereoselective formation of in vitro nucleic acid adductsby 2,3-epoxy-4-hydroxynonanal. Cancer Res. 51:137–143; 1991.

[42] Marmur, J. A procedure for the isolation of deoxyribonucleic acid from micro-organisms. J. Mol. Biol. 3:208–218; 1961.

[43] Carvalho, V. M.; et al. Novel 1,N6-etheno-20-deoxyadenosine adducts fromlipid peroxidation products. Chem. Res. Toxicol. 13:397–405; 2000.

[44] Chung, F. -L.; et al. Regioselective formation of acrolein-derived cyclic 1,N2-propanodeoxyguanosine adducts mediated by amino acids, proteins, and celllysates. Chem. Res. Toxicol. 25:1921–1928; 2012.

[45] Henle, E. S.; et al. Oxidative damage to DNA constituents by iron-mediatedFenton reactions: the deoxyguanosine family. J. Biol. Chem. 271:21177–21186;1996.

[46] Wang, H.; et al. Site-specific synthesis and reactivity of oligonucleotidescontaining stereochemically defined 1,N2-deoxyguanosine adducts of the lipidperoxidation product trans-4-hydroxynonenal. J. Am. Chem. Soc. 125:5687–-5700; 2003.

[47] Chung, F. -L.; et al. Lipid peroxidation as a potential endogenous source for theformation of exocyclic DNA adducts. Carcinogenesis 17:2105–2111; 1996.

[48] Nair, J.; et al. Apoptosis and age-dependent induction of nuclear andmitochondrial etheno-DNA adducts in Long-Evans Cinnamon (LEC) rats:enhanced DNA damage by dietary curcumin upon copper accumulation.Carcinogenesis 26:1307–1315; 2005.

[49] Yamamoto, F.; et al. Elevated level of 8-hydroxydeoxyguanosine in DNA ofliver, kidneys, and brain of Long-Evans Cinnamon rats. Cancer Sci. 84:508–511;1993.

[50] Laneuville, O.; et al. Fatty acid substrate specificities of human prostaglandin-endoperoxide H synthase-1 and -2: formation of 12-hydroxy-(9Z, 13E/Z, 15Z)-octadecatrienoic acids from alpha-linolenic acid. J. Biol. Chem. 270:19330–19336; 1995.

[51] Ishikawa, T.; et al. Role of cardiac glutathione transferase and of theglutathione S-conjugate export system in biotransformation of4-hydroxynonenal in the heart. J. Biol. Chem. 261:1576–1581; 1986.

[52] Williams, M. V.; et al. Endogenous lipid hydroperoxide-mediated DNA-adductformation in Min mice. J. Biol. Chem. 281:10127–10133; 2006.

[53] Douki, T.; et al. Predominance of the 1,N2-propano 20-deoxyguanosine adductamong 4-hydroxy-2-nonenal-induced DNA lesions. Free Radic. Biol. Med.37:62–70; 2004.

[54] Son, J.; et al. Surveying the damage: the challenges of developing nucleic acidbiomarkers of inflammation. Mol. Biosyst. 4:902–908; 2008.

[55] Watters, J. L.; et al. Comparison of three oxidative stress biomarkers in asample of healthy adults. Biomarkers 14:587–595; 2009.

[56] Zhou, X.; et al. Chemical and biological evidence for base propenals as themajor source of the endogenous M1dG adduct in cellular DNA. J. Biol. Chem.280:25377–25382; 2005.

[57] Lim, P.; et al. DNA damage and mutations induced by arachidonic acidperoxidation? Biochemistry 42:15036–15044; 2003.

[58] Zhang, W.; et al. Miscoding by the exocyclic and related DNA adducts 3,N4-etheno-20-deoxycytidine, 3,N4-ethano-20-deoxycytidine, and 3-(2-hydro-xyethyl)-20-deoxyuridine. Chem. Res. Toxicol. 8:157–163; 1995.

[59] Badjatia, N.; et al. Altered antioxidant status and lipid peroxidation in Indianpatients with urothelial bladder carcinoma. Urol. Oncol. 28:360–367; 2010.

[60] Ruch, R. J.; et al. Prevention of cytotoxicity and inhibition of intercellularcommunication by antioxidant catechins isolated from Chinese green tea.Carcinogenesis 10:1003–1008; 1989.

[61] Gupta, S.; et al. Inhibition of prostate carcinogenesis in TRAMP mice by oralinfusion of green tea polyphenols. Proc. Natl. Acad. Sci. USA 98:10350–10355;2001.

[62] Horvath, P. M.; et al. Synergistic effect of vitamin E and selenium in thechemoprevention of mammary carcinogenesis in rats. Cancer Res. 43:5335–5341; 1983.

[63] Nath, R. G.; et al. Effects of epigallocatechin gallate, L-ascorbic acid, alpha-tocopherol, and dihydrolipoic acid on the formation of deoxyguanosineadducts derived from lipid peroxidation. Nutr. Cancer 62:622–629; 2010.

[64] Yang, C. S.; et al. Molecular targets for the cancer preventive activity of teapolyphenols. Mol. Carcinog. 45:431–435; 2006.

[65] Ghosh, K. S.; et al. Studies on the interaction of copper complexes of(�)-epicatechin gallate and (�)-epigallocatechin gallate with calf thymusDNA. J. Inorg. Biochem. 102:1711–1718; 2008.

[66] Yamamoto, H.; et al. The antioxidant effect of DL-alpha-lipoic acid on copper-induced acute hepatitis in Long-Evans Cinnamon (LEC) rats. Free Radic. Res.34:69–80; 2001.

[67] Tafazoli, S.; et al. Prooxidant and antioxidant activity of vitamin E analoguesand troglitazone. Chem. Res. Toxicol. 18:1567–1574; 2005.

123456789

1011121314151617181920212223242526272829303132333435363738394041424344

45464748495051525354555657585960616263646566676869707172737475767778798081828384858687

Y. Fu et al. / Free Radical Biology and Medicine ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 9

Please cite this article as: Fu, Y; et al. In vivo detection of a novel endogenous etheno–DNA adduct derived from arachidonic acid andthe effects of antioxidants on its formation. Free Radic. Biol. Med. (2014), http://dx.doi.org/10.1016/j.freeradbiomed.2014.04.032i