In Vitro Characterization of DNAAdducts Formed by Foundry AirParticulate MatterK. Savela,l M.J. Kohan,2 D. Walsh,2 F.PR Perera,3K. Hemminki,4 and J. Lewtas2'lnstitute of Occupational Health, Molecular Dosimetry Group, Helsinki,Finland; 2U.S. Environmental Protection Agency, Research Triangle Park,North Carolina; 3Division of Environmental Sciences, ColumbiaUniversity, School of Public Health, New York, New York; 4Center ofNutrition and Toxicology, Karolinska Institute, Huddinge, Sweden

This study is part of an ongoing investigation of biomarkers in iron foundry workers exposed topolycyclic aromatic compounds. Foundry workers with the highest exposures had elevated levelsof DNA adducts in their white blood cells in previous studies. The purpose of this study was tocharacterize the nature of DNA reactive chemicals in foundry air samples through incubating thefoundry filter extract with DNA and activation enzymes. Calf thymus DNA was incubated withfoundry filter extract and activated by either rat liver activation mixture (S9 mix) or xanthineoxidase. A complex pattern of adducts was observed on thin-layer chromatography (TLC) by the32P-postlabeling assay. Two selected polycyclic aromatic hydrocarbons (PAHs)-1-NP- andanti(±)benzo[alpyrene-trans-7,8-dihydrodiol-9, 1 0-epoxide [anti (±)BPDEI-DNA adducts-wereused as marker compounds in characterizing the postlabeled DNA adducts by TLC combinedwith high-performance liquid chromatography (HPLC). After an initial separation of DNA adductsby TLC, individual spots were isolated and separated further on HPLC. HPLC analysis and spikingwith anti(±)BPDE-DNA standard confirmed the co-migration of the antil(±)BPDE-DNA standardwith one PAH adduct formed by the S9 mix-activated DCM extract in calf thymus DNA.

Environ Health Perspect 104(Suppl 3):687-690 (1996)

Key words: DNA adducts, PAH, in vitro CT DNA, foundry filter extract, 32P-postlabeling, HPLC

IntroductionHumans are exposed occupationally and exhausts (1-8). Both benzo[a]pyreneenvironmentally to polycyclic aromatic (B[a]P) and 1-nitropyrene (1-NP) inducecompounds including polycyclic aromatic mutations in bacteria and tumors in exper-hydrocarbons (PAHs) and nitrated PAHs. imental animals (9,10). B[a]P activated byPAHs are typical compounds in the atmos- oxidative metabolism leads to the forma-phere of coke oven, iron, and steel found- tion of epoxides and diol epoxides thating industries, and nitro-PAHs are found react with DNA to form a binding productin urban air, coal fly-ash, and diesel engine with the N2 atom of guanine. Reduction

This paper was developed from a poster that was presented at the 2nd International Conference onEnvironmental Mutagens in Human Populations held 20-25 August 1995 in Prague, Czech Republic.Manuscript received 22 November 1995; manuscript accepted 28 November 1995.

The authors wish to express their thanks to Raimo Liukkonen for carrying out the industrial hygienemeasurements, Marcia Nishioka of Batelle Columbus Laboratories for the 1-nitropyrene analysis, and RonWilliams of U.S. EPA for the B[a]P analysis. We also appreciate helpful discussions with Kimmo Peltonen inpreparing the anti-BPDE-DNA standard. This project was supported by NIEHS P01 ES05294 and the WorkEnvironment Fund of Finland.

Address correspondence to Dr. K. Savela, Institute of Occupational Health, Molecular Dosimetry Group,Topeliuksenkatu 41 aA, 00250 Helsinki, Finland. Telephone: 358-0-4707494. Fax: 358-0-4747208. E-mail:ksav@occuphealth.fi

Abbreviations used: anti(±)BPDE, anti(±)benzo[a]pyrene-trans-7,8-dihydrodiol-9,1 0-epoxide; TLC, thin-layer chro-matography; CT, calf thymus; EOM, extractable organic material; S9 mix, rat liver activation mixture; PAHs, poly-cyclic aromatic hydrocarbons; DCM, dichloromethane; HPLC, high-performance liquid chromatography; GC-MS,gas chromatography-mass spectrometry; Bla]P, benzo[alpyrene; 1-NP, 1-nitropyrene; WBC, white blood cells.

of 1-NP leads to the formation ofN-hydroxy- 1-aminopyrene that undergoesan acid-catalyzed reaction with DNA andgenerates the main reaction product of N-(deoxyguanosin-8-yl)1-aminopyrene; twoadditional 6- and 8-(deoxyguanosin-N2-yl)-1-aminopyrene have also been charac-terized (11,12). Oxidative metabolism of1-NP forms 1-NP 4,5-oxide and 1-NP9,10-oxide, which may form additionalDNA binding products (13).DNA adducts are important biomark-

ers that provide information about expo-sure to carcinogens, individual differencesin carcinogen metabolism, and DNArepair. In this study PAH and nitro-PAHadducts formed in vitro were analyzed bythe 32P-postlabeling assay combined withhigh-performance liquid chromatography(HPLC). Foundry air samples were col-lected by a high volume stationary air sam-pler placed near the iron casting. Theorganic matter adsorbed to the particleswas extracted with dichloromethane(DCM) and used for both chemical andDNA adduct analysis. Calf thymus (CT)DNA was modified in vitro with the DCMextract, and PAH and nitro-PAH wereactivated by rat liver activation mixture (S9mix) and xanthine oxidase. Specific 1-NPand anti(±)benzo[a]pyrene-trans-7,8-dihy-drodiol-9,10-epoxide [anti(±)BPDE]-DNA adduct standards were used to char-acterize the in vitro foundry air-derivedDNA adducts.

MethodsFoundry Samples

Air samples were collected in an ironfoundry in Finland by a high volume sta-tionary air sampler placed 3 to 5 metersaway from the iron casting process. Thetemperature of melted iron was 1380 to14000C. The samples were drawn on four20 x 20 cm Pallflex-filters (type T60A20;Pallflex Products Corp., Putnam, CT). Novapor-phase capture technique was usedin this study because the focus was to ana-lyze PAH bound to the particulate matter.Preparation of the extractable organicmaterial (EOM) with DCM and measure-ments of different PAHs and nitro-PAHsby HPLC and gas chromatography-massspectrometry (GS-MC) are describedelsewhere (14-16). The B[a]P concentra-tion in the composite extract (8.2 ng/mg)was approximately the level of 1-NP(10.9 ng/mg). The B[a]P concentration

Environmental Health Perspectives - Vol 104, Supplement 3 * May 1996 687

SAVELA ET AL.

corresponds to 7 to 41 ng/m3 B[a]Panalyzed from the air samples.

Modification ofCalfThymus DNAwith Foundry Filter Extractand 1-NitropyreneCT DNA (2 mg/ml) was incubated with100 pl of DCM extract (2.66mg/ml) and1-NP (20 pg/ml) under aerobic conditionswith S9 mix (0.5 mg/ml) and under anaer-obic conditions with xanthine oxidase (0.5U/ml) and hypoxanthine (0.6 mg/ml) for4 hr at 37°C as previously (14-16).

Anti(±)Benzo[a]pyrene-trans-7,8-dihydrodiol-9,10-epoxide-DNAStandard PreparationThe anti(±)BPDE-DNA standard wasprepared in 100 mM Tris-HCI buffer, pH7.0, and incubated at room temperaturefor 16 hr with CT DNA (1 mg/ml) with0.1 volume of anti(±)BPDE (1 mg/ml).The reaction mixture was extracted twicewith water-saturated butanol and withdiethylether.A. . ........ . . .: . . . . . . . ^ . . . . ... . . . . . ..... . . ..... . . . . ........ ...... ....... ..... . .. . ..

.: :.:... .. , ...... ... ..... . ... . ........ . ..... . :

. . . ....... ... .... .. .. .... . . . .. ... ... S. ..... S . . ..... .. .. ..... .. ... .. \.. ..... ..... .. : :. . < . :a:.:.:.::: ' : ::::: ^ ' 32P-Posdabeling Analysis

DNA was analyzed by the butanol extrac-

tion and nuclease P1-enhanced postlabel-ing assay according to Gupta (17) andReddy and Randerath (18). The labeledDNA adducts were resolved on 10 x 10 or

20 x 20 cm polyethylene imine (PEI) cellu-lose thin-layer chromatography (TLC)plates (Macherey-Nagel, Duren, Germany)using a standard solvent system previouslydescribed (19), with the exception that1.7 M sodium phosphate, pH 6, was usedfor D5. The radioactive areas of adductswere located by autoradiography usingintensifying screens, and adduct levels were

calculated from the amount of radioactivityon the chromatograms, DNA, and thespecific activity of [y-32P]ATP used for thelabeling (18).

HPLC AnalysisHPLC separation was carried out accordingto Pfau and Phillips (20) and King et al.(16) with slight modifications. The VarianModel 5500 (Varian, Walnut Creek, CA)

B

was used for HPLC, and radioactivity wasdetermined with the flow-through scintilla-tion counter Model IC/CR Flo-One\p(Radiomatic Instruments & Chemical Co.,Inc., Tampa, FL) using 2.5 ml/cell and ascintillator flow rate of 1.8 ml/min. Theradioactive spots were eluted from TLCplates by extracting in 400 pl of 4 M pyri-dinium formate, pH 4.0, for 18 hr. Theadduct residues were separated in an InerstilODS-2 column (5 pm particle size, 2.1mmx 150 mm; GL Sciences Inc.) with aflow rate of 0.45 ml/min. The solvent sys-tem used 0.3 M NaH2PO4/H3PO4, pH2.5, and 100% MeOH. HPLC sampleswere separated with two gradient systems:gradient A was increased with MeOH from10 to 50% in 0 to 20 min, 50 to 90% in20 to 40 min, 90 to 70% in 40 to 50 min,70 to 10% in 50 to 53 min, and to 10% in70 min; gradient B was increased withMeOH from 10 to 45% in 0 to 15 min, 45to 48% in 15 to 60 min, 48 to 90% in 60to 80 min, 90 to 10% in 80 to 81 min, andto 10% in 90 min.

C

Figure 1. Autoradiographs of 32P-postlabeled human DNA. CT DNA was modified in vitro with DCM extract, and PAH and nitro-PAH were activated by the S9 mix andxanthine oxidase. A and B, 1-NP-DNA adducts analyzed by the postlabeling assay enhanced by nuclease P1 and butanol extraction, respectively. C, In vitro-preparedanti(±)BPDE-DNA adduct standard. D and E, Diagonal radioactive zone of in vitro-modified CT DNA analyzed on 10 x 10 cm and 20 x 20 cm TLC plates, respectively (1-4 aremigrating DNA adducts). F, 32P-postlabeled DNA digest from WBC of an exposed foundry worker (black circle shows the area excised for the quantitation of adducts; adducts3 and 4 were not detected). All samples shown in B through Fwere analyzed by the butanol extraction. Autoradiography was carried out at room temperature for 1 min forA through Cand 5 hr for D and Eand at -70° C for 3 days for F

Environmental Health Perspectives - Vol 104, Supplement 3 * May 1996688

IN VITRO CHARACTERIZATION OF DNA ADDUCTS

Results and DiscussionDuring the 1980s, a series of biomarkerstudies was initiated in iron foundryworkers. DNA adducts were analyzed inwhite blood cells (WBC) of these workersby Phillips et al. (21) using a 32P-postlabel-ing assay and by Santella et al. (22) using anenzyme-linked immunoassay. In this study,B[a]P levels in the workplace atmosphereranged from 7 to 41 ng/m3, which are about50-fold lower than the B[a]P levels of 2 to 3xg/m3 measured earlier (1979-1980) in thesame foundry (23). The 1990 PAH mea-surements were determined by personalmonitors worn by the workers; the B[a]Plevels ranged from 2 to 60 ng/m3 (22).

The concentration of B[a]P in thecomposited extract of foundry air particleswas 8.2 ng/mg EOM. 1-NP was also quan-titated in this extract at the concentrationcomparable to B[a]P (10.9 ng/mg). FiguresIA and 1B show the autoradiographs fromthe 1-NP-DNA standard using bothnuclease P1 and butanol extraction. FigureIC shows the anti(±)BPDE-DNA adductstandard analyzed by butanol extraction.1-NP was only detected by the butanolextraction-enhanced 32P-postlabeling assay.The DNA adducts formed from a foundrysample show a diagonal radioactive zone ofadducts obtained on both lOx 10 and20 x 20 cm TLC from the in vitro withDCM extract-modified CT DNA usingthe butanol extraction-enhanced 32P-post-labeling assay (Figures 1D, IE). The chro-matography on the 20 x 20 cm TLC platesshowed separation more clearly (FigurelE). Foundry DNA adducts formed byxanthine oxidase were barely detectable(data not shown). In the presence of S9mix, relatively high levels of several DNAadducts were observed. DNA adductsdetected by the butanol extraction from a

300 - A

foundry worker's WBC DNA showed lowlevels of radioactivity, and no DNAadducts were detected on the area of 1-NP(Figure IF). However, DNA adductsmigrating close to the origin could corre-spond to adducts 1 and 2 detected on10 x 10 cm TLC plates from the in vitrofoundry DNA samples (Figure 1D).Further migrating DNA adducts 3 and 4

800 -

0

0cr0

co

0a

Cu

0

L0

0.0

cc

w

~0.C

-oco

~~~~~~~~~~~~~~~~~C,200

100 --

Cu

~0L- 200 -

100

0-

B ! 1-NP-DNAh' \adductI / \ anti(±)BPDE-DNA

adduct

0 10 20 30 40 50

Time, min

0 10 20 30 40 50

Time, min

(Figure 1D) were not detected in foundryworkers' WBC DNA.

TLC combined with HPLC was usedin this study to characterize PAH standardsand adducts formed in CT DNA incu-bated with foundry filter DCM extract.Anti(±)BPDE- and 1-NP-DNA adductswere eluting close together when gradientA was used in HPLC analysis (Figure 2A);

dduct 3 + antil±)BPDE-DNA

0 10 20 30 40 50

Time, min

Figure 2. HPLC analysis of anti(±)BPDE- (dashed line)and 1-NP-DNA (solid line) adduct standards. A, separa-tion using gradient A; B, the separation by the modifiedgradient B.

Figure 3. HPLC analysis of postlabeled adducts formed in vitro in CT DNA modified with DCM extract and S9 mixusing gradient A. A, C, E, and G show CT DNA adducts obtained on TLC. B, D, F, and H show, respectively, thesame adducts spiked with anti(±)BPDE-DNA adduct standard before the HPLC analysis.

Environmental Health Perspectives * Vol 104, Supplement 3 a May 1996

cn

0

C.)_

689

]I

SAVELA ET AL.

however, the modified gradient B separatedPAH standards distinctly (Figure 2B).Gradient A, however, was used further inthis study because no 1-NP adducts weredetected on TLC. The anti(±)BPDE-DNAadduct standard was used as a marker com-pound in HPLC analysis in characterizingPAH adducts formed in CT incubated withDCM extract (Figures 3A-H). Figure 3A,3C, 3E, and 3G show the HPLC analysisof four adducts obtained on both 10 x 10or 20 x 20 cm TLC plates (Figures 1 D, I E).

These in vitro adduct residues wereanalyzed first by HPLC; the same fourresidues were spiked before the HPLC withthe anti(±)BPDE-DNA adduct standard(Figures 3B, 3D, 3F, 3H). The main peakof adduct 2 is most probably derived fromB[a]P based on co-migration with theanti(±)BPDE-DNA standard (Figures 3C,3D). The development of the methodologyin analyzing nitro-PAH-DNA adducts inexposed humans by these two techniquesstill needs to be improved; however,

promising results have been reported inanalyzing PAH and nitro-PAH adductsfrom in vivo DNA (16,24-26). This studyconfirms that the 32P-postlabeling methodcombined with HPLC is an applicablemethod for the characterization of PAHadducts in in vitro DNA exposed to com-plex mixtures of PAHs. The ultimate objec-tives of these studies are to characterize andidentify specific PAH- and nitro-PAH-DNA adducts in human cells from exposedfoundry workers.

REFERENCES

1. IARC. IARC Monographs on the Evaluation of theCarcinogenic Risk of Chemicals to Humans. OverallEvaluations of Carcinogenicity: An Updating of IARCMonographs Volumes 1-42. Suppl 7. Lyon:InternationalAgency for Research on Cancer, 1987;224-225.

2. Grimmer G, Pott F. Occurrence of PAH. In: EnvironmentalCarcinogens: Polycyclic Aromatic Hydrocarbons. Chemistry,Occurrence, Biochemistry, Carcinogenicity (Grimmer G, ed).Boca Raton, FL:CRC Press, 1983;61-129.

3. Bjorseth A, Becher G, eds. PAH in Work Atmospheres:Occurrence and Determination. Boca Raton, FL:CRC Press,1986; 1-62.

4. IARC. IARC Monographs on the Evaluation of CarcinogenicRisk of Chemicals to Humans. Vol 34: Polynuclear AromaticCompounds, Part 3: Industrial Exposures in AluminumProduction, Coal Gasification, Coke Production, and Iron andSteel Founding. Lyon:International Agency for Research onCancer, 1984;133-190.

5. Gibson TL. Nitro derivatives of polynuclear aromatic hydrocar-bons in airborne and source particulate matter. Atmos Environ16:2037-2040 (1982).

6. Sweetman JA, Karasek FW, Schultze D. Decomposition ofnitropyrene during gas chromatographic-mass spectrometricanalysis of air particulate and fly-ash samples. J Chromatogr247:245-254 (1982).

7. IARC. IARC Monographs on the Evaluation of CarcinogenicRisk of Chemicals to Humans. No 46: Diesel and GasolineEngine Exhaust and Some Nitroarenes. Lyon:InternationalAgency for Research on Cancer, 1989.

8. Beland FA, Marques MM. DNA adducts of nitropolycyclicaromatic hydrocarbons. [ARC Sci Pub 125:229-244 (1989).

9. Mermelstein R, Kiriazides DK, Butler M, McCoy EC,Rosenkranz HS. The extraordinary mutagenicity of nitropy-renes in bacteria. Mutat Res 89:187-197 (1981).

10. Ohgaki H, Matsukura N, Morino K, Kawachi T, Sugimura T,Morita K, Tokiwa H, Hirota T. Carcinogenicity in rats of themutagenic compounds 1-nitropyrene and 3-nitrofluoranthene.Cancer Lett 15:1-17 (1981).

11. Beland FA, Smith BA, Thornton-Manning JR, Heflich RH.Metabolic activation of 1-nitropyrene to a mammalian cellmutagen and carcinogen. Xenobiotica 22(9/10):1121-1133(1992).

12. Howard PC, Heflich RH, Evans FE, Beland FA. Formation ofDNA adducts in vitro and in Salmonella typhimurium uponmetabolic reduction of the environmental mutagen 1-nitropy-rene. Cancer Res 43:2051-2058 (1983).

13. Smith BA, Heflicht RH, Ohnishi Y, Ohuchida A, Kinouchi T,Thorton-Manning JR, Beland FA. DNA adduct formationby 1-nitropyrene 4,5- and 9,10-oxide. In: Nitroarenes:Occurrence, Metabolism, and Biological Impact (Howard PC,Hecht SS, Beland FA, eds). New York:Plenum Press,1990;181-187.

14. Gallagher JE, Kohan MJ, George MH, Lewtas J. Improvementin the diagnostic potential of 3 P-postlabeling analysis demon-strated by the sensitive formation and comparative analysis ofnitrated-PAH-derived adducts arising from diesel particleextracts. Carcinogenesis 12:1685-1691 (1991).

15. Gallagher J, George M, Kohan M, Thompson G, Shank T,Lewtas J. Detection and comparison of DNA adducts after invitro and in vivo diesel emission exposures. Environ HealthPerspect 99:225-228 (1993).

16. King LC, George M, Gallagher J, Lewtas J. Separation of 32p_postiabeled DNA adducts of polycyclic aromatic hydrocarbonsand nitrated polycyclic aromatic hydrocarbons by HPLC.Chem Res Toxicol 7:503-510 (1994).

17. Gupta RC. 32P-postlabeling analysis of bulky aromatic adducts.IARC Sci Publ 124:11-23 (1993).

18. Reddy MV, Randerath K. Nuclease P1-mediated enhancementof sensitivity of 32P-postlabeling test for structurally diverseDNA adducts. Carcinogenesis 7:1543-1551 (1986).

19. Gallagher JE, Jackson MA, George MH, Lewtas J, RobertsonICG. Differences in detection ofDNA adducts in the 32P-post-labeling assay after either 1-butanol or nuclease P1 treatment.Cancer Lett 45:7-12 (1989).

20. Pfau W, Phillips DH. Improved reversed-phase high-perfor-mance liquid chromatographic separation of32P-labelled nucle-oside 3',5'-bisphosphate adducts of polycyclic aromatichydrocarbons. J Chromatogr 570:65-76 (1991).

21. Phillips DH, Hemminki K, Alhonen A, Hewer A, Grover PL.Monitoring occupational exposure to carcinogens: detection by32P-postlabelling of aromatic DNA adducts in white blood cellsfrom iron foundry workers. Mutat Res 204:531-541 (1988).

22. Santella RM, Hemminki K, Tang D-L, Paik M, Ottman R,Young TL, Savela K, Vodickova L, Dickey C, Whyatt R,Perera FP. Polycyclic aromatic hydrocarbon-DNA adducts inwhite blood cells and urinary 1-hydroxypyrene in foundryworkers. Cancer Epidemiol Biomarkers Prev 2:59-62 (1993).

23. Perera FP, Dickey C, Santella R, O'Neill JP, Albertini RJ,Ottman R, Tsai WY, Mooney LA, Savela K, Hemminki K.Carcinogen-DNA adducts and gene mutation in foundryworkers with low-level exposure to polycyclic aromatic hydro-carbons. Carcinogenesis 15:2905-2910 (1994).

24. Smith BA, Fullerton NF, Heflich RH, Beland FA. DNAadduct formation and T-lymphocyte mutation induction inF344 rats implanted with tumorigenic doses of 1,6-dinitropy-rene. Cancer Res 55:2316-2324 (1995).

25. Savela K, King L, Gallagher J, Lewtas j. 32P-postlabeling andHPLC separation of DNA adducts formed by diesel exhaustextracts in vitro and in mouse skin and lung after topical treat-ment. Carcinogenesis 16:2083-2089 (1995).

26. Moller L, Zeising M. DNA adduct formation after oral admin-istration of 2-nitrofluorene and N-acetyl-2-aminofluorene,analyzed by 32-TLC and 32-HPLC. Carcinogenesis 14:53-59(1993).

690 Environmental Health Perspectives * Vol 104, Supplement 3 * May 1996