Multi-DNA Adduct and Abasic Site Quantitation in vivo by Nano-Liquid Chromatography/High-Resolution Orbitrap Tandem Mass Spectrometry: Methodology for Biomonitoring Colorectal DNA Damage

Abstract

Epidemiological and mechanistic studies suggest that processed and red meat consumption and tobacco smoking are associated with colorectal cancer (CRC) risk. Several classes of carcinogens, including N-nitroso compounds (NOCs) in processed meats and heterocyclic aromatic amines (HAAs) and polycyclic aromatic hydrocarbons (PAHs) in grilled meats and tobacco smoke, undergo metabolism to reactive intermediates that may form mutation-inducing DNA adducts in the colorectum. Heme iron in red meat may contribute to oxidative DNA damage and endogenous NOC formation. However, the chemicals involved in colorectal DNA damage and the paradigms of CRC etiology remain unproven. There is a critical need to establish physicochemical methods for identifying and quantitating DNA damage induced by genotoxicants in human colorectum. We established robust nano-liquid chromatography/high-resolution accurate mass Orbitrap tandem mass spectrometry (LC/HRAMS²) methods to measure DNA adducts of nine meat and tobacco-associated carcinogens, and lipid peroxidation products in the liver, colon, and rectum of carcinogen-treated rats employing fresh-frozen and formalin-fixed-paraffin-embedded (FFPE) tissues. Some NOCs form O⁶-carboxymethyl-2′-deoxyguanosine and O⁶-methyl-2′-deoxyguanosine, and unstable quaternary N-linked purine/pyrimidine adducts which generate apurinic/apyrimidinic (AP) sites. AP sites were quantitated following derivatization with O-(pyridin-3-yl-methyl)hydroxylamine. DNA adduct quantitation was conducted with stable isotope-labeled internal standards, and method performance was validated for accuracy and reproducibility. Limits of quantitation ranged from 0.1 to 1.1 adducts per 10⁸ bases using 3 μg of DNA. Adduct formation in animals ranged from ~1 in 10⁸ to ~1 in 10⁵ bases, occurring at comparable levels in fresh-frozen and FFPE specimens for most adducts. AP sites increased by 25- to 75-fold in the colorectum and liver, respectively. Endogenous lipid peroxide-derived 3-(2-deoxy-β-D-erythro-pentofuranosyl)pyrimido[1,2-α]purin-10(3H)-one (M₁dG) and 6-oxo-M₁dG adduct levels were not increased by carcinogen dosing but increased in FFPE tissues. Human biomonitoring studies can implement LC/HRAMS² assays for DNA adducts and AP sites outlined in this work to advance our understanding of CRC etiology.

Graphical Abstract

Introduction

Epidemiologists have observed an association between meat consumption and colorectal cancer (CRC) risk since the 1970s.¹ In 2015, the Working Group of the International Agency for Research on Cancer (IARC) classified red meats as Group 2A carcinogens (possibly carcinogenic to humans) and cured meats as Group 1 carcinogens (carcinogenic to humans) for CRC.² These classifications are based on a compilation of epidemiological data and mechanistic evidence from human cohort studies and animal models; however, the causative chemicals responsible for CRC etiology remain uncertain. Meat products contain several classes of carcinogens, including N-nitroso compounds (NOCs) in processed meats and heterocyclic aromatic amines (HAAs) and polycyclic aromatic hydrocarbons (PAHs) in grilled red meats. These compounds are procarcinogens that undergo enzymatic bioactivation to produce reactive species that form DNA adducts. If not repaired, some DNA adducts formed in cancer-driver genes can induce mutations during cell division leading to the onset of cancer.^3,4 Some of these carcinogens and structurally similar compounds are also found in tobacco smoke; smoking is a risk factor for CRC.^5–7

Cooking meat at high temperatures forms HAAs and PAHs.⁸ Dietary exposure to benzo[a]pyrene (B[a]P), the most well-studied PAH, is associated with CRC risk.⁹ B[a]P bound to human colorectal DNA was previously identified by an HPLC/fluorescence method.¹⁰ 2-Amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), the most mass-abundant carcinogenic HAA formed in cooked meats, is a colorectal and prostate carcinogen in rodents and is implicated in human CRC.^11–13 PhIP principally forms an adduct at the C8 position of dG (N-(2′-deoxyguanosin-8-yl)-PhIP) (dG-C8-PhIP), which has been detected in human prostate by LC/HRAMS.¹⁴ Putative adducts of PhIP were also identified by gas chromatography/mass spectrometry following alkaline hydrolysis of human colorectal DNA.¹⁵ The cooked meat HAA 2-amino-9H-pyrido[2,3-b]indole (AαC) is a potent mouse lacI transgene colon mutagen¹⁶ and induces colonic aberrant crypt foci in mice.¹⁷ 2-Amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) is another mutagenic HAA and one of its major DNA adducts, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline N-(2′-deoxyguanosin-8-yl-C8-MeIQx) (dG-C8-MeIQx), was detected at low levels in human colon by ³²P-postlabeling.¹⁸ Tobacco smoke also contains several aromatic amines, including 4-aminobiphenyl (4-ABP), HAAs, and PAHs, some of which may contribute to CRC in smokers (Figure 1).^5,19

Figure 1.

DNA adducts formed with cooked meat and environmental carcinogens (dR = 2′-deoxyribose).

Processed and cured meats are often treated with sodium nitrite, preserving the meat and providing characteristic flavors and color.²⁰ This curing process can produce N-nitroso compounds (NOCs), some of which are implicated in CRC.^2,21–23 NOCs can alkylate DNA following cytochrome P450 (P450) metabolism. For example, N-nitrosodimethylamine forms the mutation-prone lesion O⁶-methyl-2′-deoxyguanosine (O⁶MedG), while nitrosated glycine derivatives form O⁶-carboxymethyl-2′-deoxyguanosine (O⁶CMdG);²⁴ both adducts may contribute to CRC (Figure 1).^25,26 A mutational signature attributed to alkylating agents was found in CRC patients and associated with high red meat intake.²⁷ Carboxymethylating agents induce mutational spectra in p53 consistent with mutations in stomach cancer and CRC patients.²⁸ Several studies have detected O⁶MedG and O⁶CMdG in human colorectum^26,29,30 and blood³¹ by immunoassays, but there was no corroboration of these lesions’ identities by specific mass spectrometry-based approaches. Heme iron is postulated to be a driver of endogenous NOC formation from red meat intake by catalyzing N-nitrosation of amines.³²

Gut inflammation caused by ulcerative colitis and Crohn’s disease is also a risk factor for CRC.³³ Chronic inflammation produces reactive oxygen species (ROS) and reactive nitrogen species (RNS).³³ Dietary heme in red meats may also promote a pro-oxidative environment and generate ROS in the gastrointestinal tract.^34,35 These ROS can oxidize lipids, forming electrophilic lipid peroxidation products (LPOs), which covalently bind to DNA, leading to mutagenic lesions. One such lesion is M₁dG,³⁶ which is also formed by base propenals generated via oxidative DNA damage.³⁷ M₁dG can be oxidized by aldehyde oxidase and xanthine oxidase to form 6-oxo-M₁dG in RKO cells, a human colon carcinoma cell line.³⁸ M₁dG has been detected in multiple human tissues by MS methods,^36–38 and in normal human colon mucosa by an immunoslot blot assay.³⁹

N7 alkylation of guanine by NOCs is typically more prevalent than adduction at the O⁶ position.^40,41 These cationic N7 adducts, and other quaternary N-linked purine and pyrimidine adducts are unstable and undergo depurination to yield apurinic/apyrimidinic (AP sites).⁴² AP sites are cytotoxic⁴³ and mutagenic⁴⁴ lesions if not repaired. Scheme 1 outlines how bioactivated NOCs form stable O⁶-guanine adducts and unstable depurinating adducts leading to AP sites. Recently, our lab developed a quantitative liquid chromatography/mass spectrometry method to measure AP sites in DNA while mitigating artifactual AP site formation during sample processing.⁴⁵

Scheme 1.

Formation of O⁶MedG, O⁶CMdG, and AP sites from NOCs following P450 bioactivation.

The measurement of DNA adducts in the human colorectum has been performed primarily by immuno-detection or ³²P-postlabeling.^{18,26,29,30,39} However, immuno-detection is prone to cross-reactivity, and ³²P-postlabeling is not a specific detection method. Neither technique provides structural data to confirm adduct identity. There is a critical need for specific mass spectrometry-based methods to identify the chemicals in the meat diet and environment which form DNA adducts that may contribute to CRC. Identifying mutation-prone DNA adducts in colorectum can improve our understanding of the chemicals involved in CRC and advance public health intervention efforts.

We have used carcinogen-treated rodent models to develop robust nano-liquid chromatography/high-resolution accurate mass Orbitrap tandem mass spectrometry (LC/HRAMS²) methods to assay nine DNA adducts formed by meat and tobacco-associated carcinogens and resulting AP sites (Figure 1). DNA adducts were measured in fresh-frozen and formalin-fixed paraffin-embedded (FFPE) liver, colon, and rectum. Fresh-frozen tissues are often not available for DNA adduct biomarker research. FFPE tissues, however, are often accessible and are an under-utilized biospecimen for carcinogen DNA adduct biomarker research. Our laboratory has pioneered methods to unravel crosslinked DNA from FFPE tissues to measure DNA adducts.^46–48 A combined approach of quantitative multi-targeted analysis of carcinogen DNA adducts and AP sites will lay the groundwork for future human biomonitoring studies to help better understand the link between chronic exposures and CRC.

Experimental Procedures

Materials.

Calf thymus (CT) DNA, RNAse A (bovine pancreas), RNase T1 (Aspergillus oryzae), Proteinase K (Tritirachium album), DNase I (type IV, bovine pancreas), alkaline phosphatase (Escherichia coli), nuclease P1 (Penicillium citrinum), azoxymethane (AOM), azaserine (AS), benzo[a]pyrene (B[a]P), and 4-aminobiphenyl were purchased from Sigma-Aldrich (St. Louis, MO). Phosphodiesterase I (Crotalus adamanteus venom) was purchased from Worthington Biochemical Corp. (Newark, NJ). Phosphate buffered saline (pH 7.4) (PBS), phenylmethylsulfonyl fluoride (PMSF), and 10% buffered formalin were purchased from Fisher Scientific (Waltham, MA). Maxwell 16 FFPE Plus LEV DNA purification kits were purchased from Promega Co. (Madison, WI). All solvents were of LC/MS grade and purchased from Thermo Fisher Scientific (Waltham, MA). 2′-Deoxyguanosine (dG) monohydrate was purchased from Alfa Aesar (Ward Hill, MA). ¹⁵N₅-dG monohydrate (98% isotopic purity) was purchased from Cambridge Isotope Laboratories (Tewksbury, MA). 2-Amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), 2-amino-9H-pyrido[2,3-b]indole (AαC), and 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) were purchased from Toronto Research Chemicals (Toronto, Ontario, CA). Chromacol fused insert glass sample vials (200 μL) were purchased from Thermo Fisher Scientific (Waltham, MA) and used for all samples analyzed by LC/HRAMS². Oasis HLB (1cc) solid-phase extraction (SPE) cartridges were purchased from Waters (Milford, MA). Strata X reversed-phase polymeric SPE (30 mg) cartridges were purchased from Phenomenex (Torrance, CA). Puregene Protein Precipitation solution was bought from Qiagen (Hilden, Germany). If not stated otherwise, all other chemicals were obtained from Sigma-Aldrich (St. Louis, MO).

Synthesis of DNA adduct standards.

Unlabeled and stable isotopically-labeled internal standards of [¹³C₁₀]-dG-C8-4-ABP, [¹³C₁₀]-dG-N²-BPDE, [¹³C₁₀]-dG-C8-AαC, [²H₃]-dG-C8-MeIQx, and [¹³C₁₀]-dG-C8-PhIP were prepared as previously described.^15,49,50 The N-acetoxy-HAA derivatives were reacted with dG or [¹³C₁₀]-dG in 100 mM potassium phosphate buffer (pH 8.0) and purified by preparative HPLC. Unlabeled and [¹⁵N₅]-labeled O⁶CMdG were synthesized as reported by Geigle and colleagues.⁵¹ Briefly, ethyl diazoacetate, CuSO₄, and sodium ascorbate were reacted with dG or [¹⁵N₅]-dG in 2-(N-morpholino)ethanesulfonic acid buffer (pH 6.0), and the product was purified by semi-preparative HPLC. Labeled and unlabeled O⁶MedG were synthesized as previously reported: 6-chloropurine was converted to O⁶-methylguanine (O⁶MeG) using metallic sodium in MeOH or [²H₃C]OH;⁵² then, the modified base was enzymatically coupled to dG using thymidine phosphorylase and purine nucleotide phosphorylase.⁵³ 3-(2-Deoxy-β-D-erythro-pentofuranosyl)pyrimido[1,2-α]purin-10(3H)-one (M₁dG) and 6-oxo-M₁dG were kindly provided by Dr. Lawrence Marnett and Dr. Philip Kingsley, Vanderbilt, University.

Animal dosing and tissue collection.

SAS Fischer 344 rats (strain 403) were bought from Charles River Lab (Wilmington, MA). Animals were acclimated for one week under standard conditions on an AIN-76A diet before dosing. In the high-dose multi-carcinogen (MC)-dosed animal study, rats were given an intraperitoneal injection of PhIP, AαC, MeIQx, 4-ABP, AOM, and AS at 5 mg/kg body weight, and B[a]P at 2.5 mg/kg in 80% DMSO. Twelve rats (six male, six female) were dosed with the carcinogens, and eight negative control animals (four male, four female) were dosed with an equivalent volume of 80% DMSO only. A follow-up study was done wherein two lower doses of the alkylating agents AS (1 mg/kg and 5 mg/kg) and AOM (10 μg/kg and 50 μg/kg) were administered in 50% DMSO. Four male rats were dosed at each level, and two negative control rats were treated with DMSO.

After 24 h, animals were euthanized by CO₂, and liver, colon, and rectum tissues were harvested. The left and right lateral lobes of the liver were separated. The colon and rectum were washed with chilled PBS (pH 7.4) containing PMSF (40 mg/L) to remove fecal debris. Half of each lateral lobe of the liver, colon, and rectum, cut lengthwise, were placed in cups of 10% neutral buffered formalin and fixed for 24 h. Following dehydration with ethanol and xylene, the tissues were embedded in paraffin blocks.⁴⁸ The remaining fresh tissues were snap-frozen on dry ice and stored at −80 °C until assayed.

DNA extraction from fresh-frozen tissues.

Fresh-frozen tissues were weighed and homogenized using a blade homogenizer (Pro Scientific, Oxford, CT) in 10 mM Tris-EDTA (50 mM Tris, 10 mM EDTA, pH 8.0; TE) buffer containing 10 mM β-mercaptoethanol (BME). Aliquots of tissue homogenate (20 mg) were centrifuged at 3,000 x g for 10 min at 4 °C and reconstituted in of TE (300 μL) containing 10 mM BME. RNase A (15 μL, 10 mg/mL) and RNase T1 (1 μL, 1000 U/mL) were added, followed by thermomixer agitation at 37 °C for 1.5 h. Proteinase K (20 μL, 20 mg/mL) and sodium dodecyl sulfate (40 μL, 10% in H₂O) were added, and the incubation continued at 37 °C for 3 h. TE buffer (300 μL) with 10 mM BME was added to each sample and placed on ice to cool. Protein Precipitation Solution (250 μL, pH 8.0) was added to each sample, then vortexed thoroughly and placed at −20 °C for 2 h. Samples were centrifuged for 30 min at 21000 x g at 4 °C, and the supernatants were added to a new Eppendorf tube. The DNA was precipitated by adding 0.1 vol of 5 M NaCl, followed by 1.5 vol of chilled isopropanol. Samples were mixed thoroughly and placed at −20 °C for 18 h. Tubes were centrifuged at 21,000 x g at 4 °C for 30 min, and the supernatant was discarded. DNA pellets were washed with 70% ethanol twice, dried, and reconstituted in LC/MS grade H₂O. DNA concentration was measured by a UV-Vis spectrophotometer. An absorbance of 1.0 at 260 nm for 50 μg/mL of double-stranded DNA was used to determine DNA concentration.

DNA extraction from FFPE tissues.

Whole tissues were removed from FFPE blocks, and excess paraffin was trimmed with a scalpel. Tissues were placed into glass vials containing p-xylene (3.5 mL) and shaken at room temperature for 1 h. The solvent was discarded, and this process was repeated once. Rehydration of deparaffinized tissues was done using an ethanol-water gradient ending with 100% H₂O, as previously reported.⁴⁶ Rehydrated tissues were weighed and homogenized as outlined above for fresh-frozen tissues. Aliquots of tissue homogenates (20 mg wet weight) were centrifuged at 3,000 x g for 10 min at 4 °C to obtain nuclear pellets. Proteinase K (20 μL, 20 mg/mL) and Incubation Buffer (180 μL, Maxwell 16 FFPE Plus LEV DNA purification kit, pH 8.0) with freshly added 10 mM BME were added. The mixture was incubated for 18 h at 50 °C with agitation. RNAse A (15 μL, 10 mg/mL) was added to each sample and incubated at room temperature for 10 min. Lysis Buffer (Maxwell 16 FFPE Plus LEV DNA purification kit, 2 vol. with 10 mM BME, pH 7.0) was added to each sample. Samples were vortexed thoroughly and allowed to sit at room temperature for 30 min before processing by the automated Maxwell 16 system (Promega, Madison, WI) using the default FFPE Plus LEV DNA purification program. The DNA concentration was measured by a UV-Vis spectrophotometer.

DNA digestion and SPE.

DNA (20 μg) containing isotopically labelled internal standards at a level of 5 adducts per 10⁸ bases ([²H₃]-O⁶MedG, [¹⁵N₅]-O⁶CMdG, [¹⁵N₅]-M₁dG, [¹⁵N₅]-6-oxo-M₁dG, [¹³C₁₀]-dG-C8-4-ABP, [¹³C₁₀]-dG-N²-BPDE, [¹³C₁₀]-dG-C8-AαC, [²H₃]-dG-C8-MeIQx, [¹³C₁₀]-dG-C8-PhIP) in 5 mM Bis-Tris (pH 7.1) and 10 mM MgCl₂ were digested with DNase I (5 μL, 1 mg/mL) and nuclease P1 (1 μL, 0.5 mg/mL) at 37 °C for 3.5 h at 750 rpm in a thermomixer. Phosphodiesterase I (1 μL, 0.05 mg/mL) and alkaline phosphatase (2 μL, 1 mg/mL) were added, and the incubation continued at 37 °C overnight. Half of each sample (10 μg DNA) was transferred to a new tube with 5 vol of chilled ethanol and kept at −20 °C for at least 2 h before centrifugation at 21,000 x g for 10 min at 4 °C. The supernatants were retrieved and dried by vacuum centrifugation and reconstituted in 20 μL 1:1 H₂O: DMSO for trapping-mode analysis. A portion of the sample (2.5%) was reserved for evaluating DNA digestion efficiency by HPLC-UV as previously described.⁵⁴

The remaining DNA samples (10 μg were diluted with 0.1% HOAc (1 mL final vol) and loaded onto Oasis HLB 1 cc SPE cartridges, which were pre-conditioned with MeOH/0.1% HOAc v/v (1 mL), and H₂O/0.1% HOAc v/v (1 mL). After application of the samples, the cartridges were sequentially washed with 0.1% HOAc v/v (1 mL) and 10% MeOH/0.1% HOAc v/v (1 mL), then eluted with MeOH (1 mL) into Eppendorf tubes. The eluents were dried by vacuum centrifugation, reconstituted in MeOH (50 μL), vortexed, centrifuged at 21,000 x g for 5 min, transferred to glass LC vials and dried by vacuum centrifugation. Samples were reconstituted in H₂O (10 μL) for direct injection analysis.

AP site derivatization, DNA extraction, and SPE.

Derivatization of AP sites was done as previously reported.⁴⁵ Nuclear pellets (the equivalent of 20 mg wet weight tissue) were reconstituted in HEPES EDTA (500 μL, 50 mM HEPES, 10 mM EDTA, pH 8.0; HE). Sodium dodecyl sulfate (10% w/v, 60 μL) and Proteinase K (20 μL, 20 mg/mL) were added, and samples were incubated at 37 °C for 1.5 h. O-(Pyridin-3-yl-methyl)hydroxylamine (32 μL, 100 mM; PMOA) was added, and incubation continued for 90 min before quenching with butyraldehyde (30 μL, 1 M in 1:1 isopropanol: H₂O). Puregene Protein Precipitation solution (250 μL, pH 8.0) was added, and samples were vortexed vigorously and placed on ice for 4 min, followed by centrifugation for 10 min at 2,000 x g at 4 °C. The supernatants were decanted and precipitated by adding NaCl (0.1 vol, 5 M) and isopropanol (1 mL) and centrifuged at 21,000 x g at 4 °C for 10 min. DNA pellets were washed twice with 70% ethanol in 5 mM HEPES buffer (pH 8.0). DNA pellets were dried and reconstituted in HE (250 μL) and incubated with RNase A (15 μL, 10 mg/mL) and RNase T1 (1 μL, 1000 U/mL) at 37 °C for 1.5 h. Enzymes were precipitated with Puregene Protein Precipitation Solution, and samples were centrifuged for 5 min at 13,000 x g at 4 °C. DNA was precipitated from the decanted supernatant with isopropanol and NaCl and washed twice, as described above. DNA was reconstituted in 5 mM HEPES buffer pH 8.0. DNA concentration was measured by a UV-Vis spectrophotometer.

DNA (20 μg) containing isotopically labelled internal standard PMOA-[¹³C₅]dR at 3.3 adducts per 10⁶ bases (5 mM HE, 15 mM MgCl₂, 5 mM CaCl₂, 100 μL) was digested with DNAse 1 (0.5 μL, 10 mg/mL) and nuclease P1 (0.5 μL, 1 mg/mL) at 37 °C for 3 h, followed by adding phosphodiesterase 1 (1 μL, 0.25 μg/ μL) and alkaline phosphatase (2 μL, 1 mg/mL) for 15 h. Then, the temperature was lowered to 25 °C, and the digest was incubated with adenosine deaminase (2.4 μL, 5 U/mL) for 1.5 h. Samples were diluted with H₂O (1 mL) for SPE.

Samples were loaded on the Strata X SPE cartridges (30 mg), which were pre-washed with MeOH (1 mL) followed by H₂O (1 mL). Cartridges were washed with 8% MeOH (2.5 mL) in 2 mM ammonium bicarbonate (pH 7.8). The derivatized AP site was eluted with (1 mL, 99:1 MeOH:2 mM ammonium bicarbonate (pH 7.8)) and vacuum centrifuged to dryness at 45 °C. Samples were reconstituted in 2 mM ammonium bicarbonate (pH 7.8, 20 μL) in glass LC vials for LC/HRAMS² analysis.

LC/HRAMS² Orbitrap analysis.

Trapping-mode analysis of hydrophobic HAA DNA adducts was performed on an Orbitrap Fusion Lumos MS (Thermo Fisher Scientific, San Jose, CA) equipped with a PicoChip nanoESI source (New Objective, Woburn, MA), interfaced with a Dionex UltiMate 3000 RSLCnano UHPLC System (Thermo Fisher Scientific, San Jose, CA). An Acclaim PepMap RP 18 (0.3 x 5 mm, 5 μm particle size, Thermo Fisher Scientific, San Jose, CA) trap column was employed for the online trapping of DNA adducts. The DNA digests (3 μg DNA) were injected and trapped for 4 min at a flow rate of 12 μL/min and then back-flushed at a flow rate of 0.6 μL/min onto a PicoChip Reprosil-Pur C18-AQ analytical column (75 μm x 100 mm, 3 μm particle size, 120 Å, New Objective, Woburn, MA). The mobile phases were A, 0.05% formic acid in H₂O, and mobile phase B, 0.05% formic acid in 95% acetonitrile containing 5% H₂O. The flow rate was held constant at 0.6 μL/min at 1% B for 4 min, followed by a linear 15 min gradient to 99% B, and held for 3 min before dropping to 1% B over two min, followed by 8 min of equilibration time.

Data were acquired by Xcalibur version 4.4 (Thermo Fisher Scientific, San Jose, CA). Adduct analysis was performed in positive ion mode at the MS² scan stage. The MS source parameters were: spray voltage, 2200 V, ion transfer tube temperature, 275 °C, RF lens, 40%, AGC target, 4 x 10⁵, maximum injection time, 50 ms. The MS² acquistion parameters were: Quadrupole isolation, isolation window, m/z 1.6, normalized higher-energy collision dissociation (HCD) energy, 25%, Orbitrap resolution, 30,000, RF lens, 40%, AGC target, 5 x 10⁴, maximum injection time, 54 ms. The targeted MS² ion transitions used to construct the extracted ion chromatograms (EICs) for adduct quantifications were: dG-C8-PhIP (m/z 490.2 → 374.1470), [¹³C₁₀]-dG-C8-PhIP (m/z 500.2 → 379.1640); dG-C8-4-ABP (m/z 435.2 → 319.1302), [¹³C₁₀]-dG-C8–4-ABP (m/z 445.2 → 324.1470); dG-C8-AαC (m/z 449.2 → 333.1207), [¹³C₁₀]-dG-C8-AαC (m/z 459.2 → 338.1375); dG-C8-MeIQx (m/z 479.2 → 363.1425), [²H₃C]-dG-C8-MeIQx (m/z 482.2 → 366.1613), dG-N²-BPDE (m/z 570.2 → 257.0959, 285.0916, 454.1510), [¹³C₁₀]-dG-N²-BPDE (m/z 580.2 → 257.0959, 285.0916, 459.1678).

Direct injection analysis was performed using the same UHPLC system, mobile phase solvents, and mass spectrometer equipped with a Nanospray Flex ion source (Thermo Fisher Scientific, San Jose, CA). DNA samples (3 μg) were injected on a home-packed Synergi Hydro RP18 (75 μm x 150 mm, 3 μm particle size, Phenomenex, Torrance, CA) analytical column. The gradient commenced with 1% B for 4 min at a flow rate of 0.6 μL/min before dropping the flow rate to 0.3 μL/min over 0.5 min. A linear gradient was then used over 12 min from 1% to 40% B, increased to 99% B over two min, and held for two min at 99% B. The flow rate was increased to 0.6 μL/min, and the B composition decreased to 1% over 1.5 mins, followed by 5 min of equilibration time.

The MS source parameters were: spray voltage, 2200 V, ion transfer tube temperature, 300 °C, RF lens, 50%, AGC target 5 x 10⁴, maximum injection time, 50 ms. MS² acquisition parameters were: quadrupole isolation, isolation window, m/z 1.6, resolution, 30,000, RF lens, 50%, AGC target, 4 x 10⁵, maximum injection time, 54 ms. Normalized HCD was 30% for all analytes except for M₁G and 6-oxo-M₁G, where 55% was used. Targeted MS² ion transitions used to construct (EICs) were: O⁶MedG (m/z 282.1 → 166.0720), [²H₃]-O⁶MedG (m/z 285.1 → 169.0912); O⁶CMdG (m/z 326.1 → 210.0622), [¹⁵N₅]-O⁶CMdG (m/z 331.1 → 215.0473). The M₁dG and 6-oxo-M₁dG glycoside linkages are labile under the in-source MS conditions and their aglycones ([M+2H-dR]⁺) were measured at the MS² scan stage: M₁G (m/z 188.0 → 97.0396), [¹⁵N₅]-M₁G (m/z 193.0 → 99.0337), 6-oxo-M₁G (m/z 204.0 → 153.0408, 135.0302), [¹⁵N₅]-6-oxo-M₁G (m/z 209.0 → 157.02761, 139.01719).

The AP site analysis was performed with the same UHPLC system and Orbitrap Fusion Lumos MS using the EASY-Spray^™ nano ion source (Thermo Fisher Scientific, San Jose, CA). Mobile phase A was 2 mM NH₄OAc pH 6.8 in H₂O, and mobile phase B was 100% acetonitrile. The flow rate was constant at 0.6 μL/min with 1% B for 4 min, followed by a linear 15-min gradient to 99% B. The solvent composition was held at 99% B for 3 min before decreasing to 1% B over two min, followed by 6 min of equilibration time. The separations were done on a Prontosil C18AQ analytical column (100 μm x 150 mm, 3 μm particle size; nano-LCMS solutions, Gold River, CA) interfaced with an EASY-Spray nano-flow emitter (Thermo Fisher Scientific, San Jose, CA).

PMOA-dR was assayed in positive ion mode at MS² scan stage with HCD. The MS source parameters were: spray voltage, 2500 V, ion transfer tube temperature, 300 °C, RF lens, 40%, AGC target, 5 x 10⁴, maximum injection time, 54 ms. The MS² acquisition parameters were: quadrupole isolation, isolation window, m/z 1.6, HCD collision energy, 40%, resolution, 30,000, RF lens, 40%, AGC target, 4 × 10⁵, maximum injection time, 54 ms. MS² ion transitions were as follows: PMOA-dR (m/z 241.1 → 108.04430), [¹³C₅]-PMOA-dR (m/z 246.1 → 108.04430).

Statistics.

Statistics on DNA adduct levels were performed with Prism 8.4.3 (GraphPad Software, La Jolla, CA). The statistical significance of DNA adduct levels between fresh-frozen and FFPE samples was determined by the unpaired Student’s t-test. The statistical significance of the AP site levels formed as a function of the carcinogen dose in fresh-frozen tissues against the untreated control group was determined by a one-way ANOVA followed by Dunnett’s Multiple Comparison. The data are reported as the mean ± SD (*p < 0.05; **p < 0.01, ***p < 0.005).

Results

DNA adduct enrichment method development.

Two different DNA adduct enrichment procedures were required to isolate the structurally diverse array of DNA adducts from the large excess of unmodified 2′-deoxynucleosides (dNs). Adducts formed by HAAs, AAs, and PAHs are bulky and hydrophobic, and relatively facile to separate from unmodified dNs. However, NOC-derived O⁶MedG and O⁶CMdG and LPO-derived M₁dG and 6-oxo-M₁dG adducts are smaller molecules and significantly more polar, with O⁶CMdG having a negative charge at neutral pH. These analytes are difficult to separate from unmodified dNs, which are present at greater than a million-fold excess in the DNA digest. Previous studies used offline HPLC fractionation⁵⁵ or immunoaffinity columns²⁵ to enrich O⁶CMdG before LC/MS. These methods are time-consuming, not amenable to large-scale human studies, and prone to contamination. We developed an SPE enrichment scheme using an acidic pH solvent, which recovers > 90% of O⁶MedG and O⁶CMdG while removing > 98% of the non-modified dNs. This technhique is the first reported SPE enrichment method for the simultaneous recovery of O⁶MedG and O⁶CMdG. Bulky DNA adducts such as dG-C8-PhIP, dG-C8-4-ABP, and dG-N²-BPDE are poorly recovered from various C18 and polymeric SPE resins (D. Konorev, unpublished data). Therefore, these adducts were enriched by online trapping.

High-purity DNA was isolated from FFPE tissues and efficiently digested to the canonical dNs using a cocktail of nucleases (Figure S1). The level of RNA contamination was < 1%. Mean DNA recovery from fresh-frozen tissues ranged from 43.6 μg in the liver to 68.5 μg in the rectum when processing 20 mg wet weight of tissue, and DNA recovery from FFPE tissues ranged from 18.8 μg to 21.7 μg per 20 mg of rehydrated tissue (Table S1).

Calibration curves and method validation.

Calibration curves for DNA adducts were constructed with synthetic unlabeled adducts added to a CT DNA digest matrix (10 μg) at five spiking levels ranging from 3 to 100 adducts per 10⁹ bases, and the stable isotopically labeled internal standards were added at a level of 5 adducts per 10⁸ bases (Figure S2). Each calibration level was done in triplicate, and the linear regression was done using ordinary least-squares with equal weightings. Regression curves were plotted as the amount and signal of response ratios (DNA adduct/internal standard). All calibration curves had a goodness-of-fit regression value of r² ≥ 0.99.

The limit of quantitation (LOQ) was determined by plotting the data as absolute adduct values vs. peak area and using the slope (S) of the regression and its standard deviation (σ) with the following formula: LOQ = 10σ/S.⁵⁶ Calculated LOQs ranged from 0.1 to 1.1 adducts per 10⁸ bases when assaying 3 μg of DNA on column (Figure S2). The intra- and inter-day performance validation were done using fresh-frozen and FFPE liver, colon, and rectum tissues from a single dosed animal on three different days (n = 4) (Table 1). Intra-day CVs ranged from 6.2% to 13.7%, and inter-day CVs were between 7.2% and 18%. The method’s accuracy, reproducibility, and performance were evaluated with FDA reference standard carcinogen-treated CT DNA containing known levels of PhIP, 4-ABP, and B[a]P adducts diluted with unmodified CT DNA to 5 adducts in 10⁸ bases and 5 adducts in 10⁹ bases (Table 2).^57–59

Table 1.

DNA adduct measurement reproducibility in rat liver.^a

Analyte	Tissue		Day 1	Day 2	Day 3	Intra-day CV (%)	Inter-day CV (%)
dG-C8-ABP	Fresh	mean	126	128	96.0	10.8	18.0
		SD	8.3	16.1	13.5
	FFPE	mean	83.5	84.4	63.5	10.4	18.0
		SD	7.3	9.2	7.4
dG-C8-AαC	Fresh	mean	72.5	72.4	63.2	10.1	12.0
		SD	5.1	7.2	8.5
	FFPE	mean	25.1	28.0	22.1	8.0	13.8
		SD	2.2	0.9	2.5
dG-C8-MeIQx	Fresh	mean	20.2	21.0	16.0	8.4	15.8
		SD	0.8	2.3	1.3
	FFPE	mean	13.0	12.6	11.1	12.2	13.4
		SD	1.1	1.3	2.1
dG-C8-PhIP	Fresh	mean	2.4	2.4	2.4	11.1	10.3
		SD	0.1	0.3	0.3
	FFPE	mean	1.6	1.4	1.3	8.5	13.2
		SD	0.1	0.1	0.1
dG-N2-BPDE	Fresh	mean	1.6	1.6	1.4	11.5	14.4
		SD	0.1	0.3	0.1
	FFPE	mean	1.9	1.9	1.8	13.7	13.4
		SD	0.3	0.3	0.1
O6MedG	Fresh	mean	769	827	718	10.6	13.6
		SD	43.5	111	75.6
	FFPE	mean	737.0	735.4	685	6.5	7.2
		SD	19.7	76.0	21.4
O6CMG	Fresh	mean	59.9	62.0	51.8	6.2	11.0
		SD	3.3	4.8	1.7
	FFPE	mean	48.8	50.8	47.8	10.9	10.4
		SD	5.1	4.0	6.6
M1dG	Fresh	mean	<LOQ	<LOQ	<LOQ	N/A	N/A
		SD	-	-	-
	FFPE	mean	4.3	4.6	4.1	10.2	10.9
		SD	0.3	0.4	0.5

^aInter- and intra-day validation of DNA adduct levels in liver tissue of an MC high-dose rat (n = 4 independent replicates). Adduct levels are reported per 10⁸ bases

Table 2.

DNA adduct measurement accuracy in CT DNA modified with known adduct levels^a

		Target value: 5 per 108 bases					Target value: 0.5 per 108 bases
		Mean	SD	CV (%)	Intra-day CV (%)	Inter-day CV (%)	Mean	SD	CV (%)	Intra-day CV (%)	Inter-day CV (%)
dG-C8-PhIP	Day 1	6.69	0.36	5.4	6.5	7.1	0.68	0.07	10.2	8.2	18.3
	Day 2	6.62	0.40	6.1			0.64	0.05	7.2
	Day 3	6.79	0.52	7.7			0.83	0.05	5.5
dG-C8-AαC	Day 1	4.24	0.43	10.1	6.9	7.9	0.52	0.06	12	13.2	12.3
	Day 2	4.69	0.21	4.4			0.54	0.07	12.9
	Day 3	4.70	0.28	6.0			0.56	0.07	13.4
dG-N2-BPDE	Day 1	6.24	0.36	5.8	6.7	12.5	0.56	0.07	12.8	14.8	17.8
	Day 2	5.32	0.46	8.7			0.66	0.1	14.9
	Day 3	5.28	0.23	4.4			0.64	0.1	14.9

^aTwo spiking levels were measured on three different days (n = 4 independent replicates). The CT DNA adduct levels were previously determined by liquid scintillation counting of radiolabeled DNA adducts.^57–59

O⁶MedG is often measured as the O⁶MeG nucleobase following acid hydrolysis of DNA. MC rat liver DNA was assayed by formic acid hydrolysis (Method S1) and enzymatic hydrolysis (see Methods), and adduct levels were comparable (p = 0.54) (Figure S3), demonstrating that O⁶MedG was recovered completely with our enzymatic hydrolysis scheme.

A calibration curve for PMOA-dR using [¹³C₅]PMOA-dR internal standard was similarly constructed using linear regression analysis with ordinary least-squares with equal weightings (Figure S4). Calibration levels were done in triplicate in CT DNA digest matrix at levels from zero, 1 AP site per 10⁷ bases to 1 AP site per 10⁵ bases. LOQ was determined by re-plotting values as absolute PMOA-dR levels and using the formula LOQ = 10σ/S. The LOQ was calculated at 7.6 AP Sites per 10⁹ bases with 3 μg of DNA assayed on column. The method’s intra- and inter-day performance was previously reported.⁴⁵

DNA adduct levels and AP sites in carcinogen-dosed fresh frozen and FFPE rat tissues.

DNA adduct levels in high-level MC-dosed rats were compared across fresh-frozen versus FFPE liver, colon, and rectum tissues (Figure 2). DNA adduct levels ranged from approximately1 per 10⁸ bases to 1 in 10⁵ bases. DNA adduct formation was highest in the liver, apart from dG-C8-PhIP, which was highest in the colon, and dG-N²-BPDE, which formed at comparable levels across all tissues. There were no statistically significant differences between negative control and MC-treated animals in the M1dG and 6-oxo-M₁dG levels. The alkylating agents combined formed considerably higher levels of O⁶MedG in the colon and rectal tissues than the principal dG adducts formed with 4-ABP, PhIP, AαC, MeIQx, and B[a]P. Notably, 4-ABP, a rodent and human bladder carcinogen present in tobacco smoke formed DNA adducts in colorectal tissue at levels comparable to PhIP, which is implicated in human CRC.

Figure 2.

DNA adduct levels in MC-dosed rats (PhIP, AαC, MeIQx, 4-ABP, AOM, and AS at 5 mg/kg body weight, and B[a]P at 2.5 mg/kg) in fresh-frozen and FFPE liver, colon, and rectum measured by LC/HRAMS² (*p < 0.05, **p < 0.01, ***p < 0.005; unpaired t-test).

We have previously shown that M₁dG is stable towards BME, which is added as a scavenger to mitigate artifactual LPO DNA adduct formation during tissue processing.⁴⁵ M₁dG and 6-oxo-M₁dG levels were the same in untreated and MC-dosed rats (D. Konorev, unpublished data). However, M₁dG and 6-oxo-M₁dG levels were elevated in FFPE tissues compared to fresh-frozen tissues (Figure 2), and thus these adducts cannot be reliably measured in FFPE specimens. Decreased dG-C8-AαC levels in FFPE tissues versus fresh-frozen were mainly due to the elevated temperature used during proteinase K digestion required to completely reverse DNA crosslinks (50 °C vs. 37 °C in fresh-frozen) (Figure S5). LC/HRAMS² chromatograms of DNA adduct analytes from MC-treated rat tissues are shown in Figure 3 (online trapping) and Figure 4 (offline SPE and direct injection). Some bulky aromatic amine adducts formed at the C8 position of dG can undergo oxidation and ring-open to form guanidine and spirohydantoin adducts.^50,60 Ring-opened adducts of dG-C8-AαC, dG-C8-PhIP, and dG-C8-4-ABP were not more abundant in FFPE tissues and were below 2% signal intensity of the intact adduct (D. Konorev, unpublished observations).

Figure 3.

LC/HRAMS2 EIC (trapping fraction) of DNA adducts in (A) negative control, (B) fresh-frozen MC-treated rat colon, and (C) FFPE MC-treated rat colon. The internal standard level was 5 adducts in 108 bases. RT: Retention time; A: Peak area

Figure 4.

LC/HRAMS² EIC (direct injection fraction) of alkylated DNA adducts in (A) negative control, (B) fresh-frozen MC-treated rat colon, and (C) FFPE MC-treated rat colon. Internal standard level 5 adducts in 10⁸ bases. *Loss of deoxyribose (116.0473 Da) in-source. RT: Retention time; A: Peak area.

The second animal study employed two lower AOM (50 μg/kg and 10 μg/kg) and AS (5 mg/kg and 1 mg/kg) doses to induce O⁶MedG and O⁶CMdG formation at adduct levels approaching the LOQ values. The recovery of O⁶MedG and O⁶CMdG adduct levels were comparable across fresh-frozen and FFPE tissues at the lower dosing levels (Figure 5). O⁶MedG and O⁶CMdG adducts formed in a dose-dependent manner. O⁶MedG levels were below the LOQ in animals dosed with the lowest dose of AOM/AS.

Figure 5.

DNA adduct levels in AOM/AS-dosed rats at two lower dosing levels (AS (1 mg/kg and 5 mg/kg) and AOM (0.01 mg/kg and 0.05 mg/kg) in fresh-frozen and FFPE liver, colon and rectum measured by LC/HRAMS² (*p < 0.05, **p < 0.01, ***p < 0.005; unpaired t-test).

AP site levels varied as a function of the dose of alkylating agents, especially AOM (Figure 6). Average AP site levels of negative control tissues were 2.11 ± 0.41 per 10⁷ bases (mean ± SD; n = 24) across all tissues; these background levels are consistent with those from other studies using the PMOA derivatization to measure AP sites.^45,61 AP site levels were highest in the liver and lowest in the colon for all dosing levels.

Figure 6.

(A) PMOA-derivatized AP site levels in liver, colon, and rectum of rats treated with AOM/AS and (B) LC/HRAMS² EIC of PMOA-dR and [¹³C₅]-PMOA-dR in negative control (NC), MC-treated (5 mg/kg) rat colon, and low level alkylating agent (0.05 mg/kg AOM; 5 mg/kg AS) treated rat colon. Internal standard level 3.3 adducts in 10⁶ bases (*p < 0.05, **p < 0.01, ***p < 0.005; one-way ANOVA with Dunnett’s test). RT: Retention time; A: Peak area.

Discussion

We have developed robust and validated LC/HRAMS² methods to quantitate a panel of chemically diverse DNA adducts of meat and tobacco-associated carcinogens, lipid peroxidation products, and AP sites using a multi-carcinogen-dosed rat model. The dosing route was intraperitoneal injection and not intragastric administration because the carcinogen mixture solvent required DMSO to solubilize the chemicals. Thus, the dosing route does not simulate human exposure to potential cancer-causing agents in the diet or cigarette smoke. However, our goal was to establish validated LC/HRAMS² methods to measure DNA adducts formed in rat colon and rectum at levels thought to occur in humans. The DNA adduct levels formed in rat organs range from ~1 in 10⁸ to ~1 in 10⁵ bases. LOQ values of the DNA adducts range from 0.1 to 1.1 adducts per 10⁸ bases when assaying 3 μg of DNA on column. We also report a novel SPE method to co-isolate O⁶MedG and O⁶CMdG and the LPO-derived DNA M₁dG and 6-oxo-M₁dG, avoiding the use of offline HPLC purification to facilitate larger-scale human studies. Our capillary-flow LC/MS-based AP site assay⁴⁵ was adapted to nano-flow LC coupled to high-resolution Orbitrap MS, resulting in increased sensitivity by over ten-fold, which allowed measurement of basal AP site levels using only 3 μg of DNA. The sensitivity of these validated analytical methods provides a comprehensive tool for screening DNA damage by meat and environmental and endogenous genotoxicants implicated in human CRC risk.

O⁶MedG was reported in the human colon at levels ranging from 0.8 – 2.0 O⁶MedG adducts per 10⁸ bases, using offline HPLC fractionation or ion-exchange chromatography followed by a radioimmunoassay with an O⁶MedG monoclonal antibody.^29,30 Povey reported that 43% of 62 individuals undergoing surgery for CRC were positive for O⁶MedG: mean adduct levels were ~1.0 ± 0.5 in 10⁸ bases in the DNA of non-cancerous samples of the rectum and sigmoid colon.²⁹ Hall et al. reported positivity in non-tumor-adjacent tissues in eight out of ten patients undergoing surgery with mean adduct levels of 2.1 ± 0.25 in 10⁸ bases.³⁰ In a study conducted with 21 healthy volunteers, the percentage of exfoliated colonocytes stained positive with an O⁶CMdG polyclonal antibody was significantly higher in individuals eating a high red meat diet, though adduct levels were not reported.²⁶

M₁dG and 6-oxo-M₁dG adducts are formed endogenously and are markers of oxidative stress and lipid peroxidation.⁶² These were the only DNA adducts detected in CT DNA and DNA from untreated control tissues of rats. Negative control animals were treated with 50% DMSO, which may induce oxidative stress⁶³ and possibly LPO adduct formation. Carcinogen dosing did not affect the M₁dG and 6-oxo-M₁dG levels. However, their levels increased by 3- to 10-fold in FFPE tissues compared to fresh-frozen tissues, signifying the increase in M₁dG and 6-oxo-M₁dG are artifacts formed during the formalin fixation process or the deparaffinization and rehydration of the FFPE samples. A previous study employing ³²P-postlabeling of fresh-frozen and FFPE liver and lung mouse and colon human tissues also observed a similar level of M₁dG artifact formation.⁶⁴ Leuratti identified M₁dG adducts in fresh-frozen normal colon mucosa at levels up to 12.3 adducts per 10⁷ bases using an immunoslot blot method with monoclonal anti-M₁dG; the assay has a limit of detection of ~0.2 adducts per 10⁷ bases.³⁹ These adduct levels are significantly higher than those observed in rodent tissues of our study, even in FFPE tissues with artifactually elevated M₁dG levels.

Adducts of HAAs and AAs were highest in the liver of carcinogen-dosed rats, except for dG-C8-PhIP, where adduct levels were markedly higher in the colon and rectum. HAA bioactivation is largely carried out by hepatic P450 1A2 to form the genotoxic N-hydroxylated-HAAs (HONH-HAAs), which can form DNA adducts. The significantly lower levels of dG-C8-PhIP in the liver compared to extrahepatic tissues, such as the colon and pancreas,^65,66 is attributed to the efficient hepatic detoxification of the bioactivated HONH-PhIP metabolite by glutathione S-transferases (GST) and uridine diphosphate glucuronosyl transferases (UGT); however, these enzymes are expressed at considerably lower levels in extrahepatic tissues.^66,67 This efficient hepatic detoxification of HONH-PhIP may explain why PhIP is not a liver carcinogen but is a colon carcinogen in rats.⁶⁸ Arlt⁶⁹ and Kaderlik⁷⁰ reported PhIP DNA adduct formation in extrahepatic tissues is dependent upon N-oxidation of PhIP in the liver. HONH-PhIP or its phase II conjugates are transported from the liver through the bloodstream to extrahepatic tissues, including the colon, to form DNA adducts. The significantly higher dG-C8-PhIP levels in the colon than in the rectum may be due to differences in expression levels of phase II enzymes involved in the bioactivation or detoxication of HNOH-PhIP in rats.^71,72 Phase II enzyme expression in humans is also variable along the gastrointestinal tract and the relationship enzyme expression and subsite tumor localization requires further investigation.^73–75 In contrast, AαC and MeIQx are rodent liver carcinogens; their genotoxic N-hydroxy metabolites are less efficiently detoxified by GST and UGT.^66,76,77 Notably, processed meat consumption is more strongly associated with tumors of the rectum or distal colon in humans,^78,79 and cooked red meat exposure containing HAAs is associated with neoplasia in the colon, but not the rectum.⁸⁰

The levels of HAA intake for people on a western diet are variable, with levels ranging from 1 to 50 μg per day, or the equivalent up to 700 ng kg body weight for a 70 kg adult,^76,81 compared to HAA doses of 5 mg/kg body weight used in our animal model studies. However, humans are exposed daily to HAAs, and human P450 1A2 is catalytically more efficient in bioactivating HAAs to DNA damaging agents than rodent P450s.^77,82 Daily dietary PAH intake is estimated at approximately 3 μg/day, and smoking 20 cigarettes a day accounts for the intake of ~ 2-5 μg.⁸³ Only a few studies have reported detecting HAA and PAH DNA adducts in the human colorectum. The HAA DNA adduct levels detected in our rodent study are comparable to the adduct levels reported in several human studies. Totsuka and colleagues detected dG-C8-MeIQx adducts by ³²P-postlabeling in two of four CRC patients at 1.4 and 1.9 adducts per 10⁹ bases.¹⁸ A small study with CRC patients reported putative adducts of PhIP in two out of six human colorectal DNA samples following hydrazine hydrolysis of DNA.¹⁵ Another study employing ³²P-postlabeling detected a putative adduct of PhIP in colon mucosa in 106 of the 150 tissues analyzed at similar levels of control non-cancer patients, polyp patients, or CRC patients at levels ranging from 0.3 to 0.9 adducts per 10⁸ bases.⁸⁴ A putative DNA adduct of B[a]P recovered as the B[a]P tetraol was detected by HPLC/fluorescence and reported in four out of seven subjects with CRC; the B[a]P tetraol levels ranged from 0.2 to 1.0 adducts per 10⁸ bases.¹⁰ Other studies have used ³²P-postlabeling to detect many putative bulky DNA adducts in non-tumor-adjacent colon tissue of CRC patients; however, the identities of these DNA adducts are unknown.⁸⁵

³²P-postlabeling and immunoassay screening technologies have made critical contributions to the chemical carcinogenesis field during the past four decades and have shown that many environmental toxicants and endogenous electrophiles damage the human genome.^86–88 However, neither method provides structural information about the presumed DNA adducts. With the advances in mass spectrometry instrumentation during the past decade, the sensitivity of LC/MS-based methods approaches the sensitivity of ³²P-postlabeling and exceeds the sensitivity of many immuno-based techniques.^86,89,90 The LC/HRAMS² methodology outlined in this work has the requisite sensitivity to characterize and identify previously reported putative DNA adducts in the human colorectum.

Conclusions

There are several hypotheses on the causative agents in meat and their roles in CRC etiology. However, drawing firm conclusions on the chemicals responsible for DNA damage in human colorectum is tenuous because most studies have not used specific MS-based methods to corroborate DNA adduct identities. Our robust, highly sensitive, and specific LC/HRAMS² assay quantitates nine structurally diverse DNA adducts formed by cooked meat and tobacco-associated genotoxicants, and AP sites postulated to contribute to CRC etiology. LC/HRAMS² can fulfill a critical unmet need for an unambiguous DNA adduct detection methodology for human biomonitoring studies. We are implementing our LC/HRAMS² technology to examine DNA adducts and AP sites in human colorectal tissue. Characterization of DNA damage from chemicals in meat and tobacco can help support unconfirmed paradigms in CRC etiology.