Targeted and Untargeted Detection of DNA Adducts of Aromatic Amine Carcinogens in Human Bladder by Ultra Performance Liquid Chromatography-High Resolution Mass Spectrometry

Abstract

Epidemiological studies have linked aromatic amines (AAs) from tobacco smoke and some occupational exposures with bladder cancer risk. Several epidemiological studies have also reported a plausible role for structurally related heterocyclic aromatic amines present in tobacco smoke or formed in cooked meats with bladder cancer risk. DNA adduct formation is an initial biochemical event in bladder carcinogenesis. We examined paired fresh-frozen (FR) and formalin-fixed paraffin-embedded (FFPE) non-tumor bladder tissues from 41 bladder cancer patients for DNA adducts of 4-aminobiphenyl (4-ABP), a bladder carcinogen present in tobacco smoke, and 2-amino-9H-pyrido[2,3-b]indole, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine, and 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline, possible human carcinogens, which occur in tobacco smoke and cooked meats. These chemicals are present in urine of tobacco smokers or omnivores. Targeted DNA adduct measurements were done by ultra-performance liquid chromatography-electrospray ionization multi-stage hybrid Orbitrap MS. N-(2′-Deoxyguanosin-8-yl)-4-ABP (N-(dG-C8)-4-ABP) was the sole adduct detected in FR and FFPE bladder tissues. Twelve subjects (29%) had N-(dG-C8)-4-ABP levels above the limit of quantification, ranging from 1.4 to 33.8 adducts per 10⁹ nucleotides (nt). DNA adducts of other human AA bladder carcinogens, including 2-naphthylamine (2-NA), 2-methylaniline (2-MA), 2,6-dimethylaniline (2,6-DMA), and lipid peroxidation (LPO) adducts were screened for in bladder tissue, by our untargeted data-independent adductomics method, termed wide-selected ion monitoring (wide-SIM)/MS². Wide-SIM/MS² successfully detected N-(dG-C8)-4-ABP, N-(2′-deoxyadenosine-8-yl)-4-ABP and the presumed hydrazo linked adduct, N-(2′-deoxyguanosin-N²-yl)-4-ABP, and several LPO adducts in bladder DNA. Wide-SIM/MS² detected multiple DNA adducts of 2-NA, 2-MA and, 2,6-DMA, when calf thymus DNA was modified with reactive intermediates of these carcinogens. However, these AA-adducts were below the limit of detection in unspiked human bladder DNA (< 1 adduct per 10⁸ nt). Wide-SIM/MS² can screen for many types of DNA adducts formed with exogenous and endogenous electrophiles and will be employed to identify DNA adducts of other chemicals that may contribute to the etiology of bladder cancer.

TOC

Introduction

Bladder cancer (BC) is one of the ten most common forms of cancer worldwide with a very high rate of recurrence.^1,2 The risk factors for developing BC include genetic and molecular abnormalities, chronic infection of the bladder, and chemical exposures.² The elevated risk of BC in factory workers of the dye, textile, and rubber industries has been linked to the occupational exposures of AAs, such as 4-ABP and 2-naphthylamine (2-NA).³ Both chemicals were used as antioxidants in the rubber industry and as dye intermediates prior to their banning by regulatory legislation.⁴ Benzidine and 4,4′-methylenebis-2-chloroaniline (MOCA), two other aromatic amines used as industrial chemicals, have also been linked to BC.⁴ Many epidemiology studies have consistently shown that cigarette smoking is an important risk factor for BC.^5,6 4-ABP and 2-NA are present in mainstream tobacco smoke at levels ranging up to several ng per cigarette.⁷ The major sources of exposure to 4-ABP are thought to be cigarette smoking,⁸ followed by combustion of fossil fuels,⁹ textile and printing related industries,¹⁰ and through the use of some hair dye and cosmetic products.^11,12 2-MA and 2,6-DMA are other AAs formed in tobacco smoke and may contribute to BC.¹³

4-ABP has been widely used as a prototypical AA for mechanistic studies on chemical carcinogenesis of the bladder.^14,15 The carcinogenic metabolites of 4-ABP, N-hydroxy-4-ABP (HONH-4-ABP) and its N-glucuronide conjugate formed in the liver, circulate in the blood stream and reach the bladder, where they can undergo solvolysis to form the reactive nitrenium ion species that covalently binds to the DNA of the uroepithelium (Figure 1).¹⁶ Alternatively, cytochrome P450s expressed in the uroepithelium can bioactivate 4-ABP and possibly other AAs to their genotoxic HONH-AA metabolites, which are capable of reacting with DNA to form DNA adducts and induce mutations.¹⁷ The major DNA adduct of 4-ABP is N-(dG-C8)-4-ABP, with lesser amounts of N-(2′-deoxyadenosine-8-yl)-4-ABP (N-(dA-C8)-4-ABP) and N-(2′-deoxyguanosin-N²-yl)-4-ABP (N-(dG-N²)-4-ABP) also formed.¹⁴

Figure 1.

Scheme of the metabolic activation of 4-ABP leading to DNA adduct formation in bladder. 4-ABP is first oxidized to the HONH-4-ABP by cytochrome P450 (P450) in liver. HONH-4-ABP may undergo circulation and reach the bladder to form DNA adducts. As a conjugation pathway, UDP-glucuronosyltransferases (UGTs) catalyze the glucuronidation of HONH-4-ABP to form N-Gluc-HONH-4-ABP, which undergoes circulation through the blood and collected in the bladder. HO-N-Gluc-4-ABP can undergo hydrolysis in urine at acidic pH and forms a nitrenium-carbenium ion intermediate that can damage DNA in the bladder epithelial cells. A small portion of unmetabolized 4-ABP can also reach bladder through circulation in the blood and undergo oxidation by P450 expressed in the bladder epithelium to form HONH-4-ABP which reacts with DNA.

While tobacco smoking is a well-known risk factor for BC,^18–20 it is uncertain which of the chemicals present in tobacco smoke are principally responsible for DNA damage of the bladder and the development of BC. AAs, including 4-ABP, are assumed to contribute to the pathogenesis of BC in smokers; however, the levels of 4-ABP in mainstream tobacco smoke occur only at 1 – 3 ng per cigarette.⁷ In a perspective written more than 20 years ago, Poirier and Beland concluded that “exposure to 4-ABP through tobacco smoke and environment pollution are too low to account for BC risk, and factors other than 4-ABP-DNA adducts, such as adducts of other carcinogens, the influence of promoters, and synergistic effects of these factors contribute substantially to smoking-related BC in humans.”²¹ There are at least 26 AAs and structurally related HAAs present in tobacco smoke;^7,8 yet, our knowledge about these and other chemicals in tobacco smoke that damage DNA of the bladder are largely limited to 4-ABP and more recently to studies conducted on the electrophile acrolein.^22,23

Non-tobacco associated BC risks have been linked to several classes of carcinogens in the diet. Well-done meat consumption has been identified as a risk factor for BC, based on the European Prospective Investigation into Cancer and Nutrition and American Association of Retired Persons Diet and Health Studies,^24–26 and case-control studies conducted in Texas and Uruguay.^27,28 Some studies have reported that red and processed meats containing HAAs but also N-nitroso compounds may be causative agents of BC.^{24–26,29–32} However, other case-control studies in Spain,³³ Sweden,³⁴ and the USA³⁵ did not find a link between meat consumption and BC risk.

Cooked meats contain more than 20 genotoxic HAAs,³⁶ which are structurally related to AAs and undergo similar pathways of metabolism and bioactivation as AAs.³⁷ All HAAs assayed thus far in rodent bioassays have been shown to be carcinogenic, inducing tumors at multiple sites.³⁶ Smokers excrete elevated levels of mutagenic compounds in urine compared to nonsmokers, and a significant portion of the mutagenicity is thought to be attributed to AAs and/or HAAs, based on the high induction of mutations in S. typhimurium strains TA98 and YG1024 that are sensitive to these classes of chemicals.^38–40 2-Amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) was reported as a major DNA-damaging agent in urine from smokers of black tobacco.⁴¹ Other HAAs, including 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) and 2-amino-9H-pyrido[2,3-b]indole (AαC) have been detected in urine of smokers or omnivores.^37,42,43 4-ABP, 2-NA, and methylated anilines, also have been detected in urine of smokers and nonsmokers and occupational factory workers exposed to diesel exhaust.^44–47 These chemicals could undergo bioactivation by P450s expressed in the bladder epithelium.¹⁷

There are a limited number of reports in the literature on the detection of DNA adducts of 4-ABP in human bladder tissues (Table 1). Talaska was the first to detect 4-ABP-DNA adducts in human bladder biopsy samples, by employing ³²P-postlabeling.⁴⁸ The levels of the presumed N-(dG-C8)-4-ABP was 16 ± 3 adducts per 10⁸ nt in current smokers (N = 13), and it was detected at significantly higher adduct levels (~4-fold) than that estimated in current non-smokers (N = 20). Thereafter, studies employing ³²P-postlabeling,^23,49 immunohistochemistry (IHC) of paraffin blocks of transurethral resection from bladder cancer patients⁵⁰ or exfoliated urothelial cells,⁵¹ gas chromatography with negative ion chemical ionization mass spectrometry (GC-NICI-MS),^52,53 and high-performance liquid chromatography (HPLC)-triple quadrupole (QqQ)-MS have reported the measurement of 4-ABP-DNA adducts in bladder.²² There is little information regarding the capacity of monocyclic AAs⁵⁴ and HAAs to form DNA adducts in human bladder, although AαC forms appreciable levels of DNA adducts in mouse bladder.⁵⁵

Table 1.

4-ABP-DNA adducts detected in human bladder tissues^*

Method of Detection	Tissue Sources	Smoking Status	Sample Preparation	LOD or LOQ	Adduct ranges	Correlation of Adduct Levels to the Smoke Status	Ref
IHC	FFPE of transurethral resection of bladder cancer patients (N = 46)	24 current smokers, 22 non-smokers,	Monoclonal antibody against 4-ABP-DNA adducts	N.R. (relative staining)	100% positivity, 1 to 7 x 10−6 nt	Significantly higher relative staining intensity in current smokers	50
	Exfoliated urinary cells (N = 20)	20 current smokers	Same as above	N.R.	N.A. (relative staining)	No correlation between no. of cigarette smoked to adduct levels	51
32P-post- labeling	Human urinary bladder biopsy samples (N = 42)	13 current smokers, 9 non-smokers, 20 former smokers	Nuclease P1 treatment or 1- butanol extraction	5 × 10−10 nt	100% positivity, 0.1 to 1.6 × 10−8 nt (total-4- ABP-DNA adducts)	Significantly higher (4-fold) adducts in current smokers	48
	Normal human urothelial mucosa (N = 19)	3 current smokers, 11 non-smokers, 4 former smokers, 1 N.A.	1-Butanol extraction	N.R.	100% positivity, 1.7 to 7.5 × 10−8 nt (normal) and 2.3 to 10.4 × 10−8 nt (tumor) (total 4-ABP- DNA adducts)	No difference in adduct levels between normal and tumor bladder tissues	23
	Bladder tumor tissues (N = 10)	3 current smokers, 6 former smokers, 1 N.A.
GC-NICI- MS	Human lung (N = 11) Human bladder urinary mucosa (N = 8)	6 current smokers, 1 non smoker, 1 former smoker, 3 N.A.	Alkaline and Thermal hydrolysis	0.32 × 10−8 nt	91% positivity, < 0.3 to 49.5 × 10−8 nt (lung). 62.5% positivity, <0.3 to 3.9 × 10−8 nt (bladder)	No correlation between adduct levels and number of cigarette smoked	52
	Bladder cancer biopsies (N = 75)	46 current smokers, 8 non-smokers, 17 former smokers, 4 N.A.	Alkaline and Thermal hydrolysis (same as above)	3.3 × 10−8 nt (100 µg DNA)	49% positivity, ND to 1.1 x 10−5 nt) (total 4-ABP- adducts)	Strong correlation between adduct levels and overall tumor grades, and between adduct levels and current smoking status in higher tumor grades	53
	Normal epithelial and submucosal balder tissues (N = 46) and bladder tumors in urothelial carcinoma patients (N = 12)	(Bladder tumor patients) 5 current smokers,** 3 non-smokers, 4 former smokers*	Acid and thermal hydrolysis	N.R.	30% positivity, N.D. to 1.4 × 10−7 nt (tumor)	No correlation between adduct levels and smoking status	54
HPLC-QqQ- MS	Bladder cancer patients (N = 27)	1 current smoker, 25 non- smokers, 1 not available	Enzymatic hydrolysis and immunoaffinity chromatography	2 × 10−9 nt (200 µg DNA)	44% positivity, N.D. to 8 x 10−8 nt (N-(dG-C8)-4- ABP)	No correlation between adduct levels and smoking status	22
nanoUPLC- ESI- Orbitrap- MS2	Tumor adjacent normal tissues from bladder cancer patients (N = 41)	6 current smokers, 11 non-smokers, 24 former smokers	Enzymatic hydrolysis and online trapping	1.4 × 10−9 nt (20 µg DNA)	29% positivity, 1.4 to 33.8 × 10−9 nt (N-(dG- C8)-4-ABP)	No correlation between adduct levels and smoking status	This study

^*: The reported adduct levels from the cited papers were converted and expressed as adducts per nucleotides (nt).

^**: Current smoker: continued smoking up to at least 4 weeks before surgery; former smokers stopped smoking 4 weeks or more prior to surgery

N.R. Not reported.

N.A. Not available.

N.D. Not detected.

We have developed online DNA adduct enrichment and ultra-performance liquid chromatography-electrospray ionization multi-stage ion trap MS (UPLC-ESI-IT-MSⁿ) to quantify and characterize DNA adducts of 4-ABP and HAAs in rodent and human tissues, and human saliva.^56,57 We have also successfully employed FFPE tissues to screen for DNA adducts of aristolochic acid-I⁵⁸ and PhIP⁵⁹ in paired FR and FFPE human renal and prostate tissues, respectively, by high-resolution accurate mass (HRAM) nanoUPLC-ESI-Orbitrap-MS². The method has a limit of quantification (LOQ) of 1 – 2 adducts per 10⁹ nt, when employing 10 – 20 µg DNA.⁵⁹ We also recently developed an untargeted, unbiased data-independent screening method, termed wide-SIM/MS², which can screen for a wide-range of DNA adducts in human tissues.⁶⁰ In this study, we have characterized and quantified DNA adducts of 4-ABP in DNA from FR and FFPE bladder tissue of BC patients, and employed wide-SIM/MS^{2 60} to screen for other DNA adducts that may form with 4-ABP, and other AAs, including 2-NA, 2-MA, 2,6-DMA, and several prominent HAAs, including PhIP, MeIQx and AαC, which may damage bladder DNA.

Experimental Procedures

Materials.

Calf thymus DNA (CT DNA), DNase I (Type IV, bovine pancreas), Benzonase nuclease ultrapure, alkaline phosphatase (Escherichia coli), nuclease P1 (from Penicillium citrinum), RNase A (bovine pancreas), RNase T1 (Aspergillus oryzae), proteinase K (Tritiachium album), C₂H₅OH for molecular biology (200 proof), Tris-HCl, BisTris, EDTA, Na₂HPO₄, citric acid, β-mercaptoethanol (βME), 4-ABP, 2-nitrosotoluene, 1,3-dimethyl-2-nitrobenzene, pyruvonitrile, triethylamine (NEt₃), Pd/C (10%), hydrazine hydrate, Zn dust, NH₄Cl, tetrahydrofuran (THF) stabilized with butylated hydroxytoluene (BHT) (250 ppm), hexane, ethyl acetate (EtOAc), p-xylene, HPLC grade NH₄CH₃CO₂, isopropanol and silica thin layer chromatography (TLC) plates, 250 µm, were purchased from Sigma-Aldrich (St. Louis, MO). Phosphodiesterase I (Crotalus adamanteus venom) was purchased from Worthington Biochemicals Corp. (Newark, NJ). CH₂Cl₂, 10% neutral buffered formalin, Optima™ LC-MS grade HCO₂H, CH₃CN, CH₃OH, and H₂O, were purchased from Thermo Fisher Scientific (Waltham, MA). HAAs were purchased from Toronto Research Chemicals (North York, Canada). [¹³C₁₀]-dG was purchased from Cambridge Isotope Laboratory (Tewksbury, MA). N-(dG-C8)-4-ABP and N-([¹³C₁₀]-dG-C8)-4-ABP, N-(2′-deoxyguanosin-8-yl)-AαC (N-(dG-C8)-AαC) and N-([¹³C₁₀]-dG-C8)-AαC, N-(2′-deoxyguanosin-8-yl)-MeIQx (N-(dG-C8)-MeIQx) and N-([²H₃C]-dG-C8)-MeIQx, N-(2′-deoxyguanosin-8-yl)-PhIP (N-(dG-C8)-PhIP) and N-([¹³C₁₀]-dG-C8)-PhIP, [²H₁₁]-6-(1-hydroxyhexanyl)-8-hydroxy-1,N²-propano-deoxyguanosine ([²H₁₁]-dG-HNE-I) were synthesized as described.^49,61–64 Microliter CapLC vials with silylated inserts were purchased from Wheaton (Millville, NJ). N-hydroxy-2-naphthylamine (HONH-2-NA) had been provided by the late Dr. Fred F. Kadlubar (National Center for Toxicological Research, Jefferson, AR). N-(2′-Deoxyguanosin-8-yl)-2-NA (N-(dG-C8)-2-NA),⁶⁵ 1-(2′-deoxyguanosin-N²-yl)-2-NA (1-(dG-N²)-2-NA), and 1-(2′-deoxyadenosin-N⁶-yl)-2-NA (1-(dA-N⁶)-2-NA) were synthesized as previously reported.^{66,67 1}H-NMR chemical shifts of 2-NA adducts are provided in supporting information Table S1 and their high resolution accurate mass spectra are discussed in the Results. PhIP- and 4-ABP-modifed CT DNA^49,63,68 were kindly provided by Dr. Frederick A. Beland (National Center for Toxicological Research, Jefferson, AR). The N-(2′-deoxyguanosin-8-yl)-2-MA (N-(dG-C8)-2-MA) and N-(2′-deoxyguanosin-8-yl)-2,6-DMA (N-(dG-C8)-2,6-DMA) synthetic standards were provided by Dr. Gabriele Sabbioni (Institute of Environmental and Occupational Toxicology, Airolo, Switzerland),⁶⁹ and 4-oxo-(2E)-nonenal (ONE) adducts dG-ONE, dA-ONE, Dr. Ian Blair (University of Pennsylvania, Philadelphia, PA).

Human Bladder Sample Collection and DNA Extraction.

The research protocol was reviewed and approved by the Institutional Review Board at University of Minnesota. De-identified human normal tumor-adjacent bladder mucosa tissues were obtained during bladder cancer surgery and snapped-frozen in liquid nitrogen and stored at −80 °C. Portions of normal, non-proliferating human bladder mucosa, located at least 2 cm away from the tumor were dissected by certified pathologists’ assistants from cystectomy specimens obtained during bladder cancer surgery. The normal tumor-adjacent bladder mucosa tissues were de-identified, snap-frozen in liquid nitrogen and stored at −80° C. The mirror section of the frozen tissue was fixed in 10% neutral buffered formalin for 24 hours at room temperature. Tissues underwent serial dehydration with C₂H₅OH followed by p-xylene wash, and embedded in paraffin by a Sakura Tissue Tech VIP 2000 (Torrance, CA).

The procedures of DNA isolation have been published.⁷⁰ Briefly, the freshly frozen bladder tissues were thawed on ice and homogenized in TE buffer (50 mM Tris-HCl buffer, pH 8.0, and 10 mM EDTA) containing 10 mM βME by a PRO 200, PRO Scientific homogenizer equipped with a 5 mm saw tooth type blade (Oxford, CT). The homogenized tissues (equivalent of 25 mg of wet tissue weight) were centrifuged at 3000 g for 10 min at 4 °C. DNA was isolated from the nuclear pellet by the phenol/chloroform extraction method, followed by digestion with RNase A and RNase T1 at 37 °C for 1.5 h and proteinase K digestion at 37 ˚C for 2 h.⁵⁸ To extract DNA from FFPE blocks, the whole tissues were carefully removed from the paraffin block with a scalpel, deparaffinized with p-xylene, and rehydrated in a gradient of C₂H₅OH. After homogenization in TE buffer with 10 mM βME, an equivalent of 25 mg dry weight of tissue was processed with the ZR FFPE DNA MiniPrep™ Kit (Zymo Research, Irvine, CA), following the manufacturer’s instructions with modifications.⁵⁸ An overnight digestion of tissue with proteinase K (0.2 mg) was performed at 50 °C to ensure complete reversal of the cross-linked DNA. The DNA concentration and purity were measured by UV spectrophotometry.

Preparation of 2-NA, 2-MA, and 2,6-DMA-modified CT DNA

2-NA-modified CT DNA.

CT DNA solution (0.5 ml of 1 mg/mL in 10 mM sodium phosphate-citrate buffer, pH 5.0) was purged under Ar, and 5 µL of freshly prepared N-hydroxy-2-naphthylamine (HONH-2-NA, 1 mg/mL in DMSO:C₂H₅OH (3:1, v:v)) was added.⁷¹ The mixture was incubated at 37 °C for 1 h, followed by three extractions with equal vol. of EtOAc to remove the unreacted HONH-2-NA. The modified CT DNA was precipitated by the addition of 0.1 vol. of 5 M NaCl, followed by two vol. of cold isopropanol. The DNA pellet washed with 70% C₂H₅OH and reconstituted in LC-MS grade H₂O.

N-Hydroxy-2-methylaniline (HONH-2-MA).

THF stabilized with BHT (250 ppm) was purged under Ar, and then 1 mL was added to 1 mg Pd/C (10%) powder. The mixture was mixed gently and stored at – 20 °C for 20 min. Hydrazine of 20 µL (0.35 mmol) was added to the solution, which was mixed and kept at – 20 °C for 5 min.⁷² 2-Nitrosotoluene (2.6 mg, 21.5 µmol) was added to the mixture and the reaction was kept at – 20 °C for 2 h with occasional mixing. The progress of the reaction was monitored by HPLC-UV using an Aquasil C18 HPLC column (4.5 × 250 mm, Thermo Fisher Scientific, Waltham, MA) with a 20 min linear gradient starting from 1% and ending at 99% B (solvent A, 10 mM NH₄CH₃CO₂ in H₂O, pH 6.8, and B, CH₃CN) at 1 mL/min flow rate, monitoring at wavelengths 260, 290 and 340 nm. After the completion of the reaction, the Pd/C pellet was removed by centrifugation, and the supernatant was used immediately for CT DNA modification (vide infra).

N-Hydroxy-2,6-dimethylaniline (HONH-2,6-DMA).

NH₄Cl (40 mg, 0.75 mmol) was dissolved in 1.6 mL of 50% CH₃OH in a 50 °C water bath. The solution was stirred vigorously before the addition of 41.6 µL 1,3-dimethyl-2-nitrobenzene (0.31 mmol) and 40 mg Zn dust (0.63 mmol).⁷³ The reaction was monitored by TLC with hexane:EtOAc (1:1, (v;v)). After 10 min, the insoluble Zn particles were removed by centrifugation, and the supernatant was concentrated under a stream of N₂ to remove CH₃OH. The HONH-2,6-DMA was extracted with an equal vol of CH₂Cl₂ three times. The combined CH₂Cl₂ extract was dried with MgSO₄, and the solvent was concentrated to dryness under N₂. THF stabilized with BHT (250 ppm) was purged under Ar and 0.5 mL was added to dissolve the HONH-2,6-DMA.

2-MA- and 2,6-DMA-modified CT DNA.

A 1.1 mol. ratio of NEt₃ was added to HONH-2-MA or HONH-2,6-DMA (20 µmol) in THF (0.5 mL). The solution was cooled to – 30 °C in a dry ice bath containing CH₃CN. A 1.1 mol. ratio of pyruvonitrile was added to convert HONH-AA to N-acetoxy-AA.^69,73 The solution was mixed and kept at – 30 °C for 2 h. CT DNA solution (1 mg/mL in 10 mM BisTris with 0.1 mM EDTA, pH 7.1) of 0.5 mL was purged under Ar, and the THF solution of N-acetoxy-AA was added and the mixture incubated at 37 °C overnight. Unreacted N-acetoxy-AAs and its decomposition products were extracted with equal vol. of EtOAc. DNA was isolated as described (vide supra).

Enzymatic Digestion of DNA.

For quantitative measurements, human bladder DNA (20 μg) was spiked with stable, isotopically labeled DNA adduct internal standards at levels of 1.25 adducts per 10⁸ nt. The N-([¹³C₁₀]-dG-C8)- adducts of 4-ABP, AαC, PhIP, and N-([²H₃]-dG-C8)-MeIQx were added just prior to the enzymatic digestion. A mixture of PhIP- and 4-ABP-modifed CT DNA of known adduct levels were used as positive controls.^49,63,68 DNA samples were incubated at 37 °C with DNase I (2 Kunitz units) (or Benzonase, 300 U for samples analyzed by wide-SIM/MS², see below) and nuclease P1 (0.1 U) for 3.5 h, followed by phosphodiesterase 1 (3.2 mU) and alkaline phosphatase (40 mU) at 37 °C overnight. The digestion solutions were dried by vacuum centrifugation and reconstituted in 50% DMSO (30 µL). The reconstituted solutions were sonicated, centrifuged at 21000 g for 5 min, and the supernatants were transferred to a silylated borosilicate glass inserts for LC-MS analysis.

For the characterization of DNA adducts by wide-SIM/MS², bladder DNA was spiked at a level of 4 adducts per 10⁸ nt (N-([¹³C₁₀]-dG-C8)- adducts of 4-ABP, AαC, PhIP, N-([²H₃C]-dG-C8)-MeIQx) and 8 adducts per 10⁸ nt ([²H₁₁]-HNE-dG) prior to the enzymatic digestion of DNA.

DNA Adducts Quantification (nanoUPLC-ESI-Orbitrap-MS²), Characterization (nanoUPLC-ESI-Orbitrap-MS³), and Screening (wide-SIM/MS²).

An UltiMate 3000 RSLC nano UHPLC System interfaced with an Orbitrap Fusion™ Tribrid™ MS (Thermo Fisher Scientific, Waltham, MA) was used for the analysis. Chromatography was performed using a New Objectives (Woburn, MA) PicoFrit emitter (75 µm x 200 mm, 10 ± 1 μm orifice) custom packed with a Luna C18 5µm stationary phase (Phenomenex Corp. Torrance, CA) and mounted in a Nanospray Flex™ ion source (Thermo Fisher Scientific, Waltham, MA). An Acclaim PrepMap trap cartridge RP C18 (0.3 × 5 mm, 5 μm, 100 Å, Thermo Fisher Scientific, Waltham, MA) was employed for online DNA adduct enrichment. The LC solvents were (A) 0.05% HCO₂H in H₂O and (B) 0.05% HCO₂H in 95% CH₃CN. The DNA digests (8 μL) were injected onto the trap column and washed with solvent A for 4 min at a flow rate of 12 µL/min by the loading pump. This procedure removed the salt, polar components in the digest, and non-modified nucleosides. After trapping, adducts were back-flushed onto the analytical column with a linear gradient: 1 to 99% B in 15 min (0.3 µL/min) (targeted-MS² and targeted-MS³),⁵⁹ or 1 to 30% B over 4 min (0.6 µL/min), followed by 30 to 99% B over 30 min (0.2 µL/min), then 0.6 µL/min for the sequential column washing and equilibrating (wide-SIM/MS²).⁶⁰

The quantification of the DNA adducts was conducted at the MS² scan stage by fragmentation of the parent ions and monitoring of the respective aglycone ions ([M+ H - 116.0473]⁺), and their corresponding internal standards ([M + H - 116.0473]⁺) for N-([²H₃C]-dG-C8)-MeIQx, or [M + H - 121.0641]⁺ for [¹³C₁₀] labeled internal standards with a 5 ppm mass tolerance). Adduct structures were confirmed at the MS³ scan stage.⁷⁰ The wide-SIM/MS² method contained a total of 20 Orbitrap scan events (with a resolution of 60,000 at m/z 200, all resolution values reported here are full width at half maximum (FWHM)). Each wide-SIM event covers a mass of 30 m/z ± 2 m/z edge overlapping (details reported below) with quadrupole isolation, and its corresponding MS² fragmentation event isolates the same mass range and a product ion detection range of m/z 100 to 550. With the 10 wide-SIM events (and the 10 corresponding MS² events), a mass range of m/z 330–630 is analyzed. A second wide-SIM/MS² method was used with a resolution of 500,000 and a single 30 m/z wide-SIM and MS² pair with a center mass of m/z 420. The DNA adducts screened for were tentatively identified as “present” when the manually extracted ion chromatogram (EIC) peaks of the precursor ions ([M + H]⁺) co-eluted with the corresponding EIC peaks of their aglycone ions ([M + H - 116.0473]⁺ or [M + H - 121.0641]⁺). The identities of the adducts were confirmed by targeted MS³ analyses (vide infra).⁷⁰ The MS was operated in the positive ionization mode with a 2200 V spray voltage and 300 ºC ion transfer tube temperature. Other MS parameters are as follows: quadrupole isolation, 2 m/z (targeted-MSⁿ) or 34 m/z (wide-SIM/MS², 30 m/z with ±2 m/z edge overlapping); RF-lens, 90%; Orbitrap resolution, 30,000 (targeted-MSⁿ), 60,000 (wide-SIM/MS²), or 500,000 (wide-SIM/MS²); HCD collision energy, 25% (MS²) and 45% (MS³); maximum injection time, 100 ms or 1014 ms (wide-SIM/MS² at 500,000 resolution); data type, profile; AGC target, 5E4 (targeted-MSⁿ, SIM in wide-SIM/MS² at 60,000 resolution), 1E5 (MS² in wide-SIM/MS²), or 2E5 (SIM in wide-SIM/MS² at 500,000 resolution). Xcalibur™ version 3.0.63 (Thermo Scientific) was used for data acquisition and analysis. Theoretical m/z of precursors, aglycones and fragments were generated based on their structures in ChemBioDraw Ultra version 13.0.2. The EICs were manually generated with a 5 ppm mass tolerance using Xcalibur’s Qualbrowser module. The MS was externally calibrated by Pierce™ LTQ ESI Positive Ion Calibration Solution (Thermo Fisher Scientific, Waltham, MA) and the internal calibrant of fluoranthene (m/z 202.0777) was employed as a lock mass and applied during the data acquisition.

Results

We sought to determine if the AA and HAA carcinogens present in tobacco smoke and/or cooked meat had formed DNA adducts in bladder tissues of bladder cancer patients. The major dG-C8 adducts formed with 4-ABP, AαC, MeIQx, and PhIP were measured in non-tumor adjacent bladder epithelium by an HR-targeted-MS² method. The method was previously validated and has an LOQ value of 1 – 2 adducts per 10⁹ nt, employing 20 µg DNA.⁵⁹ We also employed our newly developed wide-SIM/MS² scanning technology,⁶⁰ to screen for multiple DNA adducts formed with several other potential AA bladder carcinogens.

The adduct levels in paired FR and FFPE tissues of each subject, together with their demographics are listed in Table 2. Among the 41 patients, 24 were self-reported former smokers, including one current e-cigarette smoker (T16–0234) and one current smokeless tobacco user (T16–0327), and 5 subjects were current smokers. Twelve subjects had never smoked tobacco, but one subject used smokeless tobacco (T16–0198). N-(dG-C8)-4-ABP was the only adduct detected in the bladder DNA, with comparable levels found in the bladder DNA of FR and matching FFPE tissues (Table 2 and Figure 2). We previously reported that levels of DNA adducts of aristolochic acid I in the cortex of kidney from patients with upper urinary tract cancer,⁵⁸ and DNA adducts of PhIP in the prostate genome of prostate cancer patients of paired FR and FFPE were largely within 20% of each other.⁵⁹ Our current study shows that the N-(dG-C8)-4-ABP adduct is also relatively stable towards the formalin fixation and the crosslink reversal procedures. Collectively, these findings demonstrate the great potential for FFPE tissues to serve as biospecimens for the biomonitoring of environmental chemical exposures through the measurement of the relevant DNA adducts.

Table 2.

Demographic of the bladder cancer patients and N-(dG-C8)-4-ABP levels (per 10⁸ nt). LOQ is 1.4 adducts per 10⁹ nt.

Subject Code	Age	Sex	Race	Smoking Status	Cigarette Consumption, ppd	Smoking Duration, y	Quit Time	N-(dG-C8)-4-ABP level ± SD, FR tissues	N-(dG-C8)-4-ABP level ± SD, FFPE tissues	Average N-(dG- C8)-4-ABP level
T15–1251	77	M	C	Former Smoker	0.5	NA	2006	0.2 ± 0.2	0.4 ± 0.00	<LOQ
T15–1301	60	F	C	Former Smoker	NA	NA	2007	2.8 ± 0.2	1.6 ± 0.4	2.2
T16–0028	56	F	C	Former Smoker	0.5	30	2015	3.1 ± 0.1	0.9 ± 0.1	2.0
T16–0030	79	M	C	Former Smoker	6 times/d	60	2015	0.4 ± 0.0	0.5 ± 0.0	<LOQ
T16–0049	60	M	C	Current Smoker	0.5	35	--	1.2 ± 0.0	0.9 ± 0.3	<LOQ
T16–0057	70	M	C	Former Smoker	1	25	1995	0.5 ± 0.2	0.4 ± 0.0	<LOQ
T16–0069	60	M	C	Former Smoker	1.5	45	2015	1.6 ± 0.1	1.5 ± 0.0	1.6
T16–0090	74	F	C	Former Smoker	1.5	40	2012	N. D.	N. D.	<LOQ
T16–0103	47	M	C	Former Smoker	NA	NA	2016	N. D.	N. D.	<LOQ
T16–0115	82	M	C	Former Smoker	2	2	NA	N. D.	N. D.	<LOQ
T16–0144	47	M	C	Current Smoker	1	30	--	N. D.	N. D.	<LOQ
T16–0145	77	M	C	Non-smoker	--	--	--	N. D.	N. D.	<LOQ
T16–0179	65	M	C	Non-smoker	--	--	--	0.3 ± 0.3	N. D.	<LOQ
T16–0198a	59	M	C	Non-smoker	Smokeless tobacco	--	--	N. D.	0.1 ± 0.0	<LOQ
T16–0206	51	M	C	Non-smoker	--	--	--	0.7 ± 0.2	0.7 ± 0.0	<LOQ
T16–0219	81	M	C	Current Smoker	1–2	30	--	0.1 ± 0.1	N. D.	<LOQ
T16–0227	87	M	C	Former Smoker	NA	NA	2006	0.4 ± 0.1	N. D.	<LOQ
T16–0234b	60	M	C	Former Smoker	1.5	30	2014	0.8 ± 0.0	1.7 ± 0.1	<LOQ
T16–0251	66	M	C	Former Smoker	1	29	1995	2.2 ± 0.1	0.1 ± 0.0	<LOQ
T16–0298	65	M	C	Former Smoker	NA	20	NA	N. D.	N. D.	<LOQ
T16–0301	37	M	C	Non-smoker	--	--	--	N. D.	N. D.	<LOQ
T16–0306	81	M	C	Non-smoker	--	--	--	N. D.	N. D.	<LOQ
T16–0323	70	M	C	Non-smoker	--	--	--	N. D.	0.3 ± 0.1	<LOQ
T16–0327c	78	M	C	Former Smoker	1	24	1977	1.0 ± 0.1	2.2 ± 0.0	1.6
T16–0371	69	M	C	Current Smoker	0.25	NA	--	1.9 ± 0.0	3.6 ± 0.0	2.8
T16–0399	59	M	C	Former Smoker	1.5	32	1973	0.1 ± 0.1	0.3 ± 0.1	<LOQ
T16–0406	67	M	C	Non-smoker	--	--	--	0.6 ± 0.3	0.6 ± 0.0	<LOQ
T16–0416	59	M	C	Non-smoker	--	--	--	N. D.	N. D.	<LOQ
T16–0433	40	M	AA	Former Smoker	1	30	2004	0.4 ± 0.3	1.0 ± 0.5	<LOQ
T16–0467	70	M	C	Non-smoker	--	--	--	1.1 ± 0.2	1.5 ± 0.5	<LOQ
T16–0469	78	M	C	Former Smoker	1	20	1980	N. D.	N. D.	<LOQ
T16–0497	62	M	C	Current Smoker	1	35	--	4.2 ± 0.2	1.9 ± 0.2	3.1
T16–0561	69	M	C	Former Smoker	1	20	1980	N. D.	N. D.	<LOQ
T16–0573	72	M	C	Former Smoker	1	10	1996	4.1 ± 0.2	2.6 ± 0.2	3.4
T17–0015	45	M	C	Non-smoker	--	--	--	1.7 ± 0.1	2.4 ± 0.1	2.1
T17–0018	62	M	C	Former Smoker	NA	NA	NA	10.7 ± 0.4	9.5 ± 0.3	10.1
T17–0052	75	M	C	Former Smoker	NA	NA	NA	0.1 ± 0.1	0.2 ± 0.1	<LOQ
T17–0076	72	M	C	Former Smoker	1	40	2009	0.5 ± 0.0	0.6 ± 0.0	<LOQ
T17–0095	76	M	C	Former Smoker	0.25	60	NA	33.8 ± 1.9	29.3 ± 2.1	31.6
T17–0101	73	M	C	Non-smoker	--	--	--	2.3 ± 0.6	1.1 ± 0.1	1.7
T17–0119	56	M	C	Former Smoker	0.55	30	2014	2.8 ± 0.7	3.0 ± 0.1	2.9

^a: Subject T16–0198 reported to use only smokeless tobacco products, and not cigarettes.

^b: Subject T16–0234 reported to be a former smoker but currently uses e-cigarette.

^c: Subject T16–0327 reported to be a former smoker but currently uses smokeless tobacco.

Figure 2.

The detection of N-(dG-C8)-4-ABP adducts in tumor-adjacent “normal” bladder tissues from bladder cancer patients using nanoUPLC-ESI-Orbitrap-MS². Subjects that are self-reported “current smokers” are colored in blue, and non-smokers are labeled in green, including a smokeless tobacco user (T16–0198*). The remaining patients are former smokers (black), including one who is a current e-cigarette user (T16–0234**), and one smokeless tobacco user (T16–327**). The insert table summarizes the number of subjects, their smoking status, and N-(dG-C8)-4-ABP levels in FR and matching FFPE tissues. LOQ of N-(dG-C8)-4-ABP is 1.4 adducts per 10⁹ nt.

Twelve subjects were positive for N-(dG-C8)-4-ABP with levels ranging from the LOQ (1.4 adducts per 10⁹ nt) up to 31.6 per 10⁹ nt (averaged between FR and FFPE tissues). The two subjects (T17–0095 and T17–0018) who had the highest levels of N-(dG-C8)-4-ABP, at 31.6 and 10.1 adducts per 10⁹ nucleosides, were former smokers. The putative hydrazo adduct N-(dG-N²)-4-ABP,⁷⁴ an isomer of N-(dG-C8)-4-ABP, and N-(dA-C8)-4-ABP were detected by the targeted-MS² analysis in these two subjects (Figure 3). These adducts were previously detected in human hepatocytes and in bladder tissue of canines treated with 4-ABP.^14,61,75 The product ion spectra acquired at the MS³ scan stage are consistent with the proposed structures (Figure 3). Isotopically labeled internal standards of N-(dG-N²)-4-ABP and N-(dA-C8)-4-ABP were not available for quantification, but their levels were estimated to be ~10 – 20% of N-(dG-C8)-4-ABP, based on the MS responses and assuming similar ionization efficiencies (Figure 3).

Figure 3.

The EIC of 4-ABP DNA adducts measured by nanoUPLC-ESI-Orbitrap-MS², with structure confirmation by targeted-MS³. Adducts shown are (A), N-(dG-C8)-4-ABP; (B), N-(dG-N²)-4-ABP; (C) N-(dA-C8)-4-ABP; (D), N-([¹³C₁₀]-dG-C8)-4-ABP. The fragmentation schemes are shown in Figure 6. The ions used to construct EIC of targeted-MS² were m/z 319.1302 (N-(dG-C8)-4-ABP and N-(dG-N²)-4-ABP), 303.1353 (N-(dA-C8)-4-ABP), 324.1470 (N-([¹³C₁₀]-dG-C8)-4-ABP); and those of targeted-MS³ were reported in the MS³ spectra. All ions were extracted with theoretical mass at 5 ppm mass tolerance. The internal standard N-([¹³C₁₀]-dG-C8)-4-ABP contains a total of 10 ¹³C atoms on the molecule (five on guanine and five on dR)

Our wide-SIM/MS² scanning methodology⁶⁰ was employed to screen for DNA adducts formed from reactive intermediates of several other AAs, which are possible bladder carcinogens. The scanning method allows for monitoring the frequently observed neutral loss of the 2′-deoxyribose moiety (dR, 116.0473 Da) upon collision-induced dissociation (CID) of the 2′-deoxynucleosides.⁷⁶ Due to the limiting quantity of tissue, the only DNA tested was that of subjects T17–0018 and T17–0095, who had the highest levels of N-(dG-C8)-4-ABP. The adducts investigated through the generation of the parent and neutral loss EICs included: N-(dG-C8)-4-ABP and its ring-opened oxidized products,⁷⁰ N-(dA-C8)-4-ABP, dG- and/or dA- adducts of 2-NA,⁷⁷ 2-MA,⁷³ 2,6-DMA,⁷⁸ benzidine,⁷⁹ and MOCA⁸⁰ (Table S2). Wide-SIM/MS² detected N-(dG-C8)-4-ABP and N-(dG-N²)-4-ABP in bladder DNA (Figure 4A); however, DNA adducts of the other listed AAs were not detected (Figure 4C).

Figure 4.

DNA adducts detected in bladder DNA (T17–0018) by wide-SIM/MS². The co-elution of precursor ions in wide-SIM and aglycone ions in MS² suggests the presence of (A) 4-ABP and (B) several LPO adducts. The false negative detection of the precursor N-(dA-C8)-4-ABP (indicated by the red arrow) in wide-SIM is explained in the main text. (C) 2-NA, 2-MA, and 2,6-DMA adducts are not detected in human bladder DNA, except when mixed with modified-CT DNA, at (D) dG-AA ~ one per 10⁷ nt level or lower.

Previous studies using NMR have characterized the adduction products formed between the N-acetoxy intermediates of 2-MA and 2,6-DMA with dG and dA, and some of these adducts have also been observed in CT DNA treated with these N-acetoxy intermediates.^73,78,81 However, relatively few AA-DNA adducts have been characterized by mass spectrometric analysis.^82,83 In this study, we have characterized the adducts of 2-NA, 2-MA and 2,6-DMA formed with CT DNA. 2-NA, 2-MA, and 2,6-DMA adducts were biomimetically synthesized to evaluate the capacity of wide-SIM/MS² to screen for these arylamine adducts. The adducts were produced by the reaction of CT DNA under mildly acidic pH conditions with HONH-2-NA to generate its nitrenium ion,⁷¹ or by chemical derivatization of other HONH-AAs with pyruvonitrile to generate reactive N-acetoxy intermediates that undergo heterolytic cleavage to produce the nitrenium ions (vide supra).⁶⁹

The EIC, MS³ spectra, proposed adduct structures, and CID fragmentation are shown in Figure 5. Two isomeric adducts were detected for dG-2-NA, and three isomeric adducts were detected with dA-2-MA, and dG-2,6-DMA. 2-HONH-NA reacts with DNA to form 1-(2′-deoxyguanosin-N²-yl)-2-naphthylamine (1-(dG-N²)-2-NA), 1-(2′-deoxyadenosin-N⁶-yl)-2-naphthylamine (1-(dA-N⁶)-2-NA) and a ring-opened form of a purine ring-opened derivative of N-(2′-deoxyguanosin-8-yl)-2-naphthylamine (N-(dG-C8)-2-NA), tentatively identified as l-[5-(2,6-diamino-4-oxopyrimidinyl-N⁶-deoxyriboside)]-3-(2-naphthyl)urea.⁷¹ The product ion spectra that we acquired at the MS³ scan stage are consistent with the structures of the previously reported 2-NA adducts, except that we detected the intact N-(dG-C8)-2-NA and not a ring-opened form of the adduct. The product ion spectra of 2-NA, 2-MA, and 2,6-DMA are discussed below.

Figure 5.

EIC and MS³ spectra of the putative DNA adducts from the 2-NA, 2-MA, and 2,6-DMA-modified CT DNA. dG-AA adducts were estimated at ~one per 10⁷ nt level relative to the response of N-([¹³C₁₀]-dG-C8)-4-ABP. The synthetic adduct standards (ref), 1-(dG-N²)-2-NA, N-(dG-C8)-2-NA, 1-(dA-N⁶)-2-NA, N-(dG-C8)-2-MA, and N-(dG-C8)-2,6-DMA are highlighted in blue. The adducts at t_R13.5 in dA-2-MA transition, and adducts at t_R 13.1 and 14.2 in dG-2,6-DMA transition were not assigned (the MS³ spectra are shown in Figure S2).

The accurate mass of the fragment ions and observed neutral losses measured at the MS³ scan stage by Orbitrap, and the spectral pattern among the structurally similar adducts can be utilized to facilitate the assignment of the proposed structures of the AA DNA adducts. We have tabulated the characteristic fragment ions of the proposed adducts (fragmentation schemes are shown in Figure 6), and reported the measured mass to those of theoretical ones, and reported the relative mass differences in ppm (Table S3). The aglycones of all N-(dG-C8)-AA adducts undergo CID to form the following fragment ions: [A + H – NH₃]⁺ ([A + H – 17.0266]⁺), [A + H – NH₂CN]⁺ ([A + H – 42.0218]⁺), and [A + H – NH₂CN-CO]⁺ ([A + H – 70.0167]⁺),^61,83 with a mass accuracy below 5 ppm. Similarly, N-(dA-C8)-AA adducts with common fragment ions of [A + H – NH₃]⁺ ([A + H – 17.0266]⁺), [A + H – HCN]⁺ ([A + H – 27.0109]⁺), and [A + H – NH₃ – HCN]⁺ ([A – + H – 44.0375]⁺) ⁶¹ were observed for 4-ABP, and 2-MA adducts, but not for 2-NA and 2,6-DMA. 4-ABP is the only chemical that appears to form a hydrazo adduct (N-(dG-N²)-4-ABP).¹⁴ The assignment of 4-(dA-N⁶)-2-MA and 4-(dA-N⁶)-2,6-DMA are discussed in S1 in the supporting information.

Figure 6.

The proposed fragmentation scheme of the aglycone ions of the AA-DNA adducts. The MS³ spectra of the individual adducts are shown in Figure 3 and Figure 5.

On the basis of MS³ spectra assignments, the following adducts are observed: N-(dG-C8)-2-NA, 1-(dG-N²)-2-NA, and 1-(dA-N⁶)-2-NA; adducts of 2-MA, N-(dG/dA-C8)-2-MA and 4-(dA-N⁶)-2-MA; adducts of 2,6-DMA, N-(dG-C8)-2,6-DMA and 4-(dA-N⁶)-2,6-DMA. The MS³ spectra of the synthetic standards of N-(dG-C8)-2-NA, 1-(dG-N²)-2-NA, 1-(dA-N⁶)-2-NA, N-(dG-C8)-2-MA, and N-(dG-C8)-2,6-DMA (Figure 5) closely match those spectra obtained from chemically modified DNA (highlighted in blue in Figure 5) The relative amount of the major dG vs. dA adducts formed from the incubation of CT DNA at pH 5 with N-hydroxy-2-NA, or CT DNA at pH 7.1 with N-acetoxy-2-MA/2,6-DMA, were dG-2-NA ≈ dA-2-NA, dG-2-MA >> dA-2-MA (about two-fold), whereas dA-2,6-DMA >> dG-2,6-DMA (about 5 times), judging by the MS response and assuming a similar ionization efficiency among all AA-adducts.

Thereafter, the AA-treated CT DNA at levels of ~one arylamine adduct per 10⁷ (shown in Figure 4D), or at levels of ~one arylamine adduct per10⁸ nt (Figure S1B) was mixed with human bladder DNA. The levels of 2-NA-, 2-MA-, and 2,6-DMA DNA adducts were estimated based on their EICs of the dG-C8-AAs at the MS² scan stage, relative to that of N-([¹³C₁₀]-dG-C8)-4-ABP. The wide-SIM/MS² chromatograms reveal the presence of multiple isomeric AA adducts formed with dG and dA, which were further characterized by targeted-MS³. Thus, wide-SIM/MS² is capable of screening for a wide array of AA-DNA adducts in human bladder.

Wide-SIM/MS² scanning is a robust screening method for identification of DNA adducts,⁶⁰ however; false negative detections of adducts can occur, particularly in the wide-SIM scan of the precursor ions. The Orbitrap MS system utilizes automatic gain control (AGC) to tightly regulate the number of ions or charges entering the Orbitrap to avoid space-charging effects that would cause a deterioration in mass resolution and a drop-off in sensitivity.⁸⁴ When background ions are abundant, the AGC limit is rapidly reached and the signals of the DNA adducts can fail to reach a sufficient abundance for detection. Alternatively, when the resolution of the Orbitrap is insufficient to resolve the adduct ion from an adjacent background ion, the resultant merged signal will be beyond the specified mass tolerance and lead to a false negative detection.⁶⁰ An example is shown in Figure 4A, where the precursor ion of N-(dA-C8)-4-ABP in the bladder specimen was not detected in the wide-SIM scan at 60,000 resolution, even though its aglycone was a prominent feature in the MS² scan. This interference also occurred in CT DNA modified with HONH-4-ABP (Figure 7A1 and 7A2).

Figure 7.

Wide-SIM/MS² detection of N-(dA-C8)-4-ABP from 4-ABP-modified CT DNA at N-(dA-C8)-4-ABP levels ~3 adduct per 10⁹ and ~1.6 adduct per 10⁸ nt. Data were acquired at 60,000 or 500,000 resolution. (A1 - A4) the EICs of dA-C8–4-ABP precursor (m/z 419.1826) in wide-SIM and aglycones in MS² at 5 ppm mass tolerance, while (B1 – B4) shows the MS spectra of wide-SIM scan averaged from t_R 17.55 to 17.85 min between m/z 419.15 to 419.20. The peak of N-(dA-C8)-4-ABP is highlighted in blue.

We compared the signals of the precursor ion [M + H]⁺ (theoretical m/z = 419.1826) of N-(dA-C8)-4-ABP in wide-SIM with a mass tolerance of 5 ppm at 60,000 or 500,000 resolution, using 4-ABP-modified CT DNA with N-(dA-C8)-4-ABP at ~3 adducts per 10⁹ and ~1.6 per 10⁸ nt levels (Figure 7). A reproducible signal with a Gaussian chromatographic peak was obtained only at the level of 1.6 adduct per 10⁸ nt with 500,000 resolution (Figure 7A4). The averaged MS spectrum of the precursor ion from t_R 17.55 to 17.85 min revealed the presence of an interfering peak at m/z 419.1880 (500,000 resolution, Figure 7B3 and 7B4 13 ppm apart from dA-C8–4-ABP), which requires a resolving power of ~100,000 (FWHM at m/z 200) to achieve a 10% valley separation. We constructed EIC of the ions for dA-C8–4-ABP and the interferent at 5, 10, and 15 ppm mass tolerances, which further underscores the adverse effect of the isobaric interference on detection of N-(dA-C8)-4-ABP (Figure S3).

The overlaid MS spectra of N-(dA-C8)-4-ABP in wide-SIM scan at 60,000 and 500,000 resolution are shown in Figure 8 (1.6 per 10⁸ nt level). At the apex of the chromatographic peak (t_R 17.64 – 17.66 min, Figure 8A), the intensity of N-(dA-C8)-4-ABP (m/z = 419.1826) was similar to that of the interfering ion (red trace, 500,000 resolution). Thus, at 60,000 resolution (teal trace), the detected m/z was at the middle point of the two ions (reported as m/z 419.1853). In contrast, at the tail of the chromatographic peak (t_R 17.84 – 17.86 min, Figure 8B), the intensity of the N-(dA-C8)-4-ABP was 20% of the interfering ion (red trace, 500,000 resolution), and the reported m/z at 60,000 resolution shifted towards the interfering ion (m/z 419.1880). Therefore, the inability to detect the precursor ion of N-(dA-C8)-4-ABP in SIM was mainly due to insufficient resolution to separate it from the interfering ion.

Figure 8.

The overlaid MS spectra of the chromatographic peak of N-(dA-C8)-4-ABP precursor in wide-SIM scan at the apex (A, t_R 17.64–17.66 min) and at the tail (B, t_R 17.84–17.86 min) under 60,000 resolution (teal trace) or 500,000 resolution (red trace). The theoretical m/z of N-(dA-C8)-4-ABP, m/z 419.1826, is illustrated by the red dash line. The light blue box indicated the 5 ppm mass tolerance, the pink box, 10 ppm. The interfering ion, m/z 419.1880, is 13 ppm away from N-(dA-C8)-4-ABP and falls in the green box, which corresponds to 15 ppm mass tolerance.

An increase in the AGC can improve the intrascan dynamic range (the maximum abundance ratio between the most abundant and the least abundant signal observable within a given spectrum) and the LOD/LOQ,⁸⁵ but the slower transient length will result in a reduced number of scans across the peak. When the AGC value was increased from 5E4 to 2E5 in the SIM scan of the wide-SIM/MS² event and the maximum injection time was increased from 100 ms to 1014 ms to match the transient length required for 500,000 resolution, the measured injection times for the wide-SIM data acquisition, encompassing the N-(dA-C8)-4-ABP mass, ranged between 200 – 300 ms across the peak, enabled the detection of N-(dA-C8)-4-ABP at a level of 1.6 per 10⁸ nt (Figure 7A2 and 7A4). However, the DNA adduct was not detected at a level of modification of 3 adducts per 10⁹ nt at 60,000 resolution (Figure 7A1), and the peak was just above the background signal at a resolution of 500,000 (Figure 7A1 and 7A3), indicating that the limit of detection had been reached in wide-SIM.

Wide-SIM/MS² also detected several DNA adducts of endogenous LPO products in bladder DNA. The dG-HNE-I, (m/z 424.2191), dG-HNE-II (m/z 422.2034), dC-HNE (m/z 382.1973), dG-ONE (m/z 404.1929), dA-ONE (m/z 388.1079), and dC-ONE (m/z 364.1867) adducts were detected (Figure 4B). These findings are consistent with our DNA adduct analysis of LPO adducts in human prostate tissues.⁵⁹ The structures of the adducts are supported by the product ion spectra acquired at the MS³ stage (Figure S4). We are in the process of determining the potential for artifactual formation of LPO-DNA adducts generated during isolation and enzymatic digestion of DNA. Hence, any interpretation about the biological significance of the levels and the types of LPO adducts formed in bladder should be made with caution.

Discussion

We used a stable isotope dilution HR-targeted-MS² method to quantify N-(dG-C8)-4-ABP adduct in non-tumor-adjacent bladder tissues of bladder cancer patients. We have observed that N-(dG-C8)-4-ABP, a major adduct of 4-ABP,¹⁴ is present at about 5 to 10 times higher levels than those of the tentatively assigned N-(dG-N²)-4-ABP and N-(dA-C8)-4-ABP adducts (assuming similar ionization efficiencies and responses of these adducts). The range of N-(dG-C8)-4-ABP levels observed is similar to those values reported by Talaska⁴⁸ and later by Lee using a non-specific ³²P-postlabeling,²³ and by Lin and Bohm, employing GC-NICI-MS.^52,54 However, the levels of adducts reported are much lower than those values reported by Curigliano using IHC,⁵⁰ or by Airoldi with GC-NICI-MS,⁵³ and Zayas by HPLC-QqQ-MS²² (Table 1). In our pilot study, we found highly variable levels of N-(dG-C8)-4-ABP adduct among current smokers, former smokers, and non-smokers.

The dG/dA adducts of multiple AAs were successfully screened by wide-SIM/MS² and targeted-MS³ analysis with CT DNA modified with the reactive AA intermediates. The AA-modified DNA formed in vitro was spiked into human bladder DNA to confirm wide-SIM/MS² was able to detect these adducts in a human DNA matrix. The adducts could be measured at levels approaching 1 adduct per 10⁸ nt. Wide-SIM/MS² also detected LPO adducts in human bladder DNA. The infrequent detection and the low levels of 4-ABP adducts and the absence of DNA adducts of alkylanilines may be attributed in part to inaccurate self-reporting of tobacco usage at the time of bladder surgery. The smoking status, including number of cigarettes smoked per day, the duration of smoking, and the time of quitting for former smokers were self-reported (Table 2). We do not know the time interval that the current smokers had refrained from smoking cigarettes before surgery, and DNA repair may account for the low levels of 4-ABP adducts and the absence of alkylaniline DNA adducts. The bladder tissues in this study came from a biobank, and a more accurate assessment of tobacco usage at the time of surgery, such as by measurement of nicotine or cotinine in plasma or urine, should be included in future studies to better correlate DNA adduct levels and smoking status among the subjects. Moreover, a dietary questionnaire was not available, and thus, exposures to HAAs through the diet are uncertain. The measurements of urinary AA and HAA biomarkers, or possibly hair biomarkers for HAAs would provide a better understanding of meat consumption.⁸⁶

The formation of DNA adducts of ¹⁴C-labeled alkylanilines (2,6-DMA, 3,5-dimethylaniline, 3-ethylaniline) has been reported in liver and bladder of mice dosed with these chemicals, as measured by accelerator mass spectrometry;⁸⁷ however, the identities of the DNA adducts are unknown due to the nature of the detection method. Jones treated mice with 4-ABP, 4-nitrobiphenyl and a variety of methylated anilines; however, only 4-ABP and 4-nitrobiphenyl were reported to form DNA adducts.⁶⁹ In contrast, all dosed compounds formed sulfinamide adducts with hemoglobin (Hb), demonstrating that AAs underwent bioactivation through N-oxidation in rodents. Similarly, Hb sulfinamide adducts of methylated anilines have been detected in humans.⁸⁸ To our knowledge, there are no reports in the literature on the measurements of DNA adducts of alkylanilines in human bladder by LC/MS methods. There is one report on the detection of a putative 2-MA adduct in human bladder. The method of analysis employed acid hydrolysis of DNA to liberate 2-MA, followed by GC-MS, yielding remarkably frequent detection and high adduct levels (92% positivity, 1.3 to 4.6 adducts per 10⁶ nt).⁸⁹ However, this method of detection has limitations in specificity because the mass spectrometer used was a single sector instrument with nominal mass resolution.

Recent studies propose that DNA adduct damage induced by alkylanilines may not occur through direct DNA adduct formation with the HONH-AA metabolites, but rather through their para-aminophenol metabolites, which undergo redox cycling via their quinone imine intermediates, producing reactive oxygen species (ROS) as by-products that can damage DNA and induce mutations.^90–92 It is noteworthy that we detected several lipid peroxidation DNA adducts in bladder. Some of these adducts may be produced by ROS in tobacco smoke, or possibly by reactive alkylaniline quinone imine intermediates formed in the bladder, and will be the subject of further study.

Our untargeted wide-SIM/MS² scanning method in conjunction with targeted-MSⁿ are powerful approaches to screen for DNA adducts and will be further employed to advance our understanding of hazardous agents that form DNA adducts in the bladder and contribute to bladder carcinogenesis.

Abbreviations

AA

aromatic amine

HAAs

heterocyclic aromatic amines

FFPE

formalin-fixed paraffin-embedded

4-ABP

4-aminobiphenyl

AαC

2-amino-9H-pyrido[2,3-b]indole

PhIP

2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine

MeIQx

2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline

N-(dG-C8)-4-ABP

N-(2′-deoxyguanosin-8-yl)-4-ABP

N-(dA-C8)-4-ABP

N-(2′-deoxyadenosin-8-yl)-4-ABP

N-(dG-N²)-4-ABP

N-(2′-deoxyguanosin-N²-yl)-4-ABP

N-(dG-C8)-AαC

N-(2′-deoxyguanosin-8-yl)-AαC

N-(dG-C8)-MeIQx

N-(2′-deoxyguanosin-8-yl)-MeIQx

N-(dG-C8)-PhIP

N-(2′-deoxyguanosin-8-yl)-PhIP

2-NA

2-nephthylamine

2-MA

2-methylaniline

2,6-DMA

2,6-dimethylaniline

LPO

lipid peroxidation

SIM

selected ion monitoring

BC

bladder cancer

IHC

immunohistochemistry

GC-NICI-MS

gas chromatography with negative ion chemical ionization mass spectrometry

HPLC

high-performance liquid chromatography

QqQ

triple quadrupole

nanoUPLC-ESI-Orbitrap-MSⁿ

nano flow ultra-performance liquid chromatography-electrospray ionization-Orbitrap-multi-stage MS

HRAM

high-resolution accurate mass

CT DNA

calf thymus DNA

βME

β-mercaptoethanol

EtOAc

ethyl acetate

ONE

4-oxo-(2E)-nonenal

FWHM

full width at half maximum

HCD

high energy collision-induced dissociation

1-(dG-N²)-2-NA

1-(2′-deoxyguanosin-N²-yl)-2-naphthylamine

1-(dA-N⁶)-2-NA

1-(2′-deoxyadenosin-N⁶-yl)-2-naphthylamine

N-(dG-C8)-2-NA

N-(2′-deoxyguanosin-8-yl)-2-naphthylamine

N-(dG-C8)-2-MA

N-(2′-deoxyguanosin-8-yl)-2-methylaniline

N-(dA-C8)-2-MA

N-(2′-deoxyadenosin-8-yl)-2-methylaniline

4-(dA-N⁶)-2-MA

4-(2′-deoxyadenosin-N⁶-yl)-2-methylaniline

N-(dG-C8)-2,6-DMA

N-(2′-deoxyguanosin-8-yl)-2,6-dimethylaniline

4-(dA-N⁶)-2,6-DMA

4-(2′-deoxyadenosin-N⁶-yl)-2,6-dimethylaniline

ROS

reactive oxygen species