Nicotine, cotinine, and β-nicotyrine inhibit NNK-induced DNA-strand break in the hepatic cell line HepaRG

Nicotine, cotinine, and b-nicotyrine inhibit NNK-induced DNA-strand

break in the hepatic cell line HepaRG

Patricia Ordonez

a

_{, Ana Belen Sierra}

a

_{, Oscar M. Camacho}

b

_{, Andrew Baxter}

b

_{, Anisha Banerjee}

b

_,

David Waters

b

, Emmanuel Minet

b,⇑

a

Vivotecnia Research S.L., Santiago Grisolia 2, Tres Cantos, Madrid, Spain

b_{BAT, Group Research and Development, Regents Park Road, Southampton SO15 8TL, UK}

a r t i c l e

i n f o

Article history: Received 20 February 2014 Accepted 25 June 2014 Keywords: NNK Bioactivation Nicotine Cotinine DNA strand break COMET

a b s t r a c t

Recent in vitro work using purified enzymes demonstrated that nicotine and/or a nicotine metabolite could inhibit CYPs (CYP2A6, 2A13, 2E1) involved in the metabolism of the genotoxic tobacco nitrosamine NNK. This observation raises the possibility of nicotine interaction with the mechanism of NNK bioactivation. Therefore, we hypothesized that nicotine or a nicotine metabolite such as cotinine might contribute to the inhibition of NNK-induced DNA strand breaks by interfering with CYP enzymes. The effect of nicotine and cotinine on DNA strand breaks was evaluated using the COMET assay in CYP competent HepaRG cells incubated with bioactive CYP-dependent NNK and CYP-independent NNKOAc (4-(acetoxymethylnitrosoamino)-1-(3-pyridyl)-1-butanone). We report a dose-dependent reduction in DNA damage in hepatic-derived cell lines in the presence of nicotine and cotinine. Those results are discussed in the context of the in vitro model selected.

Ó 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/).

1. Introduction

The tobacco-specific nitrosamine, NNK (4-(methylnitrosami-no)-1-(3-pyridyl)-1-butanone) (Fig. 1) originates from nicotine during the curing of tobacco (nitrosation of nicotine). NNK is a class 1 IARC carcinogen, indicating that it is a human carcinogen (Cogliano et al., 2004). NNK has been strongly linked to the aetiol-ogy of lung adenocarcinoma in humans, and in lung, liver, nasal, and pancreatic tumors in rodents (Hecht et al., 1994; Hecht, 1998, 1999, 2002; Hoffmann et al., 1996).

The metabolism of NNK follows two distinct routes, namely car-bonyl reduction and

a

-hydroxylation, which occur simultaneously in the respiratory tract and the liver (Fig. 1). These reactions are catalyzed by a variety of phase I metabolic enzymes (Hecht et al., 1994; Hecht, 1998; Maser, 1998; Smith et al., 1992, 1995). Direct

a

-hydroxylation is one major bioactivation pathway for NNK, gen-erating two DNA-reactive intermediates and their final products 4-oxo-4-(3-pyridyl) 1-butanol (HPB) (Fig. 1) and 4-oxo-4-(3-pyridyl) butanone (OPB) (Fig. 1) (Smith et al., 1999, 1992). Microsome-based experiments demonstrated that CYP1A2, 2A6, 2B6, 2D6, 2E1, 3A4 and 3A5 (Hecht, 1998; Jalas et al., 2005; Smith et al., 1992) are active towards NNK in liver. CYP1A2, CYP2B6, CYP3A4, CYP2A6, and CYP2E1 can bioactivate NNK and their relative abundance in the liver suggests that they could play a role in NNK

a

-hydroxylation in this tissue (Jalas et al., 2005). Due to the selective high level of mRNA expression of CYP2A13 in airway epithelium and favorable kinetics data this enzyme has been the focus of many studies aiming at characterising the bioactivation of NNK in the lung (He et al., 2004; Su et al., 2000). NNK carcino-genicity has been clearly established in systems using doses of a single chemical (Hecht, 1998). However, bioactivation of toxicants is ultimately the result of a complex series of chemical interactions with metabolic pathways. Interestingly recent publications have highlighted the complex and potentially diverging effects of such chemical interactions. In vivo work using A/J mice has suggested a protective effect of nicotine against metabolic activation of NNK (Brown et al., 1999), however the mechanism was not identi-fied. In contrast, another in vivo study conducted with A/J mice and a K-ras mutant strain treated with NNK and nicotine concluded

http://dx.doi.org/10.1016/j.tiv.2014.06.017

0887-2333/Ó 2014 The Authors. Published by Elsevier Ltd.

This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). Abbreviations: 8-MOP, 8-methoxypsoralen; dH2O, deionosed water; HBECS,

human bronchial epithelial cells; HPB, 4-oxo-4-(3-pyridyl) 1-butanol; HPLC, high pressure liquid chromatography; IARC, International Agency for the Research on Cancer; NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone; NNKOAc, 4-(acet-oxymethylnitrosoamino)-1-(3-pyridyl)-1-butanone; OPB, 4-oxo-4-(3-pyridyl) butanone; MMS, methyl methanesulfonate; PSI, Pharmaceutical Industry Toxicol-ogy Special Interest Group; PPITC, 3-phenylpropyl isothiocyanate.

⇑ Corresponding author. Tel.: +44 (0)2380 588 997. E-mail address:Emmanuel_minet@BAT.com(E. Minet).

Toxicology in Vitro 28 (2014) 1329–1337

Contents lists available atScienceDirect

Toxicology in Vitro

that nicotine had no influence on lung tumor multiplicity, size, and progression (Maier et al., 2011; Murphy et al., 2011). Yet, a number of publications using in vitro systems have reported that nicotine or a nicotine metabolite were inhibitors of diverse CYPs involved in NNK bioactivation including CYP2A6, 2A13, and 2E1. More specifically, in vitro work using purified enzymes have demonstrated that nicotine or a nicotine metabolite could protect against CYP2A13/2A6-dependent NNK bioactivation (Bao et al., 2005; von Weymarn et al., 2006). Further investigation identified the nicotineD50(10) iminium ion as a mechanism based inhibitor of CYP2A6 and CYP2A13 (von Weymarn et al., 2012). b-nicotyrine, a nicotine related alkaloid, has also been shown to inhibit CYP2A6 and CYP2A13 in vitro (Denton et al., 2004; Kramlinger et al., 2012). CYP2E1 which has been shown to activate nitrosamines in human and rat liver (Yamazaki et al., 1992) is also inhibited by nicotine and cotinine, albeit modestly (Van Vleet et al., 2001). Interestingly cotinine is an uncompetitive inhibitor of CYP2E1 which indicates that nicotine or a nicotine metabolite does not necessarily have to be an enzyme substrate to exert an inhibitory activity (Van Vleet et al., 2001). In the context of tobacco-smoke chemical interactions and toxicological assessment, the nicotine-dependent inhibition of CYPs raises an interesting and potentially controver-sial question regarding the possibility that nicotine or nicotine metabolites could protect against some form of NNK-induced DNA damage. Therefore we hypothesized that nicotine or cotinine might inhibit CYP-dependent NNK-induced DNA single and double-strand breaks in metabolically competent cells. The effect of nicotine, cotinine and an isothiocyanate control on DNA damage was compared using the COMET assay in metabolically active

HepaRG cells incubated with bioactive CYP-independent NNKOAc and bioactive CYP-dependent NNK. The results are presented and discussed in the context of recent in vitro and in vivo data looking at the complex interaction of nicotine and nitrosamines.

2. Materials and methods 2.1. Chemicals and reagents

NNKOAc [(Acetoxymethyl)nitrosamino]-1-(3-pyridyl)-1-buta-none] (NNKOAc; CAS No. 127686-49-1) and 3-Phenylpropyl isothi-ocyanate (PPITC; CAS No. 2627-27-2) were obtained from Santa Cruz Biotechnology, Inc. (Heidelberg, Germany). Cotinine (CAS No. 486-56-6), Methyl methanesulfonate (MMS; CAS No. 66-27-3) and all other reagents required for the COMET assay were purchased from Sigma Aldrich (Madrid, Spain). Pooled lung mRNA (5 subjects) was supplied by BioChainÒ_{(Newark, USA).}

2.2. Cell line and culture conditions

Human hepatic CYP-competent HepaRGÒ_{cells were obtained}

from Merck Millipore (catalog number: MMHPR116). Seven days prior to treatment, cells were seeded on BioCoat™ Collagen I 24-well plates (Becton Dickinson, S.A.) at 0.48  106_{viable cells/}

well and maintained according to the manufacturer’s instructions. HBEC cells at passage 1 from three non-smoker male donors were obtained from Lonza (Lonza Group Ltd., Switzerland) (Donor 1 (D1): 44 years old; donor 2 (D2): 39 years old, donor 3 (D3): Fig. 1. Simplified representation of NNK metabolism detailing thea-hydroxylation route. DNA reactive intermediates are marked by a ‘#’ symbol.

24 years old) and grown to a fully differentiated culture at 28 days. A549 cell lines were obtained from the American Type Culture Collection (ATCC, USA). Cells were cultured according to the man-ufacturer’s instructions at 37 °C 5% CO2in a humidified incubator.

2.3. Quantitative RT-PCR

The procedure followed for the QRT-PCR assay is detailed in Garcia-Canton et al. (2013). A Ct threshold of 36 was selected above which a gene is considered to be not expressed. Threshold cycle (Ct) values for the genes were normalized to RPLP0, GAPDH, and B2 M.DCt values were imported into JMP Genomics 6.0 to generate the heat map using an unsupervised clustering approach. 2.4. CYP2A6 activity assay

Three replicates per condition were incubated for 4 h with 100

l

M coumarin (final data were reported as the formation of 7-hydroxycoumarin pmol/mg protein/min). For inhibited conditions, the CYP2A family inhibitor 8-methoxypsoralen was used at a final concentration of 10

l

M. Following incubation with coumarin, 250

l

l of basal medium was adjusted to pH 4.5–5.0 with hydrochloric acid and treated with ß-glucuronidase from Helix pomatia for 12 h at 37 °C. Following the glucuronidase treatment, ethoxyresorufin internal standard (10 nM final concentration) was added to the solution. 7-hydroxycoumarin was quantified using an HPLC-SCIEX/ABI 4000 Q-Trap mass spectrometer (Applied Biosystems, USA).

2.5. Cell treatment

Seven days after seeding, HepaRGÒ _{cells were incubated at}

37 °C with bioactive CYP-independent NNKOAc (50

l

M) or bioactive CYP-dependent NNK (100

l

M) dissolved in DMSO (final concentration 0.5%) in culture medium for 3 h or 12 h, respectively. Cells were also exposed to 750

l

M MMS prepared in phosphate buffered-saline (1 PBS, Invitrogen by life technologies, Spain) for 1 h, as a positive control to ensure the assay worked properly. Untreated and DMSO-treated HepaRGÒ _{cells were also assessed}

as negative controls. Additionally, in order to determine the effect of PPITC, nicotine, and cotinine inhibition on CYP-dependent NNK bioactivation, HepaRGÒ _{cells were pre-incubated with PPITC}

(10

l

M) or Cotinine (0.01

l

M, 0.1

l

M, 1.0

l

M, and 10

l

M) for 30 min before the addition of NNKOAc or NNK in order to prime the metabolic enzymes. NNKOAc or NNK were then added and the incubation continued for a total of 3 h or 12 h, respectively. 2.6. Alkaline COMET assay

The Alkaline COMET assay was based on the methods described byTice et al. (2000),Thorne et al. (2009)and the COMET assay interest group (http://www.cometassay.com) with slight modifica-tions. Cells were trypsinised (1% Trypsin-EDTA), transferred to eppendorf tubes and centrifuged (1 min, 300g). Cell pellets were resuspended in 0.6% low melting-point agarose at 37 °C. A fraction of the cell suspension was used to measure viability using the trypan blue dye exclusion method. The cell/agarose suspension was placed on SuperfrostÒ _{slides (Menzel-Glaser, Thermo Fisher}

Scientific, Germany) pre-coated with 1.0% collagen (PureColÒ_,

NUTACON B.V., Holland) to facilitate gel adherence and air-dried for 24 h prior to coating with 1.5% normal melting-point agarose. Slides were placed in lysis buffer (2.5 M NaCl, 100 mM EDTA, 10 mM Trizma, 0.2 M NaOH, 10% DMSO & 1% Triton X-100, pH 10) for at least 24 h at 4 °C then gently rinsed (3  5 min) with enzyme reaction buffer (40 mM HEPES, 0.1 M KCl, 0.5 mM EDTA, 0.2 mg/mL BSA, pH 8) at room temperature (RT). Electrophoresis

was performed for 20 min at 25 constant volts at 4 °C. After elec-trophoresis, slides were removed from the tanks and placed in a staining rack for neutralisation. Approximately 3.5 mL of neutralisation buffer (0.4 M Trizma, pH 7.5) were pipetted onto each slide. This process was repeated three times for 5 min. The excess of neutralisation buffer was removed and the slides were air dried for at least 12 h at RT prior to staining and scoring. Imme-diately prior to scoring, 20

l

L of VECTASHIELD Mounting Medium containing DAPI (1.5

l

g/mL, Vector Laboratories Inc.) was applied to slides & nuclei visualized using 20 magnification of a fluores-cence microscope fitted with an appropriate filter. Approximately 100 nuclei per slide were assessed with a minimum of two slides per tested condition performed in a minimum of 3 independent replicates. DNA damage was determined using COMET Assay IV image analysis software (Perceptive Instruments, Harverhill, UK) and the percentage of tail intensity was recorded. At least three independent assays were performed.

2.7. Statistical analysis

Data were analyzed by using the parametric statistical approach published byBright et al. (2011). The median by plate of the loga-rithmic tail intensity were analyzed in a mixed model in MinitabÒ

16 software (Minitab Ltd, Coventry, UK) with treatment as fixed effect and run as random effect, differences between the treatment variances was specified. Post-hoc multiple comparisons were adjusted by Tukey’s. Individual value plots were generated using SASÒ_JMPÒ_{Genomics 6.0 (SAS Institute, Cary (NC), USA). Mean fold}

increases were calculated from the back-transformed median of the log tail intensity for each slide divided by the mean of the back-transformed medians of the log tail intensity of the reference condition (Bright et al., 2011). Fold change significance was tested using a one sample t-test assessing no difference with 1 fold change in DNA damage relative to a reference condition. Interval plots were created with MinitabÒ_{16 statistical software.}

3. Results

3.1. Selection of CYP competent cell system

mRNA expression profiling for 41 xenobiotic metabolizing enzymes in the hepatic cell line HepaRG (3 independent repeats, 7 day cultures) and primary human bronchial epithelial cells (HBECs) (28 days culture, 3 donors) was performed to evaluate the diversity of CYPs and other metabolic genes expressed in those cell types (Fig. 2). One normal pooled human lung mRNA was also used for comparison. The heat map (Fig. 2) shows a greater diversity of CYP mRNA expression in HepaRG. Amongst others and by intensity level (DCTs), CYP3A4, 3A5, 2E1, 2B6, 2C9, 2B7, 1A1, 1B1, 1A2, 2D6 of which 3A4, 2E1, 2B6, 1A1, 1A2, have been involved with NNK bioactivation. CYP2A6 Ct values were close to the exclusion threshold for expression, but were below this threshold in all our replicate experiments in HepaRG indicating expression but at a low level. CYP2A13 was detected in HBECs but was considered not expressed in HepaRG based on high CTs values and at least one replicate over the exclusion threshold for expression. We further tested coumarin 7-hydroxylation activity, a reporter probe for CYP2A6/2A13 activity in HepaRG and compared this activity with previous data we obtained with HBECs, A549 and published data in normal hepatocytes (Table 1). Couma-rin 7-hydroxylation activity was recorded in HepaRG at day 7 in culture (1.13 pmol/mg prot/min) and was approximately 4 times lower than the median value reported for normal human hepato-cyte (Donato et al., 2000) (Table 1) and within the range of activity measured in HBECs (Table 1). 7-Hydroxycoumarin activity was also inhibited by 8-methoxypsoralen, a known selective inhibitor P. Ordonez et al. / Toxicology in Vitro 28 (2014) 1329–1337 1331

of the CYP2A enzymes 2A6 and 2A13 (Table 1). Due to the diversity of CYP expression and coumarin hydroxylation activity, HepaRG cells were chosen to study CYP-dependent NNK-induced DNA damage.

3.2. Selection of NNK and NNKOAc dose

In this work we decided to use the COMET assay to study single and double DNA strand breaks in HepaRG (7 days culture) in the Fig. 2. Hierarchical cluster representing the gene expression profiles of 41 selected metabolic genes tested in HBECs (28 days in culture) (3 subjects, 1, 2, and 3) and three independent replicates for HepaRG (7 days in culture). A representative lung profile was also included (pooled mRNA 5 subjects). Columns represent individual samples and rows represent genes. Green, black, and red indicate high signal intensity, moderate signal intensity and low to no signal in normalized gene expression data (DCt), respectively.

Table 1

Comparative coumarin 7-hydroxylation activity in primary human bronchial epithelial cells (HEBCs) from 3 donors (D1, D2, and D3), A549 cells, and HepaRG cells. HepaRG activity was measured at after 7 days recovery corresponding to the same time point used for the COMET assay. 8-methoxypsoralen (8-MOP) was used as inhibitor of CYP2A6/ CYP2A13. Results are presented as mean of three measurements ± standard deviation.

Cells 7-Hydroxycoumarin (pmol/mg prot/min) Reference

No inhibitor 8-MOP

A549 0.003 ± 0.001 0.003 ± 0.001 Newland et al. (2011)

HBECs D1 (28 days) 1.264 ± 0.05 0.275 ± 0.08 HBECs D2 (28 days) 4.289 ± 1.9 0.681 ± 0.3 HBECs D3 (28 days) 0.088 ± 0.04 0.011 ± 0.005

HepaRG (7 days) 1.13 ± 0.3 0.04 ± 0.006 This study

presence of NNK. Since no COMET assay studies were previously reported in HepaRG with NNK, we first investigated the conditions at which NNK gave statistically significant DNA damage without affecting cell viability. A 15.7% median (21% mean) tail intensity was observed for 100

l

M NNK for 12 h (Supplementary Fig. 1A and Table 1A), higher concentrations did not yield increased DNA damage. NNKOAc was also used as a CYP-independent control for NNK-induced-DNA damage. NNKOAc produces the same reactive intermediates as NNK but the reaction is catalyzed by the ubiqui-tous esterases, not by CYPs (Peterson et al., 1991). Due to the high reactivity and genotoxic driven cytotoxicity of NNKOAc a lower dose of 25

l

M NNKOAc for 3 h was selected to give DNA damage within a close range of NNK (24% median, 29% mean tail intensity) (Supplementary Fig. 1B and Table 1B). Finally, we also included a nicotine, cotinine, and b-nicotyrine (10

l

M) control only and an MMS positive control. No difference was observed between nicotine, cotinine, b-nicotyrine and the DMSO condition (Supplementary Fig. 2 and Table 2). The Trypan-blue exclusion cell viability assay for all the dose-range tested was comprised between 95% and 100%.

3.3. Inhibition of NNK-induced DNA strand breaks in the presence of nicotine, cotinine, and b-nicotyrine

To test our hypothesis that the inhibition of CYPs by nicotine, cotinine or a cotinine-derived metabolite has the potential to reduce NNK bioactivation and consequently NNK-driven DNA strand breaks, HepaRG (7 days culture) were incubated with NNK and increasing doses of nicotine, cotinine (0.01–10

l

M), and b-nicotyrine (5 and 50

l

M). A COMET assay was performed and results were compared with non-treated and DMSO controls. Single value plots for the tail intensity are presented inFig. 3A–C and show the distribution of the data points for each condition tested. Tukey group comparisons are presented in Table 2A–C. No differences in DNA damage were observed between the NNK + 10

l

M nicotine, cotinine, 5

l

M b-nicotyrine and the non-treated and DMSO controls (Fig. 3andTable 2). In order to account for the biological relevance of the changes in DNA tail intensity, fold changes and confidence intervals relative to DMSO control are also presented inFig. 4A–C. To demonstrate the CYP-depen-dency of the previous observation, HepaRG cells were pre-incu-bated with 10

l

M nicotine and cotinine for 9 h and for a further 3 h with nicotine or cotinine and NNKOAc to reach a total of 12 h. NNKOAc is an esterase-dependent, CYP-independent acti-vated form of NNK and therefore nicotine and cotinine should not inhibit NNKOAc-induced DNA damage. As hypothesised, NNKOAc-induced DNA strand break was not affected by the presence of nicotine or cotinine (Fig. 5A and B,Table 3). To further confirm that NNK-induced DNA damage was CYP2 or CYP1A dependent, a PPITC control was included. PPITC is an isothiocya-nate that efficiently inhibits multiple CYPs (2A6, 2A13, 2E1, 2B6, 1A2) (Hamilton and Teel, 1996; Nakajima et al., 2001; von Weymarn et al., 2007). NNK-induced DNA damaged was inhibited by PPITC while NNKOAc-induced DNA break was not affected by PPITC (Figs. 3–5andTables 2 and 3).

4. Discussion

In this study, we asked if the interaction of nicotine with meta-bolic processes common with NNK metabolism translates into reduced NNK-dependent DNA strand breaks in CYP-competent hepatic cell cultures. The toxicological endpoint selected was DNA strand breaks measured using the COMET assay.

A critical point in our experimental design was to select a cell system which had retained the expression and activity of

metabolic enzymes reported to be involved in NNK bioactivation such as CYPs. Since NNK is a more potent lung carcinogen than liver carcinogen, metabolically competent airway cells would be the ideal test system. Primary HBECs are better models than lung-derived cell lines because as shown in this study (Fig. 2) and others (Newland et al., 2011), they express CYP2A13 a key enzyme in nitrosamines activation in the respiratory tract (Zhang et al., 2007). However, implementing their use in a COMET assay is difficult because of interindividual variability in the metabolic activity and the morphology of the culture which forms multiple Fig. 3. Single value plots for COMET % tail intensity for HepaRG cells incubated with NNK (100lM) and a dose range of nicotine (A), cotinine (B) (0.01–10lM), and b-nicotyrine (C) (5–50lM) for 12 h. 0.5% DMSO is the vector control. Median values for % tail intensity for each condition tested are shown in red ().

layers (Garcia-Canton et al., 2013; Newland et al., 2011). Multiple cell layers is an issue in COMET due to uneven exposure of cells to the test compound. In contrast the human hepatic-derived cell line HepaRG was shown to express a diverse array of CYPs relevant to NNK-bioactivation such as CYP1A2, CYP1A1, CYP3A4, CYP2B6, CYP2E1 and low levels of CYP2A6 (Fig. 2). Coumarin 7-hydroxyl-ation activity (Table 1), bupropion hydroxylation, and ethoxyreso-rudfin o-dealkylation (Garcia-Canton et al., 2013; Gerets et al., 2012) indicative of at least CYP2A, 2B, and CYP1A activity have been demonstrated in this cell line. The coumarin 7-hydroxylation activity that was recorded in our study contrasted with the low CYP2A6 mRNA expression at 7 days in HepaRG. This could be attributed to post-transcriptional regulation of CYP2A6 which has been described previously at the mRNA and enzymatic level (Ariyoshi et al., 2001; Gilmore et al., 2001). In addition to a favor-able CYP profile, HepaRG cells are grown as a monolayer which is important for the consistency of COMET results and thus offer an advantage. Therefore, we decided to use HepaRG as a first step in the investigation of nicotine and NNK interaction by COMET. Using HepaRG, we were able to detect statistically significant DNA strand breaks when cells were incubated with 100

l

M NNK for 12 h. The damage was moderate (11–16% median tail intensity) since DNA strand break is only one of multiple mechanisms through which NNK exerts its genotoxic activity. In addition, the CYP activity of HepaRG might not reflect the activity found in normal liver. Hang et al., have reported significant NNK-mediated DNA damage (8–9%) in HepG2 but at the much lower concentration of 0.1

l

M of NNK (Hang et al., 2013). This difference in sensitivity is likely to arise from the different statistical approaches that were used to analyze the data and the cell type (Gerets et al., 2012). Our analysis followed the ‘‘Recommendations on the statistical

analysis of the COMET assay’’ designed by the Statistician in the Pharmaceutical Industry Toxicology Special Interest Group (PSI) (Bright et al., 2011). If we had applied a similar statistical method as reported in Hang et al., we would also have had a significant DNA-damage response at our lowest tested dose of NNK. We think however that the PSI method is statistically more robust since often COMET data is not normally distributed which should be Table 2

Pairwise comparison (Tukey test) for significant differences between different NNK treatments at 95% confidence. A. NNK + Nicotine (Nik), B. NNK + Cotinine (Cot), C. NNK + b-nicotyrine (b-Nic); PPITC was used as a CYP2A/1A/2E/2B control inhibitor. NT is the non-treated control. ‘‘n’’ indicates the total number of slides counted from a minimum of 3 independent experiments with the overall total nuclei counted. The ‘‘median’’ value is the median % tail intensity for all counted nuclei. Pairwise comparison was performed using the general linear model and the mean of the median of the log transformed data for each independent experiment (to account for the data distribution). Treatment results that do not share a letter (‘‘Grouping’’) are significantly different.

Treatment n Nuclei Median Grouping A 100lM NNK 14 1353 11.32 A 100lM NNK + 0.01lM Nic 14 1385 7.07 A B 100lM NNK + 0.1lM Nic 14 1348 3.92 A B C 100lM NNK + 1lM Nic 14 1372 2.73 B C D 100lM NNK + PPITC 14 1387 0.94 C D E 100lM NNK + 10lM Nic 14 1365 1.0 D E NT 14 1373 0.31 E 0.5% DMSO 14 1399 0.50 E B 100lM NNK 16 1540 16.40 A 100lM NNK + 0.01lM Cot 12 1142 3.42 A B 100lM NNK + 0.1lM Cot 12 1133 7.92 B C 100lM NNK + 1lM Cot 15 1386 2.50 B C D 100lM NNK + PPITC 21 2031 1.77 C D E 100lM NNK + 10lM Cot 15 1386 1.01 C D E NT 16 1552 1.34 D E 0.5% DMSO 14 1375 0.86 E C 100lM NNK 6 592 15.65 A 100lM NNK + 50lM b-Nic 6 581 2.22 B 100lM NNK + 10lM PPITC 6 586 2.57 B 100lM NNK + 5lM b-Nic 6 593 3.53 B NT 6 593 1.45 B 0.5% DMSO 6 596 2.14 B

Fig. 4. Interval plots: 95% confidence intervals of the mean of the medians and p-values for one sample t-test for fold change different of 1 DNA tail intensity relative to DMSO control (calculated from back-transformed data). (A) NNK + nic-otine, (B) NNK + cotinine and (C) NNK + b-nicotyrine.

factored in the statistical analysis. It is important to note that optimizing the COMET condition to obtain minimum background DNA strand break is critical to the sensitivity of the assay.

Background DNA strand break can arise from the trypsin treatment and we have seen some variation from experiment to experiment. When nicotine and cotinine were co-incubated with NNK no statistical differences in the level of DNA strand break could be detected with the DMSO control group (Table 2A and B). Nicotine and cotinine alone did not increase or decrease DNA strand break (Supplementary Table 2). The reduction in NNK-dependent DNA strand break in the presence of cotinine and nicotine was dose-dependent. Fold changes relative to the DMSO control, taking into account confidence intervals showed a 2 fold increase DNA tail intensity and a p-value of 0.13 for 100

l

M NNK + 10

l

M Nicotine treatment indicating that for this nicotine dose we cannot con-clude whether the DMSO control and the test are different (Fig. 4A). Cotinine seemed to be a stronger inhibitor of NNK-depen-dent DNA damage since no difference with control could be detected with doses as low as 1

l

M (2.7 fold change p-value = 0.1) (Fig. 4B). However with fold changes higher than 2, it is reasonable to assume that inhibition is not complete and that this can still be of biological relevance. A similar effect was observed when NNK was incubated with PPITC, a CYP inhibitor active towards CYP2A, B, E and 1A enzymes, which show fold changes in DNA tail inten-sity of no more than or lower than 2 compared to the DMSO con-trol (Fig. 4A and B). The data is robust since statistical significance was established from a minimum of 1000 cell nuclei counted from at least 3 independent experiments which included a minimum of 2 slides each. b-Nicotyrine, another reported inhibitor of the CYP2A enzymes (Kramlinger et al., 2012), was also protecting against DNA strand break in HepaRG cells at the lowest tested concentration (5

l

M) (Table 2C,Fig. 4C), but we did not test a wider dose range. In contrast, when the HepaRG cells were incubated with NNKOAc, a CYP-independent reactive form of NNK, no reduction in DNA strand break was measured in the presence of PPITC, nicotine, and cotinine as shown by the tail intensity fold changes between 1.1 and 0.91 relative to the NNKOAc only control (Fig. 5B). The con-centration of NNKOAc used (25

l

M) was below the dose that gave maximum DNA-damage (50–100

l

M) based on our dose-range assessment which meant that the absence of protection with nicotine and cotinine was not due to a plateau effect. These results indicate that nicotine and cotinine inhibit NNK-induced DNA sin-gle and double strand breaks in HepaRG and that the mechanism is CYP-dependent since NNKOAc is not affected. Those results are in agreement with in vivo results presented by Brown et al., which also showed that NNK-dependent DNA methylation was inhibited by cotinine in A/J mice (Argentin and Cicchetti, 2004; Brown et al., 1999). Importantly, we did not observe a synergistic deleterious effect on DNA damage following co-incubation of nicotine with NNK and there is no evidence that a protective effect happens in vivo, at least in the lung (Murphy et al., 2011). Our results have to be considered with caution since hepatic-derived cells were used with activity of CYPs shown to be lower than normal hepatic cells (Gerets et al., 2012). It would be of interest to optimize the COMET assay with primary human lung epithelial cells where CYP2A13 is expressed although this is likely to require extensive assay optimisation. Furthermore, due to the diverse CYP expression profile of HepaRG our experimental design did not allow the iden-tification of specific CYPs target but should be considered as a first step towards a more detailed investigation. The NNK dose that was used was relatively high, however the lowest doses of nicotine and the doses of cotinine were within ranges of what can be found in the plasma of smokers. It should also be considered that nicotine or a nicotine metabolite could have an antagonistic effect, on one hand inhibiting some CYP enzymes and on the other hand promot-ing an anti-apoptotic response through receptor bindpromot-ing (Nishioka et al., 2010). Finally, it is also important to remember that DNA-strand breaks are only one of the many mechanisms through which NNK exerts its genotoxic activity.

Fig. 5. (A) Single value plots for COMET % tail intensity for HepaRG cells pre-incubated 9 h with nicotine, cotinine, and PPITC (10lM) and a further 3 h after addition of NNKOAc (25lM). 0.5% DMSO is the vector control. Median values for % tail intensity for each condition tested are shown in red (). (B) Interval plots: 95% confidence intervals of the mean of the medians and p-values for one sample t-test for fold change different of 1 DNA tail intensity relative to NNKOAc control (calculated from back-transformed data).

Table 3

Pairwise comparison (Tukey test) for significant differences between different NNKOAc treatments at 95% confidence. DNA damage was tested after incubation with nicotine (Nic), cotinine (Cot) and PPITC; NT is the non-treated control. ‘‘n’’ indicates the total number of slides counted from a minimum of 3 independent experiments with the overall total nuclei counted. The ‘‘median’’ value is the median % tail intensity for all counted nuclei. Pairwise comparison was performed using the general linear model and the mean of the median of the log transformed data for each independent experiment (to account for the data distribution). Treatment results that do not share a letter (‘‘Grouping’’) are significantly different.

Treatment n Nuclei Median Grouping 25lM NNKOAc 6 595 26.55 A 25lM NNKOAc + 10lM Nic 6 600 23.85 A 25lM NNKOAc + 10lM Cot 6 564 26.81 A 25lM NNKOAc + 10lM PPITC 6 592 29.15 A NT 6 593 0.17 B 0.5% DMSO 6 590 0.24 B

In conclusion, we have developed and applied a robust protocol to assess DNA damage by COMET in HepaRG cells, a cell line that shares greater similarity with normal human hepatocytes. Using the COMET assay, we showed that nicotine, cotinine or a metabolite of cotinine can protect against NNK-induced DNA damage in a hepatic-derived cell line, possibly by inhibiting CYP enzymes. Importantly, this highlights the current limitations of the classical paradigm of single toxicant assessment when a complex mixture such as tobacco smoke is considered. It would be of interest to develop a similar approach using a system more physiologically relevant such as 3D primary lung epithelial cultures.

Conflict of Interest

The authors report no conflict of interest. A. Baxter, A. Banerjee, E. Minet, O.M. Camacho, and D. Waters were employees of British American Tobacco during the conduct of this study. Part of the work was contracted to Vivotecnia Research S.L., Santiago Grisolia 2, Tres Cantos, Madrid, Spain.

Transparency Document

TheTransparency documentassociated with this article can be found in the online version.

Acknowledgements

We thank Deborah Dillon, David Thorne, Annette Dalrymple, Ian Crooks, Aranxta Saiz Bautista, and Miguel Lopez Angulo for their technical assistance and advice.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, in the online version, athttp://dx.doi.org/10.1016/j.tiv.2014.06.017.

References

Argentin, G., Cicchetti, R., 2004. Genotoxic and antiapoptotic effect of nicotine on human gingival fibroblasts. Toxicol. Sci. 79, 75–81.

Ariyoshi, N., Sawamura, Y., Kamataki, T., 2001. A novel single nucleotide polymorphism altering stability and activity of CYP2a6. Biochem. Biophys. Res. Commun. 281, 810–814.

Bao, Z., He, X.Y., Ding, X., Prabhu, S., Hong, J.Y., 2005. Metabolism of nicotine and cotinine by human cytochrome P450 2A13. Drug Metab. Dispos. 33, 258–261.

Bright, J., Aylott, M., Bate, S., Geys, H., Jarvis, P., Saul, J., Vonk, R., 2011. Recommendations on the statistical analysis of the Comet assay. Pharm. Stat. 10, 485–493.

Brown, B.G., Chang, C.J., Ayres, P.H., Lee, C.K., Doolittle, D.J., 1999. The effect of cotinine or cigarette smoke co-administration on the formation of O6-methylguanine adducts in the lung and liver of A/J mice treated with 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK). Toxicol. Sci. 47, 33–39.

Cogliano, V., Straif, K., Baan, R., Grosse, Y., Secretan, B., El, G.F., 2004. Smokeless tobacco and tobacco-related nitrosamines. Lancet Oncol. 5, 708.

Denton, T.T., Zhang, X., Cashman, J.R., 2004. Nicotine-related alkaloids and metabolites as inhibitors of human cytochrome P-450 2A6. Biochem. Pharmacol. 67, 751–756.

Donato, M.T., Viitala, P., Rodriguez-Antona, C., Lindfors, A., Castell, J.V., Raunio, H., Gomez-Lechon, M.J., Pelkonen, O., 2000. CYP2A5/CYP2A6 expression in mouse and human hepatocytes treated with various in vivo inducers. Drug Metab. Dispos. 28, 1321–1326.

Garcia-Canton, C., Minet, E., Anadon, A., Meredith, C., 2013. Metabolic characterization of cell systems used in in vitro toxicology testing: Lung cell system BEAS-2B as a working example. Toxicol. In Vitro 27, 1719–1727.

Gerets, H.H., Tilmant, K., Gerin, B., Chanteux, H., Depelchin, B.O., Dhalluin, S., Atienzar, F.A., 2012. Characterization of primary human hepatocytes, HepG2 cells, and HepaRG cells at the mRNA level and CYP activity in response to inducers and their predictivity for the detection of human hepatotoxins. Cell Biol. Toxicol. 28, 69–87.

Gilmore, J., Rotondo, F., Pelletier, A.M., LaMarre, J., Alaoui-Jamali, M., Kirby, G.M., 2001. Identification of a 43-kDa protein in human liver cytosol that binds to the 30_{-untranslated region of CYP2A6 mRNA. Biochem. Pharmacol. 62, 669–678.}

Hamilton, S.M., Teel, R.W., 1996. Effects of isothiocyanates on cytochrome P-450 1A1 and 1A2 activity and on the mutagenicity of heterocyclic amines. Anticancer Res. 16, 3597–3602.

Hang, B., Sarker, A.H., Havel, C., Saha, S., Hazra, T.K., Schick, S., Jacob III, P., Rehan, V.K., Chenna, A., Sharan, D., Sleiman, M., Destaillats, H., Gundel, L.A., 2013. Thirdhand smoke causes DNA damage in human cells. Mutagenesis 28, 381– 391.

He, X.Y., Shen, J., Ding, X., Lu, A.Y., Hong, J.Y., 2004. Identification of critical amino acid residues of human CYP2A13 for the metabolic activation of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone, a tobacco-specific carcinogen. Drug Metab. Dispos. 32, 1516–1521.

Hecht, S.S., 1998. Biochemistry, biology, and carcinogenicity of tobacco-specific N-nitrosamines. Chem. Res. Toxicol. 11, 559–603.

Hecht, S.S., 1999. Tobacco smoke carcinogens and lung cancer. J. Natl. Cancer Inst. 91, 1194–1210.

Hecht, S.S., 2002. Human urinary carcinogen metabolites: biomarkers for investigating tobacco and cancer. Carcinogenesis 23, 907–922.

Hecht, S.S., Carmella, S.G., Foiles, P.G., Murphy, S.E., 1994. Biomarkers for human uptake and metabolic activation of tobacco-specific nitrosamines. Cancer Res. 54, 1912s–1917s.

Hoffmann, D., Rivenson, A., Hecht, S.S., 1996. The biological significance of tobacco-specific N-nitrosamines: smoking and adenocarcinoma of the lung. Crit. Rev. Toxicol. 26, 199–211.

Jalas, J.R., Hecht, S.S., Murphy, S.E., 2005. Cytochrome P450 enzymes as catalysts of metabolism of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone, a tobacco specific carcinogen. Chem. Res. Toxicol. 18, 95–110.

Kramlinger, V.M., von Weymarn, L.B., Murphy, S.E., 2012. Inhibition and inactivation of cytochrome P450 2A6 and cytochrome P450 2A13 by menthofuran, beta-nicotyrine and menthol. Chem. Biol. Interact. 197, 87–92.

Maier, C.R., Hollander, M.C., Hobbs, E.A., Dogan, I., Linnoila, R.I., Dennis, P.A., 2011. Nicotine does not enhance tumorigenesis in mutant K-ras-driven mouse models of lung cancer. Cancer Prev. Res. (Phila) 4, 1743–1751.

Maser, E., 1998. 11Beta-hydroxysteroid dehydrogenase responsible for carbonyl reduction of the tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone in mouse lung microsomes. Cancer Res. 58, 2996–3003.

Murphy, S.E., von Weymarn, L.B., Schutten, M.M., Kassie, F., Modiano, J.F., 2011. Chronic nicotine consumption does not influence 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone-induced lung tumorigenesis. Cancer Prev. Res. (Phila) 4, 1752–1760.

Nakajima, M., Yoshida, R., Shimada, N., Yamazaki, H., Yokoi, T., 2001. Inhibition and inactivation of human cytochrome P450 isoforms by phenethyl isothiocyanate. Drug Metab. Dispos. 29, 1110–1113.

Newland, N., Baxter, A., Hewitt, K., Minet, E., 2011. CYP1A1/1B1 and CYP2A6/2A13 activity is conserved in cultures of differentiated primary human tracheobronchial epithelial cells. Toxicol. In Vitro 25, 922–929.

Nishioka, T., Guo, J., Yamamoto, D., Chen, L., Huppi, P., Chen, C.Y., 2010. Nicotine, through upregulating pro-survival signaling, cooperates with NNK to promote transformation. J. Cell Biochem. 109, 152–161.

Peterson, L.A., Mathew, R., Murphy, S.E., Trushin, N., Hecht, S.S., 1991. In vivo and in vitro persistence of pyridyloxobutyl DNA adducts from 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone. Carcinogenesis 12, 2069–2072.

Smith, T.J., Guo, Z., Gonzalez, F.J., Guengerich, F.P., Stoner, G.D., Yang, C.S., 1992. Metabolism of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone in human lung and liver microsomes and cytochromes P-450 expressed in hepatoma cells. Cancer Res. 52, 1757–1763.

Smith, T.J., Stoner, G.D., Yang, C.S., 1995. Activation of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) in human lung microsomes by cytochromes P450, lipoxygenase, and hydroperoxides. Cancer Res. 55, 5566–5573.

Smith, G.B., Castonguay, A., Donnelly, P.J., Reid, K.R., Petsikas, D., Massey, T.E., 1999. Biotransformation of the tobacco-specific carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) in freshly isolated human lung cells. Carcinogenesis 20, 1809–1818.

Su, T., Bao, Z., Zhang, Q.Y., Smith, T.J., Hong, J.Y., Ding, X., 2000. Human cytochrome P450 CYP2A13: predominant expression in the respiratory tract and its high efficiency metabolic activation of a tobacco-specific carcinogen, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone. Cancer Res. 60, 5074–5079.

Thorne, D., Wilson, J., Kumaravel, T.S., Massey, E.D., McEwan, M., 2009. Measurement of oxidative DNA damage induced by mainstream cigarette smoke in cultured NCI-H292 human pulmonary carcinoma cells. Mutat. Res. 673, 3–8.

Tice, R.R., Agurell, E., Anderson, D., Burlinson, B., Hartmann, A., Kobayashi, H., Miyamae, Y., Rojas, E., Ryu, J.C., Sasaki, Y.F., 2000. Single cell gel/comet assay: guidelines for in vitro and in vivo genetic toxicology testing. Environ. Mol. Mutagen. 35, 206–221.

Van Vleet, T.R., Bombick, D.W., Coulombe Jr., R.A., 2001. Inhibition of human cytochrome P450 2E1 by nicotine, cotinine, and aqueous cigarette tar extract in vitro. Toxicol. Sci. 64, 185–191.

von Weymarn, L.B., Brown, K.M., Murphy, S.E., 2006. Inactivation of CYP2A6 and CYP2A13 during nicotine metabolism. J. Pharmacol. Exp. Ther. 316, 295– 303.

von Weymarn, L.B., Chun, J.A., Knudsen, G.A., Hollenberg, P.F., 2007. Effects of eleven isothiocyanates on P450 2A6- and 2A13-catalyzed coumarin 7-hydroxylation. Chem. Res. Toxicol. 20, 1252–1259.

von Weymarn, L.B., Retzlaff, C., Murphy, S.E., 2012. CYP2A6- and CYP2A13-catalyzed metabolism of the nicotine D50₍₁0_{)iminium ion. J. Pharmacol. Exp.}

Ther. 343, 307–315.

Yamazaki, H., Inui, Y., Yun, C.H., Guengerich, F.P., Shimada, T., 1992. Cytochrome P450 2E1 and 2A6 enzymes as major catalysts for metabolic activation of N-nitrosodialkylamines and tobacco-related nitrosamines in human liver microsomes. Carcinogenesis 13, 1789–1794.

Zhang, X., D’Agostino, J., Wu, H., Zhang, Q.Y., Von Weymarn, L., Murphy, S.E., Ding, X., 2007. CYP2A13: variable expression and role in human lung microsomal metabolic activation of the tobacco-specific carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone. J. Pharmacol. Exp. Ther..