Validation of a genotoxicity test based on p53R2 gene expression in human lymphoblastoid cells

Mutation Research 724 (2011) 76–85

Contents lists available at ScienceDirect

Mutation Research/Genetic Toxicology and Environmental Mutagenesis journal homepage: www.elsevier.com/locate/gentox Community address: www.elsevier.com/locate/mutres

Validation of a genotoxicity test based on p53R2 gene expression in human lymphoblastoid cells Taisei Mizota ∗ , Katsutoshi Ohno, Toshihiro Yamada Food Safety Research Institute, Nissin Foods Holdings Co., Ltd., 7-4-1 Nojihigashi, Kusatsu, Shiga 525-0058, Japan

a r t i c l e

i n f o

Article history: Received 10 November 2010 Received in revised form 11 May 2011 Accepted 12 June 2011 Available online 17 June 2011 Keywords: Genotoxicity p53 p53R2 Luciferase reporter gene assay ECVAM

a b s t r a c t Genotoxicity assessment is important for predicting the carcinogenicity of chemical substances. p53R2 is a p53-regulated gene that is induced by various genotoxic stresses. We previously developed a p53R2-dependent luciferase reporter gene assay in the MCF-7 human breast adenocarcinoma cell line, and demonstrated its ability to detect genotoxic agents. In this paper, we investigate the applicability of the p53R2-based genotoxicity test in the human lymphoblastoid cell line TK6. TK6 cells that express wild-type p53 have been widely used for genetic toxicology studies. To evaluate the performance of the test system in TK6 cells, we referred to 61 of the chemicals on the list of 20 genotoxic and 42 non-genotoxic chemicals recommended for the evaluation of modified or new mammalian cell genotoxicity tests by the European Centre for the Validation of Alternative Methods. The overall accordance, sensitivity, and specificity of our results with the ECVAM list were 90% (55/61), 85% (17/20), and 93% (38/41), respectively. These results indicate that the p53R2-based genotoxicity test can detect various types of genotoxic chemicals without compromising its specificity. This test will be a valuable tool for rapid screen for identifying chemicals that may be genotoxic to humans. © 2011 Elsevier B.V. All rights reserved.

1. Introduction To evaluate the genotoxicity of chemical substances, in vitro genotoxicity tests need to be able to provide predictions of in vivo effects. There are a number of in vitro tests available for use [1–7]. However, in vitro mammalian cell genotoxicity assays have a high rate of false positives [8–10]. Recently, an ECVAM workshop defined a set of reference chemicals for assessment of the performance of genotoxicity tests [11]. In the report, 62 chemicals were categorized into three groups based on in vivo genotoxicity and DNA reactivity. Group 1 includes 20 in vivo genotoxins, either due to DNA-reactive or non-DNA-reactive mechanisms, that should be detected as positive in in vitro mammalian cell genotoxicity tests. In Group 2 there are 23 non-DNA-reactive chemicals, including non-genotoxic carcinogens, which should give negative results in in vitro mammalian genotoxicity tests. Finally, Group 3 comprises 19 non-DNA-reactive chemicals, including non-genotoxic carcino-

Abbreviations: CDK, cyclin-dependent kinase; dNTPs, deoxyribonucleotide triphosphates; ECVAM, European Centre for the Validation of Alternative Methods; ENU, N-ethyl-N-nitrosourea; 5-FU, 5-fluorouracil; GFP, Green Fluorescent Protein; IQ, 2-amino-3-methylimidazo[4,5-f]quinoline; MMS, methyl methanesulfonate; MNNG, N-methyl-N -nitro-N-nitrosoguanidine; NADH, reduced nicotinamide adenine dinucleotide; NADPH, reduced nicotinamide adenine dinucleotide phosphate; PhIP, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine. ∗ Corresponding author. Tel.: +81 77 561 9115; fax: +81 77 561 9140. E-mail address: [email protected] (T. Mizota). 1383-5718/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.mrgentox.2011.06.003

gens, metabolic poisons and other compounds that should give negative results in in vitro mammalian genotoxicity tests, but have been reported to induce chromosomal aberrations or tk mutations in mouse lymphoma cells often at high concentrations or at high levels of cytotoxicity. The chemicals in each group were further arranged into subgroups and/or sections based on their chemical classification or pattern of results in genotoxicity tests. The Group 1 chemicals were divided into sections I, which included (i–iv) subgroups, and II (Table 1), the Group 2 chemicals were divided into subgroups (i–iii) (Table 2), and the Group 3 chemicals were divided into subgroups (i–iv) (Table 3). These reference chemicals are useful for validation of new genotoxicity tests. We previously developed a genotoxicity test based on p53R2 expression in the human MCF-7 cell line expressing wild-type p53 [12,13]. The test was also functional in p53-competent HepG2, which is a human hepatocellular liver carcinoma cell line [12]. The expression of p53R2, which encodes a small subunit of ribonucleotide reductase, is induced by phosphorylated p53 in response to DNA damage such as from ␥-ray radiation, UV irradiation, and genotoxic chemicals [14,15]. A consensus DNA-binding sequence for p53 of two copies of 5 -PuPuPuC(A/T)(A/T)GPyPyPy-3 was identified in intron 1 of p53R2 [16–18]. The product of the p53R2 gene plays a pivotal role in supplying dNTPs to the DNA repair pathway both in vivo and in vitro [14,19,20], and is also necessary for maintenance of mitochondrial DNA copy number [21,22]. Furthermore, p53R2 has been shown to interact with CDK inhibitor p21 [23], which is thought to be a biological requirement for the coordination

Table 1 In vivo genotoxins which should be detected as positive in in vitro mammalian cell genotoxicity tests. p53R2-dependent reporter gene with S9

Highest concentrationd setting

Name

Sourcea

CAS No.

Resultsb

Results

␮g/mL

I. Ames-positive in vivo genotoxins (i) O6 and N7 alkylators N Cyclophosphamide ENU S MMS A (ii) Polycyclic aromatic hydrocarbons Benzo[a]pyrene W

6055-19-2

−

759-73-9 66-27-3

+ +

50-32-8

−

+

7,12W Dimethylbenzanthracene (iii) Aromatic amines DimethylniW trosamine

57-97-6

−

+

62-75-9

−

+

LECc (␮g/mL)

250 12

Cell viability (% of Cont.)

Reason

+

3.0

50

46.4

T

+ +

120 2.0

500 50

59.9 57.4

T T

0.49

10

106.5

S

0.85

10

54.3

T

750

100.7

120

69.7

S

100

88.9

S

30

77.3

S

43

mM

Requires metabolic activation (CYP2B6) Strong gene mutagen (O6 alkylation) Strong clastogen (N7 alkylation) Requires metabolic activation (CYP1A1; 1B1, epoxide hydrolase); forms bulky adducts Requires metabolic activation (CYP1B1); forms bulky adducts Alkylating agent after activation by CYP2E1 (which is not highly expressed in rat liver S9): produces O6 - and N7 methyl guanine adducts Hydroxylated by CYP1A2 and then acetylated. Forms C8 adduct on guanine Aromatic amine, requires metabolic activation Heterocyclic amine with potent genotoxicity, requires metabolic activation Heterocyclic amine with potent genotoxicity, requires metabolic activation

2Acetylaminofluorene

W

53-96-3

+

2,4Diaminotoluen IQ

W

95-80-7

+

W

76180-96-6

+

2.9

−

PhIP

W

105650-23-5

+

0.16

+

0.14

10

88.1

S

(iv) Others Aflatoxin B1

W

1162-65-8

+

0.58

+

0.28

10

51.4

T

53.9

T

Activated by CYP3A4, which is not highly expressed in rats compared with human. Forms various adducts Inorganic carcinogen

100 1500

84.8 68.4

S mM

Cross-linking agent No adducts

1.0 30 1500 3.0

45.2 35.1 57.3 51.9

T T T T

0.30 1500

51.8 36.7

T T

Topoisomerase inhibitor Aneugen Nucleoside analogue Inorganic cacinogen; oxidant? repair inhibitor? Aneugen Clastogen that binds to DNA

Cadmium W 10108-64-2 chloride W 15663-27-1 Cisplatin p-Chloroaniline S 106-47-8 II. In vivo genotoxins negative or equivocal in Ames C 33419-42-0 Etoposide Hydroquinone W 123-31-9 Azidothymidine W 30516-87-1 W Sodium arsenite 7784-46-5 Taxol Chloramphenicol

W S

33069-62-4 56-75-7

2.6

LEC (␮g/mL)

Further information

75

− + − + − + + + +

+

+

1.1

16

− 18

0.010

5.0

− − 0.070

46 0.25

+ − − +

0.012 240

+ −

0.069

0.31

T. Mizota et al. / Mutation Research 724 (2011) 76–85

p53R2-dependent reporter gene without S9

Chemical

a These chemicals were purchased from the following sources: A, Aldrich Chemical Co., Inc. (Milwaukee, WI); C, Calbiochem, Bioscience, Inc. (Darmstadt, Germany); J, Sigma–Aldrich Japan (Tokyo, Japan); M, Merck KGaA (Darmstadt, Germany); N, Nacalai Tesque Inc. (Kyoto, Japan); T, Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan); S, Sigma–Aldrich Chemical Co. (St. Louis, MI, USA): W, Wako Pure Chemical Industries Ltd. (Osaka, Japan). b The results of p53R2 dependent reporter gene assay are recorded as positive (+) or negative (–) for the genotoxicity. The test sample was judged as positive when the relative luciferase activity of cells treated with test sample was 130% that of the cells treated with solvent only, and increased in a dose-dependent manner. c LEC (␮g/mL) indicates the lowest effective concentration of the test chemical that gave positive results in this assay. d The highest concentration of the chemical determined by its cytotoxicity or solubility as mentioned in Section 2. Unless precluded by cytotoxicity and solubility, the highest concentration was set at about 10 mM according to the current ICH S2B guideline. The reasons of the test concentration limitation are recorded as cytotoxicity (T), solubility (S), or ICH S2B guideline (mM).

77

78

Table 2 Non-DNA-reactive chemicals (including non-genotoxic carcinogens) that should give negative results in in vitro mammalian cell genotoxicity tests. p53R2-dependent reporter gene with S9

Highest concentrationd setting

Name

Sourcea

CAS No.

Resultsb

(i) Non-carcinogens with negative in vivo genotoxicity data N Ampicillin trihydrate N D-mannitol (ii) Non-carcinogens with no in vivo genotoxicity data S Phenformin HCl n-Butyl chloride W (2-Chloroethyl)trimethyl-ammonium chloride W Cyclohexanone W W N,N-dicyclohexyl thiourea W Trisodium EDTA trihydrate Ephidrine sulphate N Erythromycin stearate Fluometron W S Phenanthrene (iii) Non-genotoxic carcinogens S D-limonene Di-(2-ethylhexyl)phthalate S Amitrole W Tert-butyl alcohol W Diethanolamine W T Melamine W Methyl carbamate W Progesterone W Pyridine Tris(2-ethylhexyl)phosphate W Hexachloroethane W

Results

LEC (␮g/mL)

␮g/mL

Cell viability (% of Cont.)

Reason

7177-48-2 69-65-8

− −

− −

250 2000

115.3 109.0

S mM

834-28-6 109-69-3 999-81-5 108-94-1 1212-29-9 150-38-9 134-72-5 643-22-1 2164-17-2 85-01-8

− − − − − − NT − − −

− − − − − −

100 1500 75 1500 90 150

40.6 112.1 99.1 87.4 95.5 96.1

T mM S mM S S

− − −

200 600 300

84.4 90.0 70.6

5989-27-5 117-81-7 61-82-5 75-65-0 111-42-2 108-78-1 598-55-0 57-83-0 110-86-1 78-42-2 67-72-1

− − − − + − − − − − −

− − − − − − − − − − −

1500 1500 900 750 500 150 900 200 1000 2400 300

104.7 73.3 52.5 87.4 36.9 69.5 91.7 128.0 88.4 75.5 98.4

a b c d

LECc (␮g/mL)

NT

90

These chemicals were purchased from the sources described in footnote of Table 1. The results of p53R2 dependent reporter gene assay are recorded as described in Table 1. Ephedrine sulfate was not tested due to purchasing restrictions in Japan, and it is indicated as Not Tested (NT). LEC (␮g/mL) indicates the lowest effective concentration of the test chemical that gave positive results in this assay. The highest concentration of the chemical determined by its cytotoxicity or solubility as mentioned in Section 2. The reasons of the test concentration limitation are recorded as described in Table 1.

S S S mM S T mM T S mM S mM S S

T. Mizota et al. / Mutation Research 724 (2011) 76–85

p53R2-dependent reporter gene without S9

Chemical

Table 3 Non-DNA-reactive chemicals (including non-genotoxic carcinogens), metabolic poisons and others that should give negative results in in vitro mammalian cell genotoxicity test, but have been reported to induce chromosomal aberrations or tk mutations in mouse lymphoma cells, often at high concentrations or at high levels of cytotoxicty. Chemical

Name

Sourcea

CAS No.

a b c d

p53R2-dependent reporter gene with S9

Highest concentrationd setting

Resultsb

Results

␮g/mL

Cell viability (% of Cont.)

Reason

LECc (␮g/mL)

LEC (␮g/mL)

− − − − − − −

− − − − − − −

2000 900 100 1500 1500 1500 1500

119.6 91.8 49.3 103.4 50.5 46.2 94.7

mM S T mM T T S

− − − −

− − − −

900 50 1000 1000

91.0 30.1 63.6 124.0

S T mM mM

−

−

2000

79.8

mM

− − − − + + −

− − − − + + −

100 1000 300 1200 1200 750 500

39.8 42.8 94.4 102.7 59.5 93.3 35.1

T T S mM T mM T

54 530

350 721

T. Mizota et al. / Mutation Research 724 (2011) 76–85

(i) Non-carcinogens that are negative or equivocal for genotoxicity in vivo d,l-Menthol J 15356-70-4 W 85-44-9 Phthalic anhydride Tertiary-butylhydroquinone W 1948-33-0 W 118-92-3 o-Anthranilic acid 1,3-Dihydroxybenzene M 108-46-3 2-Ethyl-1,3-hexanediol S 94-96-2 Sulfisoxazole S 127-69-5 (ii) Non-carcinogens with no in vivo genotoxicity data Ethionamide S 536-33-4 Curcumin W 458-37-7 Benzyl alcohol W 100-51-6 Urea N 57-13-6 (iii) Non-carcinogens or carcinogenic by irrelevant (for human) mechanism W 128-44-9 Sodium saccharin (iv) Supplementary list (prediction of in vitro genotoxicity results less clear) W 121-79-9 Propyl gallate p-Nitrophenol N 100-02-7 Sodium xylene sulfonate S 1300-72-7 W 140-88-5 Ethyl acrylate Eugenol W 97-53-0 Isobutyraldehyde W 78-84-2 2,4-Dichlorophenol S 120-83-2

p53R2-dependent reporter gene without S9

These chemicals were purchased from the sources described in footnote of Table 1. The results of the p53R2 dependent reporter gene assay are recorded as described in Table 1. LEC (␮g/mL) indicates the lowest effective concentration of the test chemical that gave positive results in this assay. The highest concentration of the chemical determined by its cytotoxicity or solubility as mentioned in Section 2. The reasons of the test concentration limitation are recorded as described in Table 1.

79

80

T. Mizota et al. / Mutation Research 724 (2011) 76–85

of DNA repair and cell cycle arrest. As shown previously, the p53R2dependent luciferase reporter gene assay in MCF-7 cells was able to detect various types of genotoxic chemicals, including Amesnegative chemicals that gave positive results in in vitro mammalian cell genotoxicity tests such as mouse lymphoma test, chromosomal aberration test, and sister chromatid exchange assay [12,13]. In this study, the p53R2-based genotoxicity assay was performed using p53-competent human lymphoblastoid TK6 cells. TK6 cells have been widely used for genetic toxicology studies such as the micronucleus test [24], comet assay [25], and reporter gene assay [26], and have also been used in toxicogenomic approaches [27,28]. It is noteworthy that the tk locus mutation assay using TK6 cells is accepted by ICH guideline for pharmaceuticals (S2B-Genotoxicity: A Standard Battery for Genotoxicity Testing of Pharmaceuticals) [29]. First, we investigated the application of the system in TK6 cells with adriamycin, MNNG, 5-FU and sodium ascorbate as representative chemicals. The test system was then validated using 61 of the 62 chemicals recommended by ECVAM. 2. Materials and methods 2.1. Reagents All chemicals used in this study were of the highest purity available. Adriamycin was purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). MNNG was purchased from Nacalai Tesque, Inc. (Kyoto, Japan). 5-FU and sodium ascorbate were purchased from Sigma–Aldrich (St. Louis, MI, USA). Chemicals for the validation study were obtained according to the ECVAM recommendations, and their sources are described in footnote of Table 1. Ephedrine sulfate could not be obtained because of purchasing restrictions in Japan.

2.2. Cell culture The human lymphoblastoid cell line TK6, derived from the American Type Culture Collection (ATCC), was obtained from Summit Pharmaceuticals International Corporation (Tokyo, Japan). TK6 cells have been reported to express wild type p53 [30]. Cells were maintained in RPMI-1640 medium (Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) supplemented with 10% (v/v) fetal bovine serum (FBS, GIBCO). Cells were incubated at 37 ◦ C in a saturated atmosphere containing 5% CO2 .

2.3. Semiquantitative RT-PCR Cells (1.0 × 105 ) were plated in a 96-well microplate (Corning, Flanklin Lakes, NJ), and incubated for 1 h. Adriamycin, MNNG, 5-FU, and sodium ascorbate were then added at final concentrations of 0.13 ␮g/mL, 5.0 ␮g/mL, 33 ␮g/mL, and 100 ␮g/mL, respectively. After 16 h incubation, total RNA was extracted using an RNeasy kit (QIAGEN, Hilden, Germany). The total RNA sample (100 ng) was reverse transcribed to yield first-strand DNA using a High Capacity RNA-to-cDNATM Kit (Applied Biosystems, Foster City, CA.). After reverse transcription, PCR for semiquantitation of p53R2 expression was carried out on a LightCycler® 1.5 Instrument (Roche, Indianapolis, IN) with a sense primer (5 - TTTTTGCAGCCAGTGATGGA-3 ) and an antisense primer (5 - ACAGCGAGCCTCTGGAACCT-3 ), and 30 cycles at 95 ◦ C for 5 s, 60 ◦ C for 10 s, and 72 ◦ C for 20 s. G3PDH expression was used as an internal control. The PCR products were separated by electrophoresis on 1.8% agarose gels (Fig. 1).

Fig. 1. p53R2 mRNA expression in TK6 cells treated with adriamycin (0.13 ␮g/mL), MNNG (5.0 ␮g/mL), 5-FU (33 ␮g/mL), and sodium ascorbate (100 ␮g/mL). TK6 cells were treated with the chemicals for 16 h. Expression of G3PDH gene was used for quality control in RT-PCR experiments.

2.4. Dose level The highest test concentration of each chemical was determined by its cytotoxicity and solubility before the luciferase reporter gene assay was carried out. Cytotoxicity was investigated by seapansy luciferase activity or the WST-1 assay using the Premix WST-1 Cell Proliferation Assay System (Takara Bio Inc., Shiga, Japan). WST-1 assay monitors the activity of mitochondrial dehydrogenases, and was used as an indicator of cell growth because DNA repair is affected by cell cycle. Briefly, TK6 cells were plated at an initial concentration of 1.0 × 105 cells/well in a 96-well microplate (Corning, Flanklin Lakes, NJ), and incubated at 37 ◦ C with 5% CO2 . After 1 h incubation, test samples or solvent (final concentration, 0.3% DMSO) were added. After 16 h incubation, the precipitation of insoluble test samples was investigated by microscopic observation, and the WST-1 assay was carried out. The highest concentration of the test samples for the luciferase reporter gene assay was set to give cell viability in the WST-1 assay of about 50% that of the control cells treated with solvent only, and no precipitation of insoluble test sample. Unless precluded by cytotoxicity and solubility, the highest concentration was set at about 10 mmol/L according to the current ICH S2B guideline. In order to investigate whether the level of cytotoxicity given by the WST-1 assay is relevant to the more commonly used method, the comparison was made between the WST-1 assay and relative cell count with respect to adriamycin, MNNG, 5-FU, and sodium ascorbate. The test condition of relative cell count was identical to that of the WST-1 assay. Following chemical treatment, relative cell count was calculated as previously reported [31]. The WST-1 assay slightly underestimates the level of cytotoxicity compared to relative cell count (Fig. 2A–D). Thus, it is assumed that the concentration that gives 50% of cytotoxicity in relative cell count is within the dose range of the luciferase reporter gene assay carried out in this study.

2.5. Plasmids The p53BS-Luc plasmid used in this study was constructed as previously described [12,13]. Briefly, three tandem repeats of the p53 binding site from intron 1 of human p53R2 were inserted into the KpnI-BglII site of the PGV-P2 plasmid (TOYO INK MFG, Tokyo, Japan). Before use in the reporter gene assay, p53BS-Luc plasmid was digested with BamHI and XmnI and the larger fragment was purified. pRL-SV40 internal control plasmid encoding the seapansy-derived luciferase gene downstream of the SV40 enhancer and early promoter elements was purchased from Promega (Promega, Madison, WI, USA).

2.6. Luciferase reporter gene assay and criteria for a judgment of genotoxicity TK6 cells (1.0 × 107 ) were transiently transfected with 500 ng of p53BS-Luc plasmid, and 10 ng pRL-SV40 plasmid using the Nucleofector Solution V and the X005 program on a Nucleofector® II Device (Amaxa biosystems, Cologne, Germany). Transfected cells were plated at an initial concentration of 1.0 × 105 cells/well in white, clear-bottomed 96-well microplates (Corning, Flanklin Lakes, NJ), and incubated at 37 ◦ C with 5% CO2 . One hour after transfection, test samples or solvent (final concentration, 0.3% DMSO) were added in triplicate. Following 16 h incubation, firefly and seapansy luciferase activities were measured with a Picagene Dual Seapansy Luciferase Kit (TOYO INK MFG, Tokyo, Japan) using a luminometer (Micro Lumat P96V; Berthold, Wildbad, Germany). The transfection efficiency of reporter plasmids was determined using luciferase activity from the seapansy luciferase plasmid. For data normalization, the luciferase activity of p53BS-Luc was divided by that of pRL-SV40 as an internal control in each well. Relative p53R2dependent luciferase activity was expressed as a % relative to control cells treated with solvent only. Each assay was triplicated and repeated atleast twice on different microplates. Benzo[a]pyrene (1.5 ␮g/mL) and adriamycin (0.2 ␮g/mL) were used as positive controls for assays with and without S9 treatment. The criteria for positive responses were established by the statistical analysis. In our assay, the value of the luciferase activity of the negative control treated with solvent only was about 100 ± 10%. Thirty percent induction of luciferase activity equates to approximately three times the standard deviation of the control cells, and was appeared statistically significant by the Student’s t-test analysis. Thus, the test sample was judged as positive for genotoxicity when the relative p53R2-dependent luciferase activity was 130% or greater and responded in a dose dependent manner. The internal control luciferase activity is under a constitutively active promoter, and correlates with the ability of the cells to express protein [32]. In cases of cytotoxicity where the internal luciferase activity falls below 76.9%, the relative p53R2-dependent luciferase activity may reach 130% without any increase in reporter luciferase activity derived from p53BS-Luc plasmid. To avoid a falsepositive result and surely detect the true-positive response in that case, the test sample was only judged as positive when both the relative p53R2-dependent luciferase activity and the reporter luciferase activity were induced over 130% in a dose dependent manner, and their increases were statistically significant. Tables 1–3 provide the summaries of the results from the p53R2-dependent luciferase reporter gene assay with respect to the ECVAM reference set. More detailed data for each chemical are shown in Appendices (supplementary data, Fig. A–C).

T. Mizota et al. / Mutation Research 724 (2011) 76–85

B

140 WST-1

Adriamycin

RCC

120 100 80 60 40 20 0 0.001

0.01

140 WST-1

Cell viability (WST-1, RCC, % of Control)

Cell viability (WST-1, RCC, % of Control)

A

0.1

100 80 60 40 20 0

1

0.01

0.1

D Cell viability (WST-1, RCC, % of Control)

Cell viability (WST-1, RCC, % of Control)

5-FU

RCC

100 80 60 40 20 0

Relative p53R2-dependent luciferase activity (% of Control)

10

140 WST-1

Sodium ascorbate

RCC

120 100 80 60 40 20 0

0.1

1

10

100

0.1

1

10

100

1000

Concentration (mg/mL)

Concentration (mg/mL)

E

1

Concentration (μg/mL)

C 140 120

MNNG

RCC

120

Concentration (μg/mL)

WST-1

81

600

500

*

Adriamycin MNNG 5-FU Sodium ascorbate

400

* 300

* *

200

* *

*

*

*

*

*

*

**

100

0 0.001

0.01

0.1

1

10

100

1000

Concentration (mg/mL) Fig. 2. Cell viability and relative p53R2-dependent luciferase activity in the TK6 cells. Cell viability of (A) adriamycin, (B) MNNG, (C) 5-FU, and (D) sodium ascorbate was measured by WST-1 assay (WST-1; closed circle) and relative cell count (RCC; open circle). (E) The relative p53R2-dependent luciferase activity treated with adriamycin (closed circles), MNNG (closed triangles), 5-FU (open circles), and sodium ascorbate (open triangles). TK6 cells transfected with p53BS-Luc plasmid were treated with the chemicals for 16 h. Each value represents the mean ± S.D. of triplicate assays. The dashed line represents the genotoxicity threshold. * P < 0.05 (vs. control, Student’s t-test). 2.7. S9 mix A rat liver S9 fraction, which was derived from livers of seven-week old male Sprague–Dawley [Crj: CD] rats that had been treated with a combination of phenobarbital and 5,6-benzoflavone that was purchased from Oriental Yeast Co., Ltd. (Tokyo, Japan). Handling of the S9 mix was modified to reduce its cytotoxic effect to TK6 cells. Ten microliters of the S9 fraction was

mixed with 90 ␮L of Co-factor I (Oriental Yeast Co., Ltd.) containing 8 mmol/L MgCl2 , 33 mmol/L KCl, 5 mmol/L glucose-6-phosphate, 4 mmol/L NADPH, 4 mmol/L NADH, and 100 mmol/L sodium phosphate buffer in 100 ␮L of culture medium. Test samples (6.0 ␮L) were added to the S9 mix and incubated at 37 ◦ C for 30 min, and an 11 ␮L aliquot of the S9 reaction mixture was added to each well. Cells were then incubated for a further 16 h with a final concentration of 0.5% S9.

82

T. Mizota et al. / Mutation Research 724 (2011) 76–85

2.8. Statistical analysis Values are expressed as the mean ± S.D. The Student’s t-test was used to determine the statistical significance of differences between groups, and P < 0.05 is considered significant.

3. Results 3.1. RT-PCR In order to investigate expression from the p53R2 gene induced by DNA damage in TK6 cells, we carried out RT-PCR experiment using adriamycin, MNNG, 5-FU, and sodium ascorbate (Fig. 1). Adriamycin is reported to induce DNA double strand breaks by stabilizing covalent complexes between topoisomerase II and DNA [33,34]. MNNG is a DNA alkylator that induces gene mutations [35,36], and 5-FU is a pyrimidine analogue that activates p53 by perturbation of the nucleotide pool [37]. All three types of genotoxins induced p53R2 mRNA expression (Fig. 1). Sodium ascorbate, commercially available vitamin C, was used as a reference nongenotoxin [38]. As expected, it did not induce the expression of p53R2 gene (Fig. 1). 3.2. p53R2-dependent luciferase reporter gene assay in TK6 cells In order to assess the applicability of p53R2-dependent luciferase reporter gene assay system in TK6 cells, the luciferase activities induced by adriamycin, MNNG, 5-FU, and sodium ascorbate were examined. Prior to conducting the luciferase assay, cytotoxicity was investigated by the WST-1 assay (Fig. 2A–D), and the highest concentration of the test samples was determined as described in Section 2. Dose-dependent and statistically significant increases in relative luciferase activities were observed in cells treated with adriamycin, MNNG, and 5-FU, to varying degrees (Fig. 2E). Sodium ascorbate did not induce luciferase activity, which is consistent with the initial RT-PCR experiment. These results show that the p53R2-dependent luciferase reporter gene assay can be applied in TK6 cells, and various types of genotoxic chemicals can be detected by this system. The standard deviation of control cells treated with solvent only was approximately 10% in triplicate measurements. The robustness of the TK6-based assay provided a statistically significant increase when the luciferase activity exceeded 120% that of the control cells. For this reason, we set the criteria for genotoxicity assessment in subsequent studies as follows: the relative luciferase activity of cells treated with the test sample was greater than or equal to 130% that of the control cells, and increased in a dose-dependent manner; and the difference between luciferase activities of treated and control cells was statistically significant. 3.3. p53R2-dependent luciferase reporter gene assay in TK6 cells for Group 1 chemicals Group 1 in the ECVAM reference set for genotoxicity tests consists of 20 in vivo genotoxins that should be detected as positive in in vitro mammalian cell genotoxicity tests. When these 20 chemicals were tested using the p53R2-dependent luciferase reporter gene assay in TK6 cells, 17 were judged as positive (Table 1), giving an assay sensitivity of 85% (17/20). Of the Section I chemicals in Group 1, the Ames-positive in vivo genotoxins, cadmium chloride and p-chloroaniline gave negative results, and cyclophosphamide, benzo[a]pyrene, 7,12dimethylbenzanthracene, and dimethylnitrosamine gave positive results only in the presence of the S9 mix. The latter four chemicals are known to require metabolic activation for their genotoxic effect [11]. Conversely, 2-acetylaminofluorene, 2,4-diaminotoluene, IQ,

PhIP, and aflatoxin B1 gave positive results without S9 pretreatment, even though these five chemicals require metabolic activation [11]. Of the Section II chemicals in Group 1, which include the in vivo genotoxins that give negative or equivocal results in the Ames test, only hydroquinone gave a negative result. 3.4. p53R2-dependent luciferase reporter gene assay for Group 2 chemicals Group 2 of the ECVAM reference set consists of 23 non-DNA reactive chemicals, which includes non-genotoxic carcinogens that should give negative results in in vitro mammalian cell genotoxicity tests. Twenty-one of the 22 chemicals tested from this group gave negative results in the p53R2-dependent luciferase reporter gene assay using TK6 (Table 2). Subgroup (ii) of Group 2 includes noncarcinogens with no in vivo genotoxicity data, and ephedrine sulfate is included in this group. No data is presented for this compound as it was not tested in this study because of restrictions surrounding its purchase. Thus, the specificity of the assay in this group is 95% (21/22). Subgroup (iii) of Group 2 includes the non-genotoxic carcinogens, and only diethanolamine from this subgroup gave a positive result. 3.5. p53R2-dependent luciferase reporter gene assay for Group 3 chemicals Group 3 of the ECVAM reference set consists of 19 nonDNA-reactive chemicals, including non-genotoxic carcinogens, metabolic poisons and other substances that should give negative results in in vitro mammalian cell genotoxicity tests [11]. The chemicals in this group have been reported to induce chromosomal aberrations or tk mutations in mouse lymphoma cells, often at high concentrations or at high levels of cytotoxicity. Seventeen of the 19 chemicals in this group gave negative results using our p53R2-based assay (Table 3). Subgroup (iv) of Group 3 includes chemicals that give unclear results for the prediction of in vitro genotoxicity. Eugenol and isobutyraldehyde from this subgroup gave positive results in our test. When the results of Group 2 and Group 3 are considered together, the specificity of the assay is 93% (38/41). 4. Discussion In this work we have demonstrated that a previously developed genotoxicity test system based on p53R2 gene expression is applicable in human lymphoblastoid TK6 cells. TK6 cells have been widely used for genetic toxicology studies [24–26], and the tk locus mutation assay using TK6 cells is accepted by ICH guideline for pharmaceuticals (S2B-Genotoxicity: A Standard Battery for Genotoxicity Testing of Pharmaceuticals) [29]. Pre-validation studies suggested that TK6 cells were able to induce both p53R2 expression and p53R2-dependent luciferase activity in response to genotoxins (Figs. 1 and 2). Validation against the 61 available chemicals from the ECVAM lists showed that the overall accordance, sensitivity, and specificity were 90%, 85%, and 93%, respectively. Chemicals from Group 3 have been reported to induce chromosomal aberrations or tk mutations in mouse lymphoma cells, often at high concentrations or at high levels of cytotoxicity [11]. Thus, if we re-assess performance with the exception of Group 3, the assay achieved a higher specificity of 95% (21/22). This indicates that our assay can detect various types of genotoxic chemicals with high specificity. Of the 20 genotoxins in the ECVAM lists, our genotoxicity assay using TK6 cells detected 17 chemicals as positive. However,

T. Mizota et al. / Mutation Research 724 (2011) 76–85

cadmium chloride, p-chloroaniline, and hydroquinone gave false negative results. Despite the fact that cadmium chloride was defined as an Ames-positive chemical in the ECVAM lists, Ames tests in the TOXNET database were all negative in both the presence and absence of S9 metabolic activation [39]. Similarly, the dye intermediate p-chloroaniline has variable Ames-test results in the TOXNET database [39], and also conflicting data in in vitro mammalian cell genotoxicity tests [11]. Hydroquinone is a benzene metabolite, and is negative or equivocal in the Ames tests [11]. The Ames test data suggests that the chemicals that gave negative results in our assay have weak DNA reactivity. Westerink et al. reported a new genotoxicity test system based on Rad51C expression using HepG2 cells [40]. The Rad51C protein plays an important role in repair of DNA double strand breaks via homologous recombination [41]. Cadmium chloride, p-chloroaniline, and hydroquinone all tested negative, and Westerink et al. proposed that these chemicals did not activate the repair pathway for DNA double strand breaks. We previously showed that the p53R2-based genotoxocity assay system has high reactivity to DNA double strand break agents like topoisomerase II inhibitors [13]. Taken together, this suggests that negative results for cadmium chloride, p-chloroaniline, and hydroquinone in our assay may be due to their mechanism of action. There are nine chemicals that are known to require or be activated by metabolic activation in the ECVAM lists. Five of them, 2-acetylaminofluorene, 2,4-diaminotoluen, IQ, PhIP, and aflatoxin B1, gave positive results without pre-treatment using S9 mix in our assay (Table 1). Recently, Birrell et al. reported that four of these five chemicals, with the exception of 2-acetylaminofluorene, produced a positive genotoxic response even without S9 mix in Gadd45abased genotoxicity tests using TK6 cells [42]. Their assay system monitors the expression of the GADD45a gene, which is a p53regulated gene induced by DNA damage [43], using a GFP reporter [26,42]. It was suggested that TK6 cells themselves may possess the ability to activate 2,4-diaminotoluene, IQ, PhIP, and aflatoxin B1 without S9 treatment. Even though our assay used a different p53-regulated gene, the ability of TK6 to activate the chemicals may have contributed to our positive results in the absence of S9 treatment. 2-Acetylaminofluorene was reported to give positive results regardless of the presence of metabolic activation in various in vitro mammalian tests with different endpoints, including gene mutations, chromosome aberrations, and sister chromatid exchanges [44]. Considering this data, it is not surprising that our assay positively detected 2-acetylaminofluorene in the absence of S9 treatment. Of the 41 non-genotoxins in ECVAM lists, our genotoxicity assay using TK6 cells judged 38 chemicals as negative, but diethanolamine, eugenol, and isobutyraldehyde were false positives. Diethanolamine, an alkanolamine used widely in industry, is known to show carcinogenicity by inducing choline deficiency [45,46]. Choline deficiency has been associated with the induction of p53 and p21WAF1/CIP1 gene expression [47] and DNA damages [48]. Consequently, diethanolamine may also induce expression of the p53R2 gene, and lead to the false positive result. Eugenol, a major ingredient of herbs, is reported to elicit anti-proliferative activity on mast cells by inducing apoptosis via translocation of phosphorylated p53 into mitochondria [49]. The authors observed phosphorylation of ser-15 of p53, the same site that is phosphorylated in response to DNA damaging agents, including ionizing irradiation, UV irradiation, and genotoxins [50,51]. Considering the mode of action both of diethanolamine and eugenol, it is plausible that these chemicals induced luciferase activity in our assay. Isobutyraldehyde, one of the alkyl aldehydes, produced a positive response at the high concentration of >530 ␮g/mL (corresponding to 7.4 mmol/L). It has also been reported to give positive

83

results in a mouse lymphoma gene mutation assay in L5178Y cells, and induce chromosomal aberrations in in vivo and in vitro [52]. In light of this data, the positive result for isobutyraldehyde from our assay is not surprising. There are two reports evaluating in vitro mammalian cell genotoxicity tests carried out with the chemicals on the ECVAM lists [40,42]. One of them is a genotoxicity test system based on Rad51C expression using HepG2 cells reported by Westerink et al. [40]. In addition to Rad51C-based test system, Westerink et al. developed the reporter gene assay using p53 response element in HepG2 cells [40]. The overall accordance, sensitivity, and specificity of their p53 response element reporter were 92%, 85%, and 93% respectively. Although these scores were almost same as those of our p53R2based genotoxicity assay, the chemicals that gave the inconsistent results with respect to ECVAM lists were quite different. Westerink et al. used four tandem repeats of the 10-bp p53 binding sequence, which is different from our reporter sequence. Moreover, their assays were conducted without S9 mix and the highest test concentration was limited at 1 mmol/L. Thus, it is suggested that these factors may have led to the different responses between our assay and Westerink’s p53 response element reporter assay. Other test system is Gadd45a-based genotoxicity test using TK6 cells reported by Birrell et al. [42]. There are several differences between the two test’s results. For example, cadmium chloride and hydroquinone produced a positive genotoxic response in Gadd45a-based genotoxicity test. On the other hand, chloramphenicol gave a positive result only in p53R2-based genotoxicity test. The top test concentration of these three chemicals is much higher in Gadd45a-based assay compared to our p53R2-based assay. Thus, different results between the two tests are not solely due to top dose and/or toxicity. This notion is supported by the fact that the lowest efficiency concentration where these chemicals produced positive responses in Gadd45a-based assay was within a dose range in p53R2-based assay. Thus, it is suggested that the different outcome between the two test systems may be due to the factor such as difference in the regulation of the reporter gene expression. The studies conducted by Westerink et al., Birrell et al., and us achieved high specificity even in different systems, although it should be noted that the tests are all based on the expression of the genes induced by DNA damage. Alternatively, in vitro mammalian cell genotoxicity assays, such as chromosomal aberrations test and micronucleus test, can be performed by monitoring for altered phenotype. However, phenotypic changes often need high concentrations or high toxicity to be observed, and may lead to a high rate of false positives [8]. In conclusion, we applied the genotoxicity test system based on p53R2 expression, which was initially developed using MCF-7 cells, to TK6 cells. The overall accordance, sensitivity, and specificity were 90%, 85%, and 93%, respectively, in respect to the ECVAM lists. When the two cell lines were compared, the luciferase activity induced by adriamycin was higher in MCF-7 cells than in TK6 cells. The reactivity was also slightly different for the two cell lines. For example, antimetabolites such as 5-FU gave positive results only in TK6 cells. On the other hand, MCF-7 cells could detect the genotoxicity of cadmium chloride. These results indicate that further modification of the test system is necessary to ensure complete detection of all the genotoxins in each cell. Each newly developed assay needs to be evaluated using multiple chemicals. While only a small number of chemicals were tested in this study, the performance of our assay performed as judged by the overall accordance, sensitivity, and specificity for the ECVAM list was good in comparison with other tests that evaluated hundreds of chemicals [40,42]. Our assay is advantageous because results can be obtained within a day with the test chemicals, the operating system is simple, and throughput is high. The genotoxicity test developed in this study can be used to detect various genotoxic chemicals without

84

T. Mizota et al. / Mutation Research 724 (2011) 76–85

compromising the specificity. This test will be a valuable tool for rapid screen for identifying chemicals that may be genotoxic to humans. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.mrgentox.2011.06.003. References [1] B.N. Ames, F.D. Lee, W.E. Durston, An improved bacterial test system for the detection and classification of mutagens and carcinogens, Proc. Natl. Acad. Sci. U. S. A. 70 (1973) 782–786. [2] P. Quillardet, O. Huisman, R. d’Ari, M. Hofnung, SOS-Chromotest, a directl assay of induction of a SOS function in Escherichia coli K-12 to measure genotoxicity, Proc. Natl. Acad. Sci. U. S. A. 79 (1982) 5971–5975. [3] Y. Oda, S. Nakamura, I. Oki, T. Kato, H. Shinagawa, Evaluation of the new system (umu-test) for the detection of environmental mutagens and carcinogens, Mutat. Res. 147 (1985) 219–229. [4] S.M. Galloway, A.D. Bloom, M. Resnick, B.H. Margolin, F. Nakamura, P. Archer, E. Zeiger, Development of a standard protocol for in vitro cytogenetic testing with Chinese hamster ovary cells: comparison of results for 22 compounds in two laboratories, Environ. Mutagen. 7 (1985) 1–51. [5] D. Clive, J.F.S. Spector, Laboratory procedure for assessing specific locus mutations at the TK locus in cultured L5178Y mouse lymphoma cells, Mutat. Res. 31 (1975) 17–29. [6] N.P. Singh, M.T. McCoy, R.R. Tice, E.L. Schneider, A simple technique for quantitation of low levels of DNA damage in individual cells, Exp. Cell Res. 175 (1988) 184–191. [7] T. Matsushima, M. Hayashi, A. Matsuoka, M. Ishidate Jr., K.F. Miura, H. Shimizu, Y. Suzuki, K. Morimoto, H. Ogura, K. Mure, K. Koshi, T. Sofuni, Validation study of the in vitro micronucleus test in a Chinese hamster lung cell line (CHL/IU), Mutagenesis 14 (1999) 569–580. [8] D. Kirkland, S. Pfuhler, D. Tweats, M. Aardema, R. Corvi, F. Darroudi, A. Elhajouji, H. Glatt, P. Hastwell, M. Hayashi, P. Kasper, S. Kirchner, A. Lynch, D. Marzin, D. Maurici, J.R. Meunier, L. Müller, G. Nohynek, J. Parry, E. Parry, V. Thybaud, R. Tice, J. van Benthem, P. Vanparys, P. White, How to reduce false positive results when undertaking in vitro genotoxicity testing and thus avoid unnecessary follow-up animal tests: report of an ECVAM workshop, Mutat. Res. 628 (2007) 31–55. [9] D. Kirkland, M. Aardema, L. Henderson, L. Müller, Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and non-carcinogens I. Sensitivity, specificity and relative predictivity, Mutat. Res. 584 (2005) 1–256. [10] E.J. Matthews, N.L. Kruhlak, M.C. Cimino, R.D. Benz, J.F. Contrera, An analysis of genetic toxicity, reproductive and developmental toxicity, and carcinogenicity data: I. Identification of carcinogens using surrogate endpoints, Regul. Toxicol. Pharmacol. 44 (2006) 83–96. [11] D. Kirkland, P. Kasper, L. Müller, R. Corvi, G. Speit, Recommended lists of genotoxic and non-genotoxic chemicals for assessment of the performance of new or improved genotoxicity tests: a follow-up to an ECVAM workshop, Mutat. Res. 653 (2008) 99–108. [12] K. Ohno, Y. Tanaka-Azuma, Y. Yoneda, T. Yamada, Genotoxicity test system based on p53R2 gene expression in human cells: examination with 80 chemicals, Mutat. Res. 588 (2005) 47–57. [13] K. Ohno, K. Ishihata, Y. Tanaka-Azuma, T. Yamada, A genotoxicity test system based on p53R2 gene expression in human cells: assessment of its reactivity to various classes of genotoxic chemicals, Mutat. Res. 656 (2008) 27–35. [14] H. Tanaka, H. Arakawa, T. Yamaguchi, K. Shiraishi, S. Fukuda, K. Matsui, Y. Takei, Y. Nakamura, A ribonucleotide reductase gene involved in a p53-dependent cell-cycle checkpoint for DNA damage, Nature 404 (2000) 42–49. [15] O. Guittet, P. Hakansson, N. Voevodskaya, S. Fridd, A. Graslund, H. Arakawa, Y. Nakamura, L. Thelander, Mammalian p53R2 protein forms an active ribonucleotide reductase in vitro with the R1 protein, which is expressed both in resting cells in response to DNA damage and in proliferating cells, J. Biol. Chem. 276 (2001) 40647–40653. [16] W.S. El-Deiry, S.E. Kern, J.A. Pietenpol, K.W. Kinzler, B. Vogelstein, Definition of a consensus binding site for p53, Nat. Genet. 1 (1992) 45–49. [17] W.D. Funk, D.T. Pak, R.H. Karas, W.E. Wright, J.W. Shay, A transcriptionally active DNA-binding site for human p53 protein complexes, Mol. Cell. Biol. 12 (1992) 2866–2871. [18] Y. Wang, J.F. Schwedes, D. Parks, K. Mann, P. Tegtmeyer, Interactions of p53 with its consensus DNA-binding site, Mol. Cell. Biol. 15 (1995) 2157–2165. [19] Y. Nakamura, Isolation of p53-target genes and their functional analysis, Cancer Sci. 95 (2004) 7–11. [20] T. Kimura, S. Takeda, Y. Sagiya, M. Gotoh, Y. Nakamura, H. Arakawa, Impaired function of p53R2 in Rrm2b-null mice causes severe renal failure through attenuation of dNTP pools, Nat. Genet. 34 (2003) 440–445.

[21] M.A. Lebedeva, J.S. Eaton, G.S. Shadel, Loss of p53 causes mitochondrial DNA depletion and altered mitochondrial reactive oxygen species homeostasis, Biochim. Biophys. Acta 1787 (2009) 328–334. [22] A. Bourdon, L. Minai, V. Serre, J.P. Jais, E. Sarzi, S. Aubert, D. Chrétien, P. de Lonlay, V. Paquis-Flucklinger, H. Arakawa, Y. Nakamura, A. Munnich, A. Rötig, Mutation of RRM2B, encoding p53-controlled ribonucleotide reductase (p53R2), causes severe mitochondrial DNA depletion, Nat. Genet. 39 (2007) 776–780. [23] L. Xue, B. Zhou, X. Liu, Y. Heung, J. Chau, E. Chu, S. Li, C. Jiang, F. Un, Y. Yen, Ribonucleotide reductase small subunit p53R2 facilitates p21 induction of G1 arrest under UV irradiation, Cancer Res. 67 (2007) 16–21. [24] L.S. Zhang, M. Honma, M. Hayashi, T. Suzuki, A. Matsuoka, T. Sofuni, A comparative study of TK6 human lymphoblastoid and L5178Y mouse lymphoma cell lines in the in vitro micronucleus test, Mutat. Res. 347 (1995) 105–115. [25] L. Henderson, A. Wolfreys, J. Fedyk, C. Bourner, S. Windebank, The ability of the Comet assay to discriminate between genotoxins and cytotoxins, Mutagenesis 13 (1998) 89–94. [26] P.W. Hastwell, L.L. Chai, K.J. Roberts, T.W. Webster, J.S. Harvey, R.W. Rees, R.M. Walmsley, High-specificity and high-sensitivity genotoxicity assessment in a human cell line: validation of the GreenScreen HC GADD45a-GFP genotoxicity assay, Mutat. Res. 607 (2006) 160–175. [27] H. Ellinger-Ziegelbauer, J.M. Fostel, C. Aruga, D. Bauer, E. Boitier, S. Deng, D. Dickinson, A.C. Le Fevre, A.J. Fornace Jr., O. Grenet, Y. Gu, J.C. Hoflack, M. Shiiyama, R. Smith, R.D. Snyder, C. Spire, G. Tanaka, J. Aubrecht, Characterization and interlaboratory comparison of a gene expression signature for differentiating genotoxic mechanisms, Toxicol. Sci. 110 (2009) 341–352. [28] S.A. Amundson, K.T. Do, L. Vinikoor, C.A. Koch-Paiz, M.L. Bittner, J.M. Trent, P. Meltzer, A.J. Fornace Jr., Stress-specific signatures: expression profiling of p53 wild-type and -null human cells, Oncogene 24 (2005) 4572–4579. [29] International Conference on Harmonization (ICH), S2B, Genotoxicity: a standard battery for genotoxicity testing of pharmaceuticals. http://www.fda. gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm074929.pdf, 2010 (accessed October 2010). [30] F. Xia, X. Wang, Y.H. Wang, N.M. Tsang, D.W. Yandell, K.T. Kelsey, H.L. Liber, Altered p53 status correlates with differences in sensitivity to radiationinduced mutation and apoptosis in two closely related human lymphoblast lines, Cancer Res. 55 (1995) 12–15. [31] E. Lorge, M. Hayashi, S. Albertini, D. Kirkland, Comparison of different methods for an accurate assessment of cytotoxicity in the in vitro micronucleus test. I. Theoretical aspects, Mutat. Res. 655 (2008) 1–3. [32] T. Schagat, A. Paguio, K. Kopish, Normalizing genetic reporter assays: approaches and considerations for increasing consistency and statistical significance, Cell Note (Technical Tips) 17 (2007) 9–12. [33] B.R. Abdella, J. Fisher, A chemical perspective on the anthracycline antitumor antibiotics, Environ. Health Perspect. 64 (1985) 4–18. [34] M. Binaschi, R. Farinosi, C.A. Austin, L.M. Fisher, F. Zunino, G. Capranico, Human DNA topoisomerase II alpha-dependent DNA cleavage and yeast cell killing by anthracycline analogues, Cancer Res. 58 (1998) 1886–1892. [35] E.L. Loechler, C.L. Green, J.M. Essigmann, In vivo mutagenesis by O6 methylguanine built into a unique site in a viral genome, Proc. Natl. Acad. Sci. U. S. A. 81 (1984) 6271–6275. [36] A.C. Povey, A.F. Badawi, D.P. Cooper, C.N. Hall, K.L. Harrison, P.E. Jackson, N.P. Lees, P.J. O’Connor, G.P. Margison, DNA alkylation and repair in the large bowel: animal and human studies, J. Nutr. 132 (Suppl. (11)) (2002) 3518S–3521S. [37] S.P. Linke, K.C. Clarkin, A. Di Leonardo, A. Tsou, G.M. Wahl, A reversible, p53-dependent G0 /G1 cell cycle arrest induced by ribonucleotide depletion in the absence of detectable DNA damage, Genes Dev. 10 (1996) 934–947. [38] A.T Diplock, Safety of antioxidant vitamins and ␤-carotene, Am. J. Clin. Nutr. 62 (Suppl. (6)) (1995) 1510S–1516S. [39] United States National Library of Medicine, Toxicology Data Network (TOXNET), Chemical Carcinogenesis Research Information System (CCRIS). http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen?CCRIS, 2010 (accessed October 2010). [40] W.M. Westerink, J.C. Stevenson, G.J. Horbach, W.G. Schoonen, The development of RAD51C, Cystatin A, p53 and Nrf2 luciferase-reporter assays in metabolically competent HepG2 cells for the assessment of mechanism-based genotoxicity and of oxidative stress in the early research phase of drug development, Mutat. Res. 696 (2010) 21–40. [41] S.K. Sharan, S.G. Kuznetsov, Resolving RAD51C function in late stages of homologous recombination, Cell Div. 2 (2007) 15. [42] L. Birrell, P. Cahill, C. Hughes, M. Tate, R.M. Walmsley, GADD45a-GFP GreenScreen HC assay results for the ECVAM recommended lists of genotoxic and non-genotoxic chemicals for assessment of new genotoxicity tests, Mutat. Res. 695 (2010) 87–95. [43] Q. Zhan, Gadd45a, a p53- and BRCA1-regulated stress protein, in cellular response to DNA damage, Mutat. Res. 569 (2005) 133–143. [44] National Toxicology Program (NTP), Database Search Application. http://ntpapps.niehs.nih.gov/ntp tox/index.cfm, 2010 (accessed October 2010). [45] P.M. Newberne, Choline deficiency associated with diethanolamine carcinogenicity, Toxicol. Sci. 67 (2002) 1–3. [46] L.D. Lehman-McKeeman, E.A. Gamsky, S.M. Hicks, J.D. Vassallo, M.H. Mar, S.H. Zeisel, Diethanolamine induces hepatic choline deficiency in mice, Toxicol. Sci. 67 (2002) 38–45.

T. Mizota et al. / Mutation Research 724 (2011) 76–85 [47] M.Q. Holmes-McNary, A.S. Baldwin Jr., S.H.S.H. Zeisel, Opposing regulation of choline deficiency-induced apoptosis by p53 and nuclear factor ␬B, J. Biol. Chem. 276 (2001) 41197–41204. [48] C.D. Albright, R. Liu, T.C. Bethea, K.A. Da Costa, R.I. Salganik, S.H. Zeisel, Choline deficiency induces apoptosis in SV40-immortalized CWSV-1 rat hepatocytes in culture, FASEB J. 10 (1996) 510–516. [49] B.S. Park, Y.S. Song, S.B. Yee, B.G. Lee, S.Y. Seo, Y.C. Park, J.M. Kim, H.M. Kim, Y.H. Yoo, Phospho-ser 15-p53 translocates into mitochondria and interacts with Bcl-2 and Bcl-xL in eugenol-induced apoptosis, Apoptosis 10 (2005) 193–200.

85

[50] S.Y. Shieh, M. Ikeda, Y. Taya, C. Prives, DNA damage-induced phosphorylation of p53 alleviates inhibition by MDM2, Cell 91 (1997) 325–334. [51] J.D. Siliciano, C.E. Canman, Y. Taya, K. Sakaguchi, E. Appella, M.B. Kastan, DNA damage induces phosphorylation of the amino terminus of p53, Genes Dev. 11 (1997) 3471–3781. [52] National Toxicology Program (NTP), NTP toxicology and carcinogenesis studies of isobutyraldehyde (CAS No. 78-84-2) in F344/N rats and B6C3F1 mice (Inhalation studies), Nat. Toxicol. Program Tech. Rep. Ser. 472 (1999) 1–242.