Environmental and Molecular Mutagenesis 56:468^476 (2015)
Research Article An Evaluation of 25 Selected ToxCast Chemicals in Medium-Throughput Assays to Detect Genotoxicity Andrew D. Kligerman,1* Robert R.Young,2 Leon F. Stankowski Jr.,2 Kamala Pant,2 Tim Lawlor,2 Marilyn J. Aardema,2,3 and Keith A. Houck1 1
National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 2 BioReliance Corporation, Rockville, Maryland 3 Marilyn Aardema Consulting LLC, 5315 Oakbrook Dr., Fairfield, Ohio ToxCast is a multiyear effort to develop a costeffective approach for the US EPA to prioritize chemicals for toxicity testing. Initial evaluation of more than 500 high-throughput (HT) microwell-based assays without metabolic activation showed that most lacked high specificity and sensitivity for detecting genotoxicants. Thus, EPA initiated a pilot project to investigate the use of standard genotoxicity endpoints using medium-throughput genotoxicity (MTG) assays in the context of a large testing program. Twentyfive chemicals were selected from the ToxCast program based in part on their known genotoxicity. The two MTG assays used were the Ames TM assay and 96-well In Vitro MicroFlowV II Micronucleus (MN) assay. The Ames II assay showed a reasonable correlation with published Ames test data and industry submissions, though specificity was much better than sensitivity due R
to restraints on top concentrations as prescribed by ToxCast. Overall concordance was 73% both with and without metabolic activation. The flow MN assay had concordances of 71% and 58% with and without metabolic activation, respectively, when compared to published data and submissions. Importantly, a comparison of results without S9 from the MTG assays to an HT ToxCast p53 activation assay showed a fairly good degree of concordance (67%). The results reported here indicate that assays for genotoxicity endpoints can be conducted in a MT format and have the potential to add to the interpretation of results from large-scale testing programs such as EPA’s ToxCast program. Inherent limitations such as the top concentrations used in large scale testing programs are discussed. Environ. Mol. Mutagen. 56:468– C 2014 Wiley Periodicals, Inc. 476, 2015. V
Key words: genotoxicity; high-throughput testing; Ames test; micronucleus test; ToxCast
INTRODUCTION A standard battery of short-term tests has been used for several decades to screen chemicals for genotoxicity with the aim of detecting mutagenic carcinogens as well as to
Additional Supporting Information may be found in the online version of this article. *Correspondence to: Andrew D. Kligerman, Integrated Systems Toxicology Division, Office of Research and Development, National Health and Environmental Effects Research Laboratory, MD 105-03, U.S. Environmental Protection Agency, Research Triangle Park, NC 27710, USA. E-mail: [email protected], [email protected] Disclaimer: The information in this document has been funded wholly (or in part) by the U. S. Environmental Protection Agency. It has been subjected to review by the National Health and Environmental Effects Research Laboratory and approved for publication. Approval does not signify that the contents reflect the views of the Agency nor does C 2014 Wiley Periodicals, Inc. V
detect germ cell mutagens to limit human exposure to these potentially dangerous chemicals. However, a large number of environmental and industrial chemicals have not been tested in any significant way either due to a lack of legal authority, funding, resources, will of the parties
mention of trade names or commercial products constitute endorsement or recommendation for use. Conflict of interest: BioReliance is a contract research organization that TM performs the Ames II and the 96-well In Vitro MicroFlowV Micronucleus Assays. BioReliance has global rights ti the Ames II assay with the exception that rights in Europe are co-exclusive. Ames II kits are licensed for sale by MolTox (Boone, NC). R
Received 24 July 2014; provisionally accepted 17 November 2014; and in final form 18 November 2014 DOI 10.1002/em.21934 Published online 23 December 2014 in Wiley Online Library (wileyonlinelibrary.com).
Environmental and Molecular Mutagenesis. DOI 10.1002/em Medium-Throughput Assays for Genotoxicity
involved, or a combination of these factors. An approach needs to be established that selects chemicals for testing in an efficient manner. ToxCast is a multiyear effort to develop a cost-effective approach for prioritizing thousands of data-poor chemicals for further toxicity testing [Dix et al., 2007; Kavlock et al., 2012]. Phase I of ToxCast was designed to screen for a large number of toxicity endpoints using approximately 300 well-studied chemicals, mostly pesticides, in greater than 500 highthroughput (HT), microwell plate-based assays that generally lacked metabolic activation. It was anticipated that the various endpoints measured in the ToxCast assays would be useful for predicting higher level outcomes such as genotoxicity. In the first analysis of the usefulness of these HT assays for predicting genotoxicity, Knight et al. [2009] evaluated three HT assays (GreenScreen HC GADD45a-GFP, Gentronix; CellCiphr p53, Cellumen; and CellSensor p53REbla, Invitrogen Corp.). They concluded that the assays lacked high sensitivity and accuracy; however, the assays had relatively high specificity due to the preponderance of nongenotoxicants in the Phase I chemical set. These poorer than expected results were due in part to the predefined relatively low concentrations used in the study coupled with the lack of metabolic activation in the three assays. To refine the analysis, Kligerman et al. [2012] used the proprietary EPA submissions (US EPA Office of Pesticides and Toxic Substances), California Department of Pesticide Regulation Toxicology Data Review Summaries, and the National Toxicology Program Website as primary sources; secondary sources were LeadscopeV and the open scientific literature to classify the Phase I chemicals as to whether or not they were point mutagens, clastogens, or aneugens without exogenous metabolic activation (i.e., direct-acting genotoxicants). A evaluation (call) was based on the weight of the evidence presented, the presence or absence of contradictory or equivocal results, and for the published literature the reputation of the journal and the publication record of the research group. Of the 293 Phase I chemicals examined, 288 had some data on genotoxicity in the absence of metabolic activation. Of these, 7 were point mutagens, 46 were clastogens, and 13 were deemed to be aneugens (though not mutually exclusive). The majority were either nongenotoxic, or they had contradictory or limited information available, thus precluding the making of a call on this latter group. Therefore, a chemical was assigned a call as a genotoxicant (mutagen, clastogen, and/or aneugen), a nongenotoxicant (not a mutagen, clastogen, or aneugen), or inconclusive either due to the lack of data, the contradictory nature of the literature, or the weakness of the data. Once a chemical was assigned a category, univariate analysis was performed with an examination of each chemical in each ToxCast HT assay alone or in pairs to determine whether there were assays that detected R
469
genotoxicants, either serendipitously or due to the fact that some HT assays had indeed measured endpoints related to DNA damage checkpoints, microtubule disruption, or other specific cellular responses linked to genotoxicity. Based on Dr. Kligerman’s calls on the genotoxicity of the chemicals, none of the ToxCast HT assays tested, either individually or in pairs, showed both high specificity and sensitivity for detecting genotoxicants. Those assays that correctly identified a reasonable number of suspected direct-acting genotoxicants falsely predicted a large number of nongenotoxicants to be genotoxic [Kligerman et al., 2012]. Based on the importance of the ToxCast project for predicting various endpoints of toxicity and the importance of genotoxicity to the overall prioritization of chemicals for further testing, it was decided that data from standard genotoxicity endpoints generated in a mediumthroughput format (MTG assays) would be useful to supplement the existing ToxCast data. In this pilot study, 25 Phase I and Phase II chemicals from the ToxCast program were selected for genotoxicity testing with and without metabolic activation (S9) and included chemicals that (a) were known to be direct-acting genotoxicants (captan, folpet, captafol, dimethyl sulfate), (b) were known to be genotoxic after metabolic activation (7,12dimethylbenz(a)anthracene, N-nitrosodiethylamine), (c) were compounds giving strong responses in HT assays that might be related to genetic damage (4-chloro-1,2-diaminobenzene, allethrin, carbendazim, cladribine, triglycidyl isocyanurate, ziram), or (d) produced results in the HT assays that were of interest due to limited or contradictory published information on their genotoxicity (allethrin, diniconazole, monocrotophos, oryzalin). Here, we report the results and information gleaned from genotoxicity testing of these 25 chemicals using two MTG TM platforms: the Ames II assay and the 96-Well In Vitro V MicroFlow CHO MN assay (all with and without S9). The endpoints measured were point mutations in bacteria and clastogenicity and aneugenicity in rodent cells. We also compare these MTG results to data from a HT ToxCast p53 activation assay in the HepG2 human liver hepatoma cell line, an assay that is related to genotoxicity. R
MATERIALS AND METHODS ToxCast Chemicals The chemicals tested, their CAS#, purity, and source are listed in Table I. The supplier provided purities were all greater than 93%. Consistent with all of the ToxCast HT assays, the chemicals were dissolved in dimethylsulfoxide (DMSO) at a final stock concentration of 20 mM, and coded stock formulations of each chemical in DMSO were then shipped to BioReliance under contract to the EPA (EP-D-11-081) and as prescribed under ToxCast guidelines. The 20 mM stock solution was chosen based on consideration of realistic exposure scenarios, greatly increased solubility issues both in DMSO above this concentration and in the aqueously diluted samples in the bioassay at concentrations exceeding 20–100 lM, and increased cytotoxicity again seen as
Environmental and Molecular Mutagenesis. DOI 10.1002/em 470
Kligerman et al.
TABLE I. Chemicals Tested (Name, CAS No., Source, and Purity) Bensulide 4-Chloro-1,2-diaminobenzene 7,12-Dimethylbenz(a)anthracene 9-Phenanthrol Allethrin Captafol Captan Carbendazim Cladribine Chlorophene Dimethyl sulfate Diniconazole-M Fludioxonil Folpet Methyl isothiocyanate Methylene bis(thiocyanate) Mevinphos Monocrotophos Naled N-Nitrosodiethylamine Oryzalin Thiram Triclosan Triglycidyl isocyanurate Ziram
741-58-2 95-83-0 57-97-6 484-17-3 584-79-2 2425-06-1 133-06-2 10605-21-7 4291-63-8 120-32-1 77-78-1 83657-18-5 131341-86-1 133-07-3 556-61-6 6317-18-6 7786-34-7 6923-22-4 300-76-5 55-18-5 19044-88-3 137-26-8 3380-34-5 2451-62-9 137-30-4
concentrations exceed 20–50 lM in many HT assays. Testing of large chemical libraries precludes handling chemicals on an individual basis with the understanding that not all chemicals will have optimal conditions. In all the studies, the DMSO concentration was the same in each well.
Ames II Assay
Sigma Chemical Company Sigma Chemical Company LightBiologicals Sigma Chemical Company LightBiologicals Sigma Chemical Company ChemService, Inc. LightBiologicals Sigma Chemical Company Sigma Chemical Company LightBiologicals Waterstone Technology LLC Sigma Chemical Company Sigma Chemical Company Enamine Sigma Chemical Company Sigma Chemical Company Sigma Chemical Company Sigma Chemical Company LightBiologicals LightBiologicals Sigma Chemical Company Sigma Chemical Company Sigma Chemical Company LightBiologicals
99.5 97.3 99 93.9 96.1 97 99.2 97 98.5 98 99.8 95 99.9 99.9 97 99.3 93.3 99.9 97.2 99.9 98.2 99.9 99.5 97 99.2
Controls 2-Aminoanthracene (2AA; final concentration: 0.026 mM) and a mixture of 4-nitroquinoline N-oxide (4NQO; final concentration: 0.005 mM) and 2-nitrofluorene (2NF; final concentration: 0.009 mM) were used as positive controls with and without S9, respectively. All were purchased from Sigma-Aldrich (St. Louis, MO). DMSO was the vehicle control.
The Ames II assay was performed using Salmonella typhimurium strain TA98 and mixed Ames II strains (TAMix), that consists of TA7001, TA7002, TA7003, TA7004, TA7005, and TA7006 combined into a single culture. The exposure and indicator media for the assay were purchased commercially (MolTox, Inc., Boone, NC). Overnight cultures were inoculated from the appropriate strain from frozen stocks. Bacteria were grown in nutrient broth with 0.1% ampicillin in a shaker/incubator at 250 rpm at 37 6 2 C for approximately 12–17 hr. The longer times were used to ensure sufficient growth. Cultures were monitored for growth by spectrophotometric analysis of culture turbidity at 600 nm. Cultures diluted 1:10 should have an optical density greater than 0.2.
Criteria for a Valid Assay
Exogenous Metabolic Activation System (S9)
Cell Culture
Aroclor 1254-induced rat liver S9 homogenate was purchased commercially (MolTox, Boone, NC). The S9 mix was prepared on the day of use and contained: 256 mM KCl, 6 mM MgCl26H2O, 4 mM glucose-6-phosphate, 3 mM NADP, 77 mM NaH2PO4 (pH 7.4), and 4.3% (vol/vol final concentration) rat liver homogenate.
CHO-K1 cells (repository number CCL 61) were obtained from the American Type Culture Collection (Manassas, VA), and stock cultures were maintained in complete medium (McCoy’s 5A medium, Quality Biological, Gaithersburg, MD) containing 10% (vol/vol) dialyzed and heat-inactivated fetal bovine serum, 2 mM L-glutamine, 100 U/mL penicillin, and 100 mg/mL streptomycin) under standard conditions. Approximately 24 hr prior to treatment, cells were seeded at a density of 12,000/well in 200 mL complete medium in 96-well flat-bottomed tissue culture plates and incubated under standard conditions.
Treatment Each test chemical was diluted and tested at concentrations of 6.25, 12.5, 25, 50, 100, 200, 400, and 800 mM (final), along with appropriate vehicle and positive controls with and without S9. In cases where the test article was initially toxic even at the lowest concentration evaluated, the assay was repeated at lower concentrations. The test system was exposed to the test article via the microplate liquid culture modification reported by Gee et al. [1994]. The assay scoring and evaluation of the results followed published methods [Pant, 2013].
Each replicate was based on scoring each individual treatment (concentration of 1 chemical) in 48 wells. This was done in triplicate. The mean number of positive wells from each replicate treatment (48 wells) for the vehicle control must be on average 10. In a similar manner for the positive controls, the acceptable range is 25 positive wells for each set of 48 4NQO/2NF with TAMix and TA98 and for 2AA with TA98, and 15 positive wells for 2AA with TAMix.
In Vitro 96-Well MicroFlow CHO Cell MN Assay
Exogenous Metabolic Activation System (S9) As purchased and described above, Aroclor 1254-induced rat liver S9 mix was prepared on the day of use in serum-free medium and contained (final concentration in the treatment medium) 0.8 mM NADP, 1.0
Environmental and Molecular Mutagenesis. DOI 10.1002/em Medium-Throughput Assays for Genotoxicity mM glucose-6-phosphate, 6.6 mM KCl, 1.6 mM MgCl2, 20 mM NaPO4 buffer, and 0.5% (vol/vol final concentration) liver homogenate.
Treatment The MT MN assay was performed as previously described [Bryce et al., 2011] exposing cells to nine concentrations of the test article as well as the concurrent positive and vehicle controls with and without S9. The highest test article concentration was 200 mM (the maximum possible based upon the 20 mM stock solutions provided and using a 1% [vol/vol] concentration volume). All concentration formulations were prepared in DMSO with the lower concentrations prepared using 1:1 dilutions in the initial assays; closer 70% dilution intervals were used if additional trials were necessary due to either excessive or sub-optimal toxicity. The concentration of DMSO remained constant. Cells were treated for 4 hr with S9 and 24 hr without S9. All test and positive control articles and concentrations were evaluated in duplicate wells, and the vehicle control was evaluated using 16-wells/plate. After 4 hr the cultures treated with S9 were removed from the incubator and examined for any changes in pH (based upon color change of the medium) or the presence of precipitate at low power using an inverted microscope. The cultures were washed twice with phosphatebuffered saline containing calcium and magnesium (100 mg/mL), re-fed with complete medium, and further incubated for approximately 20 hr under standard conditions.
Controls Cyclophosphamide (CP, 9 and 18 mM with S9), mitomycin C (MMC, 0.75 and 1.5 mM without S9), and vinblastine (VB, 0.028 and 0.055 mM without S9) were used as the positive control agents. DMSO was the vehicle control. CP, MMC, and VB were all from Sigma-Aldrich (St. Louis, MO), and all concentrations shown are final concentrations.
Sample Processing At 24 hr after the start of treatment, all plates were removed from the incubator, and the cultures that were treated without S9 were examined for changes in pH or precipitation as above. All plates then were processed using In Vitro MicroFlow Kits according to manufacturer recommendations (In Vitro MicroFlow Kit, manual v100223, Litron Laboratories, Rochester, NY). The cells were scored immediately or refrigerated at 2–8 C and protected from light for up to three days until analyzed.
Scoring-Flow Cytometric Analysis Sample analysis was performed on a FACSCantoTM II flow cytometer using a High Throughput Sampler and running FACSDiva software (v6.1.2; all supplied by BD Biosciences, San Jose, CA) and the template provided with the In Vitro MicroFlow Kit. Whenever possible (e.g., if not limited by toxicity) 5,000 “healthy” nuclei (SYTOX Green-positive, ethidium monoazide (EMA)-negative nuclei) were evaluated per sample. Parameters analyzed included percentage MN, percentage apoptotic/ necrotic (EMA-positive) nuclei, percentage hypodiploid nuclei, and relative survival. Relative survival was based upon nuclei:bead ratios, and all calculations were as previously described [Bryce et al., 2008].
Criteria for a Valid Assay In order for an assay to be considered valid, the MN frequencies of the vehicle controls (plate average) must be within historical negative control ranges (1.77% 6 0.51 and 1.45% 6 0.48, with and without S9, respectively). In addition, each positive control must induce twofold increase in MN frequency, and the VB positive control must also induce
471
fivefold increase in hypodiploid frequency as compared to the concurrent vehicle controls (plate average; at least one concentration each). Finally, at least three test article concentrations with acceptable cytotoxicity and solubility must be available for analysis. Any cultures with a relative survival
Research Article An Evaluation of 25 Selected ToxCast Chemicals in Medium-Throughput Assays to Detect Genotoxicity Andrew D. Kligerman,1* Robert R.Young,2 Leon F. Stankowski Jr.,2 Kamala Pant,2 Tim Lawlor,2 Marilyn J. Aardema,2,3 and Keith A. Houck1 1
National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 2 BioReliance Corporation, Rockville, Maryland 3 Marilyn Aardema Consulting LLC, 5315 Oakbrook Dr., Fairfield, Ohio ToxCast is a multiyear effort to develop a costeffective approach for the US EPA to prioritize chemicals for toxicity testing. Initial evaluation of more than 500 high-throughput (HT) microwell-based assays without metabolic activation showed that most lacked high specificity and sensitivity for detecting genotoxicants. Thus, EPA initiated a pilot project to investigate the use of standard genotoxicity endpoints using medium-throughput genotoxicity (MTG) assays in the context of a large testing program. Twentyfive chemicals were selected from the ToxCast program based in part on their known genotoxicity. The two MTG assays used were the Ames TM assay and 96-well In Vitro MicroFlowV II Micronucleus (MN) assay. The Ames II assay showed a reasonable correlation with published Ames test data and industry submissions, though specificity was much better than sensitivity due R
to restraints on top concentrations as prescribed by ToxCast. Overall concordance was 73% both with and without metabolic activation. The flow MN assay had concordances of 71% and 58% with and without metabolic activation, respectively, when compared to published data and submissions. Importantly, a comparison of results without S9 from the MTG assays to an HT ToxCast p53 activation assay showed a fairly good degree of concordance (67%). The results reported here indicate that assays for genotoxicity endpoints can be conducted in a MT format and have the potential to add to the interpretation of results from large-scale testing programs such as EPA’s ToxCast program. Inherent limitations such as the top concentrations used in large scale testing programs are discussed. Environ. Mol. Mutagen. 56:468– C 2014 Wiley Periodicals, Inc. 476, 2015. V
Key words: genotoxicity; high-throughput testing; Ames test; micronucleus test; ToxCast
INTRODUCTION A standard battery of short-term tests has been used for several decades to screen chemicals for genotoxicity with the aim of detecting mutagenic carcinogens as well as to
Additional Supporting Information may be found in the online version of this article. *Correspondence to: Andrew D. Kligerman, Integrated Systems Toxicology Division, Office of Research and Development, National Health and Environmental Effects Research Laboratory, MD 105-03, U.S. Environmental Protection Agency, Research Triangle Park, NC 27710, USA. E-mail: [email protected], [email protected] Disclaimer: The information in this document has been funded wholly (or in part) by the U. S. Environmental Protection Agency. It has been subjected to review by the National Health and Environmental Effects Research Laboratory and approved for publication. Approval does not signify that the contents reflect the views of the Agency nor does C 2014 Wiley Periodicals, Inc. V
detect germ cell mutagens to limit human exposure to these potentially dangerous chemicals. However, a large number of environmental and industrial chemicals have not been tested in any significant way either due to a lack of legal authority, funding, resources, will of the parties
mention of trade names or commercial products constitute endorsement or recommendation for use. Conflict of interest: BioReliance is a contract research organization that TM performs the Ames II and the 96-well In Vitro MicroFlowV Micronucleus Assays. BioReliance has global rights ti the Ames II assay with the exception that rights in Europe are co-exclusive. Ames II kits are licensed for sale by MolTox (Boone, NC). R
Received 24 July 2014; provisionally accepted 17 November 2014; and in final form 18 November 2014 DOI 10.1002/em.21934 Published online 23 December 2014 in Wiley Online Library (wileyonlinelibrary.com).
Environmental and Molecular Mutagenesis. DOI 10.1002/em Medium-Throughput Assays for Genotoxicity
involved, or a combination of these factors. An approach needs to be established that selects chemicals for testing in an efficient manner. ToxCast is a multiyear effort to develop a cost-effective approach for prioritizing thousands of data-poor chemicals for further toxicity testing [Dix et al., 2007; Kavlock et al., 2012]. Phase I of ToxCast was designed to screen for a large number of toxicity endpoints using approximately 300 well-studied chemicals, mostly pesticides, in greater than 500 highthroughput (HT), microwell plate-based assays that generally lacked metabolic activation. It was anticipated that the various endpoints measured in the ToxCast assays would be useful for predicting higher level outcomes such as genotoxicity. In the first analysis of the usefulness of these HT assays for predicting genotoxicity, Knight et al. [2009] evaluated three HT assays (GreenScreen HC GADD45a-GFP, Gentronix; CellCiphr p53, Cellumen; and CellSensor p53REbla, Invitrogen Corp.). They concluded that the assays lacked high sensitivity and accuracy; however, the assays had relatively high specificity due to the preponderance of nongenotoxicants in the Phase I chemical set. These poorer than expected results were due in part to the predefined relatively low concentrations used in the study coupled with the lack of metabolic activation in the three assays. To refine the analysis, Kligerman et al. [2012] used the proprietary EPA submissions (US EPA Office of Pesticides and Toxic Substances), California Department of Pesticide Regulation Toxicology Data Review Summaries, and the National Toxicology Program Website as primary sources; secondary sources were LeadscopeV and the open scientific literature to classify the Phase I chemicals as to whether or not they were point mutagens, clastogens, or aneugens without exogenous metabolic activation (i.e., direct-acting genotoxicants). A evaluation (call) was based on the weight of the evidence presented, the presence or absence of contradictory or equivocal results, and for the published literature the reputation of the journal and the publication record of the research group. Of the 293 Phase I chemicals examined, 288 had some data on genotoxicity in the absence of metabolic activation. Of these, 7 were point mutagens, 46 were clastogens, and 13 were deemed to be aneugens (though not mutually exclusive). The majority were either nongenotoxic, or they had contradictory or limited information available, thus precluding the making of a call on this latter group. Therefore, a chemical was assigned a call as a genotoxicant (mutagen, clastogen, and/or aneugen), a nongenotoxicant (not a mutagen, clastogen, or aneugen), or inconclusive either due to the lack of data, the contradictory nature of the literature, or the weakness of the data. Once a chemical was assigned a category, univariate analysis was performed with an examination of each chemical in each ToxCast HT assay alone or in pairs to determine whether there were assays that detected R
469
genotoxicants, either serendipitously or due to the fact that some HT assays had indeed measured endpoints related to DNA damage checkpoints, microtubule disruption, or other specific cellular responses linked to genotoxicity. Based on Dr. Kligerman’s calls on the genotoxicity of the chemicals, none of the ToxCast HT assays tested, either individually or in pairs, showed both high specificity and sensitivity for detecting genotoxicants. Those assays that correctly identified a reasonable number of suspected direct-acting genotoxicants falsely predicted a large number of nongenotoxicants to be genotoxic [Kligerman et al., 2012]. Based on the importance of the ToxCast project for predicting various endpoints of toxicity and the importance of genotoxicity to the overall prioritization of chemicals for further testing, it was decided that data from standard genotoxicity endpoints generated in a mediumthroughput format (MTG assays) would be useful to supplement the existing ToxCast data. In this pilot study, 25 Phase I and Phase II chemicals from the ToxCast program were selected for genotoxicity testing with and without metabolic activation (S9) and included chemicals that (a) were known to be direct-acting genotoxicants (captan, folpet, captafol, dimethyl sulfate), (b) were known to be genotoxic after metabolic activation (7,12dimethylbenz(a)anthracene, N-nitrosodiethylamine), (c) were compounds giving strong responses in HT assays that might be related to genetic damage (4-chloro-1,2-diaminobenzene, allethrin, carbendazim, cladribine, triglycidyl isocyanurate, ziram), or (d) produced results in the HT assays that were of interest due to limited or contradictory published information on their genotoxicity (allethrin, diniconazole, monocrotophos, oryzalin). Here, we report the results and information gleaned from genotoxicity testing of these 25 chemicals using two MTG TM platforms: the Ames II assay and the 96-Well In Vitro V MicroFlow CHO MN assay (all with and without S9). The endpoints measured were point mutations in bacteria and clastogenicity and aneugenicity in rodent cells. We also compare these MTG results to data from a HT ToxCast p53 activation assay in the HepG2 human liver hepatoma cell line, an assay that is related to genotoxicity. R
MATERIALS AND METHODS ToxCast Chemicals The chemicals tested, their CAS#, purity, and source are listed in Table I. The supplier provided purities were all greater than 93%. Consistent with all of the ToxCast HT assays, the chemicals were dissolved in dimethylsulfoxide (DMSO) at a final stock concentration of 20 mM, and coded stock formulations of each chemical in DMSO were then shipped to BioReliance under contract to the EPA (EP-D-11-081) and as prescribed under ToxCast guidelines. The 20 mM stock solution was chosen based on consideration of realistic exposure scenarios, greatly increased solubility issues both in DMSO above this concentration and in the aqueously diluted samples in the bioassay at concentrations exceeding 20–100 lM, and increased cytotoxicity again seen as
Environmental and Molecular Mutagenesis. DOI 10.1002/em 470
Kligerman et al.
TABLE I. Chemicals Tested (Name, CAS No., Source, and Purity) Bensulide 4-Chloro-1,2-diaminobenzene 7,12-Dimethylbenz(a)anthracene 9-Phenanthrol Allethrin Captafol Captan Carbendazim Cladribine Chlorophene Dimethyl sulfate Diniconazole-M Fludioxonil Folpet Methyl isothiocyanate Methylene bis(thiocyanate) Mevinphos Monocrotophos Naled N-Nitrosodiethylamine Oryzalin Thiram Triclosan Triglycidyl isocyanurate Ziram
741-58-2 95-83-0 57-97-6 484-17-3 584-79-2 2425-06-1 133-06-2 10605-21-7 4291-63-8 120-32-1 77-78-1 83657-18-5 131341-86-1 133-07-3 556-61-6 6317-18-6 7786-34-7 6923-22-4 300-76-5 55-18-5 19044-88-3 137-26-8 3380-34-5 2451-62-9 137-30-4
concentrations exceed 20–50 lM in many HT assays. Testing of large chemical libraries precludes handling chemicals on an individual basis with the understanding that not all chemicals will have optimal conditions. In all the studies, the DMSO concentration was the same in each well.
Ames II Assay
Sigma Chemical Company Sigma Chemical Company LightBiologicals Sigma Chemical Company LightBiologicals Sigma Chemical Company ChemService, Inc. LightBiologicals Sigma Chemical Company Sigma Chemical Company LightBiologicals Waterstone Technology LLC Sigma Chemical Company Sigma Chemical Company Enamine Sigma Chemical Company Sigma Chemical Company Sigma Chemical Company Sigma Chemical Company LightBiologicals LightBiologicals Sigma Chemical Company Sigma Chemical Company Sigma Chemical Company LightBiologicals
99.5 97.3 99 93.9 96.1 97 99.2 97 98.5 98 99.8 95 99.9 99.9 97 99.3 93.3 99.9 97.2 99.9 98.2 99.9 99.5 97 99.2
Controls 2-Aminoanthracene (2AA; final concentration: 0.026 mM) and a mixture of 4-nitroquinoline N-oxide (4NQO; final concentration: 0.005 mM) and 2-nitrofluorene (2NF; final concentration: 0.009 mM) were used as positive controls with and without S9, respectively. All were purchased from Sigma-Aldrich (St. Louis, MO). DMSO was the vehicle control.
The Ames II assay was performed using Salmonella typhimurium strain TA98 and mixed Ames II strains (TAMix), that consists of TA7001, TA7002, TA7003, TA7004, TA7005, and TA7006 combined into a single culture. The exposure and indicator media for the assay were purchased commercially (MolTox, Inc., Boone, NC). Overnight cultures were inoculated from the appropriate strain from frozen stocks. Bacteria were grown in nutrient broth with 0.1% ampicillin in a shaker/incubator at 250 rpm at 37 6 2 C for approximately 12–17 hr. The longer times were used to ensure sufficient growth. Cultures were monitored for growth by spectrophotometric analysis of culture turbidity at 600 nm. Cultures diluted 1:10 should have an optical density greater than 0.2.
Criteria for a Valid Assay
Exogenous Metabolic Activation System (S9)
Cell Culture
Aroclor 1254-induced rat liver S9 homogenate was purchased commercially (MolTox, Boone, NC). The S9 mix was prepared on the day of use and contained: 256 mM KCl, 6 mM MgCl26H2O, 4 mM glucose-6-phosphate, 3 mM NADP, 77 mM NaH2PO4 (pH 7.4), and 4.3% (vol/vol final concentration) rat liver homogenate.
CHO-K1 cells (repository number CCL 61) were obtained from the American Type Culture Collection (Manassas, VA), and stock cultures were maintained in complete medium (McCoy’s 5A medium, Quality Biological, Gaithersburg, MD) containing 10% (vol/vol) dialyzed and heat-inactivated fetal bovine serum, 2 mM L-glutamine, 100 U/mL penicillin, and 100 mg/mL streptomycin) under standard conditions. Approximately 24 hr prior to treatment, cells were seeded at a density of 12,000/well in 200 mL complete medium in 96-well flat-bottomed tissue culture plates and incubated under standard conditions.
Treatment Each test chemical was diluted and tested at concentrations of 6.25, 12.5, 25, 50, 100, 200, 400, and 800 mM (final), along with appropriate vehicle and positive controls with and without S9. In cases where the test article was initially toxic even at the lowest concentration evaluated, the assay was repeated at lower concentrations. The test system was exposed to the test article via the microplate liquid culture modification reported by Gee et al. [1994]. The assay scoring and evaluation of the results followed published methods [Pant, 2013].
Each replicate was based on scoring each individual treatment (concentration of 1 chemical) in 48 wells. This was done in triplicate. The mean number of positive wells from each replicate treatment (48 wells) for the vehicle control must be on average 10. In a similar manner for the positive controls, the acceptable range is 25 positive wells for each set of 48 4NQO/2NF with TAMix and TA98 and for 2AA with TA98, and 15 positive wells for 2AA with TAMix.
In Vitro 96-Well MicroFlow CHO Cell MN Assay
Exogenous Metabolic Activation System (S9) As purchased and described above, Aroclor 1254-induced rat liver S9 mix was prepared on the day of use in serum-free medium and contained (final concentration in the treatment medium) 0.8 mM NADP, 1.0
Environmental and Molecular Mutagenesis. DOI 10.1002/em Medium-Throughput Assays for Genotoxicity mM glucose-6-phosphate, 6.6 mM KCl, 1.6 mM MgCl2, 20 mM NaPO4 buffer, and 0.5% (vol/vol final concentration) liver homogenate.
Treatment The MT MN assay was performed as previously described [Bryce et al., 2011] exposing cells to nine concentrations of the test article as well as the concurrent positive and vehicle controls with and without S9. The highest test article concentration was 200 mM (the maximum possible based upon the 20 mM stock solutions provided and using a 1% [vol/vol] concentration volume). All concentration formulations were prepared in DMSO with the lower concentrations prepared using 1:1 dilutions in the initial assays; closer 70% dilution intervals were used if additional trials were necessary due to either excessive or sub-optimal toxicity. The concentration of DMSO remained constant. Cells were treated for 4 hr with S9 and 24 hr without S9. All test and positive control articles and concentrations were evaluated in duplicate wells, and the vehicle control was evaluated using 16-wells/plate. After 4 hr the cultures treated with S9 were removed from the incubator and examined for any changes in pH (based upon color change of the medium) or the presence of precipitate at low power using an inverted microscope. The cultures were washed twice with phosphatebuffered saline containing calcium and magnesium (100 mg/mL), re-fed with complete medium, and further incubated for approximately 20 hr under standard conditions.
Controls Cyclophosphamide (CP, 9 and 18 mM with S9), mitomycin C (MMC, 0.75 and 1.5 mM without S9), and vinblastine (VB, 0.028 and 0.055 mM without S9) were used as the positive control agents. DMSO was the vehicle control. CP, MMC, and VB were all from Sigma-Aldrich (St. Louis, MO), and all concentrations shown are final concentrations.
Sample Processing At 24 hr after the start of treatment, all plates were removed from the incubator, and the cultures that were treated without S9 were examined for changes in pH or precipitation as above. All plates then were processed using In Vitro MicroFlow Kits according to manufacturer recommendations (In Vitro MicroFlow Kit, manual v100223, Litron Laboratories, Rochester, NY). The cells were scored immediately or refrigerated at 2–8 C and protected from light for up to three days until analyzed.
Scoring-Flow Cytometric Analysis Sample analysis was performed on a FACSCantoTM II flow cytometer using a High Throughput Sampler and running FACSDiva software (v6.1.2; all supplied by BD Biosciences, San Jose, CA) and the template provided with the In Vitro MicroFlow Kit. Whenever possible (e.g., if not limited by toxicity) 5,000 “healthy” nuclei (SYTOX Green-positive, ethidium monoazide (EMA)-negative nuclei) were evaluated per sample. Parameters analyzed included percentage MN, percentage apoptotic/ necrotic (EMA-positive) nuclei, percentage hypodiploid nuclei, and relative survival. Relative survival was based upon nuclei:bead ratios, and all calculations were as previously described [Bryce et al., 2008].
Criteria for a Valid Assay In order for an assay to be considered valid, the MN frequencies of the vehicle controls (plate average) must be within historical negative control ranges (1.77% 6 0.51 and 1.45% 6 0.48, with and without S9, respectively). In addition, each positive control must induce twofold increase in MN frequency, and the VB positive control must also induce
471
fivefold increase in hypodiploid frequency as compared to the concurrent vehicle controls (plate average; at least one concentration each). Finally, at least three test article concentrations with acceptable cytotoxicity and solubility must be available for analysis. Any cultures with a relative survival
