Rat Pig-a mutation assay responds to the genotoxic carcinogen ethyl carbamate but not the non-genotoxic carcinogen methyl carbamate

Abstract

Determination of the mode of action of carcinogenic agents is an important factor in risk assessment and regulatory practice. To assess the ability of the erythrocyte-based Pig-a mutation assay to discriminate between genotoxic and non-genotoxic modes of action, the mutagenic response of Sprague Dawley rats exposed to methyl carbamate (MC) or ethyl carbamate (EC) was investigated. EC, a potent carcinogen, is believed to induce DNA damage through the formation of a DNA-reactive epoxide group, whereas the closely structurally related compound, MC, cannot form this epoxide and its weaker carcinogenic activity is thought to be secondary to inflammation and promotion of cell proliferation. The frequency of Pig-a mutant phenotype cells was monitored before, during, and after 28 consecutive days of oral gavage exposure to either MC (doses ranging from 125 to 500mg/kg/day) or EC (250mg/kg/day). Significant increases in the frequency of mutant reticulocytes were observed from Days 15 through 43, with a peak mean frequency of 19.9×10⁻⁶ on Day 29 (i.e. 24.9-fold increase relative to mean vehicle control across all four sampling times). As expected, mutant erythrocyte responses lagged behind mutant reticulocyte responses, with a maximal mean frequency of 8.2×10⁻⁶ on Day 43 (i.e. 16.4-fold increase). No mutagenic effects were observed with MC. A second indicator of in vivo genotoxicity, peripheral blood micronucleated reticulocytes, was also studied. This endpoint was responsive to EC (3.3-fold mean increase), but not to MC. These results support the hypothesis that genotoxicity contributes to the carcinogenicity of EC but not of MC, and illustrates the value of the Pig-a assay for discriminating between genotoxic and non-genotoxic modes of action.

Introduction

Understanding a chemical’s mode of action (MoA) is a critical component of cancer and genotoxicity risk assessment (1,2). To evaluate the ability of the Pig-a mutation assay to discriminate between carcinogens that act by a genotoxic MoA and those that act by a non-genotoxic mechanism, we monitored the frequencies of mutant phenotype reticulocytes (RET^CD59−) and mutant phenotype erythrocytes (RBC^CD59−) in rats exposed to ethyl carbamate (EC, also known as urethane) or the structurally related chemical methyl carbamate (MC). The incidence of micronucleated reticulocytes (MN-RET) was studied as an additional endpoint of genotoxicity to hematopoietic cells.

As described in the IARC monograph and elsewhere (3), EC is generally negative in in vitro genotoxicity assays, findings that are likely due to insufficiencies with exogenous metabolic activation systems. Conversely, EC is genotoxic in a variety of in vivo test systems, including transgenic rodent mutation and erythrocyte micronuclei studies (3–8). It is thought that in vivo metabolism involves oxidation of the ethyl group followed by dehydration to yield vinyl carbamate, which can undergo further oxidation to the DNA-reactive form vinyl carbamate epoxide (9). Whereas EC is negative in the Ames bacterial mutagenicity test, vinyl carbamate is an S9-requiring Ames positive chemical, and vinyl carbamate epoxide is directly mutagenic to bacteria (9,10). More recently, Stankowski et al. (11) reported dose-dependent increases in the frequency of RBC^CD59− and RET^CD59− of EC-exposed Sprague Dawley rats. EC is a multispecies carcinogen that causes malignancies in several tissues, including lung, hematopoietic system and liver (3).

In contrast, MC is a non-genotoxicant that is structurally similar to EC (12), but its methyl group is inhibited from undergoing transformation to an epoxide (13). This is believed to explain why MC has yielded consistently negative results when tested for genotoxicity, including in vivo assessments (12,14). The modest level of rodent carcinogenicity observed with high dose MC exposure is thought to be the result of bioaccumulation and consequential inflammation and hyperplasia of the liver (12). Given this profile, an ECVAM publication listed MC among non-genotoxicants that are valuable for validating new genotoxicity assays (15).

This report describes in vivo genotoxicity results for male rats exposed to these two agents for 28 days. The 1-month exposure schedule was selected for three important reasons. First, it is a common test duration in general toxicology and therefore represents a model design for integrating genetic toxicology endpoints (16–18). Second, because Pig-a mutant phenotype cells accumulate with repeated treatments with a mutagen (19), the 28-day period represented an effective way of studying the effects from a weak mutagen (EC) and a non-mutagen (MC). Third, the 28-day schedule focused our analyses on an early stage of cancer initiation, a period when mutational events play an important role in the MoA of genotoxic carcinogens (20).

Materials and methods

Reagents

EC (CAS no. 51-79-6) and MC (CAS no. 598-55-0) were purchased from Sigma–Aldrich, St. Louis, MO, USA. Lympholyte®-Mammal cell separation reagent was purchased from CedarLane, Burlington, NC, USA. Anti-PE microBeads, LS columns and a QuadroMACS™ Separator were from Miltenyi Biotec, Bergisch Gladbach, Germany. CountBright™ absolute count beads and fetal bovine serum (FBS) were purchased from Invitrogen, Carlsbad, CA, USA. Anticoagulant solution, buffered salt solution, nucleic acid dye solution (contains SYTO^® 13), Anti-CD59-PE and Anti-CD61-PE were from In Vivo Rat MutaFlow® Kits (Litron Laboratories, Rochester, NY). Reagents used for flow cytometric micronucleus scoring (Anticoagulant solution, balanced salt solution, stock propidium iodide solution, anti-CD71-FITC, anti-CD61-PE, stock RNase solution and Malaria Biostandards) were from In Vivo Rat MicroFlow® Kits (Litron Laboratories).

Animals, treatments, blood harvests

Experiments were conducted with the oversight of the University of Rochester’s Institutional Animal Care and Use Committee. Male Sprague-Dawley rats were purchased from Charles River Laboratories, Wilmington, MA, USA. Rodents were allowed to acclimate for ~1 week, and their age at the start of treatment was 7 weeks. Water and food were available ad libitum throughout the acclimation and experimental periods. EC and MC were prepared in deionized water each day of treatment, and administered via oral gavage in a volume of 10ml/kg body weight/administration. Exposures occurred once a day at ~24h intervals for 28 consecutive days. MC was administered at 0, 125, 250 or 500mg/kg/day (n = 6 per group), and EC at 250mg/kg/day (n = 5). For MC the high dose level was based on a preliminary dose-range finding experiment. As described in the OECD Guideline for the Testing of Chemicals, No. 407 (21) the aim was to find a top dose level that induced toxic effects but not death or severe morbidity. The EC dose was based on a report by Stankowski et al. (11), and is a dose that was well-tolerated and induced significant in vivo genotoxic responses.

Peripheral blood was collected on Days −5 (i.e. 5 days before the first administration), 4, 15, 29 and 43. Pre-treatment frequencies were compared to the distribution of 756 historical controls animals, and rats that exceeded a one-sided 95% upper tolerance interval, alpha 0.05, were considered outliers and excluded from the study (22). Animals that did not exceed these threshold values (i.e. 2.86 RBC^CD59− × 10⁻⁶ total RBC and/or 2.97 RET^CD59− × 10⁻⁶ total RET) were randomized across treatment groups. For this study, 34 animals were ordered, and 5 were eliminated based on these criteria (hence the use of 5 rats rather than 6 in the EC treatment group).

Blood was obtained by nicking a lateral tail vein with a surgical blade after animals were warmed briefly under a heat lamp. Approximately 100–200 µl of free-flowing blood were collected directly into heparinized capillary tubes (Fisher Scientific, Waltham, MA, USA; CAT no. 22-260-950). For the Pig-a endpoint, 80 µl of each blood sample were transferred to tubes containing 100 µl kit-supplied heparin solution where they remained at room temperature for <2h until leukodepletion as described later. For the MN-RET endpoint, 30 µl of each whole blood sample were transferred to tubes containing 100 µl kit-supplied anticoagulant solution where they remained at room temperature for <2h, after which they were fixed with ultracold methanol as described previously (23).

MN-RET: Sample Preparation, Data Acquisition

MN-RET frequencies were determined for blood samples collected on Days 4 and 29. To prepare samples for flow cytometric analysis, blood was washed out of methanol fixative and incubated with antibodies and other reagents as specified in the In Vivo Rat MicroFlow Kit manual. This methodology has been described in detail elsewhere (23,24). Kit-supplied malaria-infected erythrocytes served as biological standards and guided instrument settings on each day of analysis (25,26). The frequency of MN-RET was determined upon the acquisition of 16000 to 20000 CD71-positive reticulocytes (RET) per blood sample using a BD FACSCalibur flow cytometer running CellQuest Pro v5.2.

Pig-a Mutation: Sample Preparation, Data Acquisition

As described previously (27,28) and in accordance with MutaFlow kit manual v1404003 instructions, Pig-a analyses were performed on blood samples collected on Days −5, 15, 29 and 43. These experiments utilized the so-called 96 well plate-based method, whereby microtiter plates facilitated multisample/parallel labeling and washing procedures, as well as walk-away data acquisition through the use of a BD High Throughput Sampler (HTS) instrument that automatically delivered well contents to the flow cytometer. ‘Pre-column’ samples consisted of a small aliquot of each fully labeled and stained sample which was analysed for 1min in order to provide %RET measurements as well as RBC to counting bead and RET to counting bead ratios. The majority of each sample was then used for an immunomagnetic separation procedure, and the resulting ‘post-column’ eluates were analysed for 3min to provide mutant phenotype RBC to counting bead and mutant phenotype RET to counting bead ratios. As described previously, pre- and post-column data were used to calculate RBC^CD59− and RET^CD59− frequencies (27). For the current studies >3×10⁶ RET and >127×10⁶ RBC equivalents per rat per time point were consistently scored.

An Instrument Calibration Standard was generated on each day of data acquisition. As approximately one half of these erythrocytes were not incubated with anti-CD59-PE, these samples contained a high prevalence of mutant-mimic cells and provided a means to define the location of GPI anchor-deficient erythrocytes (29). A BD FACSCanto™ II flow cytometer running Diva™ v6.1.2 software was used for Pig-a data acquisition and analysis.

Calculations, Statistical Analyses

The formulas used to calculate RBC^CD59− and RET^CD59− frequencies based on data from pre- and post-immunomagnetic column blood analyses have been described previously (27). The incidence of RET and MN-RET is expressed as frequency percent, while the incidence of mutant phenotype cells is expressed as number per 10⁶ cells. All %RET, %MN-RET, mutant phenotype cell frequencies, averages and standard error calculations were performed with Excel Office X for Mac® (Microsoft, Seattle, WA, USA).

For statistical evaluations, MN-RET, RBC^CD59− and RET^CD59− frequencies were log(10) transformed. Since zero RET^CD59− values were occasionally observed, a 0.1 offset was added to each RET^CD59− value prior to log transformation. Each time point was studied separately, where the effect of treatment on these transformed MN-RET, RBC^CD59− and RET^CD59− data was compared to vehicle control using Dunnett’s multiple comparison t-tests in the context of a one-way analysis of variance (ANOVA) model (JMP®, v8.0.1, SAS Institute Inc., Cary, NC, USA). Significance was evaluated at the 5% level using a one-tailed test for increases relative to vehicle control. These same analyses were performed with %RET frequencies, however in these cases data were not log transformed, and the tests were two-tailed.

As recommended by the 2013 IWGT Pig-a working group (30), statistical significance was not the sole determinant of a positive Pig-a result. For a statistically significant increase in Pig-a mutant phenotype cells to be considered biologically relevant, mean RBC^CD59− and mean RET^CD59− frequencies also needed to exceed a lab-specific historical negative control distribution (31).

Results and discussion

Over the course of the 28-day treatment period, rats receiving 250mg EC/kg/day or the highest dose level of MC tested (500mg/kg/day) gained less weight than vehicle controls, 19% and 40% less, respectively. No other gross effects were noted. As expected, age-related reductions to mean %RET were observed (Figures 1 and 2). Bone marrow exposure to MC was indicated by further reductions of RET frequencies. This was apparent from Pig-a analyses conducted on Day 15, where the high dose of MC caused an average reduction of 41% relative to vehicle control rats (Figures 1 and 2). Day 43 data are also suggestive of bone marrow exposure, as compensatory erythropoiesis is apparent from elevated %RET (Figures 1 and 2). At the one EC dose tested no statistically significant changes to %RET were observed, although Day 4 values associated with micronucleus analyses were suggestive of a modest effect (22% reduction; Figure 3).

Fig. 1.

Mean RET^CD59− frequencies are graphed for each of four time points (Y-axis). MC RET^CD59− data appear as grey bars, and EC data appear as white bars. Percent reticulocytes (%RET) are also graphed on the YY-axis (black line). All error bars are SEM. Asterisks associated with mean RET^CD59− values indicate significance compared to same-day vehicle control values (based on a one-sided Dunnett’s test, P < 0.05; mean also needed to exceed historical control 90% tolerance interval, upper limit, alpha 0.1). Daggers indicate a significant difference in mean %RET compared to same-day vehicle control values (i.e. two-sided Dunnett’s test, P < 0.05).

Fig. 2.

Mean RBC^CD59− frequencies are graphed for each of four time points (Y-axis). MC RBC^CD59− data appear as grey bars, and EC data appear as white bars. Percent reticulocytes (%RET) are also graphed on the YY-axis (black line). All error bars are SEM. Asterisks associated with mean RBC^CD59− values indicate significance compared to same-day vehicle control values (based on a one-sided Dunnett’s test, P < 0.05; mean also needed to exceed historical control 90% tolerance interval, upper limit, alpha 0.1). Daggers indicate a significant difference in mean %RET compared to same-day vehicle control values (i.e. two-sided Dunnett’s test, P < 0.05).

Fig. 3.

Mean percent micronucleated reticulocyte (%MN-RET) frequencies are graphed for each of two time points (Y-axis). MC MN-RET data appear as grey bars, and EC data appear as white bars. Percent reticulocytes (%RET) are also graphed on the YY-axis (black line). All error bars are SEM. Asterisks associated with mean MN-RET values indicate significance compared to same-day vehicle control values (based on a one-sided Dunnett’s test, P < 0.05).

EC caused significantly elevated RET^CD59− frequencies on Day 15, the first Pig-a sampling time after treatment was initiated (Figure 1). Maximal RET^CD59− frequencies occurred on Day 29, when a mean value of 19.9±4.1×10⁻⁶ was observed. This value is very similar to that observed by Stankowski et al. (11) who reported mean RET^CD59− values of 21 and 16×10⁻⁶ for Days 15 and 29, respectively. For comparison, our average RET^CD59− frequency for vehicle controls across the four blood collection days was 0.8±0.4×10⁻⁶. Elevated EC-induced frequencies persisted through Day 43, the last time point studied. Unlike EC, exposure to MC was not associated with increased RET^CD59− frequencies.

The mean vehicle control RBC^CD59− frequency averaged 0.5±0.1×10⁻⁶ across the four time points studied. As expected from the kinetics of RBC formation, EC-induced RBC^CD59− frequencies increased more gradually than RET^CD59− (Figure 2). The first time point that showed a statistically significant increase in RBC^CD59− was Day 29 (4.9±0.8×10⁻⁶). This value is similar to that reported by Stankowski et al., who observed mean RBC^CD59− values of 4 and 6×10⁻⁶ for Days 15 and 29, respectively. Unlike Stankowski et al., the current study included an additional sampling time (Day 43), and in this case a higher mean value was evident (8.2±0.9×10⁻⁶). MC did not cause an increase in RBC^CD59− frequencies.

As shown in Figure 3, orally administered EC increased %MN-RET on Day 4, but not on Day 29. It is not clear why MN-RET frequencies were not elevated at Day 29. This differs from the findings of Stankowski et al. who did observe a response for 250mg EC/kg/day at this time using the same rat model (Sprague Dawley; different supplier), and MacGregor et al. (8), who showed that in mice elevated micronucleus frequencies are maintained at steady state for up to 90 days during continued daily exposure to EC. No significant increases in %MN-RET were observed at either sampling time in MC-treated rats.

Collectively, our results demonstrate that the Pig-a and micronucleus assays identify the expected genotoxic activity of EC during the period of exposure critical for initiation of the carcinogenic process, and distinguish it from MC, which does not exhibit any mutagenic or clastogenic activity in these assays up to a limiting toxic dose. These results also show that these assays are valuable for providing information about genotoxic versus non-genotoxic MoA when carcinogenicity has been demonstrated, information that is necessary for cancer risk assessments as described by the U.S. EPA (1). In both cases, this represents a better use of animals and other costly resources compared to the current practice of conducting separate studies, as it provides genetic toxicology data in the context where they can best be understood and interpreted—with concurrent pharmacokinetic, blood chemistry, histopathology and other important in vivo data (8,16,32). These results illustrate the value of the routine use of these endpoints for evaluating chemicals’ genotoxic potential in the context of easily integrated short-term repeat-dose toxicology studies. Further work with diverse chemical classes are planned to form a deeper understanding of assay performance.

Funding

This work was funded by a grant from the National Institute of Health/National Institute of Environmental Health Sciences (NIEHS ; number R44ES018017). The contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS or the institutions with which the authors are affiliated.