Environmental and Molecular Mutagenesis 56:322^332 (2015)

Research Article Evaluation of the In Vivo Mutagenicity of Isopropyl Methanesulfonate in Acute and 28-Day Studies Stephanie L. Coffing,* Michelle O. Kenyon, Joel I. Ackerman, Thomas J. Shutsky, and Krista L. Dobo Pfizer Worldwide Research and Development, Genetic Toxicology, Groton, Connecticut Understanding the mutagenic dose response could prove beneficial in the management of pharmaceutically relevant impurities. For most alkyl ester impurities, such as isopropyl methanesulfonate (IPMS), little in vivo mutagenicity data exist for dose analysis. The likelihood of a sublinear dose response for IPMS was assessed by comparing the Swain Scott constant, the SN1/SN2 reaction mechanism and the O6:N7 guanine adduct ratio to that of more wellknown alkyl esters. Based on available information, IPMS was predicted to have a mutagenic profile most like ethyl nitrosourea. To test this hypothesis, mature male Wistar Han rats were administered IPMS using acute (single administration at 3.5 to 56 mg/kg) or subchronic (28 days at 0.125 to 2 mg/kg/day) exposures. The in vivo Pig-a mutation assay was used to identify mutant phenotype reticulocyte (Ret) and

red blood cell (RBC) populations. The maximum mutant response occurred approximately 15 and 28 days after the last dose administration in the mutant Ret and RBC populations respectively in the acute study and on Day 29 and 56 in the mutant Ret and RBC populations, respectively, in the subchronic study. A comparison of RBC mutant frequencies from acute and subchronic protocols suggests a sublinear response; however, this was not substantiated by statistical analysis. A No Observed Effect Level (NOEL) of 0.25 mg/kg/day resulted in a Permitted Daily Exposure equivalent to the Threshold of Toxicological Concern. An estimate of the NOEL based on the previously mentioned factors, in practice, would have pre-empted further investigation of the potent mutagen IPMS. Environ. Mol. Mutagen. 56:322–332, C 2014 Wiley Periodicals, Inc. 2015. V

Key words: In vivo mutation assay; in vivo genotoxicity; impurities; alkylating agent; Pig-a mutation assay

INTRODUCTION Alkyl esters are often formed when alcohols interact with sulfonic acids, such as methane sulfonic acid, toluene sulfonic acid or benzene sulfonic acid during saltforming steps in pharmaceutical synthesis [Glowienke et al., 2005; Elder et al., 2012]. The presence of alkyl esters as impurities in pharmaceuticals is important to consider during the drug development process as they pose a potential mutagenic and carcinogenic risk. Therefore, residual amounts of alkyl esters must be controlled such that human exposure is minimized to a very low dose of 1.5 mg/day, i.e. the threshold of toxicological concern (TTC). However, when a sublinear response can be demonstrated with experimental evidence that is supported by the mutagenic mechanism, guidelines allow for the calculation of a Permitted Daily Exposure (PDE) based on the NOEL in the most relevant animal study with the addition of safety factors [EMA, 2006]. This could result in a higher allowable limit for an alkyl ester C 2014 Wiley Periodicals, Inc. V

impurity as was demonstrated for ethyl methanesulfonate (EMS) using in vivo transgenic mutagenicity data [Gocke et al., 2009]. One pharmaceutically relevant alkyl ester is isopropyl methanesulfonate (IPMS), which is a mutagen in vitro [Glowienke et al., 2005] and in mouse germ cells in vivo [Mattison et al., 1997; Provost et al., 1997; Shelby and Tindall, 1997; Tinwell et al., 1997; Vilarino-Guell et al.,

Grant sponsor: NIH-NIEHS; Grant number: R44ES018017. *Correspondence to: Stephanie L. Coffing, Pfizer Worldwide Research and Development, Eastern Point Rd., Bldg. 274, Groton, CT 06340, USA. E-mail: [email protected] Received 29 April 2014; provisionally accepted 21 August 2014; and in final form 22 August 2014 DOI 10.1002/em.21910 Published online 17 September 2014 in Wiley Online Library (wileyonlinelibrary.com).

Environmental and Molecular Mutagenesis. DOI 10.1002/em In Vivo Mutagenecity of Isopropyl Methanesulfonate

323

TABLE I. Summary of acute and sub-chronic studies conducted in rats treated with IPMS Study Acute Study Sub-chronic Study 1 Sub-chronic Study 2

Dose levels (mg/kg/day)

Treatment Duration (days)

3.5, 7, 14, 28, 56 0.125, 0.25, 0.5, 1, 2 0.1, 0.2, 0.3, 0.4, 1

1 28 28

2003]. IPMS also induces micronucleus formation in mice [Adler et al., 1996] and the formation of DNA adducts in vitro [Li et al., 1990]. However, there is very little somatic cell in vivo mutation data for IPMS in rodents when compared to other alkyl esters such as ethyl nitrosourea (ENU), EMS, methyl nitrosourea (MNU) and methyl methanesulfonate (MMS), all of which have been studied extensively over the years. Available in vivo models to study mutation data and sublinear dose responses are limited and can be expensive, especially when using transgenic animals. One in vivo mutation assay model that shows promise for dose response analysis is the in vivo Pig-a mutation assay. The in vivo Pig-a mutation assay [Bryce et al., 2008; Miura et al., 2008] characterizes the frequency of CD59negative reticulocyte (Ret) and total red blood cell (RBC) populations using flow cytometry. Lack of CD59 cell surface antigen is characteristic of a mutation in the endogenous gene, phosphatidylinositol glycan complementation group A (Pig-a). An improvement to the method [Dertinger et al., 2011a] that utilizes a magnetic column, enriches samples for mutant phenotype cells (hereafter referred to as mutants). This increases the power of the assay to detect rare mutational events, making it attractive for low dose studies. The increase in statistical power should allow for the detection of weak mutagens or weakly mutagenic exposures with repeat dosing since mutations persist in the hematopoietic precursor cells and accumulate in the peripheral blood cells [Miura et al., 2009; Phonethepswath et al., 2010]. Because blood can be analyzed at any time without the need to euthanize the animal, it is possible to assess how an accumulating dose affects mutant frequency, as well as study the time course of the mutant induction. Similar to IPMS, little in vivo data are available for many alkyl esters relevant to pharmaceutical synthesis. Therefore, it would be helpful if one could predict the likely mutation response or likelihood of a sublinear dose response for these lesser known alkyl esters by comparing them to EMS, ENU, MMS and MNU based on their Swain-Scott constant, their mechanism of reactivity (SN1 or SN2) and their O6:N7 guanine adduct ratio. This could help to prioritize conduct of in vivo dose response studies to focus on alkyl esters that are likely to demonstrate a Point of Departure (PoD) above the TTC. The goal of this work was to estimate the mutagenic potency and likelihood of a sublinear dose response for IPMS by comparing the afore mentioned factors to those

Blood collection times Days 22, 15, 22, 28, 56 Days 22, 15, 21, 29, 49, 56,77 Days 29, and 56

for ENU, EMS, MMS, and MNU, and then to test that prediction by characterizing the dose response relationship using the Pig-a assay. Mutant frequencies from acute (single administration) and subchronic (28-day) dosing regimens were investigated for NOELs and then compared to one another to investigate the effects of cumulative dosing. Results from this work should help strengthen the accuracy of predicting dose responses based on the Swain-Scott constant, SN1/SN2 reactivity or O6:N7 guanine adduct ratio. MATERIALS AND METHODS Reagents Isopropyl methanesulfonate (IPMS, CAS 926-06-7) was purchased from Acros Organics, Fair Lawn, NJ and was formulated daily in sterile distilled water. Anticoagulant, balanced salt solution, anti-CD61 phyR 13 dye choerythrin (PE) solution, anti-CD59 PE solution and SYTOV R Pig-a mutation assay kit by were supplied as a prototype MutaFlowV Litron Laboratories Ltd., Rochester, N.Y. Anti-PE paramagnetic beads R magnetic separation columns were purchased from Milteand MACSV nyi Biotec, Auburn, CA. CountBrightTM counting beads were purchased R cell separation media was from Invitrogen, Eugene, OR. LympholyteV purchased from Cedarlane, Burlington, NC and was used for depletion of leukocytes. Heat inactivated fetal bovine serum was purchased from Sigma, St. Louis, MO.

Animals Six- to 8-week-old male Wistar-Han IGS [Crl: WI (Hans)] rats (150–200 g) from Charles River were used in all experiments. Animals received certified pelleted rodent diet and municipal drinking water ad libitum. Animals were housed individually in polycarbonate boxes in a room with relative humidity of 50 6 20%, temperature of 20 to 26 C and a 12-hr light/dark cycle. All animal care and experimental procedures were conducted in compliance with Pfizer’s Institutional Animal Care and Use Committee.

IPMS Treatment Initial dose selection and route of administration were based on information obtained from literature and preliminary range-finding Pig-a mutagenicity studies (data not shown). In all studies, IPMS was administered by intraperitoneal injection once daily in a dose volume of 10 mL/ kg. A summary of all experimental designs are included in Table I. In the acute study, five animals per group were dosed by a single intraperitoneal injection of IPMS (3.5, 7, 14, 28, and 56 mg/kg) or distilled water. In the first subchronic study, five animals per group were dosed by intraperitoneal injection with IPMS (0.125, 0.25, 0.5, 1, and 2 mg/kg/day) or distilled water once a day for 28 consecutive days. The doses in this subchronic study are equivalent to the acute doses fractionated over 28 days. In the second subchronic study, five animals per group were dosed by intraperitoneal injection with IPMS (0.1, 0.2, 0.3,

Environmental and Molecular Mutagenesis. DOI 10.1002/em 324

Coffing et al.

0.4, and 1 mg/kg/day) or distilled water once a day for 28 consecutive days.

Blood Collection All blood collection occurred via jugular venipuncture. Approximately 0.5 to 1 mL of blood was collected and transferred to a microtainer tube containing K2EDTA to prevent clotting. Eighty microliters of blood was transferred from the microtainer tube to a microcentrifuge tube containing 100 mL of anticoagulant and mixed. This sample was used for Pig-a mutation analysis. Blood collection time points are shown in Table I.

Sample Labeling, Processing, and Flow Analysis Sample labeling, processing and flow analysis were conducted according to methods described by Dertinger et al. (2011) with slight modifications (Instruction manual for prototype High Throughput Pig-a mutation assay, ver. 120209 and ver. 120910). Briefly, on the day of blood collection, blood samples were depleted of leukocytes and platelets, washed, and stored overnight at approximately 4 C. Cell labeling with anti-CD59 and anti-CD61 antibodies, sample processing and flow analysis, occurred approximately 24 hr post blood collection. Cells were labeled with anti-CD 61 and anti-CD 59-PE antibodies, and incubated with Anti-PE magnetic beads. A subset (20 mL) of these labeled samples was designated as pre-column samples, incubated with counting beads and nucleic acid dye and then analyzed by flow cytometry. The remainder of the labeled samples was run over a magnetic column to deplete the wild type cells, incubated with counting beads and nucleic acid dye and then analyzed by flow cytometry. These samples were designated as postcolumn samples. Both the pre-column and the post-column samples were analyzed on a BD FACSCanto flow cytometer equipped with an argon laser and FACSDiva software. Precolumn samples were acquired using a high fluidics rate with a forward scatter (FSC) threshold for approximately 1 min or until at least 1,000 counting beads were counted. Post-column samples were acquired under a high fluidics rate for approximately 3 min with a FSC threshold until the sample was nearly exhausted. On average, the equivalent of approximately 4 million Rets and 160 million RBCs per rat were interrogated for mutant phenotype on each day of analysis.

Animal Selection Two days prior to the first dose administration (Day -2) of the acute and first subchronic study, main study animals plus three to five extra animals were analyzed to determine background Pig-a mutant frequency. Animals with background mutant frequencies below the 5th and above the 95th percentile based on historical controls (Rets 5 0 to 1.8 mutants/ million; RBCs 5 0.3 to 1.35 mutants/million), were removed from the study. If more animals than the extra number had background frequencies outside the established range, animals with background mutant frequencies farthest from the 95th percentile were removed first. In the second (low dose) subchronic study, predose analysis was performed 5 days prior to the first dose administration (Day -5); however, due to antibody labeling errors, none of the predose data was acceptable and animals were randomized without use of predose background frequencies. After predose analysis of the mutant frequency, a random list of animal numbers was generated to assign animals to each dose group. A one-way analysis of variance analysis (ANOVA) followed by a Tukey’s R 5 for Windows version Multiple Comparison Test (GraphPad PrismV 5.02) was performed to ensure that there was no statistical difference in RBC or Ret mutant frequency between the treatment groups.

significant trends by performing a one way ANOVA and a single degree of freedom linear contrast. Prior to analyzing the Pig-a data for significance, mutant Ret and RBC frequencies for each animal group were analyzed for outliers at the R Software Quick Calcs; Grubb’s test). 0.05 significance level (GraphPadV Outliers were excluded from the statistical analysis and graphs. The number of outliers averaged approximately 2 out of 30 animals per analysis with a range of 1 to 4 outliers, with the exception of Day 15 of the first subchronic study where technical errors in labeling resulted in eight animals being excluded from analysis. For Pig-a mutant analyses, percent reticulocytes (Ret), the number of mutant Rets and the number of mutant red blood cells (RBC) were analyzed for significant trends by performing a one-way ANOVA to compare treatment groups. The analysis was performed on log transformed (base 10) number of mutant Rets, log transformed (base 10) number of mutant RBCs and untransformed percent Ret. In order to avoid taking the log of zero, a small constant (0.1) was added to each measured value for the mutant frequencies. The log transformation was found to best reduce the right-skewness of the mutant Ret and mutant RBC frequencies, bringing the underlying distribution to normality. The analysis for significant trends was performed using a sequential trend test based on linear contrasts from the ANOVA. The trend test for both mutant Rets and mutant RBCs was one sided, testing for increasing trend, whereas the trend test for % Rets was two sided. Subsequent trend tests were one sided. The trend test was followed by a pair-wise comparison using the Dunnett’s t-test, which was performed so that statistical significance would not be missed in the case of an inverted U-shaped dose–response curve. The trend test was used to determine statistical significance of a response in all cases in these experiments given the shape of the dose–response curve. To determine the day on which the maximum mutant frequency occurred in a given experiment, day to day mutant RBC/Ret frequency comparisons were made using a linear mixed model to study the dose and day effects as well as the interaction between day and dose with animals being the random block effect [Lettell et al., 2006]. If the dose-by-day interaction was not significant, a post hoc Tukey’s test was used to determine what day(s) had the highest response. If the dose-by-day interaction was significant, a linear contrast was used to study the day trend at each dose and the differences between the trends were evaluated. For the acute to subchronic comparison, two types of adjustments were used. One adjustment involved subtracting the average control value from the average value for each treated group. The other adjustment calculated the fold change of each treated group to the control group. Each set of adjusted data was then analyzed by using an ANOVA model with factors for Group (2 to 6), study (acute and subchronic) and group-study interaction. The acute to subchronic comparison was done using a t-test. It is necessary to ensure equal variance across doses for the sublinearity dose analysis; therefore, the data was first tested for equal variance across doses using a Levene’s test [Levene, 1960]. The data did not have equal variance; therefore; the responses were transformed using a square root transformation to achieve homogeneity before applying the bi-linear model [Lutz and Lutz, 2009]. The NOEL for Ret and RBC mutant induction on a given day was defined as the highest non-statistically significant dose in the sequential trend test. The NOEL for the entire test was considered to be the lowest NOEL identified across all days throughout the study.

RESULTS

Statistical Analysis

Pig-a Mutation Induction in Rats After One Administration of IPMS

Statistical analysis was performed using SAS, version 9.2 except where otherwise noted. Percent body weight change was analyzed for

Frequencies of mutant Rets and RBCs from rats treated acutely with IPMS at doses of 3.5 to 56 mg/kg are shown

Environmental and Molecular Mutagenesis. DOI 10.1002/em In Vivo Mutagenecity of Isopropyl Methanesulfonate

325

Fig. 1. a: Number of CD 59- (mutant) reticulocytes per 106 reticulocytes in the peripheral blood of rats treated acutely with IPMS at doses of 0, 3.5, 7, 14, 28, and 56 mg/kg. Bars represent mutant frequency, and error bars represent 1 SD. Sequential trend test for mutant reticulocytes; *P < 0.05; **P < 0.01. b: Number of CD 59- (mutant) RBCs per 106 RBCs in

the peripheral blood of rats treated acutely with IPMS at doses of 0, 3.5, 7, 14, 28, and 56 mg/kg. Bars represent mutant frequency, error bars represent 1 SD and lines represent the percent reticulocytes for each day of analysis. Sequential trend test for mutant RBCs; *P < 0.05; **P < 0.01.

in Figures 1a and 1b, respectively. The results for the average percent Rets for each day of the study are shown in Figure 1b. Toxicity was observed as a statistically significant dose-related decrease in percent Rets compared to control animals on Day 15 at doses of 14 to 56 mg/kg. For the remainder of the analysis days, the percentage of Rets for the treatment groups was similar to that for the control. Body weights were recorded on Days 1, 15, 21, and 29 as a possible indication of treatment related clinical signs. All animals in all dose groups gained weight; however, there was a statistically significant decreasing trend (trend test) in the percentage

of body weight gained over the dose range (data not shown). A statistically significant dose related increase in mutant Ret frequency was observed on each analysis day, with the maximum response observed on Day 15 (linear mixed model) with a 58-fold increase over the controls at the dose of 56 mg/kg. There was a statistically significant trend on all analysis days (3.5 to 56 mg/kg). There was a statistically significant sequential trend from 3.5 to 28 mg/kg, observed on Days 15 through 28 and from 3.5 to 14 mg/kg on Day 21. Based on the highest nonstatistically significant mutant Ret frequency observed

Environmental and Molecular Mutagenesis. DOI 10.1002/em 326

Coffing et al.

throughout the study, the NOEL for the induction of mutant Rets after a single dose of IPMS was defined as 7 mg/kg. A statistically significant dose response in the mutant RBC frequency was observed for each analysis day with a maximum accumulation of mutants occurring on Day 28 (linear mixed model) with a 28-fold increase over the controls at 56 mg/kg. A statistically significant trend was observed at doses of 3.5 to 56 mg/kg on all analysis days, and a statistically significant sequential trend from 3.5 to 28 mg/kg was observed on all analysis days. A statistically significant sequential trend was observed at 3.5 to 14 mg/kg on Days 15 through 28. A NOEL of 7 mg/kg was identified for the induction of mutant RBCs after a single dose, which was equivalent to the NOEL for the mutant Rets. Pig-a Mutation Induction in Rats After 28 Daily Administrations of IPMS Results for the mutant Ret and RBC frequencies from rats treated subchronically with IPMS doses ranging from 0.125 to 2.0 mg/kg/day are shown in Figures 2a and 2b. The results for the average percent Rets for each day of the study are shown in Figure 2b. RBC toxicity was not evident with subchronic dosing as there were no doserelated statistically significant changes in percent Rets on any of the analysis days. Body weights were recorded on Days 1, 5, 8, 12, 15, 19, 22, and 26 as a possible indication of treatment related clinical signs. All animals in all dose groups gained weight and there was no statistically significant difference in the percentage of body weight gained over the dose range (data not shown). There was a statistically significant dose related increase in the mutant Ret frequency on each analysis day except Day 77. The maximum response occurred on Day 29 (linear mixed model) with a fivefold increase at 2 mg/kg/day. A statistically significant trend was observed from 0.125 to 2 mg/kg/day on Days 15 through 56. A significant sequential trend was achieved from 0.125 to 1 mg/kg/day only on Days 21 and 56. In this experiment, the NOEL for the induction of mutant phenotype Rets was defined as 0.5 mg/kg/day. For mutant RBCs, a statistically significant dose related increase was observed on Days 21 through 77 with the maximum response observed on Day 56 of the study (linear mixed model) with an 11-fold increase at 2 mg/kg/ day. There was a statistically significant trend from 0.125 to 2 mg/kg/day on Days 21 through 77, while a statistically significant sequential trend was achieved at 0.125 to 1 mg/kg/day on Days 29 through 77. A statistically significance sequential trend from 0.125 to 0.5 mg/kg/day was only achieved on Day 29. In this experiment, a NOEL was identified at 0.25 mg/kg/day which was twofold lower than the NOEL observed for the mutant Rets.

Pig-a Mutation Induction in Rats After 28 Daily Administrations of Low Doses of IPMS A second subchronic study was conducted to further evaluate the shape of the dose response curve at low doses. Doses of 0.1, 0.2, 0.3, 0.4, and 1 mg/kg/day were evaluated on Days 29 and 56, where the maximum mutant response for Rets and RBCs, respectively, was observed in the first subchronic test. Results for the mutant Ret and RBC frequencies are shown in Figures 3a and 3b. The results for the average percent Rets for each day of the study are shown in Figure 3b. There was no statistically significant trend on Day 29 for RBC toxicity. Although there was a statistically significant decreasing trend on Day 56 at doses of 0.4 and 1 mg/kg/day, there was no statistical difference when those doses where compared to the controls (Dunnett’s test). Body weights were recorded on Day 1 and 28 as a possible indication of treatment related clinical signs. All animals in all dose groups gained weight and there was no statistically significant difference in the percentage of body weight gained over the dose range (data not shown). A statistically significant dose related trend in mutant Rets was not observed on either Day 29 or 56; therefore, the NOEL for mutant Rets in this experiment was identified as 1 mg/kg/day. Although there was not a significant trend, the dose of 1 mg/kg/day was significantly elevated compared to the control on Day 56. The first indication of a statistically significant dose related increase in mutant RBCs at doses of 0.4 and 1 mg/kg/day was observed on Day 56 with a sixfold maximum increase at 1 mg/kg/day. A NOEL for mutant RBC induction in this experiment was identified at 0.3 mg/kg/day, which was similar to that identified in the first subchronic study and approximately threefold lower than the NOEL observed for mutant Rets in this study. Comparison of Acute and Subchronic Mutant RBC Induction To examine the dose response for sublinearity, a comparison of the acute and subchronic mutant RBC frequencies was conducted and is shown in Figure 4. Mutant RBC frequencies from Day 28 (maximum mutant response observed) of the acute study was compared to mutant RBC frequencies from Day 56 (maximum mutant response, which occurred 28 days after the last IPMS dose) in the subchronic study. There was a statistically significant difference between the acute mutant RBC frequencies at the top two doses of 28 and 56 mg/kg and the subchronic mutant RBC frequencies at the equivalent fractionated doses of 1 and 2 mg/kg/day. The acute response at 28 and 56 mg/kg was four and threefold over the subchronic responses at 1 and 2 mg/kg/day, respectively. The RBC mutant frequency for the acute mid-dose of 14 mg/kg was not statistically different from the corresponding subchronic dose of 0.5 mg/kg/day; however, it

Environmental and Molecular Mutagenesis. DOI 10.1002/em In Vivo Mutagenecity of Isopropyl Methanesulfonate

327

Fig. 2. a: Number of CD 59- (mutant) reticulocytes per 106 reticulocytes in the peripheral blood of rats treated sub-chronically with IPMS at doses of 0, 0.125, 0.25, 0.5, 1, and 2 mg/kg/day. Bars represent mutant frequency, and error bars represent 1 SD. Sequential trend test for mutant reticulocytes; *P < 0.05; **P < 0.01. b: Number of CD 59- (mutant)

RBCs per 106 RBCs in the peripheral blood of rats treated subchronically with IPMS at doses of 0, 0.125, 0.25, 0.5, 1, and 2 mg/kg/ day. Bars represent mutant frequency, error bars represent 1 SD and lines represent the percent reticulocytes for each day of analysis. Sequential trend test for mutant RBCs; *P < 0.05; **P < 0.01.

was 3.4-fold higher than the corresponding subchronic dose. Additionally, the 14 mg/kg mutant RBC frequency of 5.4 mutants/million cells was outside of the historical control 95% confidence interval of 0 to 2.35 mutants/million cells at 99 % confidence, while the mutant frequency at 0.5 mg/kg/day (1.6 mutants/million cells) was within the tolerance interval. There was no statistical difference between acute and fractionated doses observed at the lower dose levels.

Application of the Lutz and Lutz Bilinear Model The mutant RBC frequencies from both subchronic studies were combined. The Lutz and Lutz Bilinear model [Lutz and Lutz, 2009] was then applied to test for sublinearity. Sublinearity was rejected by this model. Calculation of the PDE Based on the subchronic NOEL for mutant RBCs of 0.25 mg/kg/day with safety factors of 5 for extrapolation

Environmental and Molecular Mutagenesis. DOI 10.1002/em 328

Coffing et al.

Fig. 4. Comparison of mutant frequencies observed in RBCs of rats 28day post treatment from a single acute dose of IPMS vs. 28 days of subchronic treatment. Doses are expressed as the cumulative dose rats received in both studies.

Fig. 3. a: Number of CD 59- (mutant) reticulocytes per 106 Rets in the peripheral blood of rats treated sub-chronically with IPMS at doses of 0, 0.1, 0.2, 0.3, 0.4, and 1 mg/kg/day. Bars represent mutant frequency, and error bars represent 1 SD. Sequential trend test for mutant reticulocytes *P < 0.05; **P < 0.01. b: Number of CD 59- (mutant) RBCs per 106 RBCs in the peripheral blood of rats treated subchronically with IPMS at doses of 0, 0.1, 0.2, 0.3, 0.4, and 1 mg/kg/day. Bars represent mutant frequency, error bars represent 1 SD and lines represent the percent reticulocytes for each day of analysis. Sequential trend test for mutant RBCs; *P < 0.05; **P < 0.01.

from rat to human, 10 for human variability, 10 for shortterm study duration and 10 for a serious toxicity; the calculated PDE is 2.5 mg/day for a 50 kg human, which is not appreciably higher than the TTC.

DISCUSSION Little is known about the in vivo mutagenicity of IPMS, and the limited data is derived primarily from studies in mouse germ cells [Mattison et al., 1997; Provost et al., 1997; Shelby and Tindall, 1997; Tinwell et al., 1997; Vilarino-Guell et al., 2003]. The main purpose of this study was to characterize the dose response relationship of IPMS in rats using the Pig-a mutation assay and

to compare it to the previously reported Pig-a responses for EMS, ENU, MMS and MNU. In the current study, IPMS induced the Pig-a mutant phenotype in the Ret and RBC populations of rat peripheral blood in a dose related manner following both acute and subchronic dosing regimens. The maximum RBC mutant response for both the acute and subchronic studies occurred approximately 28 days after the last dose administration and was observed approximately 1 to 2 weeks after the corresponding maximum mutant response for Rets. This difference in temporality of the maximum mutant frequency between Rets and RBCs is expected and can be explained by the kinetics of erythropoiesis and by the model of Pig-a mutation induction proposed by Miura et al. [2009]. In Miura’s model, mutations in the Pig-a gene can occur in the stem cell population thereby creating mutant reticulocytes and ultimately mutant RBCs that remain in the RBC pool for approximately 60 days [Miura et al., 2009; Peruzzi et al., 2010] As time progresses, mutant RBCs replace wild type RBCs increasing the mutant RBC response. A similar time course has been observed with other alkyl esters and with other classes of compounds [Phonethepswath et al., 2010; Dertinger et al., 2011b; Dobo et al., 2011; Dertinger et al., 2012]. In the subchronic studies, the NOEL for mutant Rets was two and threefold higher than the respective NOEL for mutant RBCs. This difference in NOELs is expected since the assay has greater power for detecting mutant RBCs than mutant Rets based on the number of cell equivalents scored for each population. For example, there is an 88% chance of detecting mutant RBCs compared to a 55% chance of detecting mutant Rets when the mutant frequency is 3 fold greater than the control [Torous et al., 2012]. There is greater animal to animal

Environmental and Molecular Mutagenesis. DOI 10.1002/em In Vivo Mutagenecity of Isopropyl Methanesulfonate

variability in the mutant Ret population as compared to the mutant RBC population, which is likely due to greater scoring error based on the smaller number of Ret equivalents. It is also possible that there could be greater inherent variability in mutant Rets. Although both mutant populations provide important information, the mutant RBC population will likely be more useful when investigating dose responses since it appears to be more sensitive for the for the detection of weak increases than the mutant Ret population. The identification of a sublinear dose response for IPMS could be beneficial in its management as a potential pharmaceutical impurity, allowing for the calculation of a compound-specific PDE based on in vivo genotoxicity data [ICH, 2011]. In the case of a sublinear dose response, it is expected that the subchronic mutant response would not be additive due to the repair capabilities of the animal at lower doses. Therefore, the subchronic mutant response from a fractionated dose would likely be lower than the corresponding acute mutant response as was demonstrated for EMS [Gocke et al., 2009; Dobo et al., 2011].With a linear response, one would expect the subchronic and the acute mutant responses to be approximately equal, or at the very least, to have a linear relationship similar to that demonstrated for ENU by Gocke et al. [2009]. It is possible that the dose response for IPMS may not be strictly additive given the lack of a sustained mutant frequency throughout the study. For ENU, the mutant frequency was sustained for up to 94 days [Phonethepswath et al., 2010; Dertinger et al., 2011c] which suggests that the mutations occurred in stem cells. The IPMS response, however, started decreasing on Days 57 and 77 of the acute and subchronic studies, respectively. This suggests that erythroid cells with less self-renewal capacity relative to stem cells, were the target cell for mutation induction as was described by [Dertinger et al., 2010] for MNU. An initial observation of the dose response of IPMS illustrates that when the dose is doubled, the mutant frequency also doubles. This holds especially true for the high doses, in both the acute and subchronic studies and is in line with previous reports for ENU [Phonethepswath et al., 2010; Cammerer et al., 2011] The fact that the IPMS acute doses of 28 and 56 mg/kg were statistically elevated compared to the fractionated equivalents of 1 and 2 mg/kg and that the 14 mg/kg dose was elevated, although not statistically significant, over the 0.5 mg/kg suggests that a sublinear response is likely. It is interesting, therefore, that the Lutz and Lutz bilinear model did not reject linearity for the subchronic dose response data. It has, however, been demonstrated that PoDs can often be defined using other models, such as the benchmark dose approach, even when the bilinear model was not able to identify a breakpoint [Johnson et al., 2014] Although the Lutz and Lutz bilinear model did not reject linearity, there was interest in evaluating the PDE

329

based on the assumption that the response was sublinear with a NOEL of 0.25 mg/kg/day. The PDE is calculated using the NOEL from the most relevant animal study and a set of safety factors [EMA, 2006]. This approach was used by Roche to understand the risk to patient from EMS contamination of Viracept [Muller and Gocke, 2009] Since IPMS is a potent mutagen, the calculation of a PDE value does not prove especially beneficial. However, this exercise could prove beneficial when levels of alkyl esters, that are less potent than IPMS, cannot be decreased to the TTC limit. For example, the calculated PDE for MNU, EMS and MMS are 5- to 50-fold higher and would be well above the TTC. A comparison of Pig-a results for IPMS, ENU, EMS, MMS and MNU are shown in Table II. When comparing the mutant RBC NOEL for IPMS to those previously identified for the various alkyl esters tested in the Pig-a assay [Dobo et al., 2011; Lynch et al., 2011; Dertinger et al., 2012], the rank order from least potent to most potent is EMS < MMS < MNU < IPMS  ENU. This rank order, based solely on Pig-a data, is somewhat similar to one established using benchmark dose (BMD) instead of NOEL. The BMD rank order is EMS < MMS < ENU < MNU [Johnson et al., 2014], which used data from multiple in vitro and in vivo endpoints, (including Pig-a data) but was not exclusively Pig-a data unlike the data presented here. Unlike IPMS, for two of the alkylators, EMS and ENU, an in vivo sublinear dose response was statistically substantiated using the bilinear model [Dobo et al., 2011]. There were, however, differences in route of compound administration (oral versus intraperitoneal injection), Pig-a protocol (basic versus high throughput), and statistical analysis of the data (non-transformed versus transformed) between the studies that could account for the difference in how the shape of the doseresponse curve was fit statistically. Even though IPMS was dosed by intraperitoneal injection versus oral gavage for ENU and EMS, it was unlikely that this difference contributed greatly to the difference in the shape of the dose response curves. Adler et al previously reported no difference in the bone marrow micronucleus frequency in male CBA mice when dosed with IPMS at 200 mg/kg using oral or intraperitoneal administration [Adler et al., 1996]. In contrast to route of administration, it is likely that the Pig-a protocol used in the current IPMS study, which employed magnetic bead enrichment of mutant populations, had an influence on the observed response. This enrichment enables the scoring of 10 times more Rets and 100 times more RBCs than the basic protocol used for ENU and EMS [Dertinger et al., 2011a]. With the increase in the number of Rets and RBCs analyzed comes increased statistical power as described by Torous et al., [2012]. Using six animals per dose group, they were able to show that a 4.6-fold increase in mutant Rets and a 2.6-

Environmental and Molecular Mutagenesis. DOI 10.1002/em 330

Coffing et al.

TABLE II. Comparison of Pig-a results, Swain-Scott constant, O6:N7 guanine adduct ratio and alkylating mechanism of various alkyl esters

CAS No. Pig-a protocol Route of administration Rat strain Doses (mg/kg/day) Days of Pig-a analyses NOELd (28-day treatment) Point of Departure Bilinear model Permitted Daily Exposured Swain Scott Constanti,j Major alkylation mechanismj,k O6:N7 Guanine Adduct Ratioj,l

ENUa

IPMS

MNUb

EMSa

MMSc

759-73-9 Basic Oral Sprague Dawley 0.043, 0.25, 1.25, 5 and 10 15, 29/31, 57

684-93-5 Basic Oral Sprague Dawley 0.1, 0.3, 0.5, 0.6, 0.9, 1.25, 1.5, 2, 2.5, 5, and 10 1, 15, and 29 (GSK) 1, 4, 15, 29, 43 and 57 (BMS) 1.25 mg/kg/dayf Not determined

62-50-0 Basic Oral Sprague Dawley 6.25, 12.5, 25, 50 and 100 15, 29 and 55

66-27-3 High Throughput Oral Sprague Dawley 7.5, 15, and 30

0.25 mg/kg/daye 0.88 mg/kg/day

926-06-7 High Throughput Intraperitoneal Wistar-Han 0.1, 0.125, 0.2, 0.25, 0.3, 0.4, 0.5, 1, and 2 15, 21, 28/29, 49, 56 and 77 0.25 mg/kg/day Not sub-linear

12.5 mg/kg/dayg 21.9 mg/kg/day

7.5 nig/kg/dayh Not determined

0.43 lg/day 0.26 Sn1

2.5 lg/day 0.29 Sn1

12.5 lg/day 0.42 Sn1

125 lg/day 0.67 Sn1/ Sn2

75 lg/day 0.83 Sn2

0.6

030

0.1

0.03

0.004

15, 29 and 42

a

Dobo et al. 2011. Lynch et al. 2011. c Dertinger et al. 2012. d Determination and calculation of NOELs and PDEs was conducted by the authors using the data available in each respective paper cited above e There was a statistically significant increase in mutant RBCs on Day 29 but no statistical increase on Day 56. There was a statistically significant upward trend in mutant Rets on Day 15 but this did not reproduce at later time points. f Based on Day 29 analysis for both Rets and RBCs; Day 56 not performed g There was a statistically significant increase in mutant Rets only on Day 55 with questionable relevance. h There was a statistically significant increase in mutant Rets only on Day 29; no increase in mutant RBCs at this dose on any day. i Vogel et al. 1990. j Beranek 1990. k Shelby and Tindall 1997. l Lawley et al. 1975. b

fold increase in mutant RBC would be detected with 80% probability [Torous et al., 2012]. Therefore, it is likely that the enriched Pig-a protocol used for IPMS was more sensitive for mutant detection at lower doses than was possible in the earlier study with ENU and EMS. Finally, there was a difference in how the bilinear statistical model [Lutz and Lutz, 2009] was applied to the IPMS Pig-a data for determining whether a sublinear response existed when compared to its use for the EMS and ENU data. With IPMS, the data was square root transformed prior to applying the bilinear model in order to equalize the variance across groups [Levene, 1960]. The EMS and ENU data was not transformed despite apparent unequal variance. It is possible that linearity would not have been rejected for the EMS and ENU Piga data had it been transformed to equalize the variance. Due to the variety of potential alkyl ester impurities, it would take a substantial amount of time and money to investigate each impurity for the presence of a sublinear dose response. While the in vivo Pig-a mutation assay provides a less expensive and resource intensive alternative to the transgenic mutation assays, the assay still requires time for measurable accumulation of mutants. Comparison of the Swain-Scott constant, the O6:N7 gua-

nine adduct ratio and of the SN1 and SN2 reaction mechanism of an alkyl ester of interest to that of well characterized alkyl esters would provide an understanding of the relative potency. This could aid in the decision of which impurities to further investigate in vivo focusing efforts on those likely to be less potent mutagens with a PoD sufficiently above the TTC. Since the Swain-Scott constant (s value) is a measurement of an alkyl ester’s reactivity to a nucleophile, it can be used to predict the nucleophile preference of that compound [Swain, 1953]. Alkylating agents with a high s value and/or SN2 mechanism favor the N7 of guanine. This results in a lower O6:N7 guanine adduct ratio than those with a low s value and/or SN1 mechanism, which favor the O6 of guanine [Hemminki, 1983; Vogel and Ashby, 1994; Doak et al., 2007]. The amount of N7 and O6 adducts produced after exposure to an alkylating agent can also affect the shape of the dose response curve and can be used to help predict the mutagenic potency of a compound. Due to their short half- life, N7 adducts are unstable and therefore nonmutagenic [Boysen et al., 2009]. In contrast, O6 adducts are known to produce pro-mutagenic lesions [Singer, 1985; La and Swenberg, 1996; Hemminki et al., 2000].

Environmental and Molecular Mutagenesis. DOI 10.1002/em In Vivo Mutagenecity of Isopropyl Methanesulfonate

Therefore, alkylating agents that have a SN2 mechanism would be expected to produce mainly N7 adducts, which would be more likely to result in a sublinear response being observed at higher doses given the weakly mutagenic nature. Alkylating agents that have a SN1 mechanism would be expected to produce a higher proportion of O6 adducts, which would more likely result in either a very low dose PoD or the inability to detect sublinearity statistically. Since mutations are not exclusively induced by O6 and N7 adducts, the O6:N7 guanine adduct ratio is best used to predict the potency of a compound in relation to better-known compounds. Based on the SN1/SN2 reaction type and similarity between the Swain Scott constants, one would expect IPMS to have a very low sublinear dose and NOEL similar to ENU. Even though ENU has a higher O6:N7 guanine adduct ratio translating to a higher carcinogenic potential [Vogel and Ashby, 1994] than IPMS, exposure to IPMS results in bulkier adducts which are slower to repair than those induced by ENU [Sega et al., 1976; Morimoto et al., 1985; Pegg et al., 1985]. In fact, the NOEL for IPMS was 0.25 mg/kg/day, which is the same as the NOEL of 0.25 mg/kg/day for ENU [Dobo et al., 2011]. Similarly, it is assumed that it is not useful to perform in vivo mutation dose-response analyses on MNU based on the SN1 alkylation mechanism and the Swain Scott constant. On the contrary, it might prove beneficial to study the dose-response of MMS based on similarities to EMS. In conclusion, IPMS induced Pig-a mutant phenotype Ret and RBC populations following both acute and subchronic in vivo exposures with the maximum mutant Ret frequency occurring prior to the mutant RBC frequency. A comparison of the dose response relationships between the acute and subchronic exposures suggests a sublinear dose response since the actual RBC mutant frequencies were less than additive. However, this was not supported by statistical analysis using the Lutz and Lutz bilinear model. The comparison of the Swain Scott constant and the SN1/ SN2 mechanism aided in predicting the low NOEL observed for IPMS. The resulting PDE, based on the low NOEL, was not appreciably higher than the TTC, and therefore would not benefit in the impurity-related management of IPMS. The accurate estimation of the NOEL based on the previously mentioned factors, in practice, would have pre-empted further investigation of the potent mutagen, IPMS. Estimation of the PDE based on a predicted NOEL would benefit future investigations of lesser known alkyl esters, helping to identify those less potent than IPMS that merit in vivo dose response analysis.

ACKNOWLEDGMENTS The authors thank William Gunther for in vivo support during dosing and blood collection as well as Susan Por-

331

tugal and Dingzhou Li for assistance with the statistical analyses of the Pig-a data. The authors also wish to thank Litron Laboratories Ltd. for providing technical support and the Pig-a kits for some of this work. AUTHOR CONTRIBUTIONS Stephanie Coffing and Michelle Kenyon designed the studies. Stephanie Coffing procured the animals, and collected and processed samples. Michelle Kenyon collected and analyzed the data, conducted statistical analysis and prepared draft figures. Joel Ackerman collected data and reviewed all figures. Thomas Shutsky dosed animals and collected and processed samples. Stephanie Coffing prepared the manuscript draft with important intellectual input from Michelle Kenyon and Dr. Krista Dobo. All authors approved the final manuscript and had complete access to the study data. None of the authors have a conflict of interest in regard to this manuscript. REFERENCES ICoH (ICH). 2011. ICH Q3C Impurities: Guidelines for Residual Solvents (R5). pp 29. Available at: www.ich.org/fileadmin/Public_ Web_Site/ICH_Products/Guidelines/Quality/Q3C/Step4/Q3C_R5_ Step4.pdf. Adler ID, Tinwell H, Kliesch U, Ashby J. 1996. Activity of iPMS and nPMS in mouse bone marrow micronucleus assays: comparison with mouse dominant lethal assay data. Mutat Res 349:241–247. Beranek DT. 1990. Distribution of methyl and ethyl adducts following alkylation with monofunctional alkylating agents. Mutat Res 231: 11–30. Boysen G, Pachkowski BF, Nakamura J, Swenberg JA. 2009. The formation and biological significance of N7-guanine adducts. Mutat Res 678:76–94. Bryce SM, Bemis JC, Dertinger SD. 2008. In vivo mutation assay based on the endogenous Pig-a locus. Environ Mol Mutagen 49:256– 264. Cammerer Z, Bhalli JA, Cao X, Coffing SL, Dickinson D, Dobo KL, Dobrovolsky VN, Engel M, Fiedler RD, Gunther WC, Heflich RH, Pearce MG, Shaddock JG, Shutsky T, Thiffeault CJ, Schuler M. 2011. Report on stage III Pig-a mutation assays using Nethyl-N-nitrosourea-comparison with other in vivo genotoxicity endpoints. Environ Mol Mutagen 52:721–730. Dertinger SD, Bryce SM, Phonethepswath S, Avlasevich SL. 2011a. When pigs fly: immunomagnetic separation facilitates rapid determination of Pig-a mutant frequency by flow cytometric analysis. Mutat Res 721:163–170. Dertinger SD, Phonethepswath S, Avlasevich SL, Torous DK, Mereness J, Bryce SM, Bemis JC, Bell S, Weller P, Macgregor JT. 2012. Efficient monitoring of in vivo pig-a gene mutation and chromosomal damage: summary of 7 published studies and results from 11 new reference compounds. Toxicol Sci 130:328–348. Dertinger SD, Phonethepswath S, Franklin D, Weller P, Torous DK, Bryce SM, Avlasevich S, Bemis JC, Hyrien O, Palis J, MacGregor JT. 2010. Integration of mutation and chromosomal damage endpoints into 28-day repeat dose toxicology studies. Toxicol Sci 115:401–411. Dertinger SD, Phonethepswath S, Weller P, Avlasevich S, Torous DK, Mereness JA, Bryce SM, Bemis JC, Bell S, Portugal S, Aylott M, MacGregor JT. 2011b. Interlaboratory Pig-a gene mutation assay trial: Studies of 1,3-propane sultone with immunomagnetic

Environmental and Molecular Mutagenesis. DOI 10.1002/em 332

Coffing et al.

enrichment of mutant erythrocytes. Environ Mol Mutagen 52: 748–755. Dertinger SD, Phonethepswath S, Weller P, Nicolette J, Murray J, Sonders P, Vohr HW, Shi J, Krsmanovic L, Gleason C, Custer L, Henwood A, Sweder K, Stankowski LF Jr, Roberts DJ, Giddings A, Kenny J, Lynch AM, Defrain C, Nesslany F, van der Leede BJ, Van Doninck T, Schuermans A, Tanaka K, Hiwata Y, Tajima O, Wilde E, Elhajouji A, Gunther WC, Thiffeault CJ, Shutsky TJ, Fiedler RD, Kimoto T, Bhalli JA, Heflich RH, MacGregor JT. 2011c. International Pig-a gene mutation assay trial: evaluation of transferability across 14 laboratories. Environ Mol Mutagen 52:690–698. Doak SH, Jenkins GJ, Johnson GE, Quick E, Parry EM, Parry JM. 2007. Mechanistic influences for mutation induction curves after exposure to DNA-reactive carcinogens. Cancer Res 67:3904–3911. Dobo KL, Fiedler RD, Gunther WC, Thiffeault CJ, Cammerer Z, Coffing SL, Shutsky T, Schuler M. 2011. Defining EMS and ENU dose-response relationships using the Pig-a mutation assay in rats. Mutat Res 725:13–21. Elder D, Facchine KL, Levey JN, Parsons R, Ridge D, Semo L, Teasdale A. 2012. An approach to control strategies for sulfonate ester formation in pharmaceutical manufacturing based on recent scientific understanding. Org Process Dev 16:1707–1710. EMA. 2006. Evaluation of Medicines for Human Use: Guideline on the Limits of Genotoxic Impurities. London, England: EMA. Glowienke S, Frieauff W, Allmendinger T, Martus HJ, Suter W, Mueller L. 2005. Structure-activity considerations and in vitro approaches to assess the genotoxicity of 19 methane-, benzene- and toluenesulfonic acid esters. Mutat Res 581:23–34. Gocke E, Ballantyne M, Whitwell J, Muller L. 2009. MNT and MutaMouse studies to define the in vivo dose response relations of the genotoxicity of EMS and ENU. Toxicol Lett 190:286–297. Hemminki K. 1983. Nucleic acid adducts of chemical carcinogens and mutagens. Arch Toxicol 52:249–285. Hemminki K, Koskinen M, Rajaniemi H, Zhao C. 2000. DNA adducts, mutations, and cancer 2000. Regul Toxicol Pharmacol 32:264–275. Johnson GE, Soeteman-Hernandez LG, Gollapudi BB, Bodger OG, Dearfield KL, Heflich RH, Hixon JG, Lovell DP, Macgregor JT, Pottenger LH, Thompson CM, Abraham L, Thybaud V, Tanir JY, Zeiger E, van Benthem J, White PA. 2014. Derivation of point of departure (PoD) estimates in genetic toxicology studies and their potential applications in risk assessment. Environ Mol Mutagen. DOI 10.1002/em.21870. La DK, Swenberg JA. 1996. DNA adducts: biological markers of exposure and potential applications to risk assessment. Mutat Res 365:129–146. Lawley PD, Orr DJ, Jarman M. 1975. Isolation and identification of products from alkylation of nucleic acids: ethyl- and isopropylpurines. Biochem J 145:73–84. Lettell R, Milken GA, Stroup WW, Wolfinger RD. 2006. A Setting for Mixed Models Applications: Randomized Blocks Designs. SAS for Mixed Models. Cary, NC: SAS Inst Inc. pp 18–56. Levene H. 1960. Robust test for equality of variances.In: Olkin I, Ghurye SG, Hoefding W, Madow WG, Mann HB, editors. Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling. Stanford, CA: Stanford University Press. pp 278–292. Li F, Solomon JJ, Mukai F, Segal A. 1990. In vitro reactions of isopropyl methanesulfonate with DNA and with 2’-deoxyribonucleosides. Cancer Biochem Biophys 11:253–264. Lutz WK, Lutz RW. 2009. Statistical model to estimate a threshold dose and its confidence limits for the analysis of sublinear doseresponse relationships, exemplified for mutagenicity data. Mutat Res 678:118–122. Lynch AM, Giddings A, Custer L, Gleason C, Henwood A, Aylott M, Kenny J. 2011. International Pig-a gene mutation assay trial (stage III): results with N-methyl-N-nitrosourea. Environ Mol Mutagen 52:699–710.

Mattison JD, Penrose LB, Burlinson B. 1997. Preliminary results of ethylnitrosourea, isopropyl methanesulphonate and methyl methanesulphonate activity in the testis and epididymal spermatozoa of Muta Mice. Mutat Res 388:123–127. Miura D, Dobrovolsky VN, Kimoto T, Kasahara Y, Heflich RH. 2009. Accumulation and persistence of Pig-A mutant peripheral red blood cells following treatment of rats with single and split doses of N-ethyl-N-nitrosourea. Mutat Res 677:86–92. Miura D, Dobrovolsky VN, Kasahara Y, Katsuura Y, Heflich RH. 2008. Development of an in vivo gene mutation assay using the endogenous Pig-A gene: I. Flow cytometric Detection of CD59-Negative Peripheral Red Blood Cells and CD48-Negative Spleen T-Cells from the rat. Environ Mol Mutagen 49:614–621. Morimoto K, Dolan ME, Scicchitano D, Pegg AE. 1985. Repair of O6propylguanine and O6-butylguanine in DNA by O6-alkylguanineDNA alkyltransferases from rat liver and E. coli. Carcinogenesis 6:1027–1031. Muller L, Gocke E. 2009. Considerations regarding a permitted daily exposure calculation for ethyl methanesulfonate. Toxicol Lett 190:330–332. Pegg AE, Dolan ME, Scicchitano D, Morimoto K. 1985. Studies of the repair of O6-alkylguanine and O4-alkylthymine in DNA by alkyltransferases from mammalian cells and bacteria. Environ Health Perspect 62:109–114. Peruzzi B, Araten DJ, Notaro R, Luzzatto L. 2010. The use of PIG-A as a sentinel gene for the study of the somatic mutation rate and of mutagenic agents in vivo. Mutat Res 705:3–10. Phonethepswath S, Franklin D, Torous DK, Bryce SM, Bemis JC, Raja S, Avlasevich S, Weller P, Hyrien O, Palis J, Macgregor JT, Dertinger SD. 2010. Pig-a mutation: kinetics in rat erythrocytes following exposure to five prototypical mutagens. Toxicol Sci 114:59–70. Provost GS, Rogers BJ, Dycaico MJ, Carr G. 1997. Evaluation of the transgenic Lambda/LacI mouse model as a short-term predictor of heritable risk. Mutat Res 388:129–136. Sega GA, Owens JG, Cumming RB. 1976. Studies on DNA repair in early spermatid stages of male mice after in vivo treatment with methyl-, ethyl-, propyl-, and isopropyl methanesulfonate. Mutat Res 36:193–212. Shelby MD, Tindall KR. 1997. Mammalian germ cell mutagenicity of ENU, IPMS and MMS, chemicals selected for a transgenic mouse collaborative study. Mutat Res 388:99–109. Singer B. 1985. In vivo formation and persistence of modified nucleosides resulting from alkylating agents. Environ Health Perspect 62:41–48. Swain C, Scott CB. 1953. Quantitative correlation of relative rates. Comparison of hydroxide ion with other nucleophilic reagents toward alkyl halides, esters, epoxides and acyl halides. J Am Chem Soc 75:141–147. Tinwell H, Lefevre P, Williams CV, Ashby J. 1997. The activity of ENU, iPMS and MMS in male mouse germ cells using the Muta Mouse positive selection transgenic mutation assay. Mutat Res 388:179–185. Torous DK, Phonethepswath S, Avlasevich SL, Mereness J, Bryce SM, Bemis JC, Weller P, Bell S, Gleason C, Custer LL, MacGregor JT, Dertinger SD. 2012. In vivo flow cytometric Pig-a and micronucleus assays: highly sensitive discrimination of the carcinogen/ noncarcinogen pair benzo(a)pyrene and pyrene using acute and repeated-dose designs. Environ Mol Mutagen 53:420–428. Vilarino-Guell C, Smith AG, Dubrova YE. 2003. Germline mutation induction at mouse repeat DNA loci by chemical mutagens. Mutat Res 526:63–73. Vogel EW, Ashby J. 1994. Structure-activity relationships: Experimental approaches. Scope 52:231–254. Vogel EW, Barbin A, Nivard MJ, Bartsch H. 1990. Nucleophilic selectivity of alkylating agents and their hypermutability in Drosophila as predictors of carcinogenic potency in rodents. Carcinogenesis 11:2211–2217.