Bone marrow micronucleus frequencies in the rat after oral administration of cyclophosphamide, hexamethylphosphoramide or gemifloxacin for 2 and 28 days

Abstract

Cyclophosphamide (CPA), hexamethylphosphoramide (HMPA) and gemifloxacin (GF) were administered orally to Han Wistar rats for 2 and 28 days up to their maximum tolerated doses. For CPA, the sensitivity, as measured by the increases at the lowest active dose, was greater after 28 days. However, published data show that blood levels at 28 days are the same as at 2 days. The sensitivity to HMPA was also greater at 28 days but toxicokinetic analysis showed a reduction in blood levels. Although apparently anomalous, this probably reflected increased metabolism of HMPA to its genotoxic metabolite, formaldehyde. In contrast, GF gave a clear increase in micronucleus frequency after 2 doses but not after 28 and the maximum dose tolerated for 28 days gave blood levels 80% greater than the same dose after 2 days. It is apparent that for these three compounds there are differences between responses after dosing for 2 and 28 days that cannot be ascribed simply to differences in blood levels. It is important to note that although the responses were quantitatively different, the two established rodent carcinogens gave positive results at both sampling times. Because rodent oncogenicity data are not available for GF, it is not known whether the positive or negative result in the rat bone marrow micronucleus test is predictive. The purpose of this work was not to support the use of either acute or sub-chronic dosing for the rat micronucleus assay but to point out that sampling times can give both qualitatively and quantitatively different results.

---

Introduction

The most recent ICH guideline for genotoxicity testing¹ allows integration of the bone marrow micronucleus test into repeat-dose toxicology studies but this idea is not new and has been widely discussed over the last 20 years.^2–5 There are advantages to integration such as reduction in animal usage, and obtaining toxicokinetic and toxicity information from animals in addition to genotoxicity information.^3,4 However, there have been few studies comparing the activity of genotoxins in the micronucleus test in the rat after acute and sub-chronic dosing. In a collaborative study organised by the Environmental Mutagen Society of Japan, results from 28 day studies were compared with published results for 14 chemicals using 1 or 2 doses.⁶ One compound, phenacetin, was reported to be inactive after acute dosing but gave increased micronucleus frequencies after 28 days. In contrast, three rodent carcinogens, dimethylhydrazine, mitomycin C and monocrotaline, which were detected after acute dosing, gave no increases in micronucleus frequencies after 28 days. However, it was noted that the maximum doses tolerated were generally lower for 28 days and blood levels were not measured. A more recent international collaborative trial investigated whether the comet assay when integrated into 2- or 4- week rat studies could complement the micronucleus assay in the peripheral blood or bone marrow.⁷ Only 7 compounds were tested for micronucleus induction in the bone marrow, both acutely and after 2 or 4 weeks, and two of these, pyrimethamine and gemifloxacin mesylate, were positive only after acute administration, but again, blood levels were not measured. In order to investigate further possible differences in bone marrow micronucleus frequencies between acute and sub-chronic dosing, studies were performed in which cyclophosphamide (CPA), hexamethylphosphoramide (HMPA) and gemifloxacin mesylate (GF) were orally administered to rats for periods up to 28 days. Toxicokinetic analyses were also performed for HMPA and GF.

CPA has been in widespread clinical use for over 50 years as a chemotherapeutic agent and immunosuppressant. It is a known human carcinogen and an increased incidence of chromosome aberrations has been seen in lymphocytes from patients treated for both malignant and non-malignant diseases.^8,9 The genotoxicity of CPA has been reviewed by Anderson et al. and it is active in numerous in vivo and in vitro systems and is also teratogenic; consequently it has been used routinely as a positive control in various genotoxicity tests.¹⁰ It requires metabolic activation via CYP-mediated hydroxylation to 4-hydroxycyclophosphamide and subsequent breakdown to form two cytotoxic metabolites, phosphoramide mustard and acrolein in both humans and rats.¹⁰ The systemic clearance of CPA is more rapid in the rat than in man with plasma half-lives of 1.1 and 6–12 hours, respectively, but the volume of distribution is comparable.¹¹ There appear to be several DNA adducts formed by reactions between phosphoramide mustard and guanosine residues including a cross-linked dimer, two monoalkylated adducts and a phosphodiester adduct.

HMPA is an industrial solvent for processing aromatic polyamide fibres and has unique properties as a solvent for both water and organic materials. It is a potent nasal carcinogen in rats by inhalation,¹² the only species examined, but has not been adequately tested by oral administration.¹³ The available genotoxicity data have been reviewed and possible metabolic pathways have been suggested.¹⁴ HMPA is a weak bacterial mutagen but is only detected using a suspension exposure method up to 40 mg mL⁻¹ in the presence of S9 with S. typhimurium TA100 showing the clearest response; it is not detected by TA1535, TA1537, TA97, TA98 or TA100 in standard plate incorporation or liquid pre-incubation tests up to 15 000 μg per plate.¹⁴ It is weakly active in in vitro mammalian cell assays with metabolic activation and is clearly positive in several in vivo tests including the mouse bone-marrow micronucleus¹⁵ and dominant lethal tests.¹⁶ It is metabolised by successive demethylations giving unstable intermediates each of which decompose with the release of formaldehyde and it is considered that the genotoxic activity of HMPA in vitro and in vivo is due to formaldehyde generation. In vitro studies with microsomes from rat olfactory tissues and measuring DNA adducts in rat nasal epithelial cells indicate that DNA–protein cross-links formed by formaldehyde release may be important in the nasal toxicity of HMPA.¹⁷

GF is a fluoroquinolone antibacterial agent which targets bacterial topoisomerase IV. It is a racemic compound with its individual enantiomers having equivalent antibacterial activity. Bacterial topoisomerase IV is a type II topoisomerase enzyme, hence it would be considered that GF would have cross reactivity with mammalian topoisomerase II. Plasma pharmacokinetic profiles for the enantiomers are virtually identical in rats and dogs after oral and intravenous administration. GF also has a high volume of distribution and rapid elimination, with no accumulation in any tissues or organ and limited metabolism in both rat and dog.¹⁸ The principal metabolites of gemifloxacin were the E-isomer and the acyl-glucuronide of gemifloxacin in both rat and dog and the N-acetyl gemifloxacin in the rat.¹⁸ GF as Factive® (gemifloxacin mesylate) has been examined in a comprehensive battery of in vitro and in vivo tests for genotoxicity.¹⁹ It is not mutagenic in S. typhimurium TA1535, TA1537, TA98 or TA100 but its bactericidal activity limited the maximum testable levels to <1 μg per plate. However, in common with other fluoroquinolone antibiotics, GF is mutagenic for TA102 (C. Mee, AstraZeneca R&D; personal communication) and also induces mutations in the mouse lymphoma Tk assay and increases chromosome aberrations in human peripheral blood lymphocytes and Chinese hamster ovary cells.¹⁹ Unscheduled DNA synthesis was not seen in hepatocytes removed from rats 2–4 or 12–14 hours after single oral doses of 1000 or 2000 mg kg⁻¹, and no evidence of increased micronucleus frequency was seen in the bone marrow of mice given 2 daily doses of 50 mg kg⁻¹ by intraperitoneal injection.¹⁹ In contrast to the lack of activity shown in these in vivo assays, GF was found to increase bone marrow micronucleus frequency in Sprague Dawley rats.¹⁹ After two oral doses, increases were seen at 800 and 1600 mg kg⁻¹ with a no-effect level (NOEL) of 400 mg kg⁻¹; after two intravenous (iv) doses, increases were seen after 40 and 80 mg kg⁻¹ with a NOEL of 20 mg kg⁻¹. It was noted that bone marrow toxicity as indicated by reduced immature erythrocyte numbers was seen at the doses with increased micronucleus frequencies in both studies. It was also noted that the relationship between plasma levels of GF and increased micronucleus frequency was not simple. The NOEL in the iv study, 20 mg kg⁻¹, gave a C_max approximately 3-fold higher than an oral dose of 800 mg kg⁻¹, which increased micronucleus frequency. The lowest iv dose increasing micronucleus frequency resulted in an AUC_(0–24) comparable with that at the oral NOEL, 400 mg kg⁻¹.¹⁹ Standard rodent oncogenicity tests have not been performed with GF but ciprofloxacin, which has a similar pattern of activity in tests for genotoxicity, has been reported to be negative in both rats and mice.²⁰

Materials and methods

All chemicals and reagents were purchased from Sigma (Dorset, UK) unless specified.

Animal husbandry

Male Han Wistar rats, substrain HsdHan (WIST) obtained from Harlan UK were 9–10 weeks old at the start of dosing and were housed three or four per cage. Environmental controls were set to maintain conditions of 19–23 °C and 40–70% relative humidity, with a 12-hour light/dark cycle. All animals were treated in accordance with approved UK Home Office licence requirements.

Treatment with CPA

CPA is the routine positive control used in this laboratory and 20 mg kg⁻¹ was known to be well tolerated. Preliminary work showed that 5 mg kg⁻¹ would be tolerated for 28 days. CPA (cyclophosphamide monohydrate, Sigma Aldrich, ≥98%) was formulated in sterile water and groups of 5 rats were given oral doses of 0.5, 1.7 or 5 mg kg⁻¹ daily for 28 days. A group of 3 rats was given a single dose of 20 mg kg⁻¹ on day 28 to serve as a positive control. For comparison with the doses used for 28 days, groups of 5 rats were given two oral doses of 0.5, 1.7, 5, 10 or 20 mg kg⁻¹. From the historical control data for CPA in this laboratory, it was considered that reducing the numbers of animals in each group to 5 from 7, the number used routinely in this laboratory would still provide adequate statistical power.²¹

Treatment with HMPA

Following preliminary work to establish doses that should be tolerated, HMPA (hexamethylphosphoramide, Sigma Aldrich, 99%) was formulated in sterile water and groups of 7 rats were given oral doses of 50, 100, or 200 mg kg⁻¹ daily for 2, 14 and 28 days. The highest dose group was initially given 200 mg kg⁻¹ but, due to body weight loss, this was reduced to 150 mg kg⁻¹ after 15 doses, then after 19 doses the rats were maintained off dose until day 29. Satellite groups of 3 rats were given the same doses for toxicokinetic analyses. A group of 3 rats was given a single dose of CPA at 20 mg kg⁻¹ on day 28 to serve as a positive control.

Treatment with GF

Following preliminary work to establish doses that should be tolerated, GF (gemifloxacin mesylate, Yes Pharma Ltd, 98%) was formulated in sterile water and groups of 7 rats were given oral doses of 160, 600 or 1200 mg kg⁻¹ day⁻¹ for 2 days. From the results obtained, it was decided that 1200 mg kg⁻¹ would not be tolerated for 28 days. Subsequently groups of 7 rats were given 160, 300 or 600 mg kg⁻¹ day⁻¹ for 28 days. In both studies, a group of 3 rats was given a single dose of CPA at 20 mg kg⁻¹ on the last day to serve as a positive control. In the second study, satellite groups of 3 rats were given doses of 160, 300, 600 or 1200 mg kg⁻¹ for toxicokinetic analyses; the group given 1200 mg kg⁻¹ was terminated after collection of the samples taken after 2 doses.

Bone marrow micronucleus assay

In all treatment schedules, the rats were killed for bone marrow analysis 24 hour following the final dose. The femur or humerus was removed from each rat, the ends clipped to expose the bone marrow and smears prepared and stained essentially as described by Tinwell and Ashby.²² The femoral cells were flushed out with foetal bovine serum containing 25 mmol L⁻¹ ethylenediaminetetraacetic acid (EDTA), filtered through a 150 μ bolting cloth and centrifuged at 200g for 5 minutes. The supernatant was removed, the cell pellet carefully mixed by repeated aspiration with a Pasteur pipette and a small drop spread onto a clean microscope slide, air dried then fixed with methanol. After drying again, slides were dipped in fresh phosphate buffer pH 6.4 and stained with 12.5% acridine orange then rinsed in two changes of fresh phosphate buffer and air-dried. Prior to analysis, slides were coded, wet mounted in phosphate buffer and examined by fluorescence microscopy at ×400 magnification using an Olympus Bx51 fluorescent microscope with a triple band pass filter. The cells were identified by the following staining properties of acridine orange; mature erythrocytes (E) stain dull khaki-green, immature erythrocytes (IE) fluoresce bright red/orange and micronuclei (MIE) fluoresce bright yellow/green.

For all animals, 2000 IE's were scored and the number of MIE recorded. The IE to E ratio was also determined in 1000 cells and is presented as the percentage IE.

Statistics

The experiments conducted in this paper have all been designed and carried out under the AstraZeneca principles of Good Statistical Practice including the use of parallel group designs, vehicle control groups included and used in all comparisons and trend tests, positive controls groups used to monitor the reproducibility of animal models over time compared to the vehicle control, animals balanced with respect to their weight across the groups, dosing performed in ascending dose order and slides scored blind to the animal or group from which they originated.

Prior to analysis, the micronuclei counts were subjected to an average square root transformation and transformed counts were analysed in all cases.²³ Analysis of variance, with appropriate linear contrasts, was used to test for dose-related trends in micronuclei and was determined in the following way; where a significant increase was observed across all test doses then the trend test was reapplied in a cascade – first excluding the top dose, then the intermediate dose(s) and finally with only the control and the lowest dose. Pair-wise comparisons between control and test groups were performed and were one-sided at the 1% level; for HMPA, comparisons were made with concurrent control group and for CPA, comparisons were made with pre-dose values for each group. Positive controls were compared to vehicle controls using Fisher's exact test and was one-sided at the 5% level.

Toxicokinetic analysis

For HMPA, blood for toxicokinetic evaluation was collected from the tail vein of rats before and 1, 3, 6 and 24 hours following dosing for 2, 14 and 28 days.

For GF, blood was collected at 1, 2, 3, 6 and 24 hours following dosing for 2 and 28 days.

For both compounds, approximately 0.2 mL samples were taken into EDTA tubes, cooled on ice, centrifuged and plasma removed for analysis. Plasma profiles of both HMPA and GF were determined using high performance liquid chromatography (HPLC) with tandem mass spectrometry detection. C_max and AUC_(0–6) were recorded.

Results

As both femurs are currently used for pathology endpoints in the 28-day studies, the first step was to establish a method to sample bone marrow from an alternative source i.e. the humerus. We found that the micronucleus frequency was the same in the humerus and femur bone marrow samples, therefore the humerus could readily be used as the alternative bone marrow source allowing integration into 28-day studies. In addition, a control database from 28-day studies was established. The historical control values at the time of these studies were, for the femur in male Han Wistar rats in acute assays, 2.1 MIE per 2000 IE and, in the humerus from 28-day studies, 2.0 MIE per 2000 IE.

CPA micronucleus frequency

Following 2 doses, highly statistically significant dose related linear increases in MIE were observed at 5, 10 and 20 mg kg⁻¹ representing 2.7, 8.5 and 12 fold increases over the control, respectively (Table 1). No increases in MIE frequency were detected at 0.5 and 1.7 mg kg⁻¹. After 28 days, 0.5 and 1.7 mg kg⁻¹ showed statistically significant increases over the concurrent control. Furthermore, the increase in MIE at 5 mg kg⁻¹ after 28 days was approximately 14-fold over control, as compared to only 3 fold after 2 days. There was little evidence of any reduction in the IE : E ratios in the groups given CPA for either 2 or 28 days, so there was no indication that the differences in MIE frequency were influenced by excessive bone marrow cytotoxicity. Toxicokinetic analysis was not performed in this study but, as discussed later, published data indicate that CPA is rapidly cleared and blood levels in rats do not increase with repeated dosing.

Table 1 Bone marrow micronucleus frequency after oral administration of cyclophosphamide for 2 or 28 days

Treatment	Mean number of MIE in 2000 IE	% IE
a MIE scored 24 hours after the final dose. b A single oral dose of 20 mg kg^−1 cyclophosphamide given 24 hours before sampling on day 2. **: Statistically significant p < 0.01, ***: statistically significant p < 0.001.
2 doses
Control (water)	4.0	75.3
0.5 mg kg^−1	2.0	56.8
1.7 mg kg^−1	2.7	55.8
5 mg kg^−1	10.7***	65.9
10 mg kg^−1	35.3***	72.3
20 mg kg^−1	48.3***	64.7
28 doses
Control (water)	2.2	68.4
0.5 mg kg^−1	4.8**	61.8
1.7 mg kg^−1	6.6***	66.9
5 mg kg^−1	30.4***	63.2
Positive control^b	56.7***	58.4

---

HMPA micronucleus frequency and toxicokinetics

No increase in MIE was observed at the lowest dose, 50 mg kg⁻¹, after 2 or 14 doses but a statistically significant increase was seen after 28 days. At 100 mg kg⁻¹ day⁻¹, no increase in MIE was seen after 2 doses, but there were statistically significant increases after both 14 and 28 doses of 3- and 6-fold, respectively (Table 2). The group given 200 mg kg⁻¹ day⁻¹ showed statistically significant increases in MIE after 2 and 14 doses of 2- and 6-fold, respectively. The dose was reduced to 150 mg kg⁻¹ day⁻¹ after 15 doses due to bodyweight loss then discontinued after 19 doses and the group were left to recover and sampled at the scheduled end of the study. No significant increase in MIE over the control was observed which is consistent with the MIE moving out of the bone marrow compartment and demonstrates no lasting induction of MIE from erythroblasts exposed many days before. There were no changes in the IE : E ratio in any group except a slight reduction at 200 mg kg⁻¹ day⁻¹ after 14 doses.

Table 2 Bone marrow micronucleus frequency and toxicokinetics after oral administration of hexamethylphosphoramide for 2, 14 or 28 days

Treatment	Mean number of MIE in 2000IE	% IE	C max (SD) μg mL^−1		AUC (SD) μg h mL^−1
C max, maximum plasma concentration; AUC, area under the curve 0–6 hours after dosing; SD, standard deviation; blq, below level of quantification; nd, not done.a MIE scored 24 hours after the final dose; toxicokinetic samples taken before and 1, 2, 3, 6 and 24 hours after dosing.b Rats given 16 doses at 200 mg kg^−1, 3 more at 150 mg kg^−1 then maintained off dose until killed at the same time as the groups given 28 doses.c A single oral dose of 20 mg kg^−1 cyclophosphamide given 24 hours before sampling on day 2. **: Statistically significant p < 0.01, ***: statistically significant p < 0.001.
2 doses
Control (water)	2.7	66.2	blq		blq
50 mg kg^−1	2.1	63.9	24.3	(2.34)	50.7	(1.88)
100 mg kg^−1	3.4	66.1	64.2	(13.39)	146.7	(38.62)
200 mg kg^−1	5.6***	62.8	145.3	(14.57)	327.8	(51.05)
14 doses
Control (water)	2.6	62.6	blq		blq
50 mg kg^−1	2.5	65.4	25.3	(2.46)	47.5	(5.14)
100 mg kg^−1	7.5***	61.8	51.6	(10.45)	96.7	(24.44)
200 mg kg^−1	16.2***	40.0	114.1	(28.37)	220.3	(78.10)
28 doses
Control (water)	1.1	72.2	blq		blq
50 mg kg^−1	3.7**	77.3	18.9	(2.63)	35.5	(4.45)
100 mg kg^−1	13.0***	69.0	29.3	(2.46)	60.8	(8.00)
200 mg kg^−1	4.9^b	71.4	blq		blq
Positive control^c	66.3***	69.8	nd		nd

---

The toxicokinetic data for HMPA (Table 2) show that the C_max and AUC_(0–6) values were greatest on Day 1 and decreased by Day 14 and were lowest on Day 28 at all dose levels.

GF micronucleus frequency and toxicokinetics

After 2 doses, GF gave highly statistically significant increases in MIE at 600 and 1200 mg kg⁻¹ of 3.5- and 10.5-fold, respectively. Following 28 days administration, no increase in MIE was observed at any dose (Table 3). It was surprising that 600 mg kg⁻¹ gave negative results because the same dose had been shown to induce MIE after 2 days. Consequently, a repeat experiment was performed which confirmed that no induction of MIE was seen at this dose (data not shown). There was no evidence of any reduction in the IE : E ratios in any treated group after 2 or 28 doses.

Table 3 Bone marrow micronucleus frequency and toxicokinetics after oral administration of gemifloxacin for 2 or 28 days

Treatment	Mean number of MIE in 2000IE	% IE	C max (SD) μg mL^−1		AUC (SD) μg h mL^−1
C max, maximum plasma concentration; AUC, area under the curve 0–6 hours after dosing; SD, standard deviation; blq, below level of quantification; nd, not done.a MIE scored 24 hours after the final dose; toxicokinetic samples taken 1, 2, 3, 6 and 24 hours after dosing.b A single oral dose of 20 mg kg^−1 cyclophosphamide given 24 hours before sampling. ***: Statistically significant p < 0.001.
2 doses
Control (water)	2.4	76.0	blq		blq
160 mg kg^−1	3.9	67.9	2.9	(0.80)	12.4	(4.37)
300 mg kg^−1	nd	nd	5.1	(1.76)	28.1	(13.95)
600 mg kg^−1	8.4***	73.9	7.3	(1.83)	60.2	(15.33)
1200 mg kg^−1	25.1***	63.1	25.3	(14.6)	274.0	(267.3)
Positive control^b	57.0***	63.9
28 doses
Control (water)	2.3	77.8	blq		blq
160 mg kg^−1	3.2	72.4	2.4	(0.46)	9.0	(1.47)
300 mg kg^−1	3.5	72.7	8.1	(1.73)	33.5	(8.74)
600 mg kg^−1	4.5	73.4	13.4	(6.25)	107.0	(8.60)
Positive control^b	34.7***	54.2

---

Toxicokinetic data showed that C_max was higher after 28 doses than 2 at both 300 and 600 mg kg⁻¹.

Discussion

It is generally considered that increases in bone marrow micronucleus frequency in immature erythrocytes reflect genetic damage induced only in the 24 to 48 hours prior to sampling, so it would be expected that the frequencies induced by any particular genotoxin should be the same at any time of sampling provided that similar blood levels are achieved. However, results in the present study indicate the relationship may be more complex.

The lowest dose of CPA which gave significant increases in MIE after 2 days was 5 mg kg⁻¹ but significant increases were seen after 28 doses at 0.5 mg kg⁻¹. Further, at 5 mg kg⁻¹ the increase was much greater after 28 doses and was directly comparable with that after 2 doses at 10 mg kg⁻¹. In the absence of toxicokinetic data, it cannot be concluded with certainty that this increase is due to accumulated damage rather than increased blood levels, but published data indicate that steady state would have been achieved after a single dose because CPA and its metabolites have been shown to be cleared quickly in the rat. After an intraperitoneal injection of 50 mg kg⁻¹, the half-lives of CPA and 4-hydroxy-CPA were 55 and 30 min, respectively.²⁴ After intravenous injection of 25–100 mg kg⁻¹, the half-lives of CPA, 4-hydroxy-CPA and phosphoramide mustard were 29, 14 and 14 min, respectively.²⁵ After single doses of 50 mg kg⁻¹, the half-life after oral administration was slightly longer than either intravenous or intraperitoneal dosing with values of 58, 41 and 34 min, respectively.²⁵ It is not clear why CPA should result in greater increases in numbers of MIE after 28 days, but it is possible that its DNA cross-linking activity delays the cell cycle so peak numbers are not achieved 24 hours after the second of two doses, as used in this study. It is possible that the MIE frequency reaches a maximum after a few doses so it would be interesting to follow the frequency between 2 and 28 doses. If cell cycle delay is occurring, MIE frequency might be expected to plateau after a few more doses but, if the frequency progressively increases with duration of dosing, this would imply that DNA damage may be induced earlier in the erythrocyte generation process than generally believed. It is interesting to note that toxicity to the bone marrow as indicated by IE : E ratio was not seen in any treated group which would imply cell cycle delay is not an important factor.

Similar to CPA, the MIE frequencies induced by HMPA increase with duration of treatment, and the lowest doses inducing significant increases were 200, 100 and 50 mg kg⁻¹ after 2, 14 or 28 doses, respectively. However, the blood levels of HMPA in terms of both C_max and AUC_(0–6) progressively decreased throughout the study and it is likely the more rapid clearance is due to increased rate of metabolism. This would lead to greater amounts of formaldehyde being generated, which would explain the increase in MIE frequency. The group initially given 200 mg kg⁻¹ was maintained without further dosing for the last 10 days of the study. The MIE frequency was not statistically significantly different from the control showing that there was no lasting DNA damage in the bone marrow.

In contrast to CPA and HMPA, GF was clearly positive after 2 doses but did not induce MIE after 28 doses. Blood levels in terms of C_max and AUC_(0–24) were increased after 28 doses at 600 mg kg⁻¹ and both were about 80% higher than after 2 doses in contrast to the work of Ramji et al.¹⁸ that did not see any accumulation in the rat. Further, there was apparent toxicity to the bone marrow indicated by reduced IE : E ratio. Given these observations, there appears to be no obvious reason why GF should give positive results after 2 doses but not after 28. However, factors affecting MIE induction in the rat appear to be complex and, for Factive®, it was noted that the relationship between plasma levels of GF and increased micronucleus frequency was not simple.⁹ The NOEL in the iv study, 20 mg kg⁻¹, gave a C_max approximately 3-fold higher than an oral dose of 800 mg kg⁻¹, which increased micronucleus frequency. The lowest iv dose increasing micronucleus frequency resulted in an AUC_(0–24) comparable with that at the oral NOEL, 400 mg kg⁻¹. These apparently anomalous results might have been related to the effects of toxicity to the bone marrow but, unlike the studies with Factive® there was no evidence of toxicity to the bone marrow in the present study.

In conclusion, for the three compounds tested here, there are differences between responses after dosing for 2 and 28 days that cannot be ascribed simply to differences in blood levels. It is important to note that, although the responses were quantitatively different, the two established rodent carcinogens gave positive results at both sampling times. Because rodent oncogenicity data are not available for GF, it is not known whether the positive or negative result in the rat bone marrow micronucleus test is predictive, although it should be noted that ciprofloxacin which has a similar pattern of activity in tests for genotoxicity is not carcinogenic. The purpose of this work was not to support acute or sub-chronic dosing for the micronucleus assay but to point out that sampling times can give both qualitatively and quantitatively different results.

Acknowledgements

The authors would like to thank Sean Evans for his excellent technical assistance.

References

1. ICH (2011) ICH topic S2(R1), Genotoxicity: guidance on genotoxicity testing and data interpretation for pharmaceuticals intended for human use. Recommended for adoption at step 4 of the ICH process. In International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Geneva, June 11, 2009. November 9, 2011. http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Safety/S2_R1/Step4/S2R1_Step4.pdf.
2. J. T. MacGregor, J. D. Tucker, D. A. Eastmond and A. J. Wyrobek, Integration of cytogenetic assays with toxicology studies, Environ. Mol. Mutagen., 1995, 25, 328–337.
3. M. Hayashi, J. T. MacGregor, D. G. Gatehouse, I.-D. Adler, D. H. Blakey, S. D. Dertinger, G. Krishna, T. Morita, A. Russo and S. Sutou, In vivo erythrocyte micronucleus assay. II. Some aspects of protocol design including repeated treatments, integration with toxicity testing and automated scoring, Environ. Mol. Mutagen., 2000, 35, 234–252.
4. EP, EC, Regulation (EC) No 1907/2006 of the European Parliament and of the council of 18 December concerning the Registration, Evaluation, Authorisation and restriction of Chemicals (REACH), establishing a European Chemicals Agency. Amending Directive 1999/45/EC and repealing Council Regulation (EEC) No 793/93 and Commission Regulation (EC) No 1488/94 as well as Council Directive 76/769/EEC and Commission Directives 91/155/EEC, 93/67/EEC, 93/105/EEC and 2000/21/EC. Official Journal EU Law, 2006b, 396(1), 849. Last accessed 5 March 2013.
5. S. Pfuhler, D. Kirkland, P. Kasper, M. Hayashi, P. Vanparys, P. Carmichael, S. Dertinger, D. Eastmond, A. Elhajouji, C. Krul, A. Rothfuss, G. Schoening, A. Smith, G. Speit, C. Thomas, J. van Benthem and R. Corvi, Reduction of use of animals in regulatory genotoxicity testing: identification and implementation opportunities – Report from an ECVAM workshop, Mutat. Res., 2009, 680, 31–42.
6. S. Hamada, S. Sutou, T. Morita, A. Wakata, S. Asanami, S. Hosoya, S. Ozawa, K. Kondo, M. Nakajima, H. Shimada, K. Osawa, Y. Kondo, N. Asano, S. Sato, H. Tamura, N. Yajima, R. Marshall, C. Moore, D. H. Blakey, L. M. Schechtman, J. L. Weaver, D. K. Torous, R. Proudlock, S. Ito, C. Namiki and M. Hayashi, Evaluation of the rodent micronucleus assay by a 28 day treatment protocol: summary of the 13^th Collaborative Study Group for the Micronucleus Test (CSGMT)/Environmental Mutagen Society of Japan (JEMS) – Mammalian Mutagenicity Study Group (MMS), Environ. Mol. Mutagen., 2001, 37, 93–110.
7. A. Rothfuss, M. O'Donovan, M. De Boeck, D. Brault, A. Czich, L. Custer, S. Hamada, U. Plapper-Helbig, M. Hayashi, J. Howe, A. Kraynak, B.-J. van der Leede, M. Nakajima, C. Priestley, V. Thybaud, K. Saigo, S. Sawant, J. Shi, R. Storer, M. Struwe, E. Vock and S. Galloway, Collaborative study on fifteen compounds in the rat-liver comet assay integrated into 2- and 4-week repeat-dose studies, Mutat. Res., 2010, 702, 40–69.
8. International Agency for Research on Cancer, IARC Monographs on the evaluation of carcinogenic risks to humans. Some antineoplastic and immunosuppressive agents, 1981, vol. 26, pp. 165–202.
9. International Agency for Research on Cancer, IARC Monographs on the evaluation of carcinogenic risks to humans, 1987, (Supplement 7), pp. 182–183.
10. D. Anderson, J. B. Bishop, R. C. Garner, P. Ostrosky-Wegman and P. B. Selby, Cyclophosphamide: review of its mutagenicity for an assessment of potential germ cell risks, Mutat. Res., 1985, 330, 115–181.
11. M. R. Z. Talha, H. J. Rodgers and J. R. Trounce, Distribution and pharmacokinetics of cyclophosphamide in the rat, Br. J. Cancer, 1980, 41, 140–143.
12. International Agency for Research on Cancer, IARC Monographs on the evaluation of carcinogenic risks to humans. Re-evaluation of Some Organic Chemicals, Hydrazine and Hydrogen Peroxide, 1999, vol. 71, pp. 1465–1481.
13. K. P. Lee and H. J. Trochimowicz, Induction of nasal tumours in rats exposed to hexamethylphosphoramide by inhalation, J. Natl. Cancer Inst., 1982, 68, 157–171.
14. A. M. Sarrif, D. F. Krahn, S. M. Donovan and R. M. O'Neil, Evaluation of hexamethylphosphoramide for gene mutations in Salmonella typhimurium using plate incorporation, preincubation and suspension assays, Mutat. Res., 1997, 380, 167–177.
15. J. Ashby, C. R. Richardson, P. A. Lefevre, R. D. Callander and J. A. Styles, Chloracetamide-N-metholol: an example of an in vitro and in vivo clastogen which is non-mutagenic to Salmonella, Mutat. Res., 1985, 156, 19–32.
16. R. J. Srám, V. Benes and Z. Zudová, Induction of dominant lethals in mice by TEPA and HEMPA, Folia Biol. (Praha), 1970, 16, 407–415.
17. J. R. Kuykendall, B. A. Trela and M. S. Bogdanffy, DNA–protein crosslink formation in rat nasal epithelial cells by hexamethylphosphoramide and its correlation with formaldehyde production, Mutat. Res., 1995, 343, 209–218.
18. J. V. Ramji, N. E. Austin, G. W. Boyle, M. H. Chalker, G. Duncan, A. J. Fairless, F. J. Hollis, D. F. McDonnell, T. T. Musick and P. C. Shardlow, The disposition of gemifloxacin, a new fluoroquinolone antibiotic, in rats and dogs, Drug Metab. Dispos., 2001, 29, 435–442.
19. Food and Drug Administration. Center for Drug Evaluation and Research, Factive® (Gemifloxacin). Pharmacology/Toxicology Review and Evaluation, 2003 NDA Number 21–158.
20. R. D. Snyder and J. W. Green, A review of the genotoxicity of marketed pharmaceuticals, Mutat. Res., 2001, 488, 151–169.
21. J. Hayes, A. T. Doherty, D. J. Adkins, K. Oldman and M. R. O'Donovan, The rat bone marrow micronucleus test – study design and statistical power, Mutagenesis, 2009, 24, 419–424.
22. H. Tinwell and J. Ashby, Comparison of acridine orange and giemsa stains in several mouse bone marrow micronucleus assays—including a triple dose study, Mutagenesis, 1989, 4, 476–481.
23. D. P. Lovell, R. Albanese, G. Clare, M. Richold, J. R. K. Savage, D. Anderson, G. E. Amphlett, R. Ferguson and D. G. Papworth, Statistical analysis of in vivo cytogenetic assays, in Statistical Evaluation of Mutagenicity Data, ed. D. J. Kirkland, Cambridge University Press (pubs), 1989, pp. 184–232.
24. J. F. Powers and N. E. Sladek, Cytotoxic activity relative to 4-hydroxycyclophosphamide and phosphoramide mustard concentrations in the plasma of cyclophosphamide-treated rats, Cancer Res., 1983, 43, 1101–1106.
25. N. E. Sladek, J. F. Powers and G. M. Grage, Half-life of oxazaphosphorines in biological fluids, Drug Metabol. Dispos., 1984, 12, 553–559.

---