- Email: [email protected]
− mice
Mutation Research 570 (2005) 227–235
Micronucleated erythrocyte frequency in control and azidothymidine-treated Tk+/+, Tk+/− and Tk−/− mice Vasily N. Dobrovolskya,∗ , Lynda J. McGarritya , Linda S. VonTungelnb , Roberta A. Mittelstaedta , Suzanne M. Morrisa , Frederick A. Belandb , Robert H. Heflicha a
Division of Genetic and Reproductive Toxicology, U.S. Food and Drug Administration/National Center for Toxicological Research, HFT-120, 3900 NCTR Rd., Jefferson, AR 72079, USA b Division of Biochemical Toxicology, U.S. Food and Drug Administration/National Center for Toxicological Research, HFT-110, 3900 NCTR Rd., Jefferson, AR 72079, USA Received 17 August 2004; received in revised form 3 November 2004; accepted 20 November 2004 Available online 21 January 2005
Abstract The first step in the activation of the anti-retroviral nucleoside analogue azidothymidine (AZT) involves its conversion to a 5 -monophosphate. In this study, we have evaluated the role of cytosolic thymidine kinase (Tk), the major enzyme involved in phosphorylating thymidine and its analogues, in the nuclear DNA damage produced by AZT in neonatal mice. Tk+/+ , Tk+/− and Tk−/− mice were treated intraperitoneally with 200 mg/kg/day of AZT on postnatal days 1 through 8, and micronuclei were measured in peripheral blood 24 h after the last dose. AZT treatment increased the micronucleus (MN) frequencies to similar extents in both the reticulocytes (RETs) and normochromatic erythrocytes (NCEs) of Tk+/+ and Tk+/− mice; AZT did not increase the frequency of micronucleated RETs (MN-RETs) or micronucleated NCEs (MN-NCEs) in Tk−/− mice. Unexpectedly, neonatal Tk−/− mice treated with the vehicle had significantly elevated MN frequencies for both RETs and NCEs relative to Tk+/+ and Tk+/− mice (e.g., ∼3.4% MN-RETs and ∼4.8% MN-NCEs in Tk−/− mice versus ∼0.7 and ∼ 0.6% MN-RETs and MN-NCEs in neonatal Tk+/+ mice). Additional assays performed on untreated Tk−/− mice showed that elevated spontaneous MN frequencies persisted until at least 20 weeks of age, which approaches the average lifespan of Tk−/− mice. These results indicate that metabolism by Tk is necessary for the genotoxicity of AZT in neonatal mice; however, the genotoxicity of AZT is not altered by reducing the Tk gene dose by half. The elevated spontaneous MN frequencies in Tk−/− mice suggest the presence of an endogenous genotoxic activity in these mice. © 2004 Elsevier B.V. All rights reserved. Keywords: Erythrocyte frequency; Azidothymidine; Micronucleus; Thymidine kinase; Nucleoside analog reverse-transcriptase inhibitor; DNA damage; Pyrimidine metabolism
∗
Corresponding author. Tel.: +1 870 543 7549; fax: +1 870 543 7393. E-mail address: [email protected] (V.N. Dobrovolsky).
0027-5107/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.mrfmmm.2004.11.006
228
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
1. Introduction Azidothymidine (AZT; also known as Retrovir® , Zidovudine, ZDV) is the most widely used nucleoside analog reverse-transcriptase inhibitor (NRTI) for reducing the mother-to-child transmission of human immunodeficiency virus I (HIV) [1,2]. Upon entering the cell, AZT undergoes several rounds of phosphorylation. AZT 5 -triphosphate (AZT-TP) incorporates into nascent pro-viral DNA during reverse-transcription of the viral RNA genome. The incorporation results in chain termination, thus inhibiting the synthesis of pro-virus DNA and protecting the cell from infection by the virus. The active metabolite, AZT-TP, is also a substrate for endogenous DNA polymerases and, as a result, may be incorporated into nascent nuclear genomic DNA as a chain terminator [3]. The genotoxic properties of AZT are of particular concern because the current regimens for reducing the vertical transmission of HIV from infected mothers to their children call for exposure of the fetus in utero and the newborn to high doses of AZT (0.5 g daily between weeks 14 and 28 of gestation to pregnant mothers; 12 mg/kg daily to newborn babies for 6 weeks [GlaxoSmithKline Retrovir® prescribing information sheet, 2003]). Several in vivo studies indicate that exposure to AZT is mutagenic and carcinogenic. Female mice and rats exposed to large doses of AZT as adults develop vaginal tumors [4], and AZT is a carcinogen by transplacental exposure of mice and rats (reviewed in [5]). AZT damages chromosomes, as evidenced by elevated micronucleus (MN) frequencies in the erythrocytes of mice [6–9]. Also, the treatment of neonatal Tk+/− mice with high doses of AZT increases Tk gene mutant frequency, with a large proportion of the induced mutants having loss of heterozygosity (LOH) involving the wild-type Tk+ allele [9,10]. AZT also is a mutagen in vitro [4], and displays a similar LOH mutation signature [11,12]. The first step in the metabolism of AZT into a biologically active compound is believed to be 5 phosphorylation catalyzed by the endogenous cytosolic thymidine kinase (Tk1 or Tk elsewhere in the text) to produce AZT-monophosphate (AZT-MP) [13]. The rate of AZT phosphorylation is 60% that of thymidine, the natural substrate for phosphorylation by Tk1. The second and the third phosphorylations are catalyzed by the endogenous thymidylate kinase (TpK)
and nucleoside-diphosphate kinase, respectively. AZTMP is a relatively poor substrate for mammalian TpK and this is a potential limitation for the efficiency of HIV inhibition by AZT [14]. Alternative routes for in vivo AZT activation are also suspected. For instance, mitochondrial thymidine kinase (Tk2) metabolizes AZT into AZT-MP, although much less efficiently than Tk1 [15]. Human polymorphic variants of the TK1 gene were described recently [16]; however, because of the kinetics of AZT activation and the possibility of AZT activation by alternative enzymes, it is not clear if polymorphic variants will have any affect on the clinical effectiveness of AZT. In the present study, we evaluated the role of the Tk gene in the activation of AZT into genotoxic derivatives. We have used MN induction as an endpoint for determining whether alternative routes of AZT activation can result in measurable damage to nuclear DNA. For these studies, we treated neonatal Tk+/+ , Tk+/− and Tk−/− mice with AZT and measured MN frequencies in peripheral red blood cells, i.e. reticulocytes (RETs) and normochromatic erythrocytes (NCEs).
2. Materials and methods 2.1. Mice All animal procedures were approved by the Institutional Animal Care and Use Committee at the National Center for Toxicological Research. Sibling Tk+/+ , Tk+/− and Tk−/− mice were produced by breeding C57Bl/6 Tk+/− parents; the Tk genotypes were identified by allele-specific two- or three-primer PCR [17]. B6C3F1 Tk+/− mice were produced by breeding C57Bl/6 Tk+/− females from our locally maintained colony to C3H males. Tk+/− mice are available from the Mutant Mouse Regional Resource Center (http://www.mmrrc.org/strains/14/0014.html). The total number of animals involved in the study is shown in Table 1. 2.2. Treatment of mice with AZT and blood collection from treated and untreated mice Mice were injected intraperitoneally each day on postnatal days (PNDs) 1 through 8 with 5 l dimethylsulfoxide (DMSO) containing 200 mg/kg AZT (Cipla,
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
229
Table 1 Total number, genotype and sex of mice used in three micronucleus detection experiments Treatment Time course experiment (neonatal animals) Genotype DMSO
Tk+/−
AZT
Tk+/−
Total
31 ; 31 B6C3F1
Treatment Genotype experiment (neonatal animals)
Sex and number 24 h 6 ;4
5 ;7
48 h 5 ;5
5 ;5
Genotype
Sex and number
Genotype
Tk+/− Tk−/−
5 ;8 13 ; 12 4 ;9
Tk+/+
Tk+/+ Tk+/− Tk−/−
8 ;2 11 ; 11 4 ;5
72 h 5 ;5
5 ;5
DMSO
AZT
Mouse-age experiment
Tk+/+
45 ; 47 C57BI/6
Tk+/− Tk−/−
Age, sex and number 2 week 3 week 9 week 20 week 1 3 ;3 2
1 ;3 4 ;2 5 ;1
2 ;2 2 ;2 2 ;2
3 3 ;3 2 ;3
28 ; 23 C57Bl/6
In the time course and the genotype experiments the animals were treated daily with 200 mg/kg AZT or the vehicle (DMSO) on postnatal days 1 through 8. Blood samples were collected 24, 48 or 72 h after the last treatment in the time course experiment, and 24 h after the last treatment in the genotype experiment. In the mouse-age experiment animals were not treated; the ages chosen for blood collection (2, 3, 9 and 20 weeks of age) reflect the shortened lifespan of Tk−/− mice. Micronucleated erythrocytes were identified as described in Section 2.
Mumbai, India). Controls were injected with DMSO. The AZT dose in this experiment was the same as used previously for MN and Tk mutation assays [9]. Blood for MN detection was collected from AZT- and DMSOtreated neonates by cardiac puncture on PNDs 9, 10 and 11. Spontaneous MN frequencies for Tk+/+ , Tk+/− and −/− Tk mice also were assessed at 2, 3, 9, and 20 weeks of age. These time points were selected to provide MN frequencies over the average lifespan of Tk−/− mice [17]. 2.3. Detection of micronucleated cells by flow cytometry Micronucleated cells were identified using the MicroFlowplus mouse kit from Litron Laboratories (Rochester, NY) [18]. Briefly, approximately 120 l of blood were collected via cardiac puncture and diluted into 350 l of anticoagulant solution (heparin) from the kit. One hundred and eighty milliliters of the mixture were fixed in 2 ml of −80 ◦ C methanol, and stained with two fluorochromes for flow analysis; the remaining unfixed cells were used for manual MN analysis (see below). RETs were identified by FITClabeled antibodies against the CD71 mouse surface antigen, and DNA, including micronuclei, was stained with propidium iodide. Sample preparation and staining were conducted according to Litron’s suggested
protocol. Flow cytometry was performed on a FACSort (Beckton-Dickinson, San Jose, CA) equipped with a 488 nm argon ion laser and three fluorescent detectors. Calibration of the detection windows was performed according to the Litron protocol using CD71-negative and malaria-infected standards. The data were acquired and processed using CellQuestTM software and the templates included with the MicroFlowplus kit. Fluorescent data on channels FL1, FL2 and FL3 were collected in logarithmic amplification; the channels were set at 580 for FL1, 498 for FL2 and 498 for FL3. Forward and side scattering on the FSC and SSC channels were acquired in linear amplification. In order to maximize the discrimination between RETs and NCEs, the cutoff values for the bivariate plots were set at quadrant locations 13 and 5. These values were established by performing assays with and without RNAse digestion as described previously [19,20]. Data acquisition for each sample stopped automatically after 20,000 RETs were detected by the flow cytometer. 2.4. Manual detection of micronucleated cells Excess red blood cells were collected from the anticoagulant solution (see previous section) by a 3-min centrifugation at 1000 × g, and resuspended in 50 l of 50% fetal bovine serum in S¨orensen’s buffer (mixture of equal volumes of 0.1 M Na2 HPO4 and 0.1 M KH2 PO4 , pH 6.8). Ten microliters of the cell
230
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
suspension were smeared on each of four cleaned glass microscope slides and allowed to air dry. The dried slides were fixed for 15 min in fresh methanol and air dried. Fixed cells were stained for 1 min with a 10% solution of acridine orange (Sigma, St. Louis, MO) in S¨orensen’s buffer and rinsed in two changes of buffer. The dried slides were stored at 4 ◦ C. The cells were evaluated at 630× magnification under oil immersion with an Axioskop fluorescence microscope (Carl Zeiss, Thornwood, NY). The frequency of micronucleated cells among 2000 polychromatic RETs and 2000 NCEs was scored in blood smears from each animal. 2.5. Statistical analysis Statistical analysis of data was conducted using SigmaStat software (SPSS, Chicago, IL). Two-group comparisons were performed using the two-tailed homoscedastic Student’s t-test; if the data normality test failed, then the Mann–Whitney rank sum test was used. Multiple-group comparisons were performed using one-way analysis of variance (ANOVA) followed by the Bonferroni t-test for pair-wise comparisons; if the data normality test failed, then ANOVA on ranks was used followed by Dunn’s test for pair-wise comparisons.
3. Results 3.1. Time course evaluation The differences in the frequency of micronucleated RETs (MN-RETs) and micronucleated NCEs (MNNCEs) in males and females for each time point and each treatment group were not statistically significant (p > 0.05), with the exception of 24 h after the last treatment in DMSO-treated animals, where the MN-RET response in males was marginally higher than in females (p = 0.045). To simplify the analysis, the data for males and females were combined for each treatment group at each time point post exposure. The frequency of micronuclei in the RETs of AZTtreated neonatal B6C3F1 Tk+/− mice was highest 24 h after the last treatment and decreased rapidly. Within 72 h after the last injection of AZT the frequency of MN-RETs decreased almost 20-fold (Fig. 1). The differences in the frequency of MN-RETs in animals treated with AZT and DMSO, however, were statistically significant (p < 0.001) for all three time points at which the measurements were taken. The frequencies of MN-RETs in DMSO-treated mice were similar for all time points (less than 1%).
Fig. 1. Frequency of micronucleated reticulocytes (MN-RETs) and micronucleated normochromatic erythrocytes (MN-NCEs) in neonatal Tk+/− mice treated on postnatal days 1 through 8 with single injections of AZT (200 mg/kg/injection). Assays were conducted 24, 48 and 72 h after the last injection. Error bars represent standard deviation. Mean values and standard deviations for each group are shown above the bar graphs. Values for treated mice increased relative to the control: * p < 0.001. The numbers of animals for each group are shown in Table 1. Micronucleated cells were detected by automated flow cytometry. Note the different scales on the y-axes of the two charts.
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
The frequency of MN-NCEs in AZT-treated mice was similar to the frequency of MN-RETs at 24 h after the last treatment, but unlike MN-RETs, the frequency of MN-NCEs remained high in AZT-treated mice for at least 72 h. In DMSO-treated animals, the frequencies of MN-RETs and MN-NCEs remained low at all measurement time points. 3.2. Genotype comparison The responses (expressed in percent MN-RETs and MN-NCEs) were essentially the same among males and females in matching genotype/treatment groups. The exception was the MN-RET data for Tk−/− animals treated with AZT, where the response in females was higher (p = 0.045) than in males due to the presence of one outlier female. For simplicity the data for males and females (including the outlier) were combined for each genotype/treatment group. At 24 h after the last injection with AZT or DMSO, the frequency of MN-RETs in treated mice depended upon the genotype of the animal. In AZT-treated animals possessing an intact copy of the Tk gene
231
(wild-type Tk+/+ and heterozygous Tk+/− ), the percentage of MN-RETs was increased almost 20-fold and was essentially the same (p > 0.05) in the two genotypes (Fig. 2). AZT-treated Tk−/− animals had a slightly higher frequency of MN-RETs than DMSO-treated Tk−/− mice, but the increase was not statistically significant (p = 0.07). In addition, AZT-treated Tk−/− mice produced significantly lower frequencies of MN-RETs than Tk+/+ or Tk+/− animals (p < 0.001). DMSO-treated Tk−/− animals, however, produced significantly higher frequencies of MN-RETs than Tk-proficient animals (p < 0.001). AZT treatment also did not increase the frequency of MN-NCEs in Tk−/− mice (Fig. 2). 24 h after the last injection, the pattern of change in the frequency of MNNCEs in AZT-treated animals of the three different Tk genotypes was qualitatively similar to the pattern for MN-RETs. 3.3. Age-related spontaneous MN frequencies Up to 20 weeks of age, which approximates the lifespan of Tk−/− animals, the percentage of sponta-
Fig. 2. Frequency of micronucleated reticulocytes (MN-RETs) and micronucleated normochromatic erythrocytes (MN-NCEs) in neonatal Tk+/+ , Tk+/− and Tk−/− mice treated on postnatal days 1 through 8 with single injections of AZT (200 mg/kg/injection). Assays were conducted 24 h after the last injection. Error bars represent standard deviation. Mean values and standard deviations for each group are shown above the bar graphs. Values for treated mice significantly different from controls of the same genotype: * p < 0.001. Values for Tk−/− mice significantly different from Tk+/+ mice ( p < 0.001) and Tk+/− mice (♣ p < 0.001). The numbers of animals for each group are shown in Table 1. Micronucleated cells were detected by automated flow cytometry. Note the different scales on the y-axes of the two charts.
232
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
neous MN-RETs and MN-NCEs remained higher in Tk−/− mice than in age-matched Tk+/+ or Tk+/− mice (Fig. 3). In addition, there were no differences in the frequency of MN-RETs or MN-NCEs between Tk+/+ and Tk+/− mice (p > 0.05). The frequency of MN-RETs and MN-NCEs was also counted manually for 9-weekold animals. The data determined microscopically and by flow cytometry had the same pattern for the three genotypes (Tk+/+ ∼ = Tk+/− < Tk−/− ). Although the percent of MN-RETs in 20-week-old Tk−/− mice appeared to be increased compared to younger Tk−/− animals, there were no statistically significant differences among the MN-RETs frequencies in the different Tk−/− age groups (ANOVA, p = 0.06).
4. Discussion Although AZT has been studied extensively in vitro and in vivo, certain mechanistic aspects of its activation remain unclear, such as whether alternative pathways for AZT phosphorylation produce biologically significant amounts of nuclear DNA damage. If genotoxicity were detectable (e.g., expressed as increased MN formation) in systems with inactivated pyrimidine salvage (the major pathway for AZT activation), it would indi-
cate that alternative routes for AZT metabolism were involved in producing nuclear DNA damage. With the creation of Tk-deficient mice, studying alternative pathways for in vivo AZT activation became possible. The induction of micronuclei in mouse peripheral blood erythrocytes (such as in NCEs) is a good measure of genotoxicity since MN-NCEs are not specifically eliminated as they are in other species; with time, induced responses slowly decrease due to the normal turnover of red blood cells (reviewed in [21]). The efficiency of measuring genotoxicity can be even higher if micronuclei are detected in immature erythrocytes, i.e. RETs. In the bloodstream, RETs quickly mature and become NCEs, where they remain until cleared from the circulation by the spleen. If the influx of RETs is low, as in adult animals (less than 5% of total red blood cells), RETs become the best choice for measuring MN induction from short-term treatments. The treatment also may affect erythropoiesis and subsequently delay the release of MN-RETs into the circulation. Therefore, for each agent, the kinetics of erythrocyte maturation may vary and the sampling time for detecting the maximum MN frequency also may vary. In our study, the maximum frequency of MN-RETs in peripheral blood was achieved within 24 h after the last treatment of the multi-injection AZT regimen and the frequency
Fig. 3. Frequency of spontaneous micronucleated reticulocytes (MN-RETs) and micronucleated normochromatic erythrocytes (MN-NCEs) in Tk-proficient and Tk-deficient mice. Error bars represent standard deviation. Mean values and standard deviations for each group are shown above the bar graphs. For each age group, frequencies in Tk−/− mice significantly different from Tk+/+ mice (* p < 0.05; p < 0.001) and Tk+/− mice (♣ p < 0.05; ♥ p < 0.001). Micronucleated cells were detected by automated flow cytometry for mice of all ages. Assays for 9-week-old animals were conducted by both flow cytometry and by manual microscopic examination. Only one 2-week-old Tk+/+ animal was used (no standard deviation). The numbers of animals for each group are shown in Table 1. Note the different scales on the y-axes of the two charts.
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
decreased rapidly thereafter. In contrast, the maximum response for single doses of other genotoxins (mitomycin C, dimethylbenzanthracene) delivered to adult mice was observed 48 h after the treatment [22]. As expected, the frequency of MN-NCEs did not decrease after cessation of the AZT treatment. Assuming that the Tk-deficiency does not interfere with erythropoiesis, these results indicate that 24 h is more appropriate than 48 h after treatment for detecting micronuclei in AZTtreated neonates. We found that AZT was a potent inducer of MNRETs and MN-NCEs in Tk+/+ and Tk+/− mice, but the AZT treatment regimen that we employed had no significant effect on MN frequencies in Tk−/− mice. These results suggest that the major pathway of AZT activation is via pyrimidine salvage and that alternative routes of metabolic activation have little effect on the production of nuclear DNA-damaging species. An additional finding was that reduction of the Tk gene dose by half had no effect on AZT-induced MN frequency, i.e. the MN frequencies in treated Tk+/+ and Tk+/− mice were similar. This observation suggests that a small difference in Tk enzymatic activity, as might occur in individuals with a polymorphism of the Tk gene, would have little effect on the biological activity of AZT. An unexpected observation was that the background MN frequency in Tk−/− mice was significantly greater than in Tk+/− or Tk+/+ mice. These results indicate that Tk−/− mice possess an endogenous genotoxic activity, which causes an approximate three-fold increase in the spontaneous frequency of MN-RETs and an eight-fold increase in the frequency of MN-NCEs in neonatal animals. Previous observations on the effect of Tk-deficiency on endogenous DNA damage were inconclusive. Although the spontaneous Hprt mutant frequency (MF) in Tk−/− animals was elevated and the frequency of MN-RETs in the bone marrow of Tk−/− mice was higher than in matching Tk-proficient animals, these increases were not statistically significant [17,23]. The failure to demonstrate a significant increase in endogenous genotoxicity in Tk−/− mice in these previous studies may be due to technical limitations of the lymphocyte assay, manual scoring of micronucleated cells in bone marrow smears, and/or the limited number of animals evaluated. For instance, the accuracy of the Hprt assay is dependent upon the cloning efficiency of primary spleen lymphocytes;
233
Tk−/− animals have relatively low lymphocyte cloning efficiencies (1–2%), which compromises the accuracy of the assay. Also, the high-throughput flow cytometry method used to measure micronuclei in the present study permits a much greater number of the cells to be evaluated than can be scored in the manual bone marrow method. The flow cytometric method allowed us to detect an increase in spontaneous MN frequency in neonatal Tk−/− mice. Further investigation confirmed that higher frequencies of MN-RETs and MN-NCEs persist through the lifespan of Tk−/− animals. The origin of the endogenous chromosomal damage in Tk−/− animals remains to be explained. It is possible that the disrupted pyrimidine salvage pathway and two- to threefold higher thymidine concentration in the blood of Tk−/− mice [17] may interfere with DNA synthesis and cause MN formation. Other possibilities for the elevated MN frequencies in Tk−/− mice are suggested by the recent discovery that the arylformamidase (Afmid, also known as kynurenine formamidase) gene is positioned next to Tk on mouse chromosome 11 [24]. Our observations indicate that Tk−/− mice are also Afmid−/− (not shown). The Afmid gene is expressed in the kidneys, and a specific pathological condition of Tk−/− mice is kidney failure due to sclerosis of the glomeruli [17]. Kidney disease can be detected in young animals and is the primary reason that Tk−/− animals die at approximately 6 months of age. It is conceivable that the early onset of kidney malfunction may affect erythropoiesis through deregulation of erythropoietin expression by kidney peritubular interstitial cells [25]. An indication of an altered maturation of erythrocytes in Tk−/− animals may be an increase in the spontaneous frequency of MN-RETs and MN-NCEs. Afmid is the key enzyme of the tryptophan degradation pathway. Afmid deficiency potentially can affect not only the immediate substrate and product of the enzyme, but also the upstream and downstream products of the pathway. It is not known to what extent tryptophan metabolism is required for erythropoiesis, although the degradation pathway itself and several downstream products of tryptophan degradation (such as kynurenic acid and quinolinic acid) have been implicated in the proper functioning of the immune and nervous systems [26,27].
234
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
In conclusion, neonatal Tk−/− mice, but not Tk+/+ or mice, are resistant to the induction of MN-RETs and MN-NCEs by AZT. These findings suggest that the Tk gene is necessary for the activation of AZT to species genotoxic for nuclear DNA. Other pathways, such as activation involving the mitochondrial Tk2 gene, do not contribute significantly to AZT-induced genotoxicity (at least in the nucleus). The observation that AZT produced the same MN-RET and MN-NCE frequencies in Tk+/+ and Tk+/− mice confirms that metabolism by Tk is not a rate-limiting step in the activation of AZT. Therefore, our data suggest that polymorphism affecting the activity of the Tk gene (e.g., decreasing activity by 50%) will have little effect on the clinical efficacy of AZT. Finally, Tk-deficient mice have increased levels of endogenous chromosome damage, which may be a consequence of imbalanced thymidine and/or tryptophan metabolism, so that Tk−/− mice may become an in vivo model for evaluating metabolic dysfunction. Tk+/−
Acknowledgement This research was supported, in part, by Interagency Agreement #224-93-001 between National Center for Toxicological Research/Food and Drug Administration (FDA) and the National Institute for Environmental Health Sciences/National Toxicology Program. The views presented in this article do not necessarily reflect those of the US FDA.
References [1] E.M. Connor, R.S. Sperling, R. Gelber, P. Kiselev, G. Scott, M.J. O’Sullivan, R. VanDyke, M. Bey, W. Shearer, R.L. Jacobson, Reduction of maternal–infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group, New Engl. J. Med. 331 (1994) 1173–1180. [2] N. Bhana, D. Ormrod, C.M. Perry, D.P. Figgitt, Zidovudine: a review of its use in the management of vertically-acquired pediatric HIV infection, Paediatr. Drugs 4 (2002) 515–553. [3] W.C. Copeland, M.S. Chen, T.S. Wang, Human DNA polymerases alpha and beta are able to incorporate anti-HIV deoxynucleotides into DNA, J. Biol. Chem. 267 (1992) 21459–21464. [4] K.M. Ayers, D. Clive, W.E. Tucker Jr., G. Hajian, P. de Miranda, Nonclinical toxicology studies with zidovudine: genetic toxicity tests and carcinogenicity bioassays in mice and rats, Fundam. Appl. Toxicol. 32 (1996) 148–158.
[5] M.C. Poirier, O.A. Olivero, D.M. Walker, V.E. Walker, Perinatal genotoxicity and carcinogenicity of anti-retroviral nucleoside analog drugs, Toxicol. Appl. Pharmacol. 199 (2004) 151–161. [6] S.D. Dertinger, D.K. Torous, K.R. Tometsko, Induction of micronuclei by low doses of azidothymidine (AZT), Mutat. Res. 368 (1996) 301–307. [7] K.L. Witt, R.R. Tice, G.W. Wolfe, J.B. Bishop, Genetic damage detected in CD-1 mouse pups exposed perinatally to 3 -azido-3 deoxythymidine or dideoxyinosine via maternal dosing, nursing, and direct gavage. II. Effects of the individual agents compared to combination treatment, Environ. Mol. Mutagen. 44 (2004) 321–328. [8] M.D. Phillips, B. Nascimbeni, R.R. Tice, M.D. Shelby, Induction of micronuclei in mouse bone marrow cells: an evaluation of nucleoside analogues used in the treatment of AIDS, Environ. Mol. Mutagen. 18 (1991) 168–183. [9] L.S. Von Tungeln, L.P. Hamilton, V.N. Dobrovolsky, M.E. Bishop, J.G. Shaddock, R.H. Heflich, F.A. Beland, Frequency of Tk and Hprt lymphocyte mutants and bone marrow micronuclei in B6C3F(1)/Tk+/− mice treated neonatally with zidovudine and lamivudine, Carcinogenesis 23 (2002) 1427–1432. [10] R.A. Mittelstaedt, L.S. Von Tungeln, J.G. Shaddock, V.N. Dobrovolsky, F.A. Beland, R.H. Heflich, Analysis of mutations in the Tk gene of Tk(+/−) mice treated as neonates with 3 -azido3 -deoxythymidine (AZT), Mutat. Res. 547 (2004) 63–69. [11] Q. Meng, A.J. Grosovsky, X. Shi, V.E. Walker, Mutagenicity and loss of heterozygosity at the APRT locus in human lymphoblastoid cells exposed to 3 -azido-3 -deoxythymidine, Mutagenesis 15 (2000) 405–410. [12] Q. Meng, T. Su, O.A. Olivero, M.C. Poirier, X. Shi, X. Ding, V.E. Walker, Relationships between DNA incorporation, mutant frequency, and loss of heterozygosity at the TK locus in human lymphoblastoid cells exposed to 3 -azido-3 -deoxythymidine, Toxicol. Sci. 54 (2000) 322–329. [13] P.A. Furman, J.A. Fyfe, M.H. St Clair, K. Weinhold, J.L. Rideout, G.A. Freeman, S.N. Lehrman, D.P. Bolognesi, S. Broder, H. Mitsuya, Phosphorylation of 3 -azido-3 -deoxythymidine and selective interaction of the 5 -triphosphate with human immunodeficiency virus reverse transcriptase, Proc. Natl. Acad. Sci. U.S.A. 83 (1986) 8333–8337. [14] R. Brundiers, A. Lavie, T. Veit, J. Reinstein, I. Schlichting, N. Ostermann, R.S. Goody, M. Konrad, Modifying human thymidylate kinase to potentiate azidothymidine activation, J. Biol. Chem. 274 (1999) 35289–35292. [15] B. Munch-Petersen, L. Cloos, G. Tyrsted, S. Eriksson, Diverging substrate specificity of pure human thymidine kinases 1 and 2 against antiviral dideoxynucleosides, J. Biol. Chem. 266 (1991) 9032–9038. [16] S.I. Gilles, S. Romain, P. Casellas, L. Ouafik, F. Fina, T. Combes, V. Vuaroquaux, J.F. Seitz, P. Bonnier, S. Galiegue, P. Carayon, P.M. Martin, Mutation analysis in the coding sequence of thymidine kinase 1 in breast and colorectal cancer, Int. J. Biol. Markers 18 (2003) 1–6. [17] V.N. Dobrovolsky, T. Bucci, R.H. Heflich, J. Desjardins, F.C. Richardson, Mice deficient for cytosolic thymidine kinase gene develop fatal kidney disease, Mol. Genet. Metab. 78 (2003) 1–10.
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235 [18] S.D. Dertinger, Y. Chen, R.K. Miller, K.J. Brewer, T. Smudzin, D.K. Torous, N.E. Hall, K.A. Olvany, F.G. Murante, C.R. Tometsko, Micronucleated CD71-positive reticulocytes: a blood-based endpoint of cytogenetic damage in humans, Mutat. Res. 542 (2003) 77–87. [19] D.K. Torous, N.E. Hall, F.G. Murante, S.E. Gleason, C.R. Tometsko, S.D. Dertinger, Comparative scoring of micronucleated reticulocytes in rat peripheral blood by flow cytometry and microscopy, Toxicol. Sci. 74 (2003) 309–314. [20] S.D. Dertinger, D.K. Torous, N.E. Hall, F.G. Murante, S.E. Gleason, R.K. Miller, C.R. Tometsko, Enumeration of micronucleated CD71-positive human reticulocytes with a single-laser flow cytometer, Mutat. Res. 515 (2002) 3–14. [21] J.L. Weaver, D. Torous, Flow cytometry assay for counting micronucleated erythrocytes: development process, Methods 21 (2000) 281–287. [22] J.T. MacGregor, C.M. Wehr, P.R. Henika, M.D. Shelby, The in vivo erythrocyte micronucleus test: measurement at steady state increases assay efficiency and permits integration with toxicity studies, Fundam. Appl. Toxicol. 14 (1990) 513–522.
235
[23] V.N. Dobrovolsky, P.B. McKinzie, J.G. Shaddock, R.A. Mittelstaedt, R.H. Heflich, B.L. Parsons, Pms2 deficiency results in increased mutation in the Hprt gene but not the Tk gene of Tk(+/−) transgenic mice, Mutagenesis 18 (2003) 365– 370. [24] B. Schuettengruber, A. Doetzlhofer, K. Kroboth, E. Wintersberger, C. Seiser, Alternate activation of two divergently transcribed mouse genes from a bidirectional promoter is linked to changes in histone modification, J. Biol. Chem. 278 (2003) 1784–1793. [25] S.T. Koury, M.C. Bondurant, G.L. Semenza, M.J. Koury, The use of in situ hybridization to study erythropoietin gene expression in murine kidney and liver, Microsc. Res. Technol. 25 (1993) 29–39. [26] J.R. Moffett, M.A. Namboodiri, Tryptophan and the immune response, Immunol. Cell Biol. 81 (2003) 247–265. [27] R. Schwarcz, R. Pellicciari, Manipulation of brain kynurenines: glial targets, neuronal effects, and clinical opportunities, J. Pharmacol. Exp. Ther. 303 (2002) 1–10.
Micronucleated erythrocyte frequency in control and azidothymidine-treated Tk+/+, Tk+/− and Tk−/− mice Vasily N. Dobrovolskya,∗ , Lynda J. McGarritya , Linda S. VonTungelnb , Roberta A. Mittelstaedta , Suzanne M. Morrisa , Frederick A. Belandb , Robert H. Heflicha a
Division of Genetic and Reproductive Toxicology, U.S. Food and Drug Administration/National Center for Toxicological Research, HFT-120, 3900 NCTR Rd., Jefferson, AR 72079, USA b Division of Biochemical Toxicology, U.S. Food and Drug Administration/National Center for Toxicological Research, HFT-110, 3900 NCTR Rd., Jefferson, AR 72079, USA Received 17 August 2004; received in revised form 3 November 2004; accepted 20 November 2004 Available online 21 January 2005
Abstract The first step in the activation of the anti-retroviral nucleoside analogue azidothymidine (AZT) involves its conversion to a 5 -monophosphate. In this study, we have evaluated the role of cytosolic thymidine kinase (Tk), the major enzyme involved in phosphorylating thymidine and its analogues, in the nuclear DNA damage produced by AZT in neonatal mice. Tk+/+ , Tk+/− and Tk−/− mice were treated intraperitoneally with 200 mg/kg/day of AZT on postnatal days 1 through 8, and micronuclei were measured in peripheral blood 24 h after the last dose. AZT treatment increased the micronucleus (MN) frequencies to similar extents in both the reticulocytes (RETs) and normochromatic erythrocytes (NCEs) of Tk+/+ and Tk+/− mice; AZT did not increase the frequency of micronucleated RETs (MN-RETs) or micronucleated NCEs (MN-NCEs) in Tk−/− mice. Unexpectedly, neonatal Tk−/− mice treated with the vehicle had significantly elevated MN frequencies for both RETs and NCEs relative to Tk+/+ and Tk+/− mice (e.g., ∼3.4% MN-RETs and ∼4.8% MN-NCEs in Tk−/− mice versus ∼0.7 and ∼ 0.6% MN-RETs and MN-NCEs in neonatal Tk+/+ mice). Additional assays performed on untreated Tk−/− mice showed that elevated spontaneous MN frequencies persisted until at least 20 weeks of age, which approaches the average lifespan of Tk−/− mice. These results indicate that metabolism by Tk is necessary for the genotoxicity of AZT in neonatal mice; however, the genotoxicity of AZT is not altered by reducing the Tk gene dose by half. The elevated spontaneous MN frequencies in Tk−/− mice suggest the presence of an endogenous genotoxic activity in these mice. © 2004 Elsevier B.V. All rights reserved. Keywords: Erythrocyte frequency; Azidothymidine; Micronucleus; Thymidine kinase; Nucleoside analog reverse-transcriptase inhibitor; DNA damage; Pyrimidine metabolism
∗
Corresponding author. Tel.: +1 870 543 7549; fax: +1 870 543 7393. E-mail address: [email protected] (V.N. Dobrovolsky).
0027-5107/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.mrfmmm.2004.11.006
228
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
1. Introduction Azidothymidine (AZT; also known as Retrovir® , Zidovudine, ZDV) is the most widely used nucleoside analog reverse-transcriptase inhibitor (NRTI) for reducing the mother-to-child transmission of human immunodeficiency virus I (HIV) [1,2]. Upon entering the cell, AZT undergoes several rounds of phosphorylation. AZT 5 -triphosphate (AZT-TP) incorporates into nascent pro-viral DNA during reverse-transcription of the viral RNA genome. The incorporation results in chain termination, thus inhibiting the synthesis of pro-virus DNA and protecting the cell from infection by the virus. The active metabolite, AZT-TP, is also a substrate for endogenous DNA polymerases and, as a result, may be incorporated into nascent nuclear genomic DNA as a chain terminator [3]. The genotoxic properties of AZT are of particular concern because the current regimens for reducing the vertical transmission of HIV from infected mothers to their children call for exposure of the fetus in utero and the newborn to high doses of AZT (0.5 g daily between weeks 14 and 28 of gestation to pregnant mothers; 12 mg/kg daily to newborn babies for 6 weeks [GlaxoSmithKline Retrovir® prescribing information sheet, 2003]). Several in vivo studies indicate that exposure to AZT is mutagenic and carcinogenic. Female mice and rats exposed to large doses of AZT as adults develop vaginal tumors [4], and AZT is a carcinogen by transplacental exposure of mice and rats (reviewed in [5]). AZT damages chromosomes, as evidenced by elevated micronucleus (MN) frequencies in the erythrocytes of mice [6–9]. Also, the treatment of neonatal Tk+/− mice with high doses of AZT increases Tk gene mutant frequency, with a large proportion of the induced mutants having loss of heterozygosity (LOH) involving the wild-type Tk+ allele [9,10]. AZT also is a mutagen in vitro [4], and displays a similar LOH mutation signature [11,12]. The first step in the metabolism of AZT into a biologically active compound is believed to be 5 phosphorylation catalyzed by the endogenous cytosolic thymidine kinase (Tk1 or Tk elsewhere in the text) to produce AZT-monophosphate (AZT-MP) [13]. The rate of AZT phosphorylation is 60% that of thymidine, the natural substrate for phosphorylation by Tk1. The second and the third phosphorylations are catalyzed by the endogenous thymidylate kinase (TpK)
and nucleoside-diphosphate kinase, respectively. AZTMP is a relatively poor substrate for mammalian TpK and this is a potential limitation for the efficiency of HIV inhibition by AZT [14]. Alternative routes for in vivo AZT activation are also suspected. For instance, mitochondrial thymidine kinase (Tk2) metabolizes AZT into AZT-MP, although much less efficiently than Tk1 [15]. Human polymorphic variants of the TK1 gene were described recently [16]; however, because of the kinetics of AZT activation and the possibility of AZT activation by alternative enzymes, it is not clear if polymorphic variants will have any affect on the clinical effectiveness of AZT. In the present study, we evaluated the role of the Tk gene in the activation of AZT into genotoxic derivatives. We have used MN induction as an endpoint for determining whether alternative routes of AZT activation can result in measurable damage to nuclear DNA. For these studies, we treated neonatal Tk+/+ , Tk+/− and Tk−/− mice with AZT and measured MN frequencies in peripheral red blood cells, i.e. reticulocytes (RETs) and normochromatic erythrocytes (NCEs).
2. Materials and methods 2.1. Mice All animal procedures were approved by the Institutional Animal Care and Use Committee at the National Center for Toxicological Research. Sibling Tk+/+ , Tk+/− and Tk−/− mice were produced by breeding C57Bl/6 Tk+/− parents; the Tk genotypes were identified by allele-specific two- or three-primer PCR [17]. B6C3F1 Tk+/− mice were produced by breeding C57Bl/6 Tk+/− females from our locally maintained colony to C3H males. Tk+/− mice are available from the Mutant Mouse Regional Resource Center (http://www.mmrrc.org/strains/14/0014.html). The total number of animals involved in the study is shown in Table 1. 2.2. Treatment of mice with AZT and blood collection from treated and untreated mice Mice were injected intraperitoneally each day on postnatal days (PNDs) 1 through 8 with 5 l dimethylsulfoxide (DMSO) containing 200 mg/kg AZT (Cipla,
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
229
Table 1 Total number, genotype and sex of mice used in three micronucleus detection experiments Treatment Time course experiment (neonatal animals) Genotype DMSO
Tk+/−
AZT
Tk+/−
Total
31 ; 31 B6C3F1
Treatment Genotype experiment (neonatal animals)
Sex and number 24 h 6 ;4
5 ;7
48 h 5 ;5
5 ;5
Genotype
Sex and number
Genotype
Tk+/− Tk−/−
5 ;8 13 ; 12 4 ;9
Tk+/+
Tk+/+ Tk+/− Tk−/−
8 ;2 11 ; 11 4 ;5
72 h 5 ;5
5 ;5
DMSO
AZT
Mouse-age experiment
Tk+/+
45 ; 47 C57BI/6
Tk+/− Tk−/−
Age, sex and number 2 week 3 week 9 week 20 week 1 3 ;3 2
1 ;3 4 ;2 5 ;1
2 ;2 2 ;2 2 ;2
3 3 ;3 2 ;3
28 ; 23 C57Bl/6
In the time course and the genotype experiments the animals were treated daily with 200 mg/kg AZT or the vehicle (DMSO) on postnatal days 1 through 8. Blood samples were collected 24, 48 or 72 h after the last treatment in the time course experiment, and 24 h after the last treatment in the genotype experiment. In the mouse-age experiment animals were not treated; the ages chosen for blood collection (2, 3, 9 and 20 weeks of age) reflect the shortened lifespan of Tk−/− mice. Micronucleated erythrocytes were identified as described in Section 2.
Mumbai, India). Controls were injected with DMSO. The AZT dose in this experiment was the same as used previously for MN and Tk mutation assays [9]. Blood for MN detection was collected from AZT- and DMSOtreated neonates by cardiac puncture on PNDs 9, 10 and 11. Spontaneous MN frequencies for Tk+/+ , Tk+/− and −/− Tk mice also were assessed at 2, 3, 9, and 20 weeks of age. These time points were selected to provide MN frequencies over the average lifespan of Tk−/− mice [17]. 2.3. Detection of micronucleated cells by flow cytometry Micronucleated cells were identified using the MicroFlowplus mouse kit from Litron Laboratories (Rochester, NY) [18]. Briefly, approximately 120 l of blood were collected via cardiac puncture and diluted into 350 l of anticoagulant solution (heparin) from the kit. One hundred and eighty milliliters of the mixture were fixed in 2 ml of −80 ◦ C methanol, and stained with two fluorochromes for flow analysis; the remaining unfixed cells were used for manual MN analysis (see below). RETs were identified by FITClabeled antibodies against the CD71 mouse surface antigen, and DNA, including micronuclei, was stained with propidium iodide. Sample preparation and staining were conducted according to Litron’s suggested
protocol. Flow cytometry was performed on a FACSort (Beckton-Dickinson, San Jose, CA) equipped with a 488 nm argon ion laser and three fluorescent detectors. Calibration of the detection windows was performed according to the Litron protocol using CD71-negative and malaria-infected standards. The data were acquired and processed using CellQuestTM software and the templates included with the MicroFlowplus kit. Fluorescent data on channels FL1, FL2 and FL3 were collected in logarithmic amplification; the channels were set at 580 for FL1, 498 for FL2 and 498 for FL3. Forward and side scattering on the FSC and SSC channels were acquired in linear amplification. In order to maximize the discrimination between RETs and NCEs, the cutoff values for the bivariate plots were set at quadrant locations 13 and 5. These values were established by performing assays with and without RNAse digestion as described previously [19,20]. Data acquisition for each sample stopped automatically after 20,000 RETs were detected by the flow cytometer. 2.4. Manual detection of micronucleated cells Excess red blood cells were collected from the anticoagulant solution (see previous section) by a 3-min centrifugation at 1000 × g, and resuspended in 50 l of 50% fetal bovine serum in S¨orensen’s buffer (mixture of equal volumes of 0.1 M Na2 HPO4 and 0.1 M KH2 PO4 , pH 6.8). Ten microliters of the cell
230
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
suspension were smeared on each of four cleaned glass microscope slides and allowed to air dry. The dried slides were fixed for 15 min in fresh methanol and air dried. Fixed cells were stained for 1 min with a 10% solution of acridine orange (Sigma, St. Louis, MO) in S¨orensen’s buffer and rinsed in two changes of buffer. The dried slides were stored at 4 ◦ C. The cells were evaluated at 630× magnification under oil immersion with an Axioskop fluorescence microscope (Carl Zeiss, Thornwood, NY). The frequency of micronucleated cells among 2000 polychromatic RETs and 2000 NCEs was scored in blood smears from each animal. 2.5. Statistical analysis Statistical analysis of data was conducted using SigmaStat software (SPSS, Chicago, IL). Two-group comparisons were performed using the two-tailed homoscedastic Student’s t-test; if the data normality test failed, then the Mann–Whitney rank sum test was used. Multiple-group comparisons were performed using one-way analysis of variance (ANOVA) followed by the Bonferroni t-test for pair-wise comparisons; if the data normality test failed, then ANOVA on ranks was used followed by Dunn’s test for pair-wise comparisons.
3. Results 3.1. Time course evaluation The differences in the frequency of micronucleated RETs (MN-RETs) and micronucleated NCEs (MNNCEs) in males and females for each time point and each treatment group were not statistically significant (p > 0.05), with the exception of 24 h after the last treatment in DMSO-treated animals, where the MN-RET response in males was marginally higher than in females (p = 0.045). To simplify the analysis, the data for males and females were combined for each treatment group at each time point post exposure. The frequency of micronuclei in the RETs of AZTtreated neonatal B6C3F1 Tk+/− mice was highest 24 h after the last treatment and decreased rapidly. Within 72 h after the last injection of AZT the frequency of MN-RETs decreased almost 20-fold (Fig. 1). The differences in the frequency of MN-RETs in animals treated with AZT and DMSO, however, were statistically significant (p < 0.001) for all three time points at which the measurements were taken. The frequencies of MN-RETs in DMSO-treated mice were similar for all time points (less than 1%).
Fig. 1. Frequency of micronucleated reticulocytes (MN-RETs) and micronucleated normochromatic erythrocytes (MN-NCEs) in neonatal Tk+/− mice treated on postnatal days 1 through 8 with single injections of AZT (200 mg/kg/injection). Assays were conducted 24, 48 and 72 h after the last injection. Error bars represent standard deviation. Mean values and standard deviations for each group are shown above the bar graphs. Values for treated mice increased relative to the control: * p < 0.001. The numbers of animals for each group are shown in Table 1. Micronucleated cells were detected by automated flow cytometry. Note the different scales on the y-axes of the two charts.
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
The frequency of MN-NCEs in AZT-treated mice was similar to the frequency of MN-RETs at 24 h after the last treatment, but unlike MN-RETs, the frequency of MN-NCEs remained high in AZT-treated mice for at least 72 h. In DMSO-treated animals, the frequencies of MN-RETs and MN-NCEs remained low at all measurement time points. 3.2. Genotype comparison The responses (expressed in percent MN-RETs and MN-NCEs) were essentially the same among males and females in matching genotype/treatment groups. The exception was the MN-RET data for Tk−/− animals treated with AZT, where the response in females was higher (p = 0.045) than in males due to the presence of one outlier female. For simplicity the data for males and females (including the outlier) were combined for each genotype/treatment group. At 24 h after the last injection with AZT or DMSO, the frequency of MN-RETs in treated mice depended upon the genotype of the animal. In AZT-treated animals possessing an intact copy of the Tk gene
231
(wild-type Tk+/+ and heterozygous Tk+/− ), the percentage of MN-RETs was increased almost 20-fold and was essentially the same (p > 0.05) in the two genotypes (Fig. 2). AZT-treated Tk−/− animals had a slightly higher frequency of MN-RETs than DMSO-treated Tk−/− mice, but the increase was not statistically significant (p = 0.07). In addition, AZT-treated Tk−/− mice produced significantly lower frequencies of MN-RETs than Tk+/+ or Tk+/− animals (p < 0.001). DMSO-treated Tk−/− animals, however, produced significantly higher frequencies of MN-RETs than Tk-proficient animals (p < 0.001). AZT treatment also did not increase the frequency of MN-NCEs in Tk−/− mice (Fig. 2). 24 h after the last injection, the pattern of change in the frequency of MNNCEs in AZT-treated animals of the three different Tk genotypes was qualitatively similar to the pattern for MN-RETs. 3.3. Age-related spontaneous MN frequencies Up to 20 weeks of age, which approximates the lifespan of Tk−/− animals, the percentage of sponta-
Fig. 2. Frequency of micronucleated reticulocytes (MN-RETs) and micronucleated normochromatic erythrocytes (MN-NCEs) in neonatal Tk+/+ , Tk+/− and Tk−/− mice treated on postnatal days 1 through 8 with single injections of AZT (200 mg/kg/injection). Assays were conducted 24 h after the last injection. Error bars represent standard deviation. Mean values and standard deviations for each group are shown above the bar graphs. Values for treated mice significantly different from controls of the same genotype: * p < 0.001. Values for Tk−/− mice significantly different from Tk+/+ mice ( p < 0.001) and Tk+/− mice (♣ p < 0.001). The numbers of animals for each group are shown in Table 1. Micronucleated cells were detected by automated flow cytometry. Note the different scales on the y-axes of the two charts.
232
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
neous MN-RETs and MN-NCEs remained higher in Tk−/− mice than in age-matched Tk+/+ or Tk+/− mice (Fig. 3). In addition, there were no differences in the frequency of MN-RETs or MN-NCEs between Tk+/+ and Tk+/− mice (p > 0.05). The frequency of MN-RETs and MN-NCEs was also counted manually for 9-weekold animals. The data determined microscopically and by flow cytometry had the same pattern for the three genotypes (Tk+/+ ∼ = Tk+/− < Tk−/− ). Although the percent of MN-RETs in 20-week-old Tk−/− mice appeared to be increased compared to younger Tk−/− animals, there were no statistically significant differences among the MN-RETs frequencies in the different Tk−/− age groups (ANOVA, p = 0.06).
4. Discussion Although AZT has been studied extensively in vitro and in vivo, certain mechanistic aspects of its activation remain unclear, such as whether alternative pathways for AZT phosphorylation produce biologically significant amounts of nuclear DNA damage. If genotoxicity were detectable (e.g., expressed as increased MN formation) in systems with inactivated pyrimidine salvage (the major pathway for AZT activation), it would indi-
cate that alternative routes for AZT metabolism were involved in producing nuclear DNA damage. With the creation of Tk-deficient mice, studying alternative pathways for in vivo AZT activation became possible. The induction of micronuclei in mouse peripheral blood erythrocytes (such as in NCEs) is a good measure of genotoxicity since MN-NCEs are not specifically eliminated as they are in other species; with time, induced responses slowly decrease due to the normal turnover of red blood cells (reviewed in [21]). The efficiency of measuring genotoxicity can be even higher if micronuclei are detected in immature erythrocytes, i.e. RETs. In the bloodstream, RETs quickly mature and become NCEs, where they remain until cleared from the circulation by the spleen. If the influx of RETs is low, as in adult animals (less than 5% of total red blood cells), RETs become the best choice for measuring MN induction from short-term treatments. The treatment also may affect erythropoiesis and subsequently delay the release of MN-RETs into the circulation. Therefore, for each agent, the kinetics of erythrocyte maturation may vary and the sampling time for detecting the maximum MN frequency also may vary. In our study, the maximum frequency of MN-RETs in peripheral blood was achieved within 24 h after the last treatment of the multi-injection AZT regimen and the frequency
Fig. 3. Frequency of spontaneous micronucleated reticulocytes (MN-RETs) and micronucleated normochromatic erythrocytes (MN-NCEs) in Tk-proficient and Tk-deficient mice. Error bars represent standard deviation. Mean values and standard deviations for each group are shown above the bar graphs. For each age group, frequencies in Tk−/− mice significantly different from Tk+/+ mice (* p < 0.05; p < 0.001) and Tk+/− mice (♣ p < 0.05; ♥ p < 0.001). Micronucleated cells were detected by automated flow cytometry for mice of all ages. Assays for 9-week-old animals were conducted by both flow cytometry and by manual microscopic examination. Only one 2-week-old Tk+/+ animal was used (no standard deviation). The numbers of animals for each group are shown in Table 1. Note the different scales on the y-axes of the two charts.
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
decreased rapidly thereafter. In contrast, the maximum response for single doses of other genotoxins (mitomycin C, dimethylbenzanthracene) delivered to adult mice was observed 48 h after the treatment [22]. As expected, the frequency of MN-NCEs did not decrease after cessation of the AZT treatment. Assuming that the Tk-deficiency does not interfere with erythropoiesis, these results indicate that 24 h is more appropriate than 48 h after treatment for detecting micronuclei in AZTtreated neonates. We found that AZT was a potent inducer of MNRETs and MN-NCEs in Tk+/+ and Tk+/− mice, but the AZT treatment regimen that we employed had no significant effect on MN frequencies in Tk−/− mice. These results suggest that the major pathway of AZT activation is via pyrimidine salvage and that alternative routes of metabolic activation have little effect on the production of nuclear DNA-damaging species. An additional finding was that reduction of the Tk gene dose by half had no effect on AZT-induced MN frequency, i.e. the MN frequencies in treated Tk+/+ and Tk+/− mice were similar. This observation suggests that a small difference in Tk enzymatic activity, as might occur in individuals with a polymorphism of the Tk gene, would have little effect on the biological activity of AZT. An unexpected observation was that the background MN frequency in Tk−/− mice was significantly greater than in Tk+/− or Tk+/+ mice. These results indicate that Tk−/− mice possess an endogenous genotoxic activity, which causes an approximate three-fold increase in the spontaneous frequency of MN-RETs and an eight-fold increase in the frequency of MN-NCEs in neonatal animals. Previous observations on the effect of Tk-deficiency on endogenous DNA damage were inconclusive. Although the spontaneous Hprt mutant frequency (MF) in Tk−/− animals was elevated and the frequency of MN-RETs in the bone marrow of Tk−/− mice was higher than in matching Tk-proficient animals, these increases were not statistically significant [17,23]. The failure to demonstrate a significant increase in endogenous genotoxicity in Tk−/− mice in these previous studies may be due to technical limitations of the lymphocyte assay, manual scoring of micronucleated cells in bone marrow smears, and/or the limited number of animals evaluated. For instance, the accuracy of the Hprt assay is dependent upon the cloning efficiency of primary spleen lymphocytes;
233
Tk−/− animals have relatively low lymphocyte cloning efficiencies (1–2%), which compromises the accuracy of the assay. Also, the high-throughput flow cytometry method used to measure micronuclei in the present study permits a much greater number of the cells to be evaluated than can be scored in the manual bone marrow method. The flow cytometric method allowed us to detect an increase in spontaneous MN frequency in neonatal Tk−/− mice. Further investigation confirmed that higher frequencies of MN-RETs and MN-NCEs persist through the lifespan of Tk−/− animals. The origin of the endogenous chromosomal damage in Tk−/− animals remains to be explained. It is possible that the disrupted pyrimidine salvage pathway and two- to threefold higher thymidine concentration in the blood of Tk−/− mice [17] may interfere with DNA synthesis and cause MN formation. Other possibilities for the elevated MN frequencies in Tk−/− mice are suggested by the recent discovery that the arylformamidase (Afmid, also known as kynurenine formamidase) gene is positioned next to Tk on mouse chromosome 11 [24]. Our observations indicate that Tk−/− mice are also Afmid−/− (not shown). The Afmid gene is expressed in the kidneys, and a specific pathological condition of Tk−/− mice is kidney failure due to sclerosis of the glomeruli [17]. Kidney disease can be detected in young animals and is the primary reason that Tk−/− animals die at approximately 6 months of age. It is conceivable that the early onset of kidney malfunction may affect erythropoiesis through deregulation of erythropoietin expression by kidney peritubular interstitial cells [25]. An indication of an altered maturation of erythrocytes in Tk−/− animals may be an increase in the spontaneous frequency of MN-RETs and MN-NCEs. Afmid is the key enzyme of the tryptophan degradation pathway. Afmid deficiency potentially can affect not only the immediate substrate and product of the enzyme, but also the upstream and downstream products of the pathway. It is not known to what extent tryptophan metabolism is required for erythropoiesis, although the degradation pathway itself and several downstream products of tryptophan degradation (such as kynurenic acid and quinolinic acid) have been implicated in the proper functioning of the immune and nervous systems [26,27].
234
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235
In conclusion, neonatal Tk−/− mice, but not Tk+/+ or mice, are resistant to the induction of MN-RETs and MN-NCEs by AZT. These findings suggest that the Tk gene is necessary for the activation of AZT to species genotoxic for nuclear DNA. Other pathways, such as activation involving the mitochondrial Tk2 gene, do not contribute significantly to AZT-induced genotoxicity (at least in the nucleus). The observation that AZT produced the same MN-RET and MN-NCE frequencies in Tk+/+ and Tk+/− mice confirms that metabolism by Tk is not a rate-limiting step in the activation of AZT. Therefore, our data suggest that polymorphism affecting the activity of the Tk gene (e.g., decreasing activity by 50%) will have little effect on the clinical efficacy of AZT. Finally, Tk-deficient mice have increased levels of endogenous chromosome damage, which may be a consequence of imbalanced thymidine and/or tryptophan metabolism, so that Tk−/− mice may become an in vivo model for evaluating metabolic dysfunction. Tk+/−
Acknowledgement This research was supported, in part, by Interagency Agreement #224-93-001 between National Center for Toxicological Research/Food and Drug Administration (FDA) and the National Institute for Environmental Health Sciences/National Toxicology Program. The views presented in this article do not necessarily reflect those of the US FDA.
References [1] E.M. Connor, R.S. Sperling, R. Gelber, P. Kiselev, G. Scott, M.J. O’Sullivan, R. VanDyke, M. Bey, W. Shearer, R.L. Jacobson, Reduction of maternal–infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group, New Engl. J. Med. 331 (1994) 1173–1180. [2] N. Bhana, D. Ormrod, C.M. Perry, D.P. Figgitt, Zidovudine: a review of its use in the management of vertically-acquired pediatric HIV infection, Paediatr. Drugs 4 (2002) 515–553. [3] W.C. Copeland, M.S. Chen, T.S. Wang, Human DNA polymerases alpha and beta are able to incorporate anti-HIV deoxynucleotides into DNA, J. Biol. Chem. 267 (1992) 21459–21464. [4] K.M. Ayers, D. Clive, W.E. Tucker Jr., G. Hajian, P. de Miranda, Nonclinical toxicology studies with zidovudine: genetic toxicity tests and carcinogenicity bioassays in mice and rats, Fundam. Appl. Toxicol. 32 (1996) 148–158.
[5] M.C. Poirier, O.A. Olivero, D.M. Walker, V.E. Walker, Perinatal genotoxicity and carcinogenicity of anti-retroviral nucleoside analog drugs, Toxicol. Appl. Pharmacol. 199 (2004) 151–161. [6] S.D. Dertinger, D.K. Torous, K.R. Tometsko, Induction of micronuclei by low doses of azidothymidine (AZT), Mutat. Res. 368 (1996) 301–307. [7] K.L. Witt, R.R. Tice, G.W. Wolfe, J.B. Bishop, Genetic damage detected in CD-1 mouse pups exposed perinatally to 3 -azido-3 deoxythymidine or dideoxyinosine via maternal dosing, nursing, and direct gavage. II. Effects of the individual agents compared to combination treatment, Environ. Mol. Mutagen. 44 (2004) 321–328. [8] M.D. Phillips, B. Nascimbeni, R.R. Tice, M.D. Shelby, Induction of micronuclei in mouse bone marrow cells: an evaluation of nucleoside analogues used in the treatment of AIDS, Environ. Mol. Mutagen. 18 (1991) 168–183. [9] L.S. Von Tungeln, L.P. Hamilton, V.N. Dobrovolsky, M.E. Bishop, J.G. Shaddock, R.H. Heflich, F.A. Beland, Frequency of Tk and Hprt lymphocyte mutants and bone marrow micronuclei in B6C3F(1)/Tk+/− mice treated neonatally with zidovudine and lamivudine, Carcinogenesis 23 (2002) 1427–1432. [10] R.A. Mittelstaedt, L.S. Von Tungeln, J.G. Shaddock, V.N. Dobrovolsky, F.A. Beland, R.H. Heflich, Analysis of mutations in the Tk gene of Tk(+/−) mice treated as neonates with 3 -azido3 -deoxythymidine (AZT), Mutat. Res. 547 (2004) 63–69. [11] Q. Meng, A.J. Grosovsky, X. Shi, V.E. Walker, Mutagenicity and loss of heterozygosity at the APRT locus in human lymphoblastoid cells exposed to 3 -azido-3 -deoxythymidine, Mutagenesis 15 (2000) 405–410. [12] Q. Meng, T. Su, O.A. Olivero, M.C. Poirier, X. Shi, X. Ding, V.E. Walker, Relationships between DNA incorporation, mutant frequency, and loss of heterozygosity at the TK locus in human lymphoblastoid cells exposed to 3 -azido-3 -deoxythymidine, Toxicol. Sci. 54 (2000) 322–329. [13] P.A. Furman, J.A. Fyfe, M.H. St Clair, K. Weinhold, J.L. Rideout, G.A. Freeman, S.N. Lehrman, D.P. Bolognesi, S. Broder, H. Mitsuya, Phosphorylation of 3 -azido-3 -deoxythymidine and selective interaction of the 5 -triphosphate with human immunodeficiency virus reverse transcriptase, Proc. Natl. Acad. Sci. U.S.A. 83 (1986) 8333–8337. [14] R. Brundiers, A. Lavie, T. Veit, J. Reinstein, I. Schlichting, N. Ostermann, R.S. Goody, M. Konrad, Modifying human thymidylate kinase to potentiate azidothymidine activation, J. Biol. Chem. 274 (1999) 35289–35292. [15] B. Munch-Petersen, L. Cloos, G. Tyrsted, S. Eriksson, Diverging substrate specificity of pure human thymidine kinases 1 and 2 against antiviral dideoxynucleosides, J. Biol. Chem. 266 (1991) 9032–9038. [16] S.I. Gilles, S. Romain, P. Casellas, L. Ouafik, F. Fina, T. Combes, V. Vuaroquaux, J.F. Seitz, P. Bonnier, S. Galiegue, P. Carayon, P.M. Martin, Mutation analysis in the coding sequence of thymidine kinase 1 in breast and colorectal cancer, Int. J. Biol. Markers 18 (2003) 1–6. [17] V.N. Dobrovolsky, T. Bucci, R.H. Heflich, J. Desjardins, F.C. Richardson, Mice deficient for cytosolic thymidine kinase gene develop fatal kidney disease, Mol. Genet. Metab. 78 (2003) 1–10.
V.N. Dobrovolsky et al. / Mutation Research 570 (2005) 227–235 [18] S.D. Dertinger, Y. Chen, R.K. Miller, K.J. Brewer, T. Smudzin, D.K. Torous, N.E. Hall, K.A. Olvany, F.G. Murante, C.R. Tometsko, Micronucleated CD71-positive reticulocytes: a blood-based endpoint of cytogenetic damage in humans, Mutat. Res. 542 (2003) 77–87. [19] D.K. Torous, N.E. Hall, F.G. Murante, S.E. Gleason, C.R. Tometsko, S.D. Dertinger, Comparative scoring of micronucleated reticulocytes in rat peripheral blood by flow cytometry and microscopy, Toxicol. Sci. 74 (2003) 309–314. [20] S.D. Dertinger, D.K. Torous, N.E. Hall, F.G. Murante, S.E. Gleason, R.K. Miller, C.R. Tometsko, Enumeration of micronucleated CD71-positive human reticulocytes with a single-laser flow cytometer, Mutat. Res. 515 (2002) 3–14. [21] J.L. Weaver, D. Torous, Flow cytometry assay for counting micronucleated erythrocytes: development process, Methods 21 (2000) 281–287. [22] J.T. MacGregor, C.M. Wehr, P.R. Henika, M.D. Shelby, The in vivo erythrocyte micronucleus test: measurement at steady state increases assay efficiency and permits integration with toxicity studies, Fundam. Appl. Toxicol. 14 (1990) 513–522.
235
[23] V.N. Dobrovolsky, P.B. McKinzie, J.G. Shaddock, R.A. Mittelstaedt, R.H. Heflich, B.L. Parsons, Pms2 deficiency results in increased mutation in the Hprt gene but not the Tk gene of Tk(+/−) transgenic mice, Mutagenesis 18 (2003) 365– 370. [24] B. Schuettengruber, A. Doetzlhofer, K. Kroboth, E. Wintersberger, C. Seiser, Alternate activation of two divergently transcribed mouse genes from a bidirectional promoter is linked to changes in histone modification, J. Biol. Chem. 278 (2003) 1784–1793. [25] S.T. Koury, M.C. Bondurant, G.L. Semenza, M.J. Koury, The use of in situ hybridization to study erythropoietin gene expression in murine kidney and liver, Microsc. Res. Technol. 25 (1993) 29–39. [26] J.R. Moffett, M.A. Namboodiri, Tryptophan and the immune response, Immunol. Cell Biol. 81 (2003) 247–265. [27] R. Schwarcz, R. Pellicciari, Manipulation of brain kynurenines: glial targets, neuronal effects, and clinical opportunities, J. Pharmacol. Exp. Ther. 303 (2002) 1–10.