Toward a Biologically Based Dose-Response Model for Developmental Toxicity of 5-Fluorouracil in the Rat: Acquisition of Experimental Data

Christopher Lau*,1, M. Leonard Mole*, M. Frank Copeland*, John M. Rogers*, Robert J. Kavlock*, Dana L. Shuey{dagger}, Annie M. Cameron{ddagger}, David H. Ellis{ddagger}, Tina R. Logsdon{ddagger}, Jennifer Merriman{ddagger} and R. Woodrow Setzer**

* Reproductive Toxicology Division and ** Experimental Toxicology Division, National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711; {dagger} DuPont Merck Pharmaceutical, Newark, Delaware; and {ddagger} Mantech Environmental Sciences, Research Triangle Park, North Carolina

Received May 23, 2000; accepted September 22, 2000


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Biologically based dose-response (BBDR) models represent an emerging approach to improving the current practice of human health-risk assessment. The concept of BBDR modeling is to incorporate mechanistic information about a chemical that is relevant to the expression of its toxicity into descriptive mathematical terms, thereby providing a quantitative model that will enhance the ability for low-dose and cross-species extrapolation. Construction of a BBDR model for developmental toxicity is particularly complicated by the multitude of possible mechanisms. Thus, a few model assumptions were made. The current study illustrates the processes involved in selecting the relevant information for BBDR modeling, using an established developmental toxicant, 5-fluorouracil (5-FU), as a prototypic example. The primary BBDR model for 5-FU is based on inhibition of thymidylate synthetase (TS) and resultant changes in nucleotide pools, DNA synthesis, cell-cycle progression, and somatic growth. A single subcutaneous injection of 5-FU at doses ranging from 1 to 40 mg/kg was given to pregnant Sprague-Dawley rats at gestational day 14; controls received saline. 5-FU was absorbed rapidly into the maternal circulation, and AUC estimates were linear with administered doses. We found metabolites of 5-FU directly incorporated into embryonic nucleic acids, although the levels of incorporation were low and lacked correlation with administered doses. On the other hand, 5-FU produced dose-dependent inhibition of thymidylate synthetase in the whole embryo, and recovery from enzyme inhibition was also related to the administered dose. As a consequence of TS inhibition, embryonic dTTP and dGTP were markedly reduced, while dCTP was profoundly elevated, perhaps due to feedback regulation of intracellular nucleotide pools. The total contents of embryonic macromolecules (DNA and protein) were also reduced, most notably at the high doses. Correspondingly, dose-related reductions of fetal weight were seen as early as GD 15, and these deficits persisted for the remainder of gestation. These detailed dose-response parameters involved in the expression of 5-FU developmental toxicity were incorporated into mathematical terms for BBDR modeling. Such quantitative models should be instrumental to the improvement of high-to-low dose and cross-species extrapolation in health-risk assessment.

Key Words: mammalian developmental toxicity; BBDR modeling; 5-fluorouracil; thymidylate synthetase inhibition; dNTP imbalance; limb dysmorphogenesis.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Mammalian development is a complex process that involves intricate genetic programs directing cell replication, migration, differentiation, and death through a myriad of biochemical signals that control inter- and intracellular communication, and a variety of regulatory and compensatory mechanisms that serve to maintain homeostasis within the embryo and fetus. Adverse developmental outcomes, such as anatomical malformations and physiological dysfunction, reflect breakdown of a number of these critical processes. On the other hand, current practice of human health risk assessment regarding the potential developmental toxicity of a chemical entails estimation of a no-observable-adverse-effect level (NOAEL) or a benchmark dose (BMD). These are based on alterations of reproductive outcomes and incidence of malformations derived from an animal model, and adjustment by a set of uncertainty factors that account for interspecies and human variability (Crump, 1984Go; U.S. Environmental Protection Agency, 1991Go). Although advances have been made to improve the description of the dose-response relationship and estimation of the toxicity threshold, risk assessment is still hampered by the need to extrapolate experimental values derived from studies using high doses, (for observable toxic responses) to realistic human exposures that are typically at low levels, and to extrapolate the same values across species from laboratory animal models to humans.

Biologically based dose-response modeling is an emerging approach to reducing the uncertainties inherent in the risk-assessment process. The concept, depicted in Figure 1Go, is to incorporate the necessary biological information into a model that captures the sequence of key events involved in disruption of homeostasis, with mathematical terms that describe the linkage of these biological processes, thereby improving the ability for low-dose and cross-species extrapolation. Ideally, physiologically based pharmacokinetic (PBPK) data (pharmacokinetic linkage) should be applied for the dose component of the model, to facilitate direct comparisons of exposure levels between tissues and across species. Pertinent biomarkers based on mechanisms of toxicity (pharmacodynamic linkage) should be evaluated for the response component to enhance the sensitivity of detection for adverse biological outcomes (O'Flaherty, 1997Go). Although this concept is appealing, the actual practice of model construction proves to be daunting. For developmental toxicity, at least a dozen or so general categories of mechanisms have been identified to elicit various untoward cellular responses, leading to tissue damage and adverse pregnancy outcomes (Wilson, 1973Go). This list of toxic mechanisms will undoubtedly be expanded as our understanding of developmental biology is advanced. It would also be naive to assume that a chemical would act through only a single mechanism. From a toxicologist's standpoint, one must ask which and how many mechanism(s) should be embraced for modeling?



View larger version (50K):
[in this window]
[in a new window]
 
FIG. 1. Schematic representation of a biologically based dose-response model. Boxes indicate parameters to be measured for construction of the model.

 
In order to make the exercise of BBDR model construction tenable, a few a priori assumptions must be made (and tested):

Applying these principles, we have previously described a BBDR model comprising the biochemical and cellular sequelae of 5-fluorouracil (5-FU) exposure during embryonic development, with specific reference to the drug-induced digital defects (Shuey et al., 1994bGo). 5-FU is a chemotherapeutic agent and a teratogen in a number of laboratory animal species, including rat, mouse, chick, hamster, and monkey (Dagg, 1960Go; Karnofsky et al., 1958Go; Shah and MacKay, 1978Go; Shuey et al., 1994bGo; Wilson, 1971Go; Wilson et al., 1969Go). Clinically, a single case of multiple congenital anomalies in a fetus exposed to 5-FU during the first trimester of pregnancy was reported by Stephens et al. (1980), and the lesions described were generally consistent with those seen in the animal models. 5-FU acts by at least 3 known mechanisms (reviews: Parker and Cheng, 1990; Pinedo and Peters, 1988; Valeriote and Santelli, 1984): first, inhibition of thymidylate synthetase, the enzyme that catalyzes the conversion of uridylate to thymidylate (Danenberg and Lockshin, 1981Go); second, incorporation of FUTP (as substitute for UTP) into RNA, producing miscoding in transcription and leading to translation of nonfunctional proteins (Dolnick and Pink, 1985Go; Glazer and Lloyd, 1982Go; Lenz et al., 1994Go), and third, incorporation of FdUTP into DNA, requiring excessive repairs and resulting in damage due to fragmentation (Kufe et al., 1983Go; Parker et al., 1987Go; Schuetz et al., 1984Go). In addition to these 3 mechanisms, drug-induced fetal anemia could also contribute to 5-FU developmental toxicity (Shuey et al., 1994cGo; Zucker et al., 1995Go). A composite scheme of putative mechanisms for 5-FU developmental toxicity is depicted in Figure 2Go. In a previous paper, Shuey et al. (1994b) constructed a BBDR model primarily for the 5-FU-induced hind-limb dysmorphogenesis. However, 5-FU produces a gamut of anatomical abnormalities and defects, each possessing a slightly different dose-response profile. Thus, instead of providing a model for each of the malformations, the question remains as to whether a BBDR model can be extended to the whole-embryo level in order to capture the embryotoxicity of 5-FU at large, rather than specific insults of one anatomical structure.



View larger version (45K):
[in this window]
[in a new window]
 
FIG. 2. A composite scheme depicting the putative mechanisms responsible for developmental toxicity of 5-FU. Shaded boxes denote the primary pathway chosen for modeling.

 
The current study therefore attempts to demonstrate the processes involved in building a BBDR model to assess the developmental toxicity of a chemical, using 5-fluorouracil as a prototype. While this paper focuses on the acquisition of an experimental database for the model, an accompanying paper (Setzer et al., 2001Go) will address the mathematical and statistical issues related to model construction.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Animal treatment.
Timed-pregnant Sprague-Dawley rats were obtained from Charles River Laboratory (Raleigh, NC), shipped to our laboratory one week after mating, and housed individually in breeding cages with food and water available ad libitum. For the initial studies, male and female pairs were left in the same cage overnight for breeding. Subsequently, other pairs were allowed to mate during a 4-h period in the evening, in order to narrow the range of gestational stages among the dams. Presence of a copulatory plug was designated as gestational day (GD) 0. On GD 13, rats were ranked by body weight and distributed randomly to various treatment groups within the weight strata. On GD 14, these animals were given 5-fluorouracil (5-FU; Sigma, St. Louis, MO) subcutaneously with doses ranging from 1 to 40 mg/kg. Controls received injections of saline equal in volume to the 5-FU injections (2 ml/kg). Experiments were divided into several blocks, each consisting of a control group and several treatment groups of various 5-FU doses. For each 5-FU dose, measurements were pooled from at least 2 separate blocks of animals to provide a total of 8–12 litters. At intervals ranging from 1 to 72 h after drug administration, rats were sacrificed by CO2 asphyxiation. In a separate study with the 4-h breeding scheme, rats were given 5-FU (20, 30, or 40 mg/kg) on GD 14 and sacrificed on GD 15, 16, 17, or 21; weights of the fetuses were compared to those of saline-treated controls.

Biochemical assays.
Four embryos were removed randomly from each dam, placed individually in 1 ml of ice-cold 125-mM Tris buffer containing 150 mM KCl (pH 7.5) and homogenized immediately by polytron. In addition, maternal spleens were dissected, weighed, and homogenized in 20 volumes of Tris buffer. A 0.2-ml aliquot of the tissue homogenate was removed for DNA and protein determination; total protein was determined in duplicate 25 µl aliquots of tissue homogenate by the dye-binding method of Bradford (1976), and DNA content was measured in duplicate 5 µl samples using a modified method of LaBarca and Paigen (1980) as described previously by Abbott and Pratt (1988). The remainder of the homogenate was sedimented for 30 min at 32,000 x g, and duplicate 50 µl aliquots of the supernatant solution were withdrawn for the thymidylate-synthetase assay, using a modified method of Roberts (1966) that measured the release of tritiated water from enzyme reaction substrate deoxyuridine 5'-monophosphate, after a methylation step to produce thymidylate. The incubation mixture contained 50 µl of reagent containing 0.86 mM tetrahydrofolate, 75 mM mercatpoethanol, 235 mM tris, 0.43% formaldehyde (pH 7.5), 25 µl of 1 M NaF, 10 µl of 2.8 mM deoxyuridine 5'-monophosphate (dUMP), 65 µl of 1 M Tris (pH 7.5) and 0.4 µCi of deoxy[3H]uridine 5'-monophosphate (specific activity 16–19 Ci/mmol., Amersham). Samples were incubated at 37°C for 1 h and pooled supernatant samples, incubated at 0°C for 1 h, served as assay blanks. Reaction was terminated on ice and by addition of 100 µl of 40% trichloroacetic acid. 50 µl of dUMP (5 mg/ml) was then added to each sample, followed by 1.8 ml of activated charcoal (100 mg/ml). The mixture was sedimented at 500 x g for 15 min, and 0.5 ml of the supernatant was withdrawn and counted by liquid scintillation spectrometry. Enzyme activity is expressed as nmols 3H2O released/mg DNA/h.

For pharmacokinetic studies, 2 blocks of 3 to 4 pregnant rats from the 4-h breeding scheme were given 10, 20, or 40 mg/kg of 5-FU on GD 14. For one block of rats, blood samples were collected from the tail vein at 5, 15, 25, and 30 min after injection; for another block, samples were collected at 10, 20, 45, and 60 min. The serum fraction was separated from whole blood and stored frozen at –80°C until analysis. Samples were thawed and extracted by the method of Barberi-Heyob et al. (1992). In brief, 500 µl of serum was added to a Chem-Elut column (Analytichem International, Harbor City, CA) for 5 min, followed by elution with 3 aliquots of ethyl-acetate-methanol (95:5, v/v). The eluent was concentrated by evaporation under nitrogen and reconstituted with 200 µl of water. The extracted samples were analyzed for 5-FU by a modified method of Schwartz et al. (1992). 5-FU was separated from its metabolites with an adsorbosphere 5 µm C18 column (250 mm x 4.6 mm, Alltech, Deerfield, IL) isocratically using a mobile phase of 5 mM tetrabutylammonium hydrogen sulfate and 5 mM potassium phosphate (pH 7), and detected by UV spectrophotometry at 268 nm. Recovery of 5-FU under these conditions was typically greater than 95%. Results were expressed as µg of 5-FU/ml serum. In an attempt to measure the 5-FU content in whole embryos, 2 embyros were pooled and deproteinized in 0.5 ml of 0.3 M perchloric acid, and the supernatant was analyzed for 5-FU by HPLC, as described above. However, an interfering peak co-eluted with 5-FU that was also detected in saline-treated controls, thereby precluding accurate assessment of 5-FU content in the embryonic tissue.

Deoxyribonucleotide (dNTP) determination in whole embryos was also performed with pregnant rats from the 4-h breeding scheme. On GD 14, 6 to 8 rats were given either 10, 20, or 40 mg/kg 5-FU, and animals were sacrificed at 0.5, 1, 2, 4, 8, or 12 h after injection. Four embryos were removed randomly from each dam, weighed, homogenized (sonication) individually, and deproteinized in 300 µl of 0.3 M perchloric acid and stored frozen at –80°C until analysis by an HPLC method described previously (Mole et al., 1998Go). Duplicate samples were measured from each dam to ensure reproducibility, and data were expressed as pmol of dNTP/mg embryonic tissue.

To evaluate the extent of 5-FU incorporation into nucleic acid, 0.5–1 mCi/kg of 5-FU-6–3H (specific activity 20 Ci/mmol, Sigma) was isotopically diluted to 20 or 40 mg/kg and given to pregnant rats on GD 14. Rats were sacrificed by asphyxiation 8 h later, and embryos were removed. Two embryos were pooled for each sample, and duplicate samples from each dam were assayed. Embryos were homogenized (polytron) in 3.5 ml of 4 M guanidine isothiocyanate buffer. A 0.75-ml aliquot of homogenate was removed and counted by liquid scintillation spectrometry. Homogenate (2.4 ml) was then loaded onto 1.7 ml of CsCl2 buffer (0.96 g/ml) for overnight centrifugation at 100,000 x g. The DNA fraction from the gradient was collected, precipitated, and washed with 6 ml of 70% ethanol, and the resultant pellet was dissolved in 1 ml of water. The remaining CsCl2 buffer was removed and the RNA pellet was dissolved in 0.5 ml of water. Aliquots of DNA and RNA extract were removed, and the nucleic acid concentrations of each sample was determined by spectrophotometry at 260 nm. The remaining fractions were counted by liquid scintillation spectrometry, and incorporation of 5-FU into nucleic acid was calculated as follows:

To determine whether developmental toxicity of 5-FU was associated with alterations of specific embryonic proteins, pregnant rats were given either saline (control) or 5-FU (20 or 40 mg/kg) on GD 14, embryos were removed either 12 or 24 h afterward, and the embryonic protein profiles were analyzed by 2-D electrophoresis. Duplicate samples of 2 embryos from each dam were homogenized in 8 volumes of buffer containing 9 M urea, 2% NP-40 detergent, 0.5% dithiothreitol, and 2% ampholytes (pH 9–11). The homogenates were stored frozen at –80°C, thawed, centrifuged at 485,000 x g for 15 min, refrozen, and shipped to Large Scale Biology Co (Rockville, MD) for analysis according to the method described by Anderson and Anderson (1978a,b). In brief, solubilized proteins were separated by a first-dimension isoelectric focusing gel using ampholytes (BDH-4–8A) and a second-dimension SDS-gradient slab gel, and stained by Coomassie Blue G-250. The staining intensity was captured by computer imaging software and processed for quantitative analysis. The protein maps from the 5-FU-treated embryos were compared to those of corresponding controls, and we discerned differences in protein expression by migration pattern and intensity of the spots.

Statistical analysis.
Data were computed with a linear model, treating each experimental day as a random effect (PROC MIXED, SAS, 1990), and expressed as means and 95% confidence intervals. Litter means were used for analysis to eliminate the variability among embryos within each dam. The cell size of each time and dose point were not balanced due to experimental limitations; typical intra- and inter-block variabilities of each measurement were reflected by the 95% confidence interval estimation in the control groups. Because of the heterogeneity of variance and to facilitate the expression of contrasts as ratios to the corresponding control means, values were subject to log-transformation before statistical evaluation, and the resulting contrasts were back-transformed for graphing. Pharmacokinetic data were analyzed as mixed-effect linear models using "lme" in S+ version 3.4 (Mathsoft, Seattle, WA), treating results from the same dam as repeated measures. Curves were fitted using the method of nonlinear mixed-effect models of Lindstrom and Bates (1990).


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
5-FU was absorbed readily into the maternal circulation, reaching peak levels between 10 and 20 min after subcutaneous injection (Fig. 3Go) and declined rapidly, becoming undetectable by 90 min (data not shown). Half-lives of 5-FU did not vary appreciably between different dose groups, ranging from 12.2 to 14.7 min; and AUC increased proportionally with administered dose.



View larger version (37K):
[in this window]
[in a new window]
 
FIG. 3. Serum concentrations of 5-FU in pregnant rats at GD 14 after a single subcutaneous injection at t = 0. Each point and bar represent means and 95% lower and upper confidence intervals of 6–8 rats. Pharmacokinetic parameters are summarized in the box.

 
Thymidylate synthetase (TS) activity in whole embryos of control rats, expressed as nmol/mg DNA, did not vary significantly from GD 14 to GD 17 (Table 1Go). Inhibition of TS by 5-FU was detected even in the lowest-dose group (1 mg/kg), although the effect was small (~14%) and transient. With increasing doses, the magnitude and duration of enzyme inhibition amplified, reaching a maximal effect (ca. 70–80% inhibition) at 40 mg/kg. This pattern of dose-dependent enzyme inhibition is best-illustrated in Figure 4Go, where samples were obtained from the breeding regimen of the 4-h evening cohabitation of male and female rats. Recovery of TS inhibition was generally seen 24 h after 5-FU treatment; while a small overshoot of enzyme activity was in fact noted in the high-dose groups (Table 1Go).


View this table:
[in this window]
[in a new window]
 
TABLE 1 Effects of 5-FU on Thymidylate Synthetase in the Whole-Rat Embryo
 


View larger version (24K):
[in this window]
[in a new window]
 
FIG. 4. Effects of 5-FU on thymidylate synthetase activity in the whole-rat embryos. Rats were bred during a 4-h period. Data are expressed as percentages of controls. TS activities in control embryos were similar to those summarized in Table 1Go. Each point represents the mean value of 6–8 dams. The bar indicates a typical 95% lower and upper confidence intervals.

 
Concentrations of the 4 deoxyribonucleotides (dNTP) in a GD 14 embryo were not uniform; however, the levels of individual dNTP remained fairly stable over the period of observation (12 h) (Table 2Go). Diverse patterns of change among dNTPs were obtained with 5-FU treatment (Fig. 5Go). dTTP and dGTP were depleted rapidly in a dose-dependent manner; reaching a nadir between 2 and 4 h, respectively, after drug administration, in a time course that corresponded to TS inhibition. However, whereas a substantial enzyme inhibition was noted in the 10 mg/kg dose group, little to no alterations of these 2 dNTPs were seen; in contrast, marked depletions (~50%) were seen in the 40 mg/kg dose group. Duration of the deficit also varied somewhat between the 2 dNTPs, with depletion of dGTP being longer-lasting; although both dNTPs appeared to recover to control levels by 12 h. In stark contrast, embryonic concentrations of dCTP were elevated sharply (~5-fold) by 5-FU, in a dose-dependent fashion. This effect could be seen even with the lower-dose group (10 mg/kg) and appeared to have reached the maximal level at 20 mg/kg. The elevations of dCTP were also longer lasting than the depletions of dTTP and dGTP; no discernable changes in dATP were observed among the 5-FU dose groups.


View this table:
[in this window]
[in a new window]
 
TABLE 2 Deoxyribonucleotide Levels in Whole Embryos of Control Rats at GD 14
 


View larger version (33K):
[in this window]
[in a new window]
 
FIG. 5. Individual profiles of dNTP in GD 14 whole rat embryos after a single exposure to 5-FU at t = 0. Data are expressed as a percentage of controls. dNTP values in control embryos at individual time points are summarized in Table 2Go. Each point represents mean value of 6–8 dams. Bars indicate typical 95% confidence intervals for the specific dNTP. Note the different scale of percent change for dCTP.

 
Total DNA in the whole embryo rose from 1.7 mg on GD 14 to 4.8 mg on GD 17 (Table 3Go), and total protein from 6 to 31 mg (Table 4Go). Although a trend toward lower DNA and protein contents was apparent in the 5-FU-exposed embryos, significant changes were noted only in the high-dose groups (30 and 40 mg/kg). Correspondingly, a dose-dependent reduction of fetus weight was noted as early as GD 15 (24 h after drug treatment), although the effect reached statistical significance only in the 40-mg/kg dose group (p < 0.05, Table 5Go). However, these weight deficits (even those in the lower-dose groups) were not recoverable during ensuing fetal growth, leading to more notable changes at term (GD 20).


View this table:
[in this window]
[in a new window]
 
TABLE 3 Effects of 5-FU on Whole Embryo DNA
 

View this table:
[in this window]
[in a new window]
 
TABLE 4 Effects of 5-FU on Whole Embryo Protein
 

View this table:
[in this window]
[in a new window]
 
TABLE 5 Effects of 5-FU on Fetus Weight
 
Direct incorporation of 5-FU into embryonic nucleic acids on GD 14 was determined 8 h after administration of the radioactive tracer. The extent of incorporation into RNA and DNA was corrected for the amount of total uptake (in the embryo homogenate) for each sample and expressed as percent incorporated. As shown in Table 6Go, less than 2% of 5-FU was incorporated into embryonic RNA. Incorporation into embryonic DNA was even lower, about 6 times less than that into RNA, and the amounts were not influenced by varying the administered dose. To determine the extent of 5-FU-induced erroneous nucleotide substitution (UTP by FUTP and dTTP by FdUTP) into embryonic nucleic acids, the RNA and DNA contents for each sample were determined, and the 4 species of NTP and dNTP were assumed to be distributed equally among RNA and DNA respectively. Under these assumptions, approximately 1 in 5–650 UTP was substituted by FUTP into RNA and 1 in 800–1600 dTTP by FdUTP into DNA of the embryo (Table 6Go).


View this table:
[in this window]
[in a new window]
 
TABLE 6 Incorporation of 5-FU Metabolites into Embryonic DNA and RNA
 
To determine if 5-FU might alter a specific set of embryonic proteins, 2-D electrophoresis was performed with rats treated with 20 or 40 mg/kg of the drug for 12 or 24 h, and results were compared to corresponding controls. From the 6 groups of rats, 199 protein spots were matched in all but one pattern of all but one group (data not shown). Overall, analysis of data generated in this study revealed relatively small drug-induced effects that were overshadowed by the extent of genetic heterogeneity. For the 12-h time point, only 2 spots from the 20-mg/kg group and 4 spots from the 40-mg/kg groups were different from the controls (p < 0.001); these two sets had no spots in common. For the 24-h time point, only 1 spot from the 20-mg/kg group and 5 spots from the 40-mg/kg group were different from the controls; these two sets of spots also did not overlap.

To compare the effects of 5-FU between embryonic and maternal tissues, TS activity in the spleen of the pregnant rats were evaluated. TS activity in maternal spleens ranging from 1.6 to 3.9 nmol/mg DNA/h (Table 7Go), was substantially lower than that in the embryo. As seen in the embryonic tissues, 5-FU produced dose-dependent enzyme inhibition at doses as low as 1 mg/kg, although prominent effects were generally observed at and above 20 mg/kg.


View this table:
[in this window]
[in a new window]
 
TABLE 7 Effects of 5-FU on Maternal Spleen Thymidylate Synthetase Activity
 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The serum half-life estimates for 5-FU in this study are consistent with previous reports with either pregnant (Boike et al., 1989Go) or nonpregnant rats (Au et al., 1983Go), whereas the lower AUC estimates seen in the current study (compared to the previous 2 reports) perhaps reflect the different routes of drug administration (subcutaneous vs. intravenous) and, therefore, different rates of drug absorption. Nonlinear kinetics has been suggested for the elimination of 5-FU from the circulation; thus, drug metabolism (hepatic and renal) can be saturated at either high doses or rapid rate of administration. In a dose-response study with pregnant rats at GD 20, Boike and co-workers (1989) showed that at lower doses (10 or 25 mg/kg) kinetic parameters in the dams were fairly linear, but at a higher dose (100 mg/kg) the half-life of 5-FU almost doubled, and the AUC exceeded the predicted value by over 50%. However, kinetic parameters from our study (using subcutaneous administration) indicated that even at 40 mg/kg, elimination of 5-FU remained linear and that AUC increased proportionally with the administered doses. Thus, for practical purposes, linear pharmacokinetics can be assumed in the maternal compartment of our model. In their study, Boike et al. (1989) made the interesting discovery that nonlinear kinetics for 5-FU elimination occurred at a lower dose in the GD 20 fetus than in the dam. Unfortunately, because of technical limitations in our study, 5-FU could not be measured accurately in the GD 14 embryos, preventing a direct comparison of results from these investigators. Should the embryo be also susceptible to nonlinear kinetics, as noted in the fetus, the actual 5-FU exposure levels in the embryonic compartment would have to be adjusted accordingly.

While ample evidence has indicated that all 4 putative mechanisms depicted in Figure 2Go are likely to contribute to the expression of 5-FU developmental toxicity, we chose only one of these for the construction of a BBDR model in accordance with the model assumptions. Although direct incorporation of the 5-FU metabolite (FUTP) into RNA has been shown in a variety of cell culture systems, concentrations of 5-FU found to be associated with cytotoxicity were typically high, in the mM range, resulting in approximately 1–2% substitution for the natural substrate UTP (Dolnick and Pink, 1985Go; Glazer and Lloyd, 1982Go). Doses employed in our in vivo study were comparatively lower, embryonic exposure being estimated at the µmol/kg range. Indeed, the extent of 5-FU tracer incorporation into embryonic RNA in our study was considerably lower, with substitution for UTP being estimated at only 0.2%. Although transcriptional miscoding leading to production of nonfunctional proteins was hypothesized with direct incorporation of FUTP into RNA, recent findings have brought this hypothesis into doubt. Takimoto and co-workers (1993) transcribed mRNA with 100% substitution with 5-FU but did not note any abnormality with the translation product. Schuetz et al. (1984) also could not correlate incorporation of 5-FU into RNA with toxicity. Furthermore, although the 2-D electrophoresis experiment in our study provided only a relatively crude screen for protein abnormalities, the paucity of changes noted between controls and 5-FU-exposed embryos also did not support any significant aberrant production of proteins by the drug treatment. Thus, while the mechanisms and impact of RNA-directed actions of 5-FU involved in the overall expression of cytotoxicity remains to be addressed, the contribution of this pathway to embryotoxicity in the current study was likely nominal. On the other hand, the extent of erroneous incorporation of FdUTP into embryonic DNA noted in the present study (300–700 fmol/µg) was comparable to that observed in CF-1 mouse bone-marrow cells where cytotoxicity was associated (Scheutz et al., 1984). However, the net effect of erroneous DNA incorporation in the whole embryo is difficult to ascertain because it is dependent on the capacity and effectiveness of the DNA repair machinery. Excessive DNA damage due to saturation of these repair mechanisms has been linked to fragmentation of newly synthesized DNA and cell death (apoptosis) (Curtin et al., 1991Go; Golos and Malec, 1989Go; Holliday, 1985Go; Snyder, 1988Go). Although 5-FU-induced apoptosis has been reported recently in several tumor cells (Guchelaar et al., 1998Go; Inada et al., 1997Go; Koshiji et al., 1997Go; Okamoto et al., 1996Go) and rat tissues (Sakaguchi et al., 1994Go), the extent of excessive cell death in embryonic tissues caused by DNA damage remains to be evaluated carefully. 5-FU-induced fetal anemia observed in previous studies (Shuey et al., 1994cGo; Zucker et al., 1995Go) was likely a secondary effect arising from fetal hepatotoxicity because the fetal liver serves as a primary source of newly synthesized erythrocytes at this stage of development. Indeed, Shuey and co-workers (1994a) demonstrated that early biochemical events that occurred rapidly within 3 h of 5-FU administration were sufficient to produce growth retardation and dysmorphogenesis. Thus, although physiological deficiency such as fetal anemia is of significant note to the overall picture of 5-FU developmental toxicity, it should probably be modeled as an adjuvant to the primary pathway of TS inhibition and nucleotide pool imbalance.

Inhibition of TS and the attendant dNTP imbalance in the whole embryos provided sensitive endpoints to monitor the dose-response effects of 5-FU, and the temporal profiles of changes afforded a concentration x time (c x t) estimation for mathematical modeling. Indeed, while 5-FU inhibited embryonic TS in a dose-dependent fashion, the manner in which enzyme activities recovered perhaps provided a more significant clue to the drug's toxic potentials. Thus, the combination of duration as well as magnitude of enzyme inhibition might confer the overall toxicity. Mathematical modeling of these data (in the accompanying paper) may shed light on this issue. It is also heartening to note that the dose-response profile of TS inhibition in the maternal spleen (a tissue undergoing rapid cell replication) resembled that in the embryos, suggesting that the enzyme was equally sensitive to 5-FU regardless of tissue or cell type, thus lending credence to extrapolation based on this mechanistic step.

While the effect of 5-FU on de novo synthesis of dNTP mediated by TS inhibition is well established, the influences of the drug on the nucleotide salvage pathways have received less attention. For instance, thymidine kinase, an enzyme involved in maintaining the level of dTTP, was reported to be inhibited by 5-FU as well (Nord and Martin, 1991Go; Tsukamoto and Kojo, 1991Go). Thus, the profiles of individual dNTPs in the whole embryo would provide a clearer and more definite picture of the 5-FU effects. Results from our present study are consistent with a previous report where embryonic dNTP was determined at a single time point after 5-fluoro-deoxyuridine treatment (Ritter et al., 1980Go). Corresponding to TS inhibition, there was an abrupt fall in the concentration of dTTP, reaching a nadir of over 50% deficit at 2 h after drug administration. Two points are noteworthy from these data. Little change in dTTP was seen with the 10 mg/kg group despite effective TS inhibition (although the extent was less than the higher-dose groups), and the time course of recovery was faster than that of TS, suggesting a significant participation of nucleotide salvage and feedback control mechanisms (review: Lau, 1997). Corresponding to the intracellular nucleotide feedback regulation, dGTP levels fell along with dTTP, while the negative feedback control of dCTP by dTTP was repressed, leading to a dramatic rise of dCTP. It is interesting to note that while dTTP is most proximal to the 5-FU action (TS inhibition), the magnitude and duration of changes in dGTP and dCTP were greater than those in dTTP. Embryonic dATP, on the other hand, was not affected significantly by 5-FU. Because little information is available to discern the relative importance of specific dNTPs in DNA synthesis and cell replication, a direct biochemical correlation of changes in an individual dNTP to embryotoxicity is not readily attainable. Nonetheless, Yoshioka and co-workers (1985, 1987) reported that cytotoxicity of fluorodeoxyuridine in mouse mammary tumor cells correlated with dNTP pool imbalance and fragmentation of parental DNA (double-strand breaks). These researchers noted that inhibition of protein synthesis prevented DNA fragmentation and attenuated the cytotoxicity, and therefore hypothesized that dNTP imbalance induced an endonuclease that destroyed the DNA and resulted in cell death. Indeed, they were able to isolate a DNA endonuclease in extracts from the cells exposed to fluorodeoxyuridine but not in the untreated cells. Interestingly, Hirota et al. (1989) also reported DNA double-strand breaks and cell death associated with dNTP imbalance caused by another agent, 2-chlorodeoxyadenosine. Thus, activation of an endonuclease by dNTP pool imbalance may be a general mechanism for a number of agents to induce apoptosis.

Total DNA and protein contents in the whole embryo appeared to be less sensitive indicators of 5-FU toxicity. Notable changes were obtained only with the high doses. However, it is worth noting that reductions of these parameters were persistent and correlated well with the deficits of fetus weights in these dose groups. Because 5-FU should affect only the dividing cells within the embryo and for a relatively short duration, the lack of dramatic changes in total macromolecule contents may not be too surprising. A more accurate measurement to indicate the 5-FU effect was, perhaps, biosynthesis of DNA and protein. However, because thymidine kinase is known to be altered by 5-FU (Bagrij et al., 1993Go; Tsukamoto and Kojo, 1991Go), results from the conventional method of labeled thymidine uptake and incorporation into DNA would be confounded. Indeed, preliminary studies in our laboratory indicated a small increase of labeled thymidine incorporated into DNA shortly after 5-FU administration (data not shown), perhaps reflecting an acute response of the salvage pathway rather than the rate of de novo DNA synthesis. Another method of assessing DNA synthesis using BrdU was limited because of the 5-FU-induced TS inhibition.

In summary, pharmacokinetic and pharmacodynamic information of a chemical, as well as the biochemical alterations leading to the breakdown of homeostasis (toxicodynamics) within a developing organism (embryo), can be evaluated and used for the construction of a BBDR model of developmental toxicity. Indeed, data derived from the maternal serum concentrations of 5-FU, the TS activities, and the dNTP levels in the embryos, as well as the fetus weights reported in this study, were selected for the construction of a mathematical model in the accompanying paper by Setzer et al. (2001) that would describe the developmental toxicity of this chemotherapeutic agent. Although 5-FU was used as a prototypic agent in this study, the correlation of molecular and cellular changes with toxicity should reflect an inherent property of the developing organism (embryo) that can be verified and extended to other potential developmental toxicants.


    ACKNOWLEDGMENTS
 
The research reported in this document has been funded wholly by the U.S. Environmental Protection Agency. The authors wish to thank Judith E. Schmid for her skillful assistance in data handling and analysis. D.L.S. was funded by the EPA/UNC Toxicology Research Program, Training Agreement T901915, with the Curriculum in Toxicology, University of North Carolina at Chapel Hill.


    NOTES
 
This report has been subjected to review by the National Health and Environmental Effects Research Laboratory and approved for publication. Approval does not signify that the contents reflect the views of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use.

1 To whom correspondence should be addressed at Mail Drop 67, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. Fax: (919) 541-4017. E-mail: lau.christopher{at}epa.gov. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Abbott, B. A., and Pratt, R. M. (1988). Influence of retinoids and EGF on growth of embryonic mouse palatal epithelia in culture. In Vitro Cell. Dev. Biol. 24, 343–352.[ISI][Medline]

Anderson, N. G., and Anderson, N. L. (1978a). Two-dimensional analysis of serum and tissue proteins: Multiple isoelectric focusing. Anal. Biochem. 85, 331–340.[ISI][Medline]

Anderson, N. L., and Anderson, N. G. (1978b). Two-dimensional analysis of serum and tissue proteins: Multiple gradient-slab electrophoresis. Anal. Biochem. 85, 341–354.[ISI][Medline]

Au, J. L., Walker, J. S., and Rustum, Y. (1983). Pharmacokinetic studies of 5-fluorouracil and 5'-deoxy-5-fluorouridine in rats. J. Pharmacol Exp. Ther. 227, 174–180.[Abstract]

Bagrij, T., Kralovanszky, J., Gyergyay, F., Kiss, E., and Peters, G. J. (1993). Influence of uridine treatment in mice on the protection of gastrointestinal toxicity caused by 5-fluorouracil. Anticancer Res. 13, 789–793.[ISI][Medline]

Barberi-Heyob, M., Merlin, J. L., and Weber, B. (1992). Analysis of 5-fluorouracil in plasma and urine by high-performance liquid chromatography. J. Chromatog. 581, 281–286.[Medline]

Boike, G. M., Deppe, G., Young, J. D., Malone, J. M., Malviya, V. K., and Sokol, R. J. (1989). Chemotherapy in a pregnant rat model: II. 5-Fluorouracil: Nonlinear kinetics and placental transfer. Gynecol. Oncol. 34, 191–194.[ISI][Medline]

Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254.[ISI][Medline]

Crump, K. S. (1984). A new method for determining allowable daily intakes. Fundam. Appl. Toxicol. 4, 854–871.[ISI][Medline]

Curtin, N. J., Harris, A. L., and Aherne, G. W. (1991). Mechanism of cell death following thymidylate synthase inhibition: 2'-Deoxyuridine-5'-triphosphate accumulation, DNA damage, and growth inhibition following exposure to CB3717 and dipyridamole. Cancer Res. 51, 2346–2352.[Abstract]

Dagg, C. P. (1960). Sensitive stages for the production of developmental abnormalities in mice with 5-fluorouracil. Am. J. Anat.106, 89–96.[ISI][Medline]

Danenberg, P. V., and Lockshin, A. (1981). Fluorinated pyrimidines as tight-binding inhibitors of thymidylate synthetase. Pharmacol. Ther. 13, 69–90.[ISI][Medline]

Dolnick, B. J., and Pink, J. J. (1985). Effects of 5-fluorouracil on dihydrofolate reductase and dihydrofolate reductase mRNA from methotrexate-resistant KB cells. J. Biol. Chem. 260, 3006–3014.[Abstract]

Glazer, R. I., and Lloyd, L. S. (1982). Association of cell lethality with incorporation of 5-fluorouracil and 5-fluorouridine into nuclear RNA in human colon carcinoma cells in culture. Mol. Pharmacol 21, 468–473.[Abstract]

Golos, B., and Malec, J. (1989). Enhancement of methotrexate-induced growth inhibition, cell killing, and DNA lesions in cultured L5178 cells by the reduction of DNA repair efficiency. Biochem. Pharmacol. 38, 1743–1748.[ISI][Medline]

Guchelaar, H. J., Vermes, I., Koopmans, R. P., Reutelingsperger, C. P., and Haanen, C. (1998). Apoptosis- and necrosis-inducing potential of cladribine, cytarabine, cisplatin, and 5-fluorouracil in vitro: a quantitative pharmacodynamic model. Cancer Chemother. Pharmacol. 42, 77–83.[ISI][Medline]

Hirota, Y., Yoshioka, A., Tanaka, S., Watanabe, K., Otani, T., Minowada, J., Matsuda, A., Ueda, T., and Wataya, Y. (1989). Imbalance of deoxyribonucleoside triphosphates, DNA double-strand breaks, and cell death caused by 2-chlorodeoxyadenosine in mouse FM3A cells. Cancer Res. 49, 915–919.[Abstract]

Holliday, R. (1985). Aspects of DNA repair and nucleotide pool imbalance. Basic Life Sci. 31, 453–460.[Medline]

Inada, T., Ichikawa, A., Kubota, T., Ogata, Y., Moossa, A. R., and Huffman, R. M. (1997). 5-FU-induced apoptosis correlates with efficacy against human gastric and colon cancer xenografts in nude mice. Anticancer Res. 17, 1965–1971.[ISI][Medline]

Karnofsky, D. A., Murphy, M. L., and Lacon, C. R. (1958). Comparative toxicologic and teratologic effects of 5-fluoro-substituted pyrimidines in the chick embryos and pregnant rat. Proc. Am. Assoc. Cancer Res. 2, 312–313.

Koshiji, M., Adachi, Y., Taketani, S., Takeuchi, K., Hioki, K., and Ikehara, S. (1997). Mechanisms underlying apoptosis induced by combination of 5-fluorouracil and interferon-gamma. Biochem. Biophys. Res. Commun. 240, 376–381.[ISI][Medline]

Kufe, D. W., Scott, P., Fram, R., and Major, P. (1983). Biologic effect of 5-fluoro-2'-deoxyuridine incorporation in L1210 deoxyribonucleic acid. Biochem. Pharmacol. 32, 1337–1340.[ISI][Medline]

Labarca, C., and Paigen, K. (1980). A simple, rapid, and sensitive DNA assay procedure. Anal. Biochem. 102, 344–352.[ISI][Medline]

Lau, C. (1997). Nucleotide pool imbalance. In Drug Toxicity in Embryonic Development I, Advances in Understanding Mechanisms of Birth Defects: Morphogenesis and Processes at Risk, (R.J. Kavlock and G.P. Daston, Eds.), pp. 341–372. Springer-Verlag, New York.

Lenz, H. J., Manno, D. J., Danenberg, K. D., and Danenberg, P. V. (1994). Incorporation of 5-fluorouracil into U2 and U6 snRNA inhibits mRNA precursor splicing. J. Biol. Chem. 269, 31962–31968.[Abstract/Free Full Text]

Lindstrom, M. J., and Bates, D. M. (1990). Nonlinear mixed effects models for repeated measures data. Biometrics 46, 673–687.[ISI][Medline]

Mole, M. L., Hunter, D. L., Gao, P., and Lau, C. (1998). Sample preparation and high-performance liquid chromatographic analysis of deoxyribonucleoside triphosphates in individual rat embryos. Anal. Biochem. 259, 245–252.[ISI][Medline]

Nord, L. D., and Martin, D. S. (1991). Loss of murine tumor thymidine kinase activity in vivo following 5-fluorouracil (FUra) treatment by incorporation of FUra into RNA. Biochem. Pharmacol. 42, 2369–2375.[ISI][Medline]

O'Flaherty, E. J. (1997). Pharmacokinetics, pharmacodynamics, and prediction of developmental abnormalities. Reprod. Toxicol. 11, 413–416.[ISI][Medline]

Okamoto, S., Sakai, M., Uchida, J., and Saito, H. (1996). 5-Fluorouracil induces apoptosis cell death with G2 phase arrest in human breast cancer grafted in nude mice. Anticancer Res. 16, 2699–2704.[ISI][Medline]

Parker, W. B., and Cheng, Y. C. (1990). Metabolism and mechanism of action of 5-fluorouracil. Pharmacol. Ther. 48, 381–395.[ISI][Medline]

Parker, W. B., Kennedy, K. A., and Klubes, P. (1987). Dissociation of 5-fluorouracil-induced DNA fragmentation from either its incorporation into DNA or its cytotoxicity in murine T-lymphoma (S-49) cells. Cancer Res. 47, 979–982.[Abstract]

Pinedo, H. M., and Peters, G. F. J. (1988). Fluorouracil: Biochemistry and pharmacology. J. Clin. Oncol. 6, 1653–1664.[Abstract]

Ritter, E. J., Scott, W. J., Wilson, J. G., Lampkin, B. C., and Neely, J. E. (1980). Effect of 5-fluoro-2'-deoxyuridine on deoxyribonucleotide pool in vivo. J. Nat. Cancer Inst. 65, 603–605.[ISI][Medline]

Roberts, D. (1966). An isotopic assay for thymidylate synthetase. Biochemistry 5, 3546–3548.[ISI][Medline]

Sakaguchi, Y., Stephens, L. C., Makino, M., Kaneko, T., Strebel, F. R., Danhauser, L. L., Jenkins, G. N., and Bull, J. M. (1994). Apoptosis in normal tissues induced by 5-fluorouracil: Comparison between bolus injection and prolonged infusion. Anticancer Res. 14, 1489–1492.[ISI][Medline]

SAS Institute, Inc. (1990) SAS/STAT User's Guide Version 6, 4th ed. SAS Institute, Cary, NC.

Schuetz, J. D., Wallace, H. J., and Diasio, R. B. (1984). 5-Fluorouracil incorporation into DNA of CF-1 mouse bone marrow cells as a possible mechanism of toxicity. Cancer Res. 44, 1358–1363.[Abstract]

Schwartz, E. L., Hoffman, M., O'Connor, C. J., and Wadler, S. (1992). Stimulation of 5-fluorouracil metabolic activation by interferon-{alpha} in human colon carcinoma cells. Biochem. Biophys. Res. Comm. 182, 1232–1239.[ISI][Medline]

Setzer, R. W., Lau, C., Mole, M. L., Copeland, M. F., Rogers, J. M., and Kavlock, R. J. (2001). Toward a biologically-based dose-response model for developmental toxicity of 5-fluorouracil in the rat: A mathematical construct. Toxicol. Sci. 59, 49–58.[Abstract/Free Full Text]

Shah, R. M., and MacKay, R. A. (1978). Teratological evaluation of 5-fluorouracil and 5-bromo-2-deoxyuridine on hamster fetuses. J. Embryol. Exp. Morphol. 43, 47–54.[ISI][Medline]

Shuey, D. L., Buckalew, A. R., Wilke, T. S., Rogers, J. M., and Abbott, B. D. (1994a). Early events following maternal exposure to 5-fluorouracil lead to dysmorphology in cultured embryonic tissues. Teratology 50, 379–386.[ISI][Medline]

Shuey, D. L., Lau, C., Logsdon, T. R., Zucker, R. M., Elstein, K. H., Narotsky, M. G., Setzer, R. W., Kavlock, R. J., and Rogers, J. M. (1994b) Biologically based dose-response modeling in developmental toxicology: Biochemical and cellular sequelae of 5-fluorouracil exposure in the developing rat. Toxicol. Appl. Pharmacol. 126, 129–144.[ISI][Medline]

Shuey, D. L., Zucker, R. M., Elstein, K. H., and Rogers, J. M. (1994c). Fetal anemia following maternal exposure to 5-fluorouracil in the rat. Teratology 49, 311–319.[ISI][Medline]

Snyder, R. D. (1988). Consequences of the depletion of cellular deoxynucleoside triphosphate pools in the excision-repair process in cultured human fibroblasts. Mutat. Res. 200, 193–199.[ISI][Medline]

Stephens, J. D., Golbus, M. S., Miller, T. R., Wilber, R. R., and Epstein, C. J. (1980). Multiple congenital anomalies in a fetus exposed to 5-fluorouracil during the first trimester. Am. J. Obstet. Gnyecol. 137, 747–749.[ISI][Medline]

Takimoto, C. H., Voeller, D. B., Strong, J. M., Anderson, L., Chu, E., and Allegra, C. J. (1993). Effects of 5-fluorouracil substitution on the RNA conformation and in vitro translation of thymidylate synthase messenger RNA. J. Biol. Chem. 268, 21438–21442.[Abstract/Free Full Text]

Tsukamoto, I., and Kojo, S. (1991). The effects of fluorouracil on thymidylate synthetase and thymidine kinase in regenerating rat liver after partial hepatectomy. Biochim. Biophys. Acta 1074, 52–55.[ISI][Medline]

U.S. Environmental Protection Agency (1991). Guidelines for developmental toxicity risk assessment. Fed. Regist. 56, 63798–63826.

Valeriote, F., and Santelli, G. (1984). 5-Fluorouracil (FUra). Pharmacol. Ther. 24, 107–132.[ISI][Medline]

Wilson, J. G. (1971). Use of rhesus monkeys in teratological studies. Fed. Proc. 30, 104–109.[ISI][Medline]

Wilson, J. G. (1973). Environment and Birth Defects. Academic Press, New York.

Wilson, J. G., Jordan, R. L., and Schumacher, H. (1969). Potentiation of the teratogenic effects of 5-fluorouracil by natural pyrimidines: I. Biological aspects. Teratology 2, 91–97.[ISI][Medline]

Yoshioka, A., Tanaka, S., Hiraoka, O., Koyama, Y., Hirota, Y., Ayusawa, D., Seno, T., Garrett, C. and Wataya, Y. (1987). Deoxyribonucleoside triphosphate imbalance: 5-Fluorodeoxyuridine-induced DNA double strand breaks in mouse FM3A cells and the mechanism of cell death. J. Biol. Chem. 262, 8235–8241.[Abstract/Free Full Text]

Yoshioka, A., Tanaka, S., Hiraoka, O., Koyama, Y., Hirota, Y. and Wataya, Y. (1985) The mechansim of dNTP-unbalanced cell death induced by 5-fluorouracil and its derivatives. Nucleic Acid. Symp. Ser. 16, 245–248.

Zucker, R. M., Elstein, K. H., Shuey, D. L., and Rogers, J. M. (1995). Flow cytometric detection of abnormal fetal erythropoiesis: Application to 5-fluorouracil-induced anemia. Teratology 51, 37–44.[ISI][Medline]