neonatal mouse tumorigenicity bioassay

0 downloads 0 Views 105KB Size Report
increases in endogenous DNA adducts, 2. endocrine disruption, and 3. induction of ...... trichloroacetic acid, and trichloroethanol in- rodents” assay, or a second ...
Drug Information Journal, Vol. 32, pp. 711–728, 1998 Printed in the USA. All rights reserved.

0092-8615/98 Copyright  1998 Drug Information Association Inc.

NEONATAL MOUSE TUMORIGENICITY BIOASSAY* PETER P. F U, PHD Research Chemist

LINDA S. VON TUNGELN Research Chemist

PING YI Visiting Scientist

QINGSU XIA Visiting Scientist Division of Biochemical Toxicology, United States Food and Drug Administration, National Center for Toxicological Research, Jefferson, Arkansas

DANIEL A. CASCIANO, PHD Director, Division of Genetic & Reproductive Toxicology, United States Food and Drug Administration, National Center for Toxicological Research, Jefferson, Arkansas

THOMAS J. FLAMMANG, PHD Special Assistant, Office of the Director, United States Food and Drug Administration, National Center for Toxicological Research, Jefferson, Arkansas

FRED F. K ADLUBAR, PHD Director, Division of Molecular Epidemiology, United States Food and Drug Administration, National Center for Toxicological Research, Jefferson, Arkansas

The International Conference on Harmonisation (ICH) has suggested the use of the newborn mouse bioassay as an alternative tumorigenicity assay. There are sufficient data to conclude that this animal model is highly sensitive to chemical carcinogens that exert their action through mechanisms involving the formation of covalently bound DNA adducts (exogenous adducts) of the administered compound and the processing of these adducts to mutations. Mechanistic studies, including metabolism, DNA adduct formation, and ras-oncogene activation, presented in this paper aid in the interpretation of tumor experiments. By comparison, the data thus far obtained suggest that this bioassay is very likely insensitive to some indirect-acting chemical carcinogens. Ongoing studies are focused on the sensitivity of this bioassay to indirect-acting carcinogens that are believed to exert their tumorigenicity through secondary mechanisms, including: 1. induction of lipid peroxidation and increases in endogenous DNA adducts, 2. endocrine disruption, and 3. induction of

Presented at the DIA Workshop “Testing Human Pharmaceuticals For Carcinogenic Potential: New Approaches after ICH 4,” October 27–28, 1997, Noordwijk, the Netherlands. Reprint address: Dr. Peter P. Fu, U.S. FDA/NCTR, 3900 NCTR Road, HFT-110, Jefferson, AR 72079. *This article is not an official Food and Drug Administration guidance or policy statement. No official support or endorsement by the Food and Drug Administration is intended or should be inferred.

711

712

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar peroxisome proliferation. A systematic approach to validate the neonatal mouse tumorigenicity bioassay as a part of a risk identification procedure is proposed. This approach combines the tumorigenicity bioassay with several simple and convenient supportive mechanistic experiments of study so that direct-acting chemical carcinogens, indirectacting carcinogens, and noncarcinogens can be distinguished. Key Words: Neonatal mouse tumorigenicity bioassay; Direct-acting genotoxic chemical carcinogen; DNA adduct; Secondary mechanism

INTRODUCTION THE FIRST REPORT that newborn mice were highly sensitive to the development of chemically-induced tumors was published in Nature in 1959 by Pietra et al. (1). A group of 25 newborn Swiss male and female mice each received a single s.c. injection of 30 µg of 7,12-dimethylbenz[a]anthracene (DMBA) on the day of birth. In the period of 11–24 weeks, eight of these 25 mice (32%) developed lymphomas. No tumors were found in the control animals. A subsequent study reported in 1961 by the same group (2) indicated that Swiss newborn mice receiving benzo[a]pyrene, 3-methylcholanthrene, dibenz[a,h]anthracene, or urethane (ethyl carbamate) developed malignant lymphoma and pulmonary tumors. Throughout the early to mid 1960s, the susceptibility of neonatal mice of different strains and dosing regimens to a variety of chemicals of various structures was widely studied (3–8,9–18). In general, it was shown that different strains of neonatal mice were highly susceptible to chemicallyinduced tumor formation. Several factors that can affect differences in tumor susceptibility have been extensively studied. These factors include mouse strain, sex, hormonal status, dose level, time of dosing, route of administration, time to tumor development, vehicle, and the nature of the chemicals. These factors can also be responsible for differences in susceptibility between neonatal and adult animals. Additional factors include differences in the absorption of the chemical into the body, distribution of the chemical to each organ, levels of metabolizing enzymes present in the target and non-

target organs, rate of transfer of the activated metabolite into the nuclei, DNA adduct formation, DNA repair capability, and the rate of cell proliferation (19–22). More recent studies have focused on: 1. The use of the neonatal mouse for assaying the tumorigenicity of a wider number of chemical classes, 2. Identification of genetic factors influencing tumorigenicity, and 3. The molecular mechanisms of chemical carcinogenesis involved in this bioassay, including metabolism, DNA binding, and mutation frequency of protooncogenes contained in the target tissues (eg, liver and lung). The tumorigenicity of different chemical classes in the neonatal mouse model has been reviewed recently (23,24). The neonatal mouse tumorigenicity bioassay utilizes a total dose of chemical that is 5,000- to 10,000-fold less than that needed for the standard chronic rodent bioassay (25,26). This feature has made this bioassay attractive for the study of the tumorigenic potential of metabolites of chemical carcinogens, which are often available only in very limited quantities. Thus, during the last 20 years, several laboratories (27–32) have frequently utilized this animal bioassay to determine the metabolic activation pathways of a variety of many DNA reactive chemical carcinogens. EXPERIMENTAL DESIGN In this bioassay protocol, mice are generally treated with carcinogens one to four times

Neonatal Mouse Tumorigenicity Bioassay

during the preweanling period (one to 22 days of age) and then sacrificed at 8–12 months of age (Figure 1) to evaluate tumor yield. The test compounds are dissolved in water, trioctanoin, acetone, or dimethylsulfoxide (DMSO) and then are administered by i.p. injection, s.c. injection, or oral gavage. The Swiss CD-1 and B6C3F1 have been the most commonly used mouse strains for the neonatal mouse tumorigenicity bioassay. The CD-1 mouse strain, which is relatively sensitive to developing chemically-induced lung and liver tumors, is used frequently by the pharmaceutical community. By comparison, the B6C3F1 strain tends to be used in academic research laboratories to facilitate direct comparison of neonatal tumorigenicity data with that of the United States National Toxicology Program’s (NTP) chronic tumorigenicity bioassay. The authors use the B6C3F1 mouse strain to conduct the neonatal mouse bioassay (33). A high frequency of ras-protooncogene mutation has been suggested to be associated with tumor development and has been extensively studied in the neonatal mouse model. In general, high mutation frequency and specific mutation patterns, formed either spontaneously or by direct-acting carcinogens in DNA from liver tumors in the B6C3F1 mouse have been observed. Thus, the authors have employed the B6C3F1 mouse for the neonatal tumorigenicity bioassay and have analyzed the chemical-induced ras-protooncogene mutational pattern and frequency to develop a better understanding of the mechanisms of tumor induction.

713

NEONATAL MOUSE TUMORIGENICITY BIOASSAY: HIGHLY SENSITIVE TO DIRECTACTING GENOTOXIC CARCINOGENIC CHEMICALS Chemical carcinogenesis has been extensively studied during the last three decades. Based on the mechanism of action, chemical carcinogens have been conventionally classified as genotoxic and nongenotoxic. In this paper, in order to easily distinguish the types of chemical carcinogens that the neonatal mouse is susceptible to, the chemical carcinogens are categorized into two classes: “direct-acting genotoxic chemical carcinogens” and “indirect-acting genotoxic chemical carcinogens.” Direct-acting genotoxic chemical carcinogens are defined as chemicals that exert tumorigenicity through mechanisms involving the formation of covalently bound DNA adducts of the administered compound (exogenous adducts) and the processing of these adducts to mutations. These compounds either can bind to DNA directly (such as N-ethyl-N-nitrosourea) or require metabolism to form the reactive metabolites that are capable of binding to DNA (such as benzo[a]pyrene). Indirect-acting genotoxic chemical carcinogens are defined as chemicals that exert tumorigenicity through secondary mechanisms, including: 1. Formation of endogenous DNA adducts from lipid peroxidation, 2. Induction of hypomethylation or hypermethylation,

FIGURE 1. Time schedule of the neonatal mouse tumorigenicity bioassay.

714

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

3. Enhancement of oxidative stress, 4. Induction of peroxisome proliferation, and 5. Modulation of sex hormones (24,33). It is recognized that carcinogenesis can also occur via non-DNA reactive mechanisms, for example, topoisomerase inhibition and aneuploidy (ie, spindle disruption). For convenience, chemical carcinogens that exert their tumorigenesis through these mechanisms are also included in the category of “indirectacting genotoxic chemical carcinogens.” The mechanism by which the direct-acting chemical carcinogens exert their tumorigenicity has been extensively described since the early 1970s through many contributions from the Millers’ research group. This body of work, as well as that by others, has convincingly shown the neonatal mouse tumorigenicity bioassay to be highly sensitive to direct-acting chemical carcinogens. Several other laboratories also have demonstrated the utility of this bioassay to dissect the metabolic pathways responsible for converting chemical carcinogens to their genotoxic agents (26,27,30–32,34–39). Direct-acting chemical carcinogen classes that have been studied so far include polycyclic aromatic hydrocarbons (PAHs) (1,2,23,32,40), halogenated PAHs (38,41), nitrated PAHs (33), food pyrolysis products (heterocyclic amines) (25, 26), complex environmental mixtures (42,43), nitrosamines (44), mycotoxins and other naturally occuring substances (29,45,46), vinyl compounds (29,46), azo dyes (30,47), aromatic amines (23,48), and allylarenes, such as estragole and safroles (49). The genotoxic mechanisms by which some of these compounds exert their tumorigenicity in the neonatal mouse also have been determined by employing the neonatal mouse for in vitro and/or in vivo studies. Mechanistic studies, which will be discussed in the following section, include: determination of metabolizing enzyme activity in the target tissues, for example, neonatal liver and lung, metabolic activation pathways leading to liver tumors in vivo; exogenous DNA adducts formed in the liver of neonates treated with the chemical; and ras-protooncogene activation present in

the liver and lung tumors (25,26,28,33,36, 37,39,41,50). The Millers and co-workers, through a wide range of studies (34,35,47,49,51–53), found that male B6C3F1 mice treated with a chemical at 12 days of age exhibited higher susceptibility than those given the same dose of the chemical at one day of age. In particular, the studies of Wiseman et al. (29) demonstrated the sensitivity of the male 12-day old B6C3F1 neonatal mouse to a variety of directacting carcinogenic chemicals. These results, shown in Table 1, demonstrate the more than a 250-fold increase in tumorigenic potency between DEN and cis-asarone, indicating that the neonatal mouse bioassay has the capacity to determine the tumorigenic potential of direct-acting genotoxic chemical carcinogens with a very wide range of activities. Studies using both the CD-1 and B6C3F1 mouse from the authors’ laboratories tested several classes of direct-acting chemical carcinogens, including PAHs (36,54), halogenated PAHs (38,41), nitrated PAHs (36,37, 39,54,55) (Fu, unpublished observations), heterocyclic amines (25,26), and environmental complex mixture samples (42). The initial studies utilized the CD-1 mouse for bioassay of a series of PAHs, nitro-PAHs, and heterocyclic amines (36) (Von Tungeln, unpublished observations). The results indicate that the CD-1 mouse was susceptible to chemical-induced lung and liver tumors. Later, for practical considerations and for comparison of neonatal tumorigenicity data with the NTP chronic tumorigenicity bioassay, the B6C3F1 strain was frequently employed for bioassay. This strain has been confirmed to be sensitive mainly to liver tumors. Both CD-1 and B6C3F1 strains exhibit distinct sex specificity for tumor development. Male mice are highly susceptible to directacting chemical carcinogens, while female animals are often insensitive to these same chemicals. The tumorigenicity of four representative direct-acting chemical carcinogens, 4-aminobiphenyl (4-ABP), benzo[a]pyrene (BaP), 6nitrochrysene (6-NC), and aflatoxin B1 (AFB1) assayed by the two-year chronic bio-

Neonatal Mouse Tumorigenicity Bioassay

715

TABLE 1 Relative Liver Tumorigenic Potency of Chemical Carcinogens Administered to 12-day-old Male B6C3F1 Micea Average Number of Hepatomas/ µmol/Chemical/g bwb

Chemical Nitrosodiethylamine (DEN) Aflatoxin B1 (AFB1) Vinyl carbamate (urethane) 1′-Hydroxy-2′,3′-dehydroestragole N-Hydroxy-2-acetylaminofluorene 1′-Hydroxy-2′,3′-dehydrosafrole N-Methyl-4-aminoazobenzene (MBA) 4-Aminoazobenzene 2-Methyl-N,N-dimethyl-4-aminoazobenzene N,N-Dimethyl-4-aminoazobenzene (DAB) 1′-Hydroxyestragole Benzo[a]pyrene 1′-Hydroxysafrole Precocene I Ethyl carbamate cis-Asarone 3′-Hydroxy-trans-anethole

1100 ± 67 350 ± 27 250 ± 13 220 ± 15 150 ± 11 110 ± 3 91 ± 3 78 ± 5 66 ± 3 52 ± 3 32 ± 3 27 ± 0.2 20 ± 2 10 ± 0.7 7 ± 0.6 4 ± 0.2 ,0

a

Data from Wiseman et al. (29). Determined from the linear slope of the regression line ± SD.

b

assay and by the neonatal B6C3F1 mouse tumorigenicity bioassay is shown in Table 2. These results indicate that the neonatal male B6C3F1 mouse bioassay exhibits similar sus-

TABLE 2 Comparison of Tumorigenicity of Direct-Acting (Genotoxic) Chemical Carcinogens Determined in the Two-Year Chronic Bioassay and the Neonatal B6C3F1 Mouse Tumorigenicity Bioassay Two-year Chronic Bioassaya Rat

Mouse

Neonatal Mouse Bioassay

Compound

M

F

M

F

M

F

4-ABP BaP 6-NC AFB1

+ + + +

+ + + +

+ + + −

+ + + −

+ + + +

− − − −

a

DHEW Publication No. (NIH)72-35, 1968.

ceptibility to 4-ABP-, BaP-, and 6-NC-induced tumorigenicity as compared to the rat and mouse chronic bioassay of both sexes. The sensitivity of the neonatal B6C3F1 female mouse to tumor incidence induced by these chemicals is very low. It is worth noting that while AFB1 is tumorigenic in the neonatal mouse, it is not carcinogenic when tested in the adult mouse chronic bioassay (Table 2). Furthermore, the authors have found that halogenated PAHs, 7-chlorobenz[a]anthracene (7-Cl-BA) and 7-bromobenz[a]anthracene (7-Br-BA), both expected to be directacting carcinogens, are potent tumorigens in the neonatal mouse bioassay (38). 7-Cl-BA and 7-Br-BA induced hepatocellular adenoma in 92 and 96% of the mice and hepatocellular carcinoma in 100 and 83% of the mice, respectively. In the chronic tumorigenicity bioassay, however, 7-Cl-BA was weakly carcinogenic and 7-Br-BA was essentially inactive (56). Together, these results clearly indicate that although both the twoyear chronic bioassay and the neonatal

716

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

mouse bioassay are sensitive to direct-acting chemical carcinogens, discrepancies in tumorigenic responses exist. The reason for differences in the tumorigenic response between these two bioassays is not known, and may presumably be due to differences in the rate of metabolic activation and detoxification, dosing regime, cell proliferation rate, efficiency in DNA repair, and gene expression. A more general comparison of tumorigenicity response between the neonatal mouse and chronic rodent bioassay has been presented by Fujii (23). In this comparison, most, if not all, of the expected direct-acting carcinogens that could be identified were tumorigenic in both bioassays.

THE NEONATAL MOUSE TUMORIGENICITY BIOASSAY— A MECHANISM-BASED BIOASSAY IN TESTING DIRECT-ACTING CHEMICAL CARCINOGENS The neonatal mouse tumorigenicity bioassay has been demonstrated to be highly sensitive to direct-acting genotoxic chemical carcinogens. To establish this animal model as part of a mechanism-based bioassay and to determine whether or not the test chemical is a direct-acting chemical carcinogen, the mechanism by which this compound exerts tumorigenicity in the target tissue should be elucidated. A general approach is to carry out metabolism in vitro with liver (target tissue) microsomes of neonatal (usually 12–15-dayold) male mice. Furthermore, studies of the identified metabolites, including determination of mutagenicity and DNA binding capability, are carried out. When possible, the tumorigenicity of the metabolites is determined by the neonatal mouse tumorigenicity bioassay and compared with that of the substrate. The enzymes responsible for metabolic activation of the carcinogenic chemicals may also be determined. The following summarizes part of the work on the identification of the metabolizing enzymes, the study of metabolism in vitro and DNA adduct formation in vivo, and the analysis of ras-

protoocongene mutation pattern and mutation frequency in target (liver) tissues. Metabolizing Enzymes As previously described (33), metabolism of the food-derived mutagen/carcinogen, PhIP, by liver microsomes of 15-day-old male B6C3F1 mice resulted in the formation of Nhydroxy-PhIP and 4′-hydroxy-PhIP, which indicates that both CYP1A1 and CYP1A2 are present in the liver microsomes of the 15-day old male B6C3F1 mouse. It is known that metabolic transformation of PAHs and aromatic amines (including heterocyclic amines) to their more carcinogenic metabolites is catalyzed by CYP1A1 and CYP1A2, respectively. Thus, it is reasonable to assume that the neonatal male B6C3F1 mouse metabolizes PAHs, arylamines, and heterocyclic amines catalyzed by CYP1A1 and CYP1A2 present in the liver. Because human liver lacks CYP1A1 isozymes, the neonatal male B6C3F1 mouse liver is suspected to be more susceptible to PAH-induced hepatotumorigenicity than humans. In many cases, metabolic activation of arylamines and heterocyclic amines to species that react covalently with cellular DNA requires O-esterification of the resulting N-hydroxylamino intermediates. The high level of both acetyltransferase and sulfotransferase in the liver of neonatal male B6C3F1 mice suggests that this animal model is highly susceptible to chemical carcinogens of this type. N-Hydroxy arylamine O-acetyltransferase activity has been measured in neonatal and adult B6C3F1 mouse liver cytosols and the activities are comparable to those found in human liver from individuals who are rapid acetylators (33). Sulfotransferase activity, on the other hand, was considerably higher in neonatal mice than in adult mice or human livers. Boberg et al. (34) compared the levels of PAPS-dependent hepatic cytosolic sulfotransferase in adult and neonatal animals. In adult animals, the levels in the female CD1 mouse were six-fold higher than the male CD-1 and male B6C3F1 mouse and two-fold higher than the female B6C3F1 mouse and

Neonatal Mouse Tumorigenicity Bioassay

male CD (Sprague-Dawley) rat. The levels in the female CD-1 mouse were also threefold higher than the female CD (Sprague Dawley) rat. By comparison, the hepatic sulfotransferase activity of both the neonatal and adult female and male B6C3F1 mouse was age dependent. Within the period from the day of birth to weaning at 22 days of age, the hepatic cytosolic sulfotransferase activities of the female and male B6C3F1 mouse were sex independent. In addition, the activities of both the female and male B6C3F1 mouse increased 3.5-fold from the day of birth to 22 days of age, reaching an activity level greater than the adult CD-1 female mouse. Starting at 22 days of age for both sexes, the sulfotransferase began to decline rapidly, and the sulfotransferase activity of the female B6C3F1 mouse at eight weeks of age declined to a level similar to that of the one-day-old mouse. Delclos et al. (47) also found that sulfotransferase activity in the liver cytosols of the 12-day old male B6C3F1 mouse was several fold higher than that of the 12-day-old Fischer rat and the 20-week-old male Sprague-Dawley rat. Taken together, the high level of sulfotransferase activity suggests the high susceptibility of the B6C3F1 mouse to chemical carcinogens and also suggests that enzymatic sulfation is a critical metabolic activation process in the tumorigenic response. Metabolic Activation Pathways Generally, the identification of potential metabolic activation pathways are studied in vitro with microsomes prepared from the target organ (liver) of neonatal mice. These experiments are then followed by comparison of the biological activities between the metabolites and the substrate, including determination of mutagenicity, DNA binding capability, and tumorigenicity. For example, the Millers’ laboratory first determined the tumorigenicity of a series of chemicals in the neonatal mouse tumorigenicity bioassay as discussed previously. They subsequently performed a series of studies to determine the

717

probable mechanisms leading to tumorigenicity of these chemicals. Accordingly, the metabolic activation pathways of direct-acting carcinogenic chemicals were examined first by metabolism with liver microsomes of 12-day-old B6C3F1 male mice, followed by analysis of the DNA adducts in vivo in the liver of neonatal mice one day and two days after final dosing of the compound (34,35,49,51,52,53). The authors have also examined the mechanisms by which a test chemical exhibits tumorigenicity in the neonatal animal model (38,41,55) (Fu et al., unpublished observations, Von Tungeln, L.S., Xia, Q., and Fu, P.P., unpublished data). To illustrate this work on the metabolic activation of directacting chemical carcinogens, the mechanistic study of the halogenated PAH, 7-Cl-BA, which was found to be a potent liver tumorigen in the neonatal male B6C3F1 mouse, is described. The studies include: 1. Metabolism with liver microsomes of 15day-old B6C3F1 male mice where 7-ClBA generated 10 identifiable metabolites, including trans-3,4-dihydrodiol, trans-5,6dihydrodiol, trans-8,9-dihydrodiol, trans10,11-dihydrodiol, and several phenolic metabolites (38,57,58), 2. Salmonella typhimurium mutagenicity bioassays where 7-Cl-BA trans-3,4-dihydrodiol was found to be more mutagenic than the parent substrate when tested in Salmonella typhimurium TA100 in the presence of S9 (57); the other metabolites exhibited lower mutagenicity than the parent substrate (7-Cl-BA), and 3. Demonstration of 7-Cl-BA covalent binding to calf thymus DNA, catalyzed by liver microsomes of 15-day-old B6C3F1 male mice, which yielded a 7-Cl-BA trans-3,4dihydrodiol derived DNA adduct (38). To complement this experiment, the trans3,4-dihydrodiol was synthetically converted to the corresponding vicinal bayregion diol-epoxide, which was then bound to calf thymus DNA to obtain an authentic DNA adduct standard. The DNA adduct formed by in vitro metabolism was

718

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

identical to the authentic DNA adduct standard by 32P-postlabeling/HPLC analysis. No DNA adducts were derived from the trans-8,9-dihydrodiols and trans-8,9dihydrodiols (38), and 4. Oncogene analyses showed that the liver tumor tissues of the B6C3F1 male mice treated with 7- Cl-BA at eight and 15 days of age contained a high frequency of rasprotooncogene mutation (this will be further described in the following section). These mechanistic studies together with the tumorigenicity of this compound described in the previous section indicate that 7-Cl-BA is a potent direct-acting genotoxic carcinogen in the neonatal B6C3F1 male mouse tumorigenicity bioassay and that bayregion diolepoxides are the ultimate metabolites that lead to DNA adduct formation and tumor initiation (Figure 2). ras-Protooncogene Activation The spontaneously-formed and chemicallyinduced liver tumors in B6C3F1 mice have been found to possess a very high frequency of mutations in ras-protooncogenes (59–62). These sequence alterations in the ras-protooncogenes frequently reflect the mutational specificity of the carcinogen (63,64). It has been found that ras-oncogene activation

occurs both at an early stage (eg, during focus formation) as well as at later stages of liver tumor development (eg, during adenoma and carcinoma formation) (65). Although the sequential and/or necessary activation (deactivation) of oncogenes in tumor formation is not fully understood, to date it has been found that more than half of the spontaneously-formed and direct-acting carcinogen-induced liver tumors in the B6C3F1 mouse possess mutations in the ras-protooncogene (60,62). A CAA to AAA transversion (G to T in the nontranscribed strand), in particular, is the principal H-ras point mutation in codon 61 occurring in both the spontaneous and the carcinogen-induced B6C3F1 mouse liver tumors (59,62,65–77). From these basic observations the authors rationalized that determination of ras-protooncogene activation might provide insight into the mechanisms by which chemical carcinogens induce liver tumors in mice, including the neonatal carcinogenesis model. The following illustrates some of the authors’ ongoing studies to determine ras-protooncogene mutations contained in the liver tumors of male B6C3F1 mice treated with directacting carcinogens. As described earlier, both 7-Cl-BA and 7-Br-BA exhibit potent tumorigenicity in the neonatal mouse bioassay. Analysis of the ras-protooncogene mutation in the liver tumors induced by 7-Cl-BA and

FIGURE 2. Proposed metabolic activation of 7-chlorobenz[a]anthracene leading to liver tumors in male B6C3F1 mice.

Neonatal Mouse Tumorigenicity Bioassay

7-Br-BA indicated that 83% (20/24) of 7-ClBA-induced and 91% (20/22) of 7-Br-BAinduced liver tumors had activated ras protooncogenes (41). In contrast to the general finding of H-ras mutations in B6C3F1 mouse liver tumors, all the mutations were at the first base of K-ras codon 13, resulting in a pattern of GGC 6 CGC (Table 3). These results demonstrate that 7-Cl-BA and 7-BrBA induce a unique type of K-ras-protooncogene activation in the liver tumors of B6C3F1 mice. The results also suggest that the formation of K-ras oncogene mutation, GGC 6 CGC, is involved in the development of liver tumors in B6C3F1 mice treated with these two chemical carcinogens. Similar striking results have been obtained from liver tumors of mice treated with other polycyclic aromatic hydrocarbons and derivatives (Fu et al., unpublished data). Combined with the previous report of a high incidence of K-ras mutations in liver tumors induced by benzo[a]pyrene (70), these data suggest that liver tumors induced by PAHs and PAH derivatives may contain a unique type of activated ras protooncogene. Since many carcinogens produce liver tumors in the neonatal mouse assay (59,60,62,65,67–70,77), regardless of their organ specificity in other carcinogene-

719

sis models, the neonatal mouse bioassay is a good model for determining whether or not K-ras mutation is a common feature of mouse liver carcinogenesis by PAHs. BIOASSAY OF INDIRECT-ACTING GENOTOXIC CARCINOGENS In contrast to direct-acting genotoxic mechanisms, “indirect-acting genotoxic” mechanisms of carcinogenesis have been studied much less and, consequently, are less clearly understood. Relatively recently, “indirectacting” mechanism has taken on expanded meanings. As described in an earlier section, for the purpose of this paper indirect-acting genotoxic chemical carcinogens are defined as those that exert their tumorigenicity through mechanisms where DNA alterations leading to mutations arise from events that do not involve the formation of exogenous carcinogen-DNA adducts. The indirect-acting mechanisms also have been described under the general term secondary mechanisms. At present, only a small number of putative indirect-acting genotoxic chemical carcinogens have been tested in the neonatal mouse tumorigenicity bioassay and the re-

TABLE 3 Pattern of Mutations in Ras-protooncogenes Contained in Liver Tumors of Male B6C3F1 Mice Treated Neonatally with Direct-Acting and Indirect-Acting Genotoxic Chemicals Number of Mutations Chemical

% of Tumors with Mutations

Direct-acting genotoxic chemicals 12/14 (86) BAa 20/24 (83) 7-Cl-BAb 20/22 (91) 7-Br-BAb Indirect-acting genotoxic chemicals 1/8 (12.5) Chloral hydratec 0/6 (0) Trichloroacetic acidc 0/6 (0) Trichloroethanolc 1/6 (17) Vehicle (DMSO)c a

K-ras (Codon 13) GGC 6 CGC

H-ras (Codon 61) CAA 6 CTA

11 19 19

1 1 1

0 0 0 0

1 0 0 1

Data from Von Tungeln et al. (Von Tungeln, L.S., Xia, Q., and Fu, P.P., Benz[a]anthracene is a potent liver tumorigen in the neonatal B6C3F1 mouse tumorigenicity bioassay. b Data from (41). c Fu et al., unpublished data.

720

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

sults suggest that the neonatal mouse model is not sensitive to indirect-acting chemical carcinogens (23,24,33). Nevertheless, to characterize the neonatal mouse tumorigenicity bioassay, it is necessary that the tumorigenicity of a series of indirect-acting chemical carcinogens be tested by the neonatal mouse bioassay. Many chemicals that may not be directacting genotoxic carcinogens have been found to give some measure of a positive (tumorigenic) response in two-year chronic bioassays, including data from the NTP bioassay (78,79,80) or from the FDA rodent carcinogenicity database containing many pharmaceutical compounds (81). Conceivably, at least part of these chemicals may be indirectacting chemical carcinogens. Thus, several compounds for the neonatal mouse bioassay have been chosen and the results compared with those of the chronic tumorigenicity bioassays. Bioassay of Benzodiazepine and Antihistamine Drugs Of the more than 300 drugs tested so far in the NTP chronic bioassay, about one half were found to be tumorigens (81), leading to criticism that either there is a selection bias or that these chemicals would be tumorigens only under the conditions of the chronic bioassay (eg, high dose levels) and thus, may not be relevant for human comparison. It is not practical or cost effective to test the tumorigenicity of all the chemicals (or drugs) in the database and then elucidate their mechanisms of action, so a series of over-thecounter benzodiazepines and antihistamines have been chosen for the neonatal mouse tumorigenicity bioassay. Many of these chemicals are suspected to involve secondary mechanisms of carcinogenicity. The benzodiazepine and antihistamine drugs selected for the neonatal mouse bioassay were oxazepam, diazepam, chlordiazepoxide, midazolam, flurazepam, methapyrilene, doxylamine, and pyrilamine. The drugs were assayed in both the male and female B6C3F1 mouse and the animals were sacri-

ficed at 12 months of age or up to 24 months of age. The neonates were administered i.p. injections of the test drug dissolved in DMSO on Days 8 and 15 with 1/3 and 2/3 of the total dose in volumes of 10 µl and 20 µl, respectively. These drugs were inactive in the neonates of both sexes (Von Tungeln et al., unpublished observations). In most cases, these results are drastically different from those in the chronic bioassay. As shown in Table 4, while these benzodiazepine drugs were found to be tumorigenic in one or more cells of the two-year chronic rat and mouse bioassay paradigm (81), they were inactive in the neonatal mouse tumorigenicity bioassays. The mechanisms by which these drugs exert tumorigenicity in the two-year chronic bioassay are not clear (82,83). Nor is it understood why these drugs show no tumor response in the neonatal mouse. One possible explanation has been that the neonatal mouse is less capable of metabolizing these drugs to the tumorigenic metabolites in comparison to the adult mouse. In order to test this possibility, the metabolism of diazepam was examined by liver microsomes of 15-day-old and adult (9-month-old) male B6C3F1 mice. The results indicate the metabolite profile as well as the metabolic rate with adult and neonatal liver microsomes, obtained under the same experimental conditions, were similar (Von Tungeln et al., unpublished observations). Thus, these results suggest that the neonatal mice are capable of metabolizing these benzodiazepine drugs. As a consequence, the authors hypothesize that these drugs are indirect-acting carcinogens and exert their tumorigenicity in the two-year chronic bioassay via an indirect-acting (secondary) mechanism. It has been suggested that many indirectacting carcinogens induce cell proliferation, which is an apparent property of some tumor promoters (84). It is thus critical to consider that in comparison to the dosing regimen of the typical chronic bioassay, which commences at six weeks of age, chemical treatment of neonates is completed during a phase of maximum cell replication, that is, the first

Neonatal Mouse Tumorigenicity Bioassay

721

TABLE 4 Tumorigenicity of Benzodiazepines Tested in the B6C3F1 Mouse by the NTP Chronic Bioassay and the Neonatal Tumorigenicity Bioassay

Compound Oxazepam Diazepam Chlordiazepoxide Midazolam

Two-year Chronic Mouse Bioassaya

Neonatal Mouse Bioassayb

Male

Female

Male

Female

+ + + −

+ + + +

− − − −

− − − −

a

Data either from NTP bioassay or from the FDA rodent carcinogenicity database (81). b The results designated as “−” indicate there was no significant difference when compared to the solvent control, DMSO.

few weeks of life. Therefore, it is likely that additional cell proliferation induced by the indirect-acting carcinogen in the neonatal mouse is negligible compared to the normal ongoing cell proliferation rates. Thus, if an indirect-acting carcinogen, such as the benzodiazepines listed in Table 4, exhibits a tumorigenic response in the chronic bioassay in part by inducing hyperplasia, this chemical (or drug) may be negative or very mildly tumorigenic in the neonatal mouse assay. In contrast, the induction of cell proliferation by a chemical in the chronic bioassay could produce a significant burden on the homeostatic levels of cell replication and DNA repair/misrepair, ultimately leading to tumor formation. Bioassay of Indirect-acting Carcinogens of Known Secondary Mechanisms To help characterize the B6C3F1 neonatal mouse as an alternative tumorigenicity animal model, it is necessary to test a large number of indirect-acting carcinogens thought to exert tumorigenicity through a known secondary mechanism other than through induction of cell proliferation. Consequently, the neonatal mouse bioassay has been used to test a series of indirect-acting carcinogens that are suspected of exerting to exert carcinogenicity in the chronic bioassay via secondary mechanisms including induction of lipid

peroxidation and formation of endogenous DNA adducts, induction of peroxisome proliferation, and modulation of sex hormones. The first set of chemicals where the authors have evidence for mechanisms of tumorigenicity are those that induce lipid peroxidation and endogenous DNA adduct formation (ie, chloral hydrate, trichloroacetic acid, trichloroethanol, carbon tetrachloride, methylene chloride, chlordane, heptachlor, paraquat, and trifluralin). The authors have previously reported that mouse liver microsomal metabolism of chloral hydrate generated free radicals and induced lipid peroxidation products (85). When metabolized in the presence of calf thymus DNA, a malondialdehyde (MDA) modified deoxyguanosinyl-DNA adduct (MDA-MG1) was formed (86). These data led to an expectation that induction of lipid peroxidation and endogenous DNA adduct formation may be involved in chloral hydrate induced tumorigenicity. Thus, the lipid peroxidation products: MDA, acrolein, crotonaldehyde and 4-hydroxy-2-nonenal, and trichloroacetic acid and trichloroethanol (two metabolites of chloral hydrate) along with chloral hydrate were selected for the neonatal mouse bioassay. With the exception of the positive control, 4-aminobiphenyl, that was shown to be a potent tumorigen, chloral hydrate, trichloroacetic acid, trichloroethanol, and all the peroxidation products were negative in the

722

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

neonatal mouse bioassay (87). Analysis by 32 P-postlabeling/HPLC of DNA isolated from livers 24 and 48 hours after the final dose indicated the MDA-MG-1 adduct level was 1.5- and 1.2-fold higher, respectively, in the chloral hydrate-treated mice compared to the DMSO-treated mice (Fu, unpublished observations). Preliminary data on the analysis of the ras-protooncogenes contained in liver tumors of mice treated with chloral hydrate, trichloroacetic acid, and trichloroethanol indicate that they exhibit very low mutation frequency (Table 3). These results suggest that the neonatal mouse tumorigenicity bioassay is not sensitive to the chemicals that induce lipid peroxidation and endogenous DNA adduct formation. Additional sets of indirect-acting carcinogens should also be examined to determine whether the results would be consistent with chloral hydrate. The authors have chosen estradiol and its hydroxylated metabolites diethylstilbestrol (DES), tamoxifen, toremifene, genistein, daidzein, and DDT, most of which are believed to exert any possible tumorigenic effect by modulation of sex hormone activity, and benzene hexachloride (BHC), another suspect endocrine disrupter. A third set of suspect indirect-acting carcinogens consists of two known peroxisome proliferators di(2-ethylhexyl)phthalate and clofibrate. Bioassay of these three sets of compounds along with the panels of antihistamines should allow determination of whether or not the neonatal tumorigenicity bioassay is sensitive to these types of chemicals and it will provide a better understanding of this animal model for predicting carcinogenicity, as well as its utility as a risk assessment model. DEVELOPMENT AND VALIDATION OF THE NEONATAL MOUSE TUMORIGENICITY BIOASSAY AND STRATEGY OF MECHANISTIC STUDIES FOR RISK IDENTIFICATION The International Conference on Harmonisation (ICH) for Technical Requirements for

the Registration of Pharmaceuticals for Human Use has provided the specific guidance, “Topic S1.B. Testing for Carcinogenicity of Pharmaceuticals,” and recommendations for selecting in vitro and in vivo methods for detecting genotoxic pharmaceuticals (88). The basic principle of the guidance allows for the use of one long-term rodent carcinogenicity study, plus one additional supplementary study, for example, the “new-born rodents” assay, or a second chronic study. The neonatal mouse tumorigenicity bioassay database contains sufficient data to conclude that this animal model is highly sensitive to direct-acting genotoxic chemical carcinogens, with development of tumors in a relatively short latency time (8–12 months). The information on mechanistic studies presented in this paper also indicates that this animal model can serve as a convenient tumorigenicity bioassay. The data thus far obtained suggest that the neonatal mouse tumorigenicity bioassay is very likely insensitive to indirect-acting chemical carcinogens. The indirect-acting carcinogens are believed, in most cases, to exert their tumorigenicity in the chronic bioassay by acting as cell proliferators (tumor promoters) (84). If this mechanism of tumorigenicity is true, this chemical class will likely test negative in the neonatal mouse tumorigenicity bioassay since the test chemical is administered to the neonate when cells are replicating at their highest rate. The level of the additional cell proliferation induced by an indirect-acting carcinogen compared with the endogenous cell proliferation will be inconsequential. In contrast, in the two-year chronic tumorigenicity bioassay, an indirectacting chemical carcinogen is administered daily to the rodent for a period of about two years. The resulting continuous additional cell proliferation induced by the indirect-acting chemical carcinogen at times of normal quiescent cell division is physiologically inappropriate and thus could result in an increased tumor incidence. Therefore, because of the marked difference in dosing regimen, it is understandable that this chemical may

Neonatal Mouse Tumorigenicity Bioassay

exert a negative or very mild tumorigenic response in the neonatal mouse assay. It is believed by many that indirect-acting chemical carcinogens are less hazardous to human health than direct-acting carcinogens, and that they may exhibit a threshold in tumor dose-responses (84). Consequently, it is important that risk assessment models incorporate the ability to distinguish between direct-acting and indirect-acting chemical carcinogens. As described above, it is suggested that the neonatal mouse tumorigenicity bioassay can help distinguish between the direct-acting and indirect-acting carcinogens, which in turn supports the recommendation of this animal model as a useful alternative tumorigenicity bioassay by the ICH. The authors have proposed a systematic approach to validate the neonatal B6C3F1 mouse tumorigenicity bioassay as a part of a risk identification procedure. This approach combines the tumorigenicity bioassay with several simple and convenient supportive mechanistic experiments of study so that direct-acting chemical carcinogens, indirectacting carcinogens, and noncarcinogens can be distinguished. As shown in Table 5, if a chemical exhibits strong tumorigenic activity in the neonatal mouse tumorigenicity bioassay, the compound should be classified as a direct-acting chemical carcinogen and its genotoxic mechanism confirmed at least by its mutagenicity in Salmonella typhimurium TA98 and/or TA100 and its high mutation frequency of ras-oncogene in liver tumors. The genotoxic mechanism can be solidly supported if metabolic activation and exogenous DNA adduct formation in vivo also are established. On the other hand, when a chemical exerts no tumorigenic activity in the neonatal mouse tumorigenicity bioassay, this chemical may be an indirect-acting chemical carcinogen or a noncarcinogen. As shown in Table 5, it is proposed that an indirect-acting chemical carcinogen can be distinguished from a noncarcinogen. A noncarcinogen should give negative responses in all of the mechanistic experiments listed in Table 5 including negative evidence of characteristics

723

of secondary mechanisms, while an indirectacting chemical carcinogen: 1. Exhibits effects through a secondary mechanism, 2. May exert mutagenicity in Salmonella typhimurium TA102 or TA104 or mouse lymphoma, and 3. Has a very low or no ras-protooncogene mutation frequency in liver tumors. In addition, some indirect-acting chemical carcinogens may also exhibit tumorigenicity in a two-year chronic tumorigenicity bioassay, perhaps acting via mutational mechanisms not detected by Salmonella or mouse lymphoma. It is known that Salmonella typhimurium TA102 and TA104 are sensitive to chemicals that exert mutations mediated by free radical formation (89,90,91). The chemically-induced free radicals formed in vivo may result in induction of lipid peroxidation and endogenous DNA adduct formation, oxidative stress, and peroxisome proliferation. The mouse lymphoma assay detects carcinogens acting through point mutational, recombinational, deletional, or clastogenic mechanisms. Therefore, further investigation of whether or not mutagenicity in Salmonella typhimurium TA102 and/or TA104 and mouse lymphoma can serve as biomarkers or complementary indexes to distinguish directacting and indirect-acting genotoxic carcinogens is needed. As described earlier, when direct-acting carcinogens induce liver tumors in B6C3F1 mice, the resulting liver tumors contain a very high ras-protooncogene mutation frequency. In contrast, the liver tumors of B6C3F1 mice that are induced by indirectacting chemical carcinogens usually contain a very low ras-protooncogene mutation frequency (eg, similar to the background level observed in the control animals), which indicates a mechanism of mutation induction other than point mutation. The neonatal mouse tumorigenicity bioassay, combined with supporting studies for mechanism of toxicity as outlined in Table 5, is suggested as

724

Genotoxic Mechanism

Neonatal Mouse Bioassay

Secondary Mechanism Evidence

Direct-acting carcinogen Indirect-acting carcinogen

+ −

+ −

− +

+ −

− ±

Noncarcinogen











Chemical

TA 98 or TA102 or TA100 TA104

Mutation Frequency in Ras-oncogene Very high Very low or − −

Chronic Bioassay

Mouse Lymphoma Bioassay

+ +

+ ±





Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

TABLE 5 Proposed Test Battery to Distinguish Direct-Acting and Indirect-Acting Genotoxic Carcinogens, and Noncarcinogens, Based on the Neonatal B6C3F1 Tumorigenicity Bioassay

Neonatal Mouse Tumorigenicity Bioassay

an excellent method to distinguish between direct-acting genotoxic and indirect-acting genotoxic carcinogens. It also may be desirable to be able to understand the mechanism of cancer induction at the molecular level. It is known that, when tested in the mouse, in most cases the majority of liver tumors induced by direct-acting chemical genotoxic carcinogens contain rasprotooncogene mutations. Furthermore, in all cases tested a majority of the liver tumors of mice treated with indirect-acting chemical carcinogens do not contain ras-protooncogene mutations. Thus, in order to validate further the neonatal mouse tumorigenicity bioassay, it is important to determine the genetic alterations in these liver tumors. The authors therefore propose that study of other genetic mechanisms, such as tumor suppressor gene mutations, other protooncogene mutations, and the causes of genomic instability and gene amplication, should be included for investigation. It is believed that the neonatal mouse tumorigenicity in combination with the other endpoints can convincingly distinguish between a direct-acting genotoxic carcinogen and an indirect-acting genotoxic carcinogen. In many cases, depending on the results from the accompanying short-term tests, this neonatal model should also be able to distinguish between an indirect-acting carcinogen and a noncarcinogen; and if not, bioassay of this compound by a chronic bioassay will be required. Acknowledgments—The authors recognize Drs. Frederick A. Beland, Joseph F. Contrera, Joseph J. DeGeorge, and K. Barry Delclos for suggestions and critical review of this manuscript.

REFERENCES 1. Pietra G, Spencer K, Shubik P. Response of newly born mice to a chemical carcinogen. Nature. 1959; 183:1689. 2. Pietra G, Rappaport H, Shubik P. The effects of carcinogenic chemicals in newborn mice. Cancer. 1961;14:308–317. 3. Fiore-Donati L, et al. Leukaemogenesis by urethan in newborn Swiss mice. Nature. 1961;190:278–279.

725 4. Klein M. Induction of lymphocytic neoplasms, hepatomas, and other tumors after oral administration of urethan to infant mice. J Natl Cancer Institute. 1962; 29(6):1035–1046. 5. Klein M. Susceptibility of strain B6AF1/J hybrid infant mice to tumorigenesis with 1,2-benzanthracene, deoxycholic acid, and 3-methylcholanthrene. Cancer Res. 1963;23:1701–1707. 6. Toth B, Shubik P. Carcinogenesis in Lewis rats injected at birth with 7,12-dimethylbenz(a)anthracene. Br J Cancer. 1963;17:540–545. 7. Toth B, Rappaport H, Shubik P. Accelerated development of malignant lymphomas in AKR mice injected at birth with 7,12-dimethylbenz(a)anthracene. Proceedings of the Society for Experimental Biology & Medicine. 1962;110:881–884. 8. Toth B. Development of malignant lymphomas by cell-free filtrates prepared from a chemically induced mouse lymphoma. Proceedings of the Society for Experimental Biology & Medicine. 1963;112: 873–875. 9. Kelly MG, Newton WL, O’Gara RW. Susceptibility of newborn germ-free mice to tumor induction by 3methylcholanthrene. Cancer Res. 1963;23:978–983. 10. Roe FJC, Mitchley BCV, Walters M. Tests for carcinogenesis using newborn mice: 1,2-benzanthracene, 2-naphthylamine, 2-naphthylhydroxylamine and ethyl methane sulphonate. Br J Cancer. 1963;17: 255–260. 11. Roe FJC, Millican D, Mallety JM. Induction of melanotic lesions of the iris in rats by urethane given during the neonatal period. Nature. 1963;199:1201– 1202. 12. Roe FJ, Waters MA. Induction of hepatoma in mice by carcinogens of the polycyclic hydrocarbon type. Nature. 1967;214(85):299–300. 13. Chieco-Bianchi L, et al. Role of age and other factors influencing the oncogenic effect of chemical carcinogens in mice. Tumori. 1965;51:53–68. 14. Vesselinovitch SD, Mihailovich N. Significance of newborn age and dose of urethan in leukemogenesis. Cancer Res. 1966;26(8):1633–1637. 15. Walters MA. The induction of lung tumors by the injection of 9,10-dimethyl-1,2-benzanthracene (DMBA)into newborn suckling and young adult mice. Br J Cancer. 1966;20:148–160. 16. Walters MA, Roe FJ. The time of appearance of lung tumours in mice injected when newly born with 9,10-dimethyl-1,2-benzanthracene (DMBA). Br J Cancer. 1966;20(1):161–167. 17. Walters MA, Roe FJ. The induction of skin tumours in mice by neonatal injection of 9,10-dimethyl-1,2benzanthracene (DMBA) followed by applications of croton oil to the skin. Br J Cancer. 1967;21(2): 358–366. 18. Walters MA, et al. Further tests for carcinogenesis using newborn mice: 2-naphthylamine, 2-naphthylhydroxylamine, 2-acetylaminofluorene and ethyl methane sulphonate. Br J Cancer. 1967;21(2):367– 372.

726

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

19. Della Porta G, Terracini B. Chemical carcinogenesis in infant animals. Progress Experimental Tumor Res. 1969;11:334–363. 20.Toth B. A critical review of experiments in chemical carcinogenesis using newborn animals. Cancer Research. 1968;28:727–738. 21. Vesselinovitch SD, Rao KV, Mihailovich N. Neoplastic response of mouse tissues during perinatal age periods and its significance in chemical carcinogenesis. Natl Cancer Institute Monographs. 1979 (51):239–250. 22. Vesselinovitch SD. Perinatal mouse liver carcinogenesis as a sensitive carcinogenesis model and the role of the sex hormonal environment in tumor development. [Review]. Progress Clin Biolog Res. 1990;331:53–68. 23. Fujii K. Evaluation of the newborn mouse model for chemical tumorigenesis. Carcinogenesis. 1991; 12(8):1409–1415. 24. Flammang TJ, et al. Neonatal mouse assay for tumorigenicity: Alternative to the chronic rodent bioassay. Regulatory Toxicol Pharmacol. 1997;26:230– 240. 25. Dooley KL, et al. Comparative carcinogenicity of the food pyrolysis product, 2-amino-5-phenylpyridine, and the known human carcinogen, 4-aminobiphenyl, in the neonatal B6C3F1 mouse. Cancer Letters. 1988;41(1):99–103. 26. Dooley KL, et al. Comparative carcinogenicity of 4-aminobiphenyl and the food pyrolysates, Glu-P1, IQ, PhIP, and MeIQx in the neonatal B6C3F1 male mouse. Cancer Letters. 1992;62(3):205–209. 27. Hecht SS, et al. Tumorigenicity of 5-methylchrysene dihydrodiols and dihydrodiol epoxides in newborn mice and on mouse skin. Cancer Res. 1985;45(4): 1449–1452. 28. Fu PP, et al. Comparative formation of DNA adducts of nitro-polycyclic aromatic hydrocarbons in mouse and rat liver microsomes and cytosols. Polycyclic Aromatic Compounds. 1996;10:187–194. 29. Wiseman RW, et al. Structure-activity studies of the hepatocarcinogenicities of alkenylbenzene derivatives related to estragole and safrole on administration to preweanling male C57BL/6J × C3H/HeJ F1 mice. Cancer Res. 1987;47(9):2275–2283. 30. Lai CC, et al. Initiation of hepatocarcinogenesis in infant male B6C3F1 mice by N-hydroxy-2-aminofluorene or N-hydroxy-2-acetylaminofluorene depends primarily on metabolism to N-sulfooxy-2-aminofluorene and formation of DNA-(deoxyguanosin-8-yl)2-aminofluorene adducts. Carcinogenesis. 1987; 8(3):471–478. 31. LaVoie EJ, et al. Tumorigenic activity of the 4,5and 9,10-dihydrodiols of benzo[j]fluoranthene and their syn- and anti-diol epoxides in newborn mice. Cancer Res. 1994;54(4):962–968. 32. Conney AH. Induction of microsomal enzymes by foreign chemicals and carcinogenesis by polycyclic aromatic hydrocarbons: G. H. A. Clowes Memorial Lecture. Cancer Res. 1982;42(12):4875–4917.

33. Fu PP, et al. Carcinogenicity in the B6C3F1 neonatal mouse: a potential alternative bioassay. In Progress in Clinical and Biological Research. LeBroeuf RA, Slaga TJ, Tennant R, Eds. New York: Wiley-Liss; 1998. 34. Boberg EW, et al. Strong evidence from studies with brachymorphic mice and pentachlorophenol that 1′sulfooxysafrole is the major ultimate electrophilic and carcinogenic metabolite of 1′-hydroxysafrole in mouse liver. Cancer Res. 1983;43(11):5163–5173. 35. Delclos KB, et al. 4-Aminoazobenzene and N,Ndimethyl-4-aminoazobenzene as equipotent hepatic carcinogens in male C57BL/6 × C3H/He F1 mice and characterization of N-(deoxyguanosin-8-yl)-4aminoazobenzene as the major persistent hepatic DNA-bound dye in these mice. Cancer Res. 1984; 44(6):2540–2550. 36. Wislocki PG, et al. Tumorigenicity of nitrated derivatives of pyrene, benz[a]anthracene, chrysene and benzo[a]pyrene in the newborn mouse assay. Carcinogenesis. 1986;7(8):1317–1322. 37. Fu PP, et al. Caloric restriction profoundly inhibits liver tumor formation after initiation by 6-nitrochrysene in male mice. Carcinogenesis. 1994;15(2):159– 161. 38. Fu PP, et al. Potent tumorigenicity of 7-chlorobenz[a]anthracene and 7-bromobenz[a]anthracene in the neonatal B6C3F1 male mouse. Cancer Letters. 1996;101(1):37–42. 39. Von Tungeln LS, et al. Inhibitory effect of caloric restriction on tumorigenicity induced by 4-aminobiphenyl and 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) in the CD1 newborn mouse bioassay. Cancer Letters. 1996;104(2):133–136. 40. LaVoie EJ, et al. Tumorigenic activity of fluoranthene, 2-methylfluoranthene and 3-methylfluoranthene in newborn CD-1 mice. Carcinogenesis. 1994; 15(10):2131–2135. 41. Xia Q, et al. Liver tumors induced in B6C3F1 mice by 7-chlorobenz[a]anthracene and 7-bromobenz[a]anthracene contain K-ras oncogene mutations. Cancer Letters. 1998;123:21–25. 42. Lewtas J, et al. Comparative tumorigenicity of complex environmental mixtures in the newborn mouse and mouse skin tumor assays. Proceedings of the Annual Meeting of the American Association for Cancer Research. 1993;34:A697. 43. Epstein SS, et al. Carcinogenicity of organic particulate pollutants in urban air after administration of trace quantities to neonatal mice. Nature. 1966;212: 1305–1307. 44. Vesselinovitch SD. Factors modulating response to carcinogenic mutagens. In Progress in Environmental Mutagensis. Alacevid ME, Ed. Amsterdam: Elsevier/North Holland Biomedical Press; 1980: 281–296. 45. Vesselinovitch SD, et al. Aflatoxin B1, a hepatocarcinogen in the infant mouse. Cancer Res. 1972; 32(11):2289–2291. 46. Park KK, et al. Vinyl carbamate epoxide, a major

Neonatal Mouse Tumorigenicity Bioassay strong electrophilic, mutagenic and carcinogenic metabolite of vinyl carbamate and ethyl carbamate (urethane). Carcinogenesis. 1993;14(3):441–450. 47. Delclos KB, et al. Sulfuric acid esters as major ultimate electrophilic and hepatocarcinogenic metabolites of 4-aminoazobenzene and its N-methyl derivatives in infant male C57BL/6J × C3H/HeJ F1 (B6C3F1) Mice. Carcinogenesis. 1986;7(2):277– 287. 48. Miller JA, et al. DNA adduct and tumor formation from N-hydroxy-2-aminofluorene and N-hydroxy2-acetylaminofluorene in the livers of infant male B6C3F1 mice. Carcinogenic and Mutagenic Responses to Aromatic Amines and Nitroarenes. King CM et al, eds. New York: Elsevier; 1988. 49. Wiseman RW. Hepatocarcinogenesis in the preweanling male B6C3F1 mouse with alkenylbenzene derivatives: Hepatic DNA adducts, structure-carcinogenicity relationships, and activating mutations in the c-Ha-ras proto-oncogene. Diss Abstr Int (Sci). 1987; 47(9):3657. 50. Wiseman RW, et al. Characterization of c-Ha-ras proto-oncogene mutations in chemically induced hepatomas of the B6C3F1 mouse. Proceedings of the Annual Meeting of the American Association of Cancer Research. 1987;28:147. 51. Fennell TR, et al. Major role of hepatic sulfotransferase activity in the metabolic activation, DNA adduct formation, and carcinogenicity of 1′-hydroxy-2′,3′dehydroestragole in infant male C57BL/6J × C3H/ HeJ F1 mice. Cancer Res. 1985;45(11 Pt 1):5310– 5320. 52. Leithauser MT, et al. 1, N6-Ethenoadenosine formation, mutagenicity and murine tumor induction as indicators of the generation of an electrophilic epoxide metabolite of the closely related carcinogens ethyl carbamate (urethane) and vinyl carbamate. Carcinogenesis. 1990;11(3):463–473. 53. Lai CC, et al. N-Sulfooxy-2-aminofluorene Is the major ultimate electrophilic and carcinogenic metabolite of N-hydroxy-2-acetylaminofluorene in the livers of infant male C57/6J × C3H/HeJ F1 (B6C3F1) mice. Carcinogenesis. 1985;6(7):1037–1045. 54. Von Tungeln LS, et al. Tumorigenicity of nitro-polycyclic aromatic hydrocarbons in the neonatal B6C3F1 male mouse. Proceedings of the Annual Meeting of the American Association of Cancer Research. 1994;35:A969. 55. Herreno-Saenz D, et al. DNA adduct formation by 2-nitrofluoranthene in Salmonella typhimurium and neonatal B6C3F1 mice. Polycyclic Aromatic Compounds. 1994;6:79–85. 56. Dipple A. Polycyclic aromatic hydrocarbon carcinogenesis. An introduction. ACS Symp Ser. 1985(283): 1–17. 57. Fu PP, et al. Comparative regioselective and stereoselective metabolism of 7-chlorobenz[a]anthracene and 7-bromobenz[a]anthracene by mouse and rat liver microsomes. Carcinogenesis. 1991;12(3):371– 378.

727 58. Fu PP, Von Tungeln LS, Chou MW. Stereoselective metabolism of 7-chlorobenz(a)anthracene by rat liver microsomes. Absolute configurations and optical purities of trans-dihydrodiol metabolites. Molecular Pharmacol. 1985;28(1):62–71. 59. Wiseman RW, et al. Activating mutations of the c-Ha-ras protooncogene in chemically induced hepatomas of the male B6C3F1 mouse. Proceedings of the National Academy of Sciences of the United States of America. 1986;83(16):5825–5829. 60. Fox TR, et al. Mutational analysis of the H-ras oncogene in spontaneous C57BL/6 × C3H/He mouse liver tumors and tumors induced with genotoxic and nongenotoxic hepatocarcinogens. Cancer Res. 1990; 50(13):4014–4019. 61. Reynolds SH, et al. Detection and identification of activated oncogenes in spontaneously occurring benign and malignant hepatocellular tumors of the B6C3F1 mouse. Proceedings of the National Academy of Sciences of the United States of America. 1986;83(1):33–37. 62. Reynolds SH, et al. Activated oncogenes in B6C3F1 mouse liver tumors: implications for risk assessment. Science. 1987;237:1309–1316. 63. Balmain A, Brown K. Oncogene activation in chemical carcinogenesis. Advances Cancer Res. 1988;51: 147–182. 64. Brown K, Burns PA, Balmain A. Transgenic approaches to understanding the mechanisms of chemical carcinogenesis in mouse skin. Toxicol Lett. 1995; 82–3:123–130. 65. Helvering LM, et al. H-ras 61st codon activation in archival proliferative hepatic lesions isolated from female B6C3F1 mice exposed to the leukotriene D4antagonist, LY171883. Carcinogenesis. 1994;15(2): 331–333. 66. Fox TR, Watanabe PG. Detection of a cellular oncogene in spontaneous liver tumors of B6C3F1 mice. Science. 1985;228(4699):596–597. 67. Richardson KK, et al. Genetic alterations in the 61st codon of the H-ras oncogene isolated from archival sections of hepatic hyperplasias, adenomas and carcinomas in control groups of B6C3F1 mouse bioassay studies conducted from 1979 to 1986. Carcinogenesis. 1992;13(6):935–941. 68. Richardson KK, et al. Temporal changes in the mutant frequency and mutation spectra of the 61st codon of the H-ras oncogene following exposure of B6C3F1 mice to N-nitrosodiethylamine (DEN). Carcinogenesis. 1992;13(7):1277–1279. 69. Sohn YW, et al. Activation of H-ras oncogenes in male B6C3F1 mouse liver tumors induced by vinthionine or 2-chloroethyl-methyl sulfide. Carcinogenesis. 1996;17(6):1361–1364. 70. Manjanatha MG, et al. H- and K-ras mutational profiles in chemically induced liver tumors from B6C3F1 and CD-1 mice. J Toxicol Environ Health. 1996;47(2):195–208. 71. Stowers SJ, et al. Detection of activated proto-oncogenes in N-nitrosodiethylamine-induced liver tu-

728

Fu, Von Tungeln, Yi, Xia, Casciano, Flammang, and Kadlubar

mors: a comparison between B6C3F1 mice and Fischer 344 rats. Carcinogenesis. 1988;9(2):271–276. 72. Dragani TA, et al. Incidence of mutations at codon 61 of the Ha-ras gene in liver tumors of mice genetically susceptible and resistant to hepatocarcinogenesis. Oncogene. 1991;6(2):333–338. 73. Buchmann A, et al. Mutational activation of the c-Ha-ras gene in liver tumors of different rodent strains:correlation with susceptibility to hepatocarcinogenesis. Proceedings of the National Academy of Sciences of the United States of America. 1991; 88(3):911–915. 74. Chen B, et al. Dose-dependent ras mutation spectra in N-nitrosodiethylamine induced mouse liver tumors and 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone induced mouse lung tumors. Carcinogenesis. 1993;14(8):1603–1608. 75. Moulds BA, Goodman JI. Spontaneous mutation at codon 61 of the Ha-ras gene in the nascent liver of B6C3F1, C3H/He and C57BL/6 mice. Mutation Research. 1994;311(1):1–7. 76. Stanley LA, et al. Ras mutations in methylclofenapate-induced B6C3F1 and C57BL/10J mouse liver tumours. Carcinogenesis. 1994;15(6):1125–1131. 77. Goodman JI, et al. Mouse liver carcinogenesis: mechanisms and relevance. Fundamental Appl Toxicol. 1991;17(4):651–665. 78. Huff JE, et al. Carcinogenesis studies: Results of 398 experiments on 104 chemicals from the U.S. National Toxicology Program. Ann New York Academy Sciences. 1988;534:1–30. 79. Ashby J, Tennant RW. Chemical structure, Salmonella mutagenicity and extent of carcinogenicity as indicators of genotoxic carcinogenesis among 222 chemicals tested in rodents by the U.S. NCI/NTP. Mutation Research. 1988;204(1):17–115. 80. Tennant RW. Relationships between in vitro genetic toxicity and carcinogenicity studies in animals. Ann New York Academy Sciences. 1988;534:127–132. 81. Contrera JF, Jacobs AC, DeGeorge JJ. Carcinogenicity testing and the evaluation of regulatory requirements for pharmaceuticals. Regulatory Toxicol Pharmacol. 1997;25(2):130–145.

82. Batzinger RP, Ou SY, Bueding E. Antimutagenic effects of 2(3)-tert-butyl-4-hydroxyanisole and of antimicrobial agents. Cancer Res. 1978;38(12): 4478–4485. 83. Stopper H, et al. An investigation of micronucleus and mutation induction by oxazepam in mammalian cells. Mutagenesis. 1993;8(5):449–455. 84. Melnick RL, Kohn MC, Portier CJ. Implications for risk assessment of suggested nongenotoxic mechanisms of chemical carcinogenesis. Environmental Health Perspect. 1996;104(Suppl 1):123–134. 85. Ni YC, et al. Mouse liver microsomal metabolism of chloral hydrate, trichloroacetic acid, and trichloroethanol leading to induction of lipid peroxidation via a free radical mechanism. Drug Metabol Disposition. 1996;24(1):81–90. 86. Ni YC, Kadlubar FF, Fu PP. Formation of malondialdehyde-modified 2′-deoxyguanosinyl adduct from metabolism of chloral hydrate by mouse liver microsomes. Biochemical & Biophysical Res Comm. 1995; 216(3):1110–1117. 87. Von Tungeln LS, et al. Tumorigenicity of chloral hydrate (CH), trichloroacetic acid (TCA), trichloroethanol (TCE), and malondialdehyde (MDA) in the B6C3F1 neonatal mouse. Proceedings of the Annual Meeting of the American Association of Cancer Research. 1997;38:354. 88. FDA/ICH. International Conference on Harmonisation; Draft Guidelines on Testing for Carcinogenicity of Pharmaceuticals. Federal Register. 1996;61(163): 43298–43300. 89. Levin DE, et al. A new Salmonella tester strain (TA102) with A × T base pairs at the site of mutation detects oxidative mutagens. Proceedings of the National Academy of Sciences of the United States of America. 1982;79(23):7445–7449. 90. Levin DE, et al. Detection of oxidative mutagens with a new Salmonella tester strain (TA102). Methods Enzymol. 1984;105:249–254. 91. Chesis PL, et al. Mutagenicity of quinones: pathways of metabolic activation and detoxification. Proceedings of the National Academy of Sciences of the United States of America. 1984;81(6):1696–1700.