In Vivo Mutagenicity and Hepatocarcinogenicity of 2-Amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) in Bitransgenic c-myc/λlacZ Mice Free

MeIQx³ is a member of a family of mutagenic/carcinogenic HCAs found in cooked meats, including beef, chicken, and fish. MeIQx is currently recognized as the second most prevalent of the HCAs in the human diet (1). In rodent carcinogenicity bioassays, MeIQx has been shown to be a strong liver carcinogen. In a previous study, the incidence of hepatic cancer in CDF1 mice exposed to 0.06% MeIQx in the diet for 84 weeks was 43% and 91% in males and females, respectively (2).

MeIQx is a procarcinogen that requires metabolic activation to reactive ester derivatives to form DNA adducts, induce mutations, and initiate the process of carcinogenesis. In experimental animals (including rats, mice, and monkeys) who were fed MeIQx (3, 4, 5, 6), MeIQx has been shown to form adducts at guanine nucleotides. Studies to determine the structures of MeIQx-DNA adducts reveal that the electrophilic metabolite of MeIQx reacts predominantly at the C8 and N² positions of guanine, forming N²-(deoxyguanosin-8-yl)-MeIQx and 5-(deoxyguanosin-N²-yl)-MeIQx, respectively (5). N²-(deoxyguanosin-8-yl)-MeIQx is the major adduct of MeIQx found in liver and other tissues of animals who were given MeIQx (3, 4).

MeIQx is known to be a potent mutagen in several in vitro assay systems, including the Ames Salmonella mutagenicity assay, and in the lacZ gene of Escherichia coli (7, 8, 9, 10). MeIQx has also been shown to be genotoxic in vivo, as assessed by an animal-mediated microbial assay (11), and in the lacZ gene of the Muta mice (12). As with other HCAs (13), in vitro mutagenicity studies support that MeIQx-DNA adducts play a role in the induction of specific genetic mutations (10).

In the paradigm of the multistep model of carcinogenesis, the accumulation of multiple genetic alterations in critical genes is essential for the development of cancer. Deregulation of the c-myc gene by overexpression or amplification is one alteration that has been implicated in the development of hepatocellular carcinoma in both experimental models and humans (14). c-Myc is a proto-oncogene involved in the regulation of DNA synthesis, differentiation, apoptosis, and cell cycle progression. Deregulation of c-myc can lead to a loss of cell cycle control, improper initiation of DNA synthesis, and genetic instability (14, 15, 16). Recent studies in a c-myc transgenic mouse model have shown, however, that hepatic overexpression of c-myc leads to a low incidence of hepatocellular carcinoma after a long latency period (17, 18). Additional cooperating genetic alterations seem to be essential for hepatocarcinogenesis under conditions of c-myc deregulation (19).

The central role of c-myc in cell cycle regulation and the requirement for multiple genetic alterations for hepatocarcinogenesis raise the question of whether the sensitivity to a chemical mutagen and hepatocarcinogen, such as MeIQx, is affected by overexpression of the c-myc oncogene. In a previous study, this laboratory described the use of a bitransgenic mouse harboring both the c-myc gene and the lacZ mutational reporter gene (12). After an acute exposure, the in vivo mutagenicity of MeIQx and other HCAs was shown to be higher in c-myc/λlacZ bitransgenic mice than in control (C57Bl/λlacZ) mice. This study suggested the possibility of an interaction between the specific HCAs and the overexpressed c-myc gene. In the current study, we evaluate the roles of MeIQx-induced DNA adducts and mutagenesis in hepatocarcinogenesis, address whether c-myc overexpression modulates the mutagenicity of MeIQx, and examine if hepatic overexpression of c-myc alters the susceptibility to MeIQx hepatocarcinogenesis.

MeIQx was purchased from Toronto Research Chemicals (North York, Ontario, Canada). AIN-76 diet was purchased from Harlan Teklad (Madison, WI). P-gal was purchased from Sigma Chemical Co. (St. Louis, MO). RNAce-It RNase mixture and Giga-Pack II were obtained from Stratagene (La Jolla, CA). The LacZ⁻ galE⁻ E. coli C strain with overexpressed galT and galK genes was a generous gift from Dr. Jan Vijg (Harvard Medical School, Boston, MA).

The Muta^TM mouse [BALB/c × DBA/2 (CD2F1)] carrying the λlacZ transgene was purchased from Hazleton Research Products (Denver, PA). Female c-myc transgenic mice (C57Bl/6J × CBA/J) were a generous gift from Dr. S. S. Thorgeirsson (18). C57Bl/λlacZ and c-myc/λlacZ mice were generated by breeding male Muta mice with female C57Bl/6J and c-myc transgenic mice. F1 mice were weaned onto AIN-76 diet containing 0.06% (w/w) MeIQx or AIN-76 (control) diet. Food and water were provided ad libitum throughout the study. Animal housing and care were in accordance with NIH guidelines.

Mice were euthanized after 30 and 40 weeks on diet by cervical dislocation and autopsied. Livers were examined macroscopically for the appearance of tumor. Portions of liver and liver tumors were fixed in 10% formalin or promptly frozen in liquid nitrogen and stored at −80°C until DNA isolation. All fixed tissues were embedded in paraffin, and sections were stained with H&E. Sections of nontumorous liver and grossly visible tumors were examined, and the latter diagnosed as hepatocellular adenomas or carcinomas (20).

For ³²P-postlabeling analysis, genomic DNA was isolated from frozen tissue essentially by the method described previously (21). The analysis of MeIQx adducts in genomic DNA was performed by ³²P-postlabeling analysis as before (3, 4), using the intensification (ATP-deficient) method (22). This method resolves [³²P]ATP-labeled bisphosphonucleotide adducts on polyethyleneimine-cellulose thin layer sheets as fingerprints after autoradiography.

High molecular weight DNA was isolated from frozen mouse liver and tumor tissues according to the protocol described earlier (23). Briefly, tissues were disaggregated and homogenized in dounce buffer [14 mm sodium phosphate (pH 8.0), 137 mm NaCl, 3 mm KCl, and 10 mm NaEDTA) containing 1/50 volume of RNAce-It RNase mixture. An equal volume of lysis buffer [2 mg/ml proteinase K, 2% SDS, and 0.1 M NaEDTA (pH 7.5)] was added. After a 3-h incubation at 50°C, the DNA was sequentially extracted with phenol/chloroform. The DNA was ethanol-precipitated, spooled onto a hooked glass Pasteur pipette, and then dissolved in 10 mm Tris buffer (pH 8.0) at 37°C to a concentration of 1–3 mg/ml.

To determine the frequency of mutations in the lacZ transgene in liver of mice, λgt10 DNA was rescued from genomic DNA by packaging into λ phage using Giga-Pack II packaging extract (23). Each packaging reaction was adsorbed to E. coli C at log phase growth. The reactions were then plated with and without P-gal. The number of plaques were counted after overnight incubation at 37°C. The lacZ ⁻ mutant frequency was determined as the ratio of lacZ ⁻ mutant plaques obtained in the presence of P-gal to the number of plaques recovered on the titer plates (23, 24). From each mouse-liver DNA sample, a minimum of 300,000 total plaques were scored on the titer plates.

LacZ mutants were sequenced as described previously (23). Briefly, mutant plaques were picked randomly from plates and expanded in liquid culture before DNA isolation. Regions of the lacZ gene were amplified by PCRs and sequenced using the previously reported sequencing primers (23). Sequencing reactions were performed using a DNA sequencing kit for dye terminator cycle reactions (Perkin-Elmer Corp., Foster City, CA). Automated sequencing was carried out on an ABI Prism 377 DNA Sequencer (Perkin-Elmer Corp.).

The data were analyzed by both two-way ANOVA (mouse strain and carcinogen treatment) and Student’s t tests using the computer program Prizm (GraphPad Software Incorporated, San Diego, CA).

³²P-postlabeling analysis was used to measure MeIQx-DNA adduct levels in hepatic DNA from c-myc/λlacZ and C57Bl/λlacZ strains of mice who received the MeIQx diet for 30 weeks. Three principle adduct spots, identical to those identified previously as guanine nucleotide adducts of MeIQx (4), were detected in DNA from both strains. The profile of MeIQx-DNA adducts was also identical to that reported previously in c-myc/λlacZ and C57Bl/λlacZ strains of mice who were given multiple doses of MeIQx (12). No qualitative differences in MeIQx-DNA adduct formation was noted among the various treatment groups (data not shown). The major adduct comprising at least 50% of total adduct levels coincided chromatographically with the 3′, 5′ bisphosphate of N²-(deoxyguanosin-8-yl)-MeIQx, the C8-guanine adduct of MeIQx. For each gender, total MeIQx-DNA adduct levels were not statistically different between c-myc/λlacZ and C57Bl/λlacZ mice, although there did seem to be a trend for slightly higher adduct levels in mice harboring the c-myc transgene (Fig. 1). Irrespective of mouse strain, MeIQx-DNA adduct levels were significantly higher in the females than in the males (P < 0.05, two-factor ANOVA). No MeIQx-DNA adducts were detected in animals on control diet.

The mutant frequency in the lacZ gene of hepatic DNA was measured in c-myc/λlacZ and C57Bl/λlacZ mice on MeIQx or control diet for 30 and 40 weeks (Fig. 2). The dietary MeIQx exposure had a striking effect on mutant frequency in the lacZ gene in both c-myc/λlacZ and C57Bl/λlacZ mice. In both strains and at both time points, MeIQx-fed mice showed at least a 40-fold higher lacZ mutant frequency than mice on control diet (Fig. 2, compare A with B and C with D). The presence of the c-myc transgene also affected mutant frequency of the lacZ transgene in both control and MeIQx treatment groups. In the MeIQx treatment groups, mutant frequency was 1.4–2.6-fold higher in mice harboring the c-myc transgene than that in the control strain (C57Bl/λlacZ; Fig. 2, B and D). This difference was statistically significant in males after 30 weeks on diet and in both males and females after 40 weeks on MeIQx diet (Fig. 2 D, compare c-myc/λlacZ and C57Bl/λlacZ strains; Student’s t test, P < 0.05). In mice on control diet, the mutant frequency in the lacZ gene was 1.4–2.7-fold higher in mice that carried the c-myc transgene than in the control strain. The differences between the c-myc/λlacZ and C57Bl/λlacZ strains was statistically higher in female mice after 30 weeks on control diet and in both male and female mice after 40 weeks on control diet (Student’s t test, P < 0.05). Gender differences in lacZ mutant frequencies were also apparent among the animal groups. Within a treatment group, females showed statistically higher mutant frequencies than the respective males (Student’s t test, P < 0.05).

Hepatic tumor incidence was recorded after 30 and 40 weeks of feeding MeIQx or control diet (Fig. 3, A and B). After 30 weeks on diet, tumors were detected only in the c-myc/λlacZ male mice who were fed MeIQx (Fig. 3 A). All of the four c-myc/λlacZ mice on MeIQx diet examined at this time point had developed several grossly visible tumor masses that were classified histologically as hepatocellular carcinomas. Grossly visible hepatic tumors were not detected in the other groups of male or female mice after 30 weeks. Microscopic changes, including the development of hepatic foci and adenomatous changes, were detected in two of four female c-myc/λlacZ mice who were fed MeIQx. Thus, at the 30-week time point, whereas neither MeIQx nor c-myc overexpression alone was sufficient to induce hepatocellular carcinomas, the combination of MeIQx treatment and c-myc overexpression was highly effective in inducing hepatocellular carcinomas in male mice.

After 40 weeks on diet, all of the groups of mice receiving MeIQx developed grossly visible hepatic tumors; however, strain and sex differences in tumor incidence were still observed (Fig. 3 B). The highest incidence of tumors at 40 weeks was observed in the male c-myc/λlacZ mice who were given MeIQx. At this time point, hepatocellular carcinoma development involved much of the liver. Grossly visible isolated hepatic tumors, identified as hepatocellular carcinomas, were also seen in 80% of female c-myc/λlacZ mice on the MeIQx diet. Hepatocellular carcinomas were also observed in 44% and 17% of male and female C57Bl/λlacZ mice on MeIQx diet for 40 weeks, respectively. Approximately 17% of male c-myc/λlacZ mice on control diet developed grossly visible tumors after 40 weeks. The lacZ mutant frequency was also analyzed in hepatocellular carcinomas from MeIQx-treated mice after 40 weeks. There were no significant differences between the mutant frequencies (mean ± SE, n) in hepatocellular carcinomas and normal tissue from C57Bl/λlacZ males (1911 ± 337, 3 versus 2433 ± 297, 3), c-myc/λlacZ males (5192 ± 907, 3 versus 4954 ± 335, 3), or c-myc/λlacZ females (8358 ± 1505, 3 versus 7136 ± 659, 3). Tumors from the other groups did not provide sufficient tissue for analysis.

Sequence analysis of the lacZ gene mutants from liver of male mice was carried out to assess the types of mutations induced in the four different groups of mice: C57Bl/λlacZ on control or MeIQx diet and c-myc/λlacZ mice on control or MeIQx diet (Table 1). Irrespective of the animal strain or MeIQx treatment, the principal sequence alteration observed in the lacZ gene was a single base substitution. The frequency of CpG mutations varied with treatment groups and mouse strains (Table 1). In all four groups, the majority of all single-base mutations were at G:C bp, which consisted of G:C to T:A transversions, G:C to A:T transitions, and G:C to C:G transversions. There was no apparent strand bias in the occurrence of base substitution mutations. In either strain, adenine mutations were observed only when mice were on the control diet.

To address the role of multiple factors in MeIQx hepatocarcinogenesis, MeIQx-DNA adduct formation, in vivo mutagenicity, and carcinogenicity were examined in mice carrying the lacZ mutational reporter gene and an overexpressed c-myc oncogene. It is generally recognized that HCA-DNA adduct formation plays a role in HCA carcinogenesis. In the current study, MeIQx-DNA adduct formation was associated with elevated mutant frequency in the liver of mice. Irrespective of mouse strain, the mutant frequency in mice on the MeIQx diet was over 40-fold higher than in control mice, and, accordingly, MeIQx-DNA adducts were detected only in MeIQx-treated mice. Comparison between male and female mice also illustrate an association between MeIQx-DNA adduct levels and in vivo mutagenicity. In either strain of mice, MeIQx-DNA adduct levels were 2-fold higher in female mice than in male mice (Fig. 1). In accordance with higher MeIQx-DNA adduct levels, the mutant frequency in the lacZ reporter gene was 1.4–2.6-fold higher in female mice than in male mice (Fig. 2). The difference in adduct levels between male and female mice is likely to be associated with a higher capacity for metabolic activation in the females (25).

In a previous study, we measured MeIQx-DNA adduct levels and mutant frequency in the lacZ gene in the liver of mice who were given 10 oral doses of MeIQx (12). Comparison with the current data shows that the level of MeIQx-DNA adducts detected in liver after a 30-week dietary exposure to MeIQx are, on average, 2-fold higher than the levels found after the 10-dose regimen. Interestingly, mutant frequency in the lacZ gene was at least 15–20-fold higher after a 30-week exposure than after a 10-day exposure, despite the mere doubling in MeIQx-DNA adduct levels. These findings suggest that the duration of exposure to MeIQx or the persistence of MeIQx-DNA adducts is a more important determinant of in vivo mutagenesis than is the total adduct level. It seems likely that the increase in mutagenic response over an extended exposure to MeIQx is the result of a greater period of time for DNA adduct formation and the fixation of mutations. The data also suggest that DNA damaging mutations accumulate with chronic exposure to MeIQx.

The mutation spectra found in the lacZ gene of MeIQx-treated mice is also consistent with the involvement of MeIQx-DNA adducts in mutagenesis in vivo. In accordance with the formation of guanine adducts, all of the base substitution mutations observed in MeIQx-treated mice, irrespective of strain, occurred at the guanine base. The contribution of the different MeIQx-DNA adducts, such as the C8- guanine adduct, N²-guanine adduct, or the role of oxidative base damage, including 8-oxo-dG formation by MeIQx in the mutation spectra of MeIQx, is not yet known. In a previous study, Solomon et al. (10) showed that base substitution mutations associated with MeIQx-DNA adducts in the lacZ gene of E. coli in vitro include primarily G to T transversions, followed by G to A transition mutations. Both of these types of mutations were also observed in MeIQx-treated mice, although in C57Bl/λlacZ mice, G to A transition mutations were more common than G to T transversions. It is noteworthy, however, that the percentage of guanine base substitution mutations in mice on control diet was similar to that observed in mice on the MeIQx diet. Adenine mutations, however, were observed only in mice (of either strain) on control diet, and these mutations have been shown to occur spontaneously in the lacZ gene of the Muta mice (26). Interestingly, c-myc overexpression seems to alter the percentage of various base substitution mutations induced either in control or in MeIQx-treated mice; the mechanisms involved in this alteration are not yet known.

The results from this study are consistent with the notion that the carcinogenicity of MeIQx is related to the accumulation of MeIQx-induced genetic mutations. Other studies have supported the general hypothesis that accumulation of mutations is pertinent to the development of cancer (27, 28). The high mutant frequency of the lacZ gene in MeIQx-treated mice paralleled the development of preneoplastic alterations and tumors in these mice. However, factors in addition to mutagenesis play a role in MeIQx-induced hepatocarcinogenesis. For example, in C57Bl/λlacZ mice on MeIQx diet for 40 weeks, lacZ mutant frequency was higher in female than in male mice, whereas tumor incidence was lower in females than in males. Sex differences in susceptibility to chemically induced hepatocarcinogenesis and sex differences in the incidence of hepatocarcinogenesis in humans (29, 30) have been noted previously. Androgen levels in the male have been shown to contribute to a higher susceptibility of males to liver cancer (30).

The use of the c-myc/λlacZ bitransgenic mouse strain for studies with MeIQx indicated that c-myc overexpression was associated with a dramatic acceleration of MeIQx-induced hepatocarcinogenesis. In fact, after 30 weeks on either MeIQx or control diet, only male c-myc mice fed MeIQx developed hepatocellular carcinoma. Neither c-myc overexpression alone nor MeIQx treatment alone induced hepatocellular carcinoma at this time point. These in vivo findings are consistent with a previous study showing that rodent embryo fibroblasts were more susceptible to in vitro transformation by chemical carcinogens when the cells expressed a high level of c-myc (31).

The results from the current study in bitransgenic mice further suggest that the mechanism of the synergism between c-myc and MeIQx in hepatocarcinogenesis involves an enhancement of MeIQx-induced mutagenesis. In mice on MeIQx diet, mutant frequency was ∼2-fold higher in the c-myc mice than in the control strain. Therefore, the relatively modest increase in mutations in the lacZ gene was associated with a 100% increase in the incidence of hepatocarcinogenesis. This finding raises the possibility that other types of genetic alterations, such as large deletion mutations or amplifications that are not detectable using the λ phage-based lacZ transgene system (23), also contribute to the enhancement of mutations and carcinogenesis. An additional explanation for this finding may be that, as a mutational reporter gene, lacZ provides an estimate of the mutations occurring in the genome, whereas mutations in specific critical genes that cooperate with c-myc are likely to lead to the enhancement of carcinogenesis.

Previous studies have shown that the rate of hepatic proliferation is significantly higher in the c-myc transgenic mice than in wild-type mice. For example, between 4 and 8 months of age, mitotic activity in mouse liver is 50–100-fold higher in male c-myc mice than in control mice not carrying the transgene (32). This high rate of hepatic proliferation in c-myc mice is likely to facilitate the fixation of MeIQx-induced mutations from MeIQx-DNA adducts. In the multistep model of carcinogenesis, the higher frequency of MeIQx mutations in the genome would be expected to contribute to the accumulation of critical genetic alterations that give rise to neoplasia. A previous study showed that MeIQx-DNA adduct levels were statistically higher in c-myc/λlacZ mice than in C57Bl/λlacZ mice after 10 doses of MeIQx (12). Although the trend for higher MeIQx-DNA adduct levels was observed with chronic feeding with MeIQx, the difference was not statistically significant. The results, therefore, suggest that the high rate of hepatic proliferation associated with c-myc overexpression rather than higher adduct levels in this strain is likely to be the major factor in determining higher mutant frequency in c-myc mice.

Recent studies have indicated that c-myc overexpression is associated with genomic instability (33, 34). For example, overexpression of c-myc in a variety of cell lines is followed by the amplification and rearrangement of the dihydrofolate reductase gene (34). The results shown here indicate that c-myc overexpression is associated with an increase in the frequency of spontaneous base substitution mutations, as well as the frequency of carcinogen/mutagen-induced base substitution mutations. In both male and female mice on control diet for 40 weeks, lacZ mutant frequency was significantly higher in mice that harbored the c-myc gene than in the control strain. The findings raise the possibility that in addition to amplification and genomic rearrangements, the genomic instability associated with c-myc overexpression also includes point mutations. To our knowledge, this is the first study showing that c-myc increases the frequency of base substitution mutations in vivo. Whereas enhanced replication rate and disruption of the G₁-S stage of the cell cycle with c-myc overexpression may partly account for the increase in mutagenesis (34), the mechanisms by which c-myc overexpression enhances the frequency of spontaneous base substitution mutations still requires further study.

Studies in transgenic mouse models have indicated that c-myc cooperates with a variety of other genes to enhance carcinogenesis. For example, the combination of c-myc with an overexpressed TGFα gene or with mutated H-ras in bitransgenic mouse models causes a striking acceleration of hepatocarcinogenesis (32, 35). The results from this study also indicate that the synergistic effect of MeIQx and c-myc is likely to involve MeIQx-induced mutations in a critical gene or series of genes that cooperate with c-myc to accelerate malignant conversion. Additional studies are required to elucidate the genes mutated by MeIQx that cooperate with the overexpressed c-myc gene to facilitate hepatocellular carcinogenesis.

We thank Dr. Snorri S. Thorgeirsson (National Cancer Institute, Bethesda, MD) for providing the c-myc transgenic mice and for helpful advice and Dr. Jan Vijg (Harvard Medical School, Boston, MA) for providing the E. coli C strain.