Frequency and Spectrum of lacI Mutations in the Liver of Big Blue Mice Following the Administration of Genotoxic Carcinogens Singly and in Series

Abstract

Transgenic mouse models offer a unique opportunity to study in vivo mutagenicity in any tissue of interest. In this study, the authors have determined the liver mutant frequency (MF) and mutational spectra (MS) of 12 week-old male Big Blue B6C3F1 transgenic mice exposed to the genotoxic carcinogens benzo[a]pyrene (B[a]P; 250 mg/kg/day), N-nitrosodimethylamine (NDMA; 7 mg/kg/day), or N-ethyl-1-nitrosourea (ENU; 50 mg/kg/day) singly (3 daily oral doses) or in series (B[a]P on day 1, NDMA on day 2, and ENU on day 3). All genotoxic agents, alone or in series, increased MF in the liver (three-to sixfold). MS analyses of liver DNA revealed a high percentage of G:C → A:T transitions in the control (88%) and the NDMA (64%) groups. In contrast, B[a]P, ENU, and the series treatment induced a high percentage (≥50%) of transversions. Significantly, 46% (19 out of 41) of the mutations in the series treatment group occurred at CpG dinucleotides, compared to less than 22% in the other treatment groups. The MS from the series exposure was most similar to B[a]P with a high percentage of transversion mutations occurring at guanine nucleotides (36%). These preliminary data suggest that genotoxic carcinogens, when exposed in series, produce a unique MS profile characterized not only by shifts in mutation class but also sequence context.

Transgenic technology has facilitated the development of several animal models that provide new and useful information about mutant frequency (MF) and mutational spectrum (MS) (Donehower et al. 1992; Gollapudi et al. 1998b; Gossen et al. 1989; Kohler et al. 1991b). The Big Blue mouse harbors a shuttle vector containing the lacI gene within its genome that is recoverable and amenable to mutation analysis (Kohler et al. 1991a, 1991b). A distinct advantage of the Big Blue assay is that spontaneous as well as treatment-related changes in MF and MS can be evaluated in vivo from any tissue (de Boer et al. 1998; Hill et al. 1999; Kohler et al. 1991a, 1991b). Furthermore, route of exposure limitations are not a factor and the model is able to detect both single and multiple base pair substitution and frameshift mutations induced by chemical carcinogens (Gollapudi, Jackson, and Stott 1998a; Sisk et al. 1994; Walker et al. 1996). MS have been characterized in this mouse model for a number of frequently used genotoxic carcinogens including benzo[a]pyrene (B[a]P), N-nitrosodimethylamine (NDMA), and N-ethyl-1-nitrosourea (ENU) (Monroe et al. 1998; Shane et al. 2000a, 2000b; Walker et al. 1996).

The majority of studies describing DNA alterations induced by a mixture of genotoxic carcinogens have all been conducted in vitro (Taylor, Setzer, and DeMarini 1995; Said et al. 1995, 1999). In these studies, the magnitude and type of DNA damage caused by the carcinogen mixtures could not be explained by simply adding the expected DNA damage caused by each carcinogen. It was concluded that the mutagenicity of a single carcinogen is often influenced in an antagonistic or synergistic manner by pretreatment with other genotoxic carcinogens (Taylor, Setzer, and DeMarini 1995; Said et al. 1995, 1999). Biological factors that may affect the mutagenicity of a specific carcinogen following pretreatment with other genotoxic agents include the modification of the spatial arrangement of the DNA target, the metabolism of the genotoxic carcinogen, and/or the DNA repair status of the target cell (Bailly, Kenani, and Waring 1997; Culp and Beland 1994; Said et al. 1995, 1999).

The primary purpose of this investigation was to examine the liver MF and MS from Big Blue B6C3F1 transgenic mice exposed to a series of mutagens (B[a]P, NDMA, and ENU) and to determine whether the MS generated from the series exposure would produce a profile more similar to only one of the exposure compounds, a blend of each, or a unique spectrum. The model genotoxicants were selected based on their distinctive MS in the Big Blue mouse model (de Boer, Mirsalis, and Glickman 1999; Hakura et al. 1998; JEMS/MMS 1996; Nishikawa et al. 1997; Shane et al. 2000a; Walker et al. 1996). Furthermore, we have selected these chemicals based on their diverse DNA adduct and metabolite profile. Although B[a]P and NDMA both require metabolic activation by different cytochrome P450 (CYP) enzymes (CYP1A1 and CYP2E1, respectively), ENU is a direct-acting mutagen (Hong and Yang 1989; Whitlock 1999). Also, NDMA and ENU are both alkylating agents, causing methyl and ethyl DNA adducts, respectively, whereas B[a]P induces bulky adducts.

MATERIALS AND METHODS

Materials

NaCl and MgSO₄·7H₂O were purchased from Fisher Scientific (Springfield, NJ). Polymerase chain reaction (PCR) primers, agarose, and yeast extract were purchased from Life Technologies (Grand Island, NY). Casein hydrolysate, pH 7.0, was purchased from ICN Biomedicals (Costa Mesa, CA). Mutant phage CM0, X-gal (5-bromo-4-chloro-3-indoyl-ß-d-galactopyranoside), and 0.25% maltose were purchased from Stratagene (La Jolla, CA). B[a]P (CAS no. 50-32-8), ENU (CAS no. 759-73-9), NDMA (CAS no. 62-75-9), and corn oil vehicle were purchased from Sigma Chemical (St. Louis, MO).

Treatment of Animals

Six week-old male transgenic Big Blue B6C3F1 mice were purchased from Stratagene. Mice were singly housed in plastic tubs with air-filtered tops and “Cell Sorb Plus” bedding material (A&W Products, New Philadelphia, OH) in rooms designed to maintain adequate environmental conditions (temperature, humidity, and photocycle). Mice were fed Purina Certified Rodent Chow 5002 (Purina Mills, St. Louis, MO) and municipal drinking water ad libitum during the prestudy and study periods. At 11 weeks of age, mice were randomized according to body weight and uniquely identified with metal ear tags. At 12 weeks, mice were administered three daily gavage doses of B[a]P (250 mg/kg/day), ENU (50 mg/kg/day), or NDMA (7 mg/kg/day) (Figure 1). An additional group received the same dose of B[a]P, NDMA, and ENU on days 1, 2, and 3, respectively. The dose levels were based upon the literature and selected to induce a clear mutagenic response in the tissue examined. All test chemicals were suspended in corn oil and administered by oral gavage at a dose volume of 10 ml/kg of body weight. Control mice received corn oil vehicle only. All treatment groups were held for an expression period of 21 days (study day 24) except for NDMA-treated animals, which were sacrificed after 11 days (study day 14) due to toxicity. Liver tissue was snap-frozen in liquid nitrogen and stored at –80°C until used for DNA isolation. Animal facilities were accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC, Frederick, MD, USA) and this study was approved by the Institutional Animal Care and Use Committee of the Dow Chemical Company (Midland, MI, USA).

Determination of Mutant Frequencies

DNA was isolated from liver tissue using the RecoverEase drop dialysis kit from Stratagene. The shuttle vector was recovered from DNA samples using two rounds of packaging with Transpack lambda packaging extract. Mutant frequencies at the lacI target were determined following the methods described by Stratagene (Big Blue Manual). Packaged DNA was added to the plating cells (Escherichia coli SCS-8) freshly grown in NZY broth (5 g/L NaCl, 2 g/L MgSO₄ ·7H₂O, 5 g/L yeast extract, 10 g/L casein hydrolysate, pH 7.0) supplemented with 0.25% maltose and 12.5 mM MgSO₄. To aid in evaluating conditions for optimal mutant detection, light blue mutant phage CM0 was plated by mixing 50 μl of the CM0 stock with 2 ml of SCS-8 cells as described below. To the mixture containing cells and phage, 35 ml of molten top agar (NZY medium with 0.7% agarose cooled down to 48°C to 50°C) containing 1.5 mg/ml X-gal was added and the contents poured into bottom agar plates (25 × 25 cm) containing approximately 250 ml of NZY medium with 1.5% agar. After the agar solidified, plates were inverted and incubated at 37°C for 16 to 20 h. Plates were scored visually using a red-filtered light box to enhance the visualization of faint blue plaques. The number of plaque forming units was estimated by counting three representative sections (each section representing 1/100 of the total area of the 25 × 25 cm plate) on two plates from each packaging reaction plated.

Preparation of λ DNA

Putative mutant plaques were verified as mutant or nonmutant by replating at low density in the presence of X-gal. Clearly defined blue mutant plaques were cored with a Pasteur pipette and the phage eluted in microfuge tubes containing 500 μl of SM buffer (5.8 g/L NaCl, 2 g/L MgSO₄·7H₂O, 0.01% gelatin (w/v) in 50 mM Tris-HCl buffer, pH 7.5) and 50 μl of chloroform. Eight microliters of phage stock was added to 100-μl PCR reactions (30 cycles of 94°C for 30 s, 53°C for 40 s, and 72°C for 75 s) that amplified a 1325-bp region of the lacI gene that included the entire translated region. The forward primer 5′-TTCTTTCCTGCGTTATCCC-3′ is located 5’ of the transcription initiation site at positions –129 to –111. The reverse primer 5′-TCCACACAACATACGAGCC-3′ is located 3′ of the translation stop codon at positions 1178 to 1196. PCR products were qualitatively analyzed by electrophoretic separation on ethid-ium bromide stained 2% agarose gels and prepared for DNA sequencing using Wizard PCR preps DNA purification system (Promega, Madison, WI). Purified PCR products were eluted with sterile ddH₂O and DNA concentrations quantified by ultraviolet (UV) spectrophotometry.

LacI Sequence Analysis

All DNA samples were sequenced by the Department of Energy (DOE) Plant Research Laboratory at Michigan State University (East Lansing, MI). Automated fluorescent sequencing was performed using Taq cycle sequencing on an ABI 377 Sequencer. Primers used for lacI sequence analysis were (1) 5′-CAAAACCTTTCGCGGTATGG-3′ located at position –33 to –14; (2) 5′-GTGTCGATGGTAGAACGAAGC-3′ located at position 314 to 334; and (3) 5′-ACTCGCAATCAAATTCAGCC-3′ located at position 644 to 663. Mutations in the lacI gene were identified by aligning mutant sequence data with the wild type lacI gene sequence using Omiga 1.1 sequence analysis software (Oxford Molecular, Oxford, England). Mutants from a minimum of three animals were analyzed in each treatment group. Seven nonmutants were also sequenced to evaluate the quality of the sequencing reactions. Once the mutant was identified from the individual plaque, no additional sequencing on that plaque was performed. DNA sequences that were ambiguous (i.e., large number of unidentified bases) were resequenced until the sequence was clearly defined. Out of 80 DNA samples sequenced, 76 were positive for mutants. This incidence is in agreement with studies reporting that in 3% to 5% of sequenced mutants, no mutation is found (Erfle et al. 1996). It has been suggested that mutations located outside the lacI coding region (i.e., operator region) may account for this observation.

Statistical Analyses

Mutant frequency data were first analyzed for normality and homogeneity of variance using Bartlett’s test. Data not showing equal variance were log transformed. For analysis, an analysis of variance (ANOVA) followed by Dunnett’s test for pairwise comparisons (control versus treated) was performed on log transformed MF data using JMP version 6.0.3 software (SAS Institute, Cary, NC, USA).

RESULTS

LacI MF from all treatment groups were significantly increased compared to control liver values (Table 1). At the dose levels examined, similar MF (4.5- to 6-fold) were observed in B[a]P, NDMA, and series treatment groups, with ENU inducing slightly lower values (3-fold).

The MS from all treatment groups are shown in Tables 2 and 3. Both the control mice and the NDMA-treated mice exhibited a high percentage (88% and 64%, respectively) of G:C → A:T transitions in the lacI gene from liver DNA samples. The majority of these transitions occurred on the coding strand, but the NDMA induced G:C → A:T transitions did not occur at CpG dinucleotides as observed in the controls. NDMA treatment also caused a small number of frameshift mutations (3 out of 11), two single-base-pair deletions, and one complex deletion involving a four-base-pair sequence (data not shown).

In contrast to NDMA, the majority of mutations observed following B[a]P and ENU exposure were transversions (56% and 50%, respectively). ENU induced primarily A:T → T:A transversions (3 out of 5); however, two G:C → T:A transversions were also identified. B[a]P exposure generated mainly G:C → C:G transversions (3 out of 4), but one G:C → T:A transversion was also observed. ENU and B[a]P treatment also induced a number of transitions and frameshifts, mostly occurring at non-CpG sites.

The series exposure produced mainly transversion mutations (58%), with G:C → T:A transversions (10 out of 24) being the most common, followed by A:T → C:G transversions (7 out of 24). A small number of G:C → C:G (5 out of 24) and A:T → T:A (2 out of 24) transversions were also identified. Furthermore, a small number of transition (25%) and frameshift mutations (17%) were observed. Interestingly, in the series treatment group, 12 out of 15 G:C → T:A and G:C → C:G transversions occurred at CpG sites, with most occurring on the coding strand of the lacI gene.

DISCUSSION

Male transgenic Big Blue B6C3F1 mice exposed singly to B[a]P induced a MS primarily consisting of G:C → C:G and G:C → T:A transversions (Table 2). This observation is consistent with commonly observed sequence changes following B[a]P exposure (Liu et al. 2005). B[a]P requires metabolic activation before generating DNA damage in the form of bulky DNA adducts. B[a]P is a polycyclic aromatic hydrocarbon (PAH) found in automobile exhaust, cigarette smoke, and charbroiled food (IARC 1983) that is primarily activated by aryl hydrocarbon receptor (AhR)-regulated cytochrome P450 (CYP) enzymes (Whitlock 1999). The AhR pathway activates genes such as CYP1a1, CYP1a2, CYP1b1, glutathione S-transferases, and UDP-glucoronosyltransferases (Nebert et al. 2000; Whitlock 1999). CYP enzymes transform PAHs to hydroxyl-containing metabolites that conjugate to glucoronides and sulphates via phase II enzymes. Several reactive intermediates, such as epoxides, are known to cause DNA adducts, predominantly N ²-deoxyguanosine (Jeffrey et al. 1976; Jennette et al. 1977; Koreeda et al. 1978; Osborne et al. 1976), whereas the radical cation forms depurinating adducts at the N7 position of adenine and guanine and the C8 position of guanine (Chen et al. 1996; Devanesan et al. 1992). B[a]P can also be enzymatically converted to a highly reactive ortho-quinone, benzo[a]pyrene-7,8-dione, that can form stable cyclic DNA adducts by reaction with 2′-deoxyguanosine and 2′-deoxyadenosine (Balu et al. 2004).

NDMA treated mice exhibited a high percentage of G:C → A:T transitions (64%); however, unlike the controls, none of these occurred at CpG sites. NDMA requires metabolic activation before inducing methyl DNA adducts (Hong and Yang 1985). Estimates indicate that air, diet, and smoking contribute to human exposure of a few micrograms per day (NTP 2005). The N ⁷-methylguanine (N7-meG) adduct is the principal DNA lesion induced by NDMA exposure. This adduct comprises approximately 70% of the total DNA methylation induced by NDMA, and is enzymatically repaired or spontaneously depurinated to produce mutagenic apurinic sites (Takeshita and Eisenberg 1994). The N7-meG adduct can potentially accumulate due to high levels of formation and slow repair (Bianchini and Wild 1994). Another adduct generated by NDMA exposure is O ⁶-methylguanine (O6-meG). This adduct is repaired by the protein O ⁶-methylguanineDNA methyltransferase, and if not repaired, can produce G:C → A:T transitions (Mirsalis, Mon-forte, and Winegar 1994; Pegg 1990; Pletsa et al. 1994; Rossi et al. 1989; Tan, Swann, and Chance 1994). The G:C → A:T transition has been associated with an increased risk of cancer (Bedell et al. 1982; Swenberg et al. 1982).

In this study, ENU exposure induced primarily transversion mutations (50%) (Table 2). Of these transversions, three were A:T → T:A, and two were G:C → T:A. ENU also induced two G:C → A:T transitions and two A:T → G:C transitions. Although the most common transversion mutation induced by ENU in vivo is A:T → T:A, ENU can induce a high percentage of A:T → C:G transversions following the induction of SOS proteins in Esherichia coli (Walker et al. 1996; Eckert, Ingle, and Drinkwater 1989) These SOS enzymes are involved in the mutagenic bypass of lesions like O ⁴-ethylthymine that block DNA replication (Eckert, Ingle, and Drinkwater 1989; Fix) 19930. Previous administration of B[a]P may induce a similar set of enzymes in mouse liver prior to ENU exposure and cause an increase in A:T → C:G transversions (Hanawalt 2001).

Similarly to NDMA, the majority of ENU induced mutations were not located at CpG sites. Unlike B[a]P and NDMA, ENU is a direct-acting mutagen that does not require chemical bioactivation. The potential for human exposure is limited because ENU is not produced or used in large quantities in the United States according to the Toxic Chemicals Release Inventory (NTP 2005). ENU reacts with oxygen in DNA, inducing modifications at the O6 position of guanine and the O2 and O4 positions of thymine. O ⁶-ethylguanine induces G:C → A:T transitions, O4-ethylthymine induces A:T → G:C transitions, and O ²-ethylthymine induces A:T → T:A transversions. The latter two adducts are more persistent in mammalian cells (Bronstein, Skopek, and Swenberg 1992; Jansen et al. 1994, 1995; Liem, Lim, and Li 1994; Mittelstaedt, Smith, and Heflich 1995; Walker et al. 1996).

The series exposure was conducted using B[a]P on day 1, NDMA on day 2, and ENU on day 3. One of the principal differences noticed between the serial and individual exposures was the occurrence of mutations at CpG sites (46% versus 12%, 9%, and 22%, respectively). The low incidence of mutation induction at CpG sites following exposure to ENU or NDMA, 8% and 15% respectively, has been demonstrated (de Boer, Mirsalis, and Glickman 1999; Walker et al. 1996). Previous studies in lacI and lacZ transgenic mice have shown B[a]P-induced mutational spectra at CpG sites in liver and other tissues (Shane et al. 2000a; Hakura et al. 1998). However, in these studies the frequency of G:C → T:A and G:C → C:G transversion mutations at CpG islands did not exceed 50%, with the exception of a study done with coal tar where these same transversion mutations at CpG sites exceeded 75% (Vogel et al. 2001), similar to our reports here. A more recent study using Big Blue mouse embryonic fibroblasts demonstrated an increase in G:C → T:A transversions at CpG sites when UVA irradiation was administered following B[a]P diol epoxide (Besaratinia and Pfeiffer 2003). It was suggested that consecutive treatments with different mutagens may significantly reduce the efficacy of nucleotide excision repair facilitating the mutagenic conversion of DNA adducts. Together these findings may suggest that B[a]P-induced mutations at CpG sites may be facilitated by neighboring adducts and/or altered DNA repair capacity caused by concurrent exposure to other mutagens.

The number of mutants examined in this study for each individual compound were relatively small (8, 11, and 9 total mutants for B[a]P, NDMA, and ENU respectively); however, the MS were similar to literature values (de Boer, Mirsalis, and Glick-man 1999; Shane et al. 2000a; Walker et al. 1996). The MS from the series exposure shared greater similarity to B[a]P than to the other individual compounds (Figure 2).

It is likely that the order of mutagen administration in the series treatment influenced the MS due to altered or enhanced DNA repair. Following exposure to B[a]P, the induction of DNA repair enzymes involved in excision repair of bulky adducts would be expected. Furthermore, exposure to NDMA may have induced repair processes similar to those stimulated by ENU (i.e., O ⁶-methyltransferase) due to the similar adducts induced by these compounds. However, it is unclear how different DNA repair processes (nucleotide excision repair vs base excision repair) might influence each other in vivo during the repair of their respective DNA adducts. Nevertheless, there are similarities between the MS observed in the series treatment of this study and that observed in a study investigating p53 mutations in smoking related human lung cancers (Yoon et al. 2001).

To summarize, all genotoxic agents, alone or in series, increased MF in the liver (three- to sixfold) A high percentage of G:C → A:T transitions were observed in the control (88%) and NDMA (64%) groups; however, none of the NDMA transitions occurred at CpG sites, unlike the control group. B[a]P, ENU, and the series treatment principally induced transversions, with G:C → C:G (3 out of 4), A:T → T:A (3 out of 5), and G:C → T:A (10 out of 24) being the most common transversions induced, respectively. Significantly, 46% (19 out of 41) of the mutations in the series treatment group occurred at CpG dinucleotides, compared to less than 22% in the other treatment groups. In conclusion, these data suggest that genotoxic carcinogens, when exposed in series versus singly, produce a unique MS profile characterized not only by shifts in mutation class but also sequence context.