Alterations in the mutagenicity and mutation spectrum induced by benzo[a]pyrene instilled in the lungs of gpt delta mice of various ages

To examine whether the mutagenic potential of lung exposure to air-borne environmental mutagens is age dependent, we administered 1 mg of benzo[a]pyrene intratracheally to 11- and 24-month old (middle-aged and old, respectively) gpt delta transgenic mice that harbor gpt (guanine phosphoribosyltransferase) genes integrated in the genomic DNA as a target for mutation detection, and then analyzed the benzo[a]pyrene-induced and spontaneous in vivo mutations and mutation spectrum in the lungs. The mutant frequencies in the lungs of the 11- and 24-month–old control (vehicle-treated) gpt delta mice were 1.14 ± 0.22 × 10−5 and 1.00 ± 0.20 × 10−5, respectively, which are significantly higher than that observed for the control 3-month–old (young) mice (0.59 ± 0.13 × 10−5) in our previous studies, indicating that spontaneous mutation in the lung increases with age. The mutant frequencies in 11- and 24-month–old mice treated with benzo [a] pyrene were 1.5- and 2.3-fold, respectively, that of the age-matched control mice, and 4.3-fold that of the 3-month–old mice in our previous studies. Analysis of mutation spectra showed that both G:C to A:T transitions and G:C to T:A transversions were predominant in the lungs of control mice at all ages. In benzo [a] pyrene-treated mice in our previous studies, G:C to T:A transversions were the predominant type of mutation (55 %) at 3 months. Here we found that their frequency was dramatically reduced to 18 % by 24 months, and the G:C to A:T transitions became the predominant type of mutation in 24-month–old mice (41 % [16 % at CpG sites]). Our findings suggest that susceptibility to benzo[a]pyrene is highest in young mice and is elevated again in old age. The elevation of G:C to A:T transitions was observed following benzo [a] pyrene administration in the lungs of aged mice, and accelerated cytidine deamination is speculated to contribute to this elevation.


Introduction
Accumulation of mutations in the genome is considered to be at least in part responsible for the phenomenon of aging. The increase in genomic mutation frequency with age is believed to be a factor in the age-related functional decline of homeostasis and resistance to stressors, which leads to diseases such as cancer [1]. However, it remains to be clarified how vulnerable groups, such as young and old individuals, are susceptible to environmental mutagens and related environmental stressors. For instance, young (3-month-old) rats are more susceptible than adult (11-month-old) rats to acrylamide-induced testicular genotoxicity [2].
Transgenic rodents that harbor exogenous genes integrated in the genomic DNA, as a target for mutation detection, are a useful system for the study of in vivo somatic mutations caused by environmental mutagens. The widely used transgenic lines, Muta mouse, Big Blue rodents, and gpt delta rodents, harbor the Escherichia coli genes, lacZ (β-galactosidase), lacI (lac repressor), and gpt (guanine phosphoribosyltransferase), respectively [3]. Use of these model animals to evaluate the level of mutations accumulated in aged animals has revealed that the mutation frequency increases spontaneously with age in various organs including liver, spleen, small intestine, kidney, and heart [4][5][6][7][8][9][10], but the magnitude of the increase is different among the organs.
From the viewpoint of interactions between the genome and the environment, the lung is a unique organ. In the lung, air-borne environmental mutagens directly contact the pulmonary epithelial cells and induce mutations in the genomic DNA, whereas mutagens reach most organs via the blood circulation system. Therefore, here we chose to use the lung to address how the susceptibility to environmental mutagens differs among age groups. We selected to use benzo[a]pyrene (B[a]P), because it is a common air-borne mutagen generated by the burning of fossil fuels. Previously, we reported that B[a]P (0.5-2 mg per animal) induces a linear dosedependent increase in mutation frequency in the lungs of 3-month-old gpt delta mice following a single intratracheal instillation [11]. Here, to determine whether the magnitude of the elevation of in vivo mutant frequency following treatment with an environmental mutagen is age dependent, we examined the mutant frequency and types of mutations in the gpt gene in B[a]P-instilled lungs of 11-and 24-month-old gpt delta mice (representative of middle-aged, and old animals, respectively), comparing with those in 3-month-old gpt delta mice (representative of young animals).

Treatment of mice
Male gpt delta mice (9-weeks old; body weight,~25 g), which carry approximately 80 copies of lambda EG10 DNA on each chromosome 17 on a C57BL/6 J background [12], were obtained from Japan SLC (Shizuoka, Japan). The mice were maintained at 24 to 26°C with 55 % to 75 % humidity and a 12-h light-dark cycle, and were fed CA-1 diet (Japan Clea Co., Tokyo, Japan) with water ad libitum, in the specific-pathogen-free (SPF) animal facility of the National Institute for Environmental Studies. The animals were anesthetized with 4 % halothane (Hoechst Japan, Tokyo, Japan) in a desiccator until they did not respond to a tactile stimulus. A single dose of B[a]P (1 mg, Wako Pure Chemical Industries, Osaka, Japan) dissolved in 50 μL of tricaprylin (Sigma-Aldrich, St. Louis, MO, USA) was intratracheally instilled via a polyethylene tube [11,13]. Control mice were treated with 50 μL of vehicle (tricaprylin) alone; this dose of B[a]P is within the range (0.5-2 mg) that causes a linear dose-dependent increase in mutant frequency in the lungs of 3-month-old gpt delta mice [11]. The mice were sacrificed 14 days after the administration, and their lungs were removed, frozen in liquid nitrogen, and stored at −80°C until the DNA was isolated. The animal studies were approved by the Animal Care and Use Committee of National Institute for Environmental Studies.

DNA isolation and in vitro packaging of DNA
High-molecular-weight genomic DNA was extracted from the lungs by using the RecoverEase DNA Isolation Kit (Agilent Technologies, Santa Clara, CA, USA). Lambda EG10 phages containing the gpt gene were recovered from the genomic DNA by using Transpack Packaging Extract (Agilent Technologies).
Mutation assay and DNA sequencing analysis of the gpt gene in 6-TG-resistant colonies The gpt mutagenesis assay was performed according to previously described methods [14]. To convert the phage DNA into plasmids, E. coli strain YG6020 expressing Cre recombinase was infected with the rescued phage. The bacteria were then spread onto M9 salts plates containing chloramphenicol (Cm) and 6-thioguanine (6-TG) [14], and incubated for 72 h at 37°C for selection of the colonies harboring a plasmid carrying the chloramphenicol acetyltransferase (cat) gene and a mutated gpt gene. The 6-TG-resistant colonies were streaked onto selection plates for confirmation of the resistant phenotype. The cells were then cultured in LB broth containing 25 mg/mL Cm at 37°C and collected by centrifugation. The bacterial pellets were stored at −80°C until DNA sequencing analysis was performed. Mutant frequencies for the gpt gene were calculated by dividing the number of colonies growing on M9 + Cm + 6-TG agar plates by the number of colonies growing on M9 + Cm agar plates. PCR and DNA sequencing analysis of 6-TG-resistant mutants were performed as previously reported [11].

Statistical analysis
All data are expressed as means ± SD. Statistical significance was evaluated by using Student's t-tests. A statistical analysis of mutational spectra was performed by using the Adams-Skopek test [15,16] and Chi-square test. P values less than 0.05 were considered to be statistically significant.

Results and discussion
Mutant frequencies in the lungs of B[a]P-treated gpt mice To examine which age group is most susceptible to a common air-borne environmental mutagen, the mutant frequency and the mutation spectrum in the lungs of 11-and 24-month-old gpt delta mice (representative of middle-aged and old animals, respectively) following treatment with B[a]P and age-matched controls (vehicletreated) were analyzed and compared with our published data for 3-month-old mice (representative of young animals) [11,17,18].
The mutant frequencies in the lungs of 11-and 24-month-old control (vehicle-treated) mice were 1.14 ± 0.22 × 10 −5 and 1.00 ± 0.20 × 10 −5 , respectively. In contrast, the mutant frequency in the lungs of 3-month-old control mice was 0.59 ± 0.13 × 10 −5 according to the combined data from our previous reports [11,17,18]. Consistent with previously reported observations in various organs such as liver and spleen [10,19,20], these observations indicate that the frequency of spontaneous mutants in the lung increased with age.
A single administration of 1 mg of B[a]P elevated the mutant frequency in 11-and 24-month-old mice to 1.68 ± 0.19 × 10 −5 and 2.25 ± 0.54 × 10 −5 , respectively, which was 1.5-and 2.3-fold the mutant frequency in the respective age-matched control mice (Table 1). We previously reported that B[a]P instillation to the lungs of 3-month-old mice elevated the mutant frequency to 2.52 ± 0.33 × 10 −5 [11] which was 4.3-fold the frequency observed in the 3-month-old control mice in our previous studies [11,17,18].
Our observations indicate that the order of the age groups in terms of highest to lowest fold of increase in mutant frequency in the lungs following instillation of B[a]P was 3-, 24-, and 11-month-old mice. These results suggest that young mice are the age-group most susceptible to B[a]P, which may be explained age-related changes of the mutant frequency by their relatively high level of DNA replicating activity and cell turnover rate [21] and/or perhaps higher levels of metabolic activation of B[a]P, as well as DNA repair activity, which promote the formation of B[a]P-DNA adducts, and that the susceptibility is elevated again in old age by several possible mechanisms, as discussed later. Age-dependent alteration in the metabolic activation of B[a]P has not been well-documented in the lung, but the activity of cytochrome 1A, a mono-oxygenase that catalyzes the metabolic activation of B[a]P, has been suggested to decline with age in the rat liver [22].

Characteristics of the gpt mutation spectrum
To analyze the age-dependent alterations in the mutation spectra of B[a]P-instilled and control lungs, we sequenced gpt mutants recovered from the lungs as shown in Table 2. We observed a significant difference between the mutation spectrum of the 24-month-old control mice studied here and that of the 3-month-old control mice we studied previously [11,17,18] (P < 0.05, Adams-Skopec test). In our published data from 3-month-old control mice, the most predominant type of base substitution was G:C to A:T transitions (45 % of total mutants), and a half of these transitions were induced at CpG sites (18 % of total mutants), while G:C to T:A (21 %) and G:C to C:G (16 %) transversions were also major base substitutions. Here, were found that the percentages of G:C to C:G transversions among total mutants were less in 11-and 24month-old control mice than in 3-month-old control mice, but G:C to A:T transitions and G:C to T:A transversions were still predominant (35 % and 14 % of total mutants, respectively) in 24-month-old control mice ( Table 2). In contrast, increased proportions of A:T to T:A transversions, A:T to C:G transversions, and base deletions were observed in the lungs of 24-month-old control mice compared with 3-month-old control mice. An increase in base substitutions at A:T has previously been reported in the intestine and spleen of 32-month-old lacZ plasmid transgenic mice, and DNA polymerase η was speculated to act as an A:T mutator in the spleen of aged animals [23]. Taken together, these results suggest that an increase in point mutations at A:T may be a common phenomenon in proliferative aged tissues.
We observed that there was a significant difference between the mutation spectra for both 3-and 11-month-old B[a]P-instilled mice (P < 0.05, Adams-Skopec test), and 3-and 24-month-old B[a]P-instilled mice (P < 0.01, Adams-Skopec test). In B[a]P-instilled groups, we reported previously that G:C to T:A transversions, a major base substitution induced by B[a]P administration, were the predominant type of mutation (55 %) in the lungs of 3-month-old mice in our previous study [11], but surprisingly in the aged mice, the percentage of these mutations was dramatically lowered to be 24 % and 18 % in 11-and 24-month-old mice, respectively ( Table 2). In contrast, the percentage of G:C to A:T transitions was observed to increase in B[a]P-instilled lungs age-dependently, and they became the predominant type of mutation in 24-monthold mice (41 % [16 % at CpG sites]). As observed for control lungs, the percentage of G:C to C:G transversions in B[a]P-instilled lungs decreased with age.
The positions of spontaneous and B[a]P-induced gpt mutations are listed in Table 3. Among the mutated sequences isolated from the control mice, 6 gpt mutations (G:C to A:T transitions) at nucleotides 64, 86, 115, and 406 in 11-month-old mice, at nucleotide 110 in 4 and 24-month-old mice, and at nucleotide 417 in 24month-old mice, were each observed in three or more mice. These positions therefore are potential hotspots, while G:C to A:T transitions at nucleotides 64 and 110 were reported to be sites of spontaneous mutation in gpt delta mice [24]. The G:C to A:T transitions at nucleotides 64 and 110 and the G:C to C:G transversion at nucleotide 340 were also hotspots in 11-month-old B[a]P-instilled mice. Regarding G:C to T:A transversions, we previously observed that nucleotides 115, 140, 143, 189, and 413 were hotspots for G:C to T:A transversions in B[a]P-instilled lungs of 3-month-old gpt delta mice [11], but these nucleotides were not hotspots for this base substitution in 11-and 24-month-old mice; rather the hotspots were nucleotides 402 and 406 in B[a]Pinstilled lungs of 11-month-old mice, but there was no hotspot in 24-month-old mice.
Our results showed that the predominant type of mutation in the lungs of the vehicle control gpt delta mice was G:C to A:T transitions in all age groups; these transitions have also been shown to be the predominant type of mutation in liver and other organs in both newborn and 23-month-old lacZ-transgenic mouse [19]. G:C to A:T transitions are recognized to be more frequently induced on CpG sites, in which cytosines tend to be     methylated, by spontaneous deamination of methylated cytosines to form thymine residues, resulting in the formation of G:T mispairs [25]. Most cytosines in CpG sites in the liver of gpt gene integrated in the genomic DNA are highly methylated [26], and are therefore hotspots of spontaneous mutation in the control mice, such as G:C to A:T transitions at nucleotides 64, 110, and 115 are at CpG sites (Table 3). Methyl-CpG Domain Protein 4 (MBD4) and thymine-DNA glycosylase (TDG) [27] are mismatch repair enzymes that correct G:T mispairs by excising the mispaired thymine [28]. We consider that decrease in these DNA glycosylases and other mismatch repair enzymes possibly contribute to the observed increase in occurrence of G:C to A:T transitions in aged animals [19,21,29]. In B[a]P-instilled lung of gpt delta mice, G:C to A:T transitions were shown to increase in an age-dependent manner ( Table 2); this type of mutation was most predominant in 24-month-old mice. In contrast, G:C to T:A transversions (a landmark mutation of B[a]P-DNA adduct formation possibly induced by translesional DNA synthesis [30]) were the major base substitution in B[a]P-instilled lungs of 3-month-old mice. As summarized in Table 4, estimation of specific mutant frequency ([Average mutant frequency in Table 1] × [% mutant of G:C to A:T transition or G:C to T:A transversion of  [11,17,18] e Our previous data [11] CpG: mutation at CpG site Our mutation spectrum analysis of B[a]P-instilled lungs revealed that not only the overall percentage of G:C to A:T transitions but the percentage of G:C to A:T transitions at non-CpG sites increased with age (8 %, 12 %, and 25 % ['the percentage of total G:C to A:T transitions' minus 'the percentage of total G:C to A:T transitions at CpG sites'] at 3 months, 11 months, and 24 months, respectively). A possible mechanism for the induction of G:C to A:T transitions in B[a]P-instilled old mice is that spontaneous deamination of cytosine at CpG sites was elevated in the lungs of these mice by instillation of B[a]P, resulting in an increase in the percentage of G:C to A:T transitions at CpG sites in 24-month-old mice (16 %). A decrease in mismatch repair [29] might also contribute to the increase in occurrence of G:C to A:T transitions in B[a]P-instilled aged animals. Another possibility is that DNA cytidine deaminase is activated in the lungs of aged mice by instillation of B[a]P. This enzyme catalyzes the conversion of cytosine to uracil at both CpG and non-CpG sites, which leads to G:U mispair formation and hence mutation of G:C to A:T. The hypermutation induced by DNA cytidine deaminase plays a role in creating antibody diversity in the variable regions [31], and expression of this enzyme causes genomic instability related to cancer and other diseases [32]. We speculate that B[a]P induces DNA cytidine deaminase in the lungs of aged animals resulting in the induction of G:C to A:T transversions at non-CpG sites.
The biological significance of the age-related increase in G:C to A:T transitions in both B[a]P-instilled and control mice remains unclear. G:C to A:T transitions are induced on codons 12 and 13 of the K-ras gene in lung tumors spontaneously induced in either p53-suppressed old mice or age-matched wild-type mice (13-to 24-month-old mice) [33]. Recently, G:C to A:T transitions were shown to occur frequently in six types of human tumors (lung adenocarcinoma, lung squamous cell carcinoma], bladder, cervix, head and neck), in which APOBEC3B, an isoform of DNA-cytidine deaminase, was upregulated [34], and lung adenocarcinoma developed in transgenic mice with constitutive expression of DNA-cytidine deaminase [35]. These observations suggest that an increase in G:C to A:T transitions may contribute to not only spontaneous but also B[a]P-induced tumorigenesis in the lungs of aged mice. However, further studies are required to reveal the mechanism that G:C to A:T transitions induced in the lung cause cancer and other diseases in the old age.

Conclusions
Our observations indicate that the order of the age groups in terms of highest to lowest fold of increase in mutant frequency in the lungs following instillation of B[a]P was 3-, 24-, and 11-month-old mice, suggesting that young mice are the age-group most susceptible to B[a]P. G:C to T:A transversion was shown to be a predominant mutation for elevation of mutant frequency in B[a]P-instilled lung of 3month-old mice, but in 24-month-old mice, induction of G:C to A:T transition as well as G:C to T:A transversion drove the elevation of mutant frequency by B[a]P instillation. We speculate that B[a]P induces DNA cytidine deaminase in the lungs of aged animals resulting in the induction of G:C to A:T transversions at non-CpG sites.