High Frequency and Heterogeneous Distribution of p53 Mutations in Aflatoxin B₁-induced Mouse Lung Tumors

INTRODUCTION

AFB₁ 3 is a toxic fungal metabolite produced by several Aspergillus species. It is considered the most toxic of the aflatoxin group and has mutagenic and carcinogenic effects in mammalian cells and tissues. The principal target for AFB₁ is the liver; however, there is evidence that exposure to AFB₁ is also linked to the development of pulmonary cancer in humans and in experimental animals including mice (1, 2, 3, 4) . For AFB₁ to assert its mutagenic and carcinogenic actions, it must be bioactivated to its reactive metabolite, AFBO (5 , 6) . AFBO preferentially binds to the N⁷ position of guanine to form an AFBO-DNA adduct that can prevent normal replication of cellular DNA, resulting in perturbation of genes that are essential to the normal function of the cell (6) .

The p53 tumor suppressor gene is the most frequently mutated gene in human tumors (7) and tumor-derived cell lines (8, 9, 10) and is an archetypal checkpoint regulator. It exists as a single copy on chromosome 11 of the mouse genome and on the short arm of chromosome 17 of the human genome. The p53 protein binds to DNA in a sequence-specific fashion (11) , is involved in negative control of the cell cycle (12) , and regulates transcription (13) . Most notably, p53 is responsible for arresting cell growth in G₁, thus allowing the cell to repair itself (i.e., DNA repair mechanisms) or to initiate apoptosis (14) . By preventing continued cell proliferation in the face of damaged DNA, p53 limits the likelihood of mutations becoming fixed, thereby acting as a “guardian of the genome” (15) .

In geographical areas where dietary exposure to AFB₁ is common, p53 mutations are found in a high percentage of human hepatocellular carcinomas and often involve G:C→T:A transversions in the third base of codon 249 (16, 17, 18, 19, 20, 21) . A homologue of this mutation has been observed in AFB₁-induced liver cancer in mice possessing a mutated p53 transgene (22) . In contrast to the high frequency of p53 alterations in many cancers, its involvement in chemically induced mouse lung tumorigenesis has rarely been observed (23, 24, 25, 26, 27, 28, 29) .

In this study, we examined AFB₁-induced mouse lung tumors for p53 alterations using immunostaining, SSCP analysis, and direct sequencing. In addition, the novel LCM technique was used to determine the p53 genotype in regions of tumor sections with different immunostaining properties.

MATERIALS AND METHODS

Mouse Lung Tumor Induction and Diagnosis.

Female 5–7-week-old AC3F1 mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Animals were dosed (150 mg/kg i.p. divided into 24 doses over 8 weeks) and sacrificed 6–14 months after dosing, as described by Donnelly et al. (30) . A portion of each tumor was frozen in liquid nitrogen and stored at −70°C, and the remainder was fixed in 10% neutral buffered formalin. Lungs were inflated with 10% formalin and immersed in fixative. Diagnosis of formalin-fixed, paraffin-embedded lung tumor sections was carried out using established diagnostic criteria (31) after staining with H&E.

Immunohistochemical Staining of Accumulated p53.

For immunohistochemistry, tissues were fixed in 10% neutral buffered formalin, processed routinely, and embedded in paraffin. Localization of p53 protein expression was investigated using a polyclonal antibody, CM5 (Novocastra Laboratories; distributed by Vector Laboratories, Burlingame CA). Sections were first depa- raffinized in xylene and hydrated through a series of aqueous ethanol solutions to 1× Automation Buffer (Biomeda, Foster City, CA). Endogenous peroxidase was then blocked using 3% H₂O₂ for 15 min. After washing in 1× Automation Buffer, sections were microwaved, cooled, and blocked with normal goat serum (5%). All antibody incubations were carried out in a humidified chamber for 30 min (excluding the primary antibody) at room temperature. The p53 primary antibody was then applied for 1 h at room temperature. Nonimmune rabbit IgG (Jackson Immunoresearch Laboratories Inc., West Grove, PA) was used as the negative control at equivalent conditions in place of the primary antibody. The bound primary antibody was visualized by streptavidin-biotin-peroxidase detection using the Biogenex Super Sensitive Animal Detection Kit (mouse and rat absorbed; Biogenex Laboratories, San Ramon, CA) according to the manufacturer’s instructions and using 3,3′-diaminobenzidine as the color-developing agent. Slides were counterstained with Harris Hematoxylin (Harelco, Gibbstown, NJ), dehydrated through a series of aqueous ethanol solutions to xylene, and coverslipped with Permount (Fisher Scientific, Norcross, GA).

DNA Isolation of Whole Lung Tumors and Amplification of p53 by PCR.

DNA was obtained from frozen whole lung tumors by proteinase digestion followed by phenol/chloroform/isoamyl alcohol extraction and ethanol precipitation. The evolutionary conserved exons 5–8 of p53 were amplified by PCR (28) . The 20-μl reaction mixture contained 20 mm Tris-HCl (pH 8.4), 50 mm KCl, 200 μm each deoxynucleotide triphosphate (dATP, dCTP, dGTP, and dTTP), 0.4 μm each primer (Table 1) ⇓ , 1 unit of Taq polymerase (Life Technologies, Inc., Burlington, Ontario, Canada), and 100 ng of genomic DNA. After initial denaturation at 94°C for 1 min, samples were subjected to 35 cycles of amplification as follows: denaturation at 94°C for 30 s; and primer annealing and extension at 71°C for 1 min for exons 5, 7, and 8 and at 70°C for 1 min for exon 6. An additional extension step (70°C or 71°C for 5 min) was conducted at the end of the 35 cycles. Zero DNA template controls were included in each set of PCR reactions to test for intertube contamination.

SSCP Analysis of Whole Lung Tumor DNA.

For SSCP analysis of p53 exons 5–8 from whole lung tumors, PCR products from the first reaction (described above) were subjected to a second amplification. The 20-μl SSCP reaction mixture was similar to that of the initial PCR amplification, except that 2 μm deoxynucleotide triphosphates and 2.1 μCi of [α-³³P]dATP (2600 Ci/mmol; Amersham) were included. The PCR conditions were also similar; however, amplification was carried out for 20 cycles. Four μl of the α-³³P-labeled PCR products were diluted with 50 μl of 0.1% SDS and 10 mm EDTA. Equal volumes of the diluted PCR products and gel loading buffer (89% formamide, 0.1% bromphenol blue, 0.1% xylene cyanol FF, and 10 mm EDTA) were mixed and heated at 95°C for 5 min. Samples were then cooled on ice before loading on a 6% nondenaturing polyacrylamide gel containing 10% glycerol. Samples were electrophoresed at 55 W for 7 h at 4°C, dried for 1 h at 80°C, and autoradiographed using Kodak Biomax MR X-ray film.

Direct DNA Sequencing of Whole Lung Tumor DNA.

Before direct DNA sequencing, whole lung tumor outer PCR products were purified with a QIAquick PCR purification kit (Amersham Life Science, Cleveland, OH) and resuspended in 50 μl of distilled water. DNA from 2.0 μl of this DNA-containing solution was sequenced with 2.0 mm of the outer reverse PCR primer and a Thermo Sequenase radiolabeled terminator cycle sequencing kit (Amersham Life Science). Samples were then electrophoresed at 55 W for approximately 2 h on an 8% polyacrylamide denaturing gel containing 7 m urea. The gel was dried and autoradiographed as described above.

LCM.

Ten paraffin-embedded tissue sections from AFB₁-induced tumors that had been stained for accumulated p53 were subjected to LCM using a PXL-100 PixCell LCM System (Arcturus Engineering Inc., Mountain View, CA) as described by Bonner et al. (32) . Areas of interest were selected by activating the Capsure LCM Transfer Film, which covers the region of tissue to be acquired, and exposing tissue sections to pulses (varying amplitude and wavelength) from a short-pulse infrared laser. The carriers of the transfer film, which are transparent caps, were then fitted onto 1.5-ml microcentrifuge tubes containing 50 μl of cell digestion buffer [0.2 mg/ml proteinase K (Sigma, St. Louis, MO)], 10 mm Tris-HCl (pH 8; Sigma), 1 mm EDTA (Sigma), and 0.05% Tween 20 (Sigma). The microcentrifuge tubes were then inverted, and the tissues were exposed to the digestion buffer overnight at 37°C. Sample-containing tubes were then incubated at 95°C for 8 min to inactivate the proteinase K. Two μl of the resulting DNA-containing mixture were used as a template source for PCR amplification (reaction conditions were as described for whole tumor DNA).

SSCP Analysis and Direct Sequencing of Microdissected DNA.

For SSCP analysis of p53 exons 5–8 from microdissected tumor regions, DNA was amplified using the outer PCR primers and conditions. Two μl of a 1:10 dilution of this product were then subjected to a second round of amplification similar to that for SSCP analysis of whole lung tumors, except that inner forward primers and outer reverse primers were used (Table 1) ⇓ . Nondenaturing PAGE and autoradiography were carried out as described above for whole tumor DNA.

Conditions for direct sequence analysis of microdissected DNA were similar to those described for whole lung tumor DNA, except that PCR products from the inner amplification were used as template.

RESULTS

Tumor Induction.

Diagnosis of tumors from AFB₁-treated AC3F1 mice was as reported by Donnelly et al. (30) . Seventy-six pulmonary adenomas and carcinomas, one normal pulmonary tissue sample, and one normal liver tissue sample were removed (Table 2) ⇓ . The fact that the vehicle-treated mice developed very few lesions (<0.5 lesion/mouse) that were primarily hyperplasias and small adenomas suggests that the tumors analyzed in the AFB₁-treated group were caused by AFB₁ and were not spontaneous.

Immunohistochemistry.

Immunhistochemical staining for accumulated p53 protein of paraffin-embedded tumor tissue sections (Table 2) ⇓ indicated a heterogeneous distribution of positive nuclear p53 staining in 56 of 71 (79%) AFB₁-induced mouse lung tumor sections. In approximately 73% of the positive-stained tumor sections, <5% of cells demonstrated positive staining; however, in the other 27% of the positive-stained tumor sections, 10–60% of cells stained positively, with staining localized primarily at the periphery of the tumor.

SSCP and Direct Sequencing Analysis of AFB₁-induced Whole Lung Tumor DNA.

SSCP analysis of exons 5–8 from whole lung tumor DNA revealed altered banding patterns suggestive of the presence of mutations in 20 of 76 (26%) tumor DNA samples, with 85% of the mutations occurring in exon 7, and 15% of the mutations occurring in exon 6 (Fig. 1 ⇓ ; Table 2 ⇓ ). It was not possible to unequivocally identify mutations by direct sequencing of exon 6 and exon 7 SSCP-positive samples because putative mutant bands were weak. This may have been due to heterogeneous p53 sequences in the DNA of the whole lung tumor samples, as suggested by heterogeneous immunostaining. Thus, DNA isolated from whole lung tumor samples could have contained mutant DNA in low amounts relative to normal DNA. Furthermore, the correlation between the immunostaining results (i.e., the presence of any positive staining) and the whole lung tumor SSCP results was low, with a concordance rate of 17%. Consequently, LCM was used to analyze tumor samples in which DNA was isolated from both p53 positive- and negative-stained regions of tumor tissue sections.

SSCP Analysis and Direct Sequencing of DNA Isolated from Microdissected AFB₁-induced Tumor Tissue Sections.

%Different regions of 10 tumor tissue sections that contained a high percentage of positive-stained nuclei for accumulated p53 or those in which positive staining was highly localized to a specific region of a tumor section were microdissected to determine the mutation spectrum of p53 (Table 3) ⇓ . SSCP analysis of exons 5–8 revealed altered banding patterns in 18 of 30 (60%) microdissected tumor samples, with mutations occurring in exons 5, 6, and 7 (Fig. 2 ⇓ ; Table 3 ⇓ ). Thirteen of the 18 LCM samples with altered banding patterns were samples in which the DNA was isolated from p53-positive staining regions, whereas the other 5 DNA samples were isolated from p53-negative staining regions. Direct sequence analysis of microdissected samples exhibiting altered SSCP banding revealed mutations at several different codons in exons 5, 6, and 7 (Fig. 3 ⇓ ; Table 3 ⇓ ). Twenty transition mutations (9 G:C→A:T and 11 A:T→G:C), 5 transversion mutations (2 G:C→T:A, 2 T:A→A:T, and 1 A:T→C:G), and 1 deletion mutation were identified. Interestingly, LCM samples 18 and 28 both had two mutations in exon 7, but one of the mutations (codon 222 in sample 18 and codon 229 in sample 28) was silent. LCM sample 22, which contained DNA from a negative staining region of a tumor section, had a silent mutation in codon 247 of exon 7. In addition, several microdissected samples contained mutations in two or more exons, and LCM sample 23 also contained mutations in the flanking introns. A base substitution and a base deletion in LCM sample 10 and a base substitution in LCM sample 23 were identified in the first codon of exon 6. LCM samples 23, 27, and 28 showed the presence of a strong signal for mutant alleles on the sequencing autoradiogram without visible normal bands at the same positions, whereas the remaining 14 microdissected samples showed the presence of two bands, one representing the normal allele, and the other representing the mutant allele. Reproducibility of the mutation analysis was confirmed by reamplifying and sequencing LCM samples 18 and 20 from original template. SSCP analysis of LCM sample 24 and sequencing of some exons of LCM samples 24, 28, and 29 were not performed because there was no remaining DNA template.

DISCUSSION

The high frequency (79%) of tumors exhibiting accumulated p53, as determined by immunohistochemistry, is the highest reported to date for chemically induced mouse lung tumors, suggesting a carcinogen-specific role for p53 in AFB₁-induced mouse lung tumorigenesis. Similar studies of chemically induced mouse lung tumors have not included staining for the protein, except for that of Hegi et al. (28) , which reported that 7% of methylene chloride-induced mouse lung tumors had positive staining (24, 25, 26, 27, 28, 29) . Although the polyclonal antibody used in the present study will bind to both the wild-type and mutant forms of p53, p53 detection by immunohistochemistry in paraffin sections is consistent with the increased half-life of mutated protein relative to wild-type p53 (33, 34, 35) . Mutation of one of the p53 alleles is adequate for positive staining; hence, the other allele can be mutated, deleted, or normal.

Homogeneous distribution of a gene mutation throughout a tumor implies that the mutation occurred before clonal expansion, whereas heterogeneous distribution implies that the mutation occurred relatively late during tumor development (36) . From the results presented, the heterogeneous staining pattern of the tissue sections suggests that p53 mutation is a relatively late event in AFB₁-induced mouse lung tumors. An early mutation of the p53 gene due to direct AFBO-DNA adduct formation at specific guanine residues in the gene can be an initiating factor in carcinogenesis; however, a late mutation suggests that loss of function of the p53 protein is not directly due to AFBO action but rather to subsequent genomic instability.

Despite earlier evidence to the contrary, positive p53 immunostaining is not 100% predictable for the presence of point mutations because there are other cellular mechanisms, such as inhibition of p53 degradation, that may also increase the half-life of p53 (37) . Therefore, molecular analysis of the tumor DNA was also used to detect alterations in the gene.

The p53 gene is composed of 11 exons, the first of which is noncoding and is localized 8–10 kb away from exons 2–11. In cross-species comparison, the p53 protein shows five highly (>90%) conserved regions among amino acid residues 13–19, 117–142, 171–181, 234–258, and 270–286 (38) . Four of these regions are contained within exons 5–8, the sequence-specific DNA binding domain, where approximately 90% of the missense mutations in p53 occur (39) . SSCP analysis of exons 5–8 of the p53 gene were examined because 90% of p53 mutations occur in this hot spot region (40 , 41) . Consistent with the immunostaining results, the relatively high frequency of molecular perturbations in whole lung tumor DNA (i.e., 26%) has not been observed in previous chemically induced mouse lung tumors studies (24, 25, 26, 27, 28, 29) , again suggesting a carcinogen-specific response.

A limitation of analysis of whole lung tumor DNA samples is that false negatives may be observed due to heterogeneity of the DNA sample (i.e., the amount of normal DNA may greatly exceed the amount of mutated DNA, thus limiting the ability of SSCP and direct sequencing to detect mutated DNA). Identification of the mutations in the whole lung tumors by direct sequencing was not informative, probably for this reason. The mutation spectrum of p53 in the microdissected samples was broad, with some samples containing point mutations at two different codons, and some of these being silent. Samples 23, 27, and 28, the sequences of which showed mutant alleles but not normal alleles, were apparently either homozygous or hemi zygous for the mutation with loss of the other allele. The latter scenario is quite possible because mutation of one of the p53 alleles and loss of the other allele is common (42) . The data also suggest that the increase in apparently random DNA damage may be a result of genomic instability.

Several DNA samples (i.e., LCM samples 20–23) originated from the same tumor section; however, different mutations were identified in the DNA from different regions, whether positively or negatively stained. The concordance rate of 72% that we found between immunostaining and molecular analysis is similar to previous studies involving immunostaining and SSCP (43 , 44) for p53 in humans and indicates that analysis of microdissected samples can reveal mutations in p53 that are not detectable through the analysis of whole lung tumor DNA.

Analogous to the whole lung tumor molecular analysis results, not all of the microdissected DNA samples isolated from positive-staining regions had detectable mutations in exons 5–8. One possible explanation is that lesions outside the exons analyzed may have been present, because a small percentage of p53 mutations can occur outside exons 5–8 (45) . Another explanation is that the distressed tumor cell is responding through activation of biochemical pathways controlled by the p53 gene. The precise mechanism by which DNA damage and other stimuli result in p53 stabilization has not yet been elucidated but may involve the action of other tumor suppressor or oncogene products. Recently, Pomerantz et al. (46) and Zhang et al. (47) illustrated that the inhibitor of the cyclin-dependent kinase 4/6-alternate reading frame (INK4a-ARF) gene locus encodes the ARF protein, which allows the stabilization of p53 by neutralizing the murine double minute gene 2 (MDM2), an oncogene that is involved in degradation of p53. Therefore, the increase in p53 immunostaining is not always necessarily due to direct perturbation in the p53 gene itself but may be secondary to changes in the expression of other genes.

Samples containing mutations in the p53 gene but not staining positively for accumulated p53 can be attributed to silent mutations, splicing mutations, nonsense mutations, deletions, and/or missense mutations outside exons 5–8 that do not yield an accumulation of p53 proteins (38) . For example, LCM sample 22 contained DNA from a negative-staining region of a tumor section; however, SSCP identified a mutation in exon 7 that was later confirmed as a silent mutation in codon 247. LCM samples 7, 12, 19, and 25 were shown to have altered banding patterns, but no mutations were identified by direct sequencing. These samples may have contained mutations in the flanking introns; however, because of the limited availability of DNA template after LCM, it was not possible to determine whether this was the case. Another possibility is that the mutations came from contaminating cells from areas other than those targeted for microdissection.

The difference in the occurrence of AFB₁-induced p53 mutations in our mouse lung tumor model compared to the human hepatocellular carcinoma data can be attributed to differences in carcinogenic mechanisms. Human hepatocellular carcinoma is associated not only with AFB₁ exposure but also with infection with the hepatitis B virus, and these factors may act synergistically (48, 49, 50) . Also, the likelihood of codon 246 being a hot spot location for AFB₁-induced mutations in mice is small because the p53 sequence is different from that of humans. As discussed by Ghebranious et al. (51) , the critical mutation for Arg→Ser requires a 2-base (CGA→TCA) change in mice, compared with a 1-base change (AGG→AGT) in the homologous codon 249 in humans (changes are represented by the italicized residues). Moreover, other animal models of AFB₁ tumorigenesis have shown neither codon 249 mutations nor the detection of significant levels of other p53 mutations in hepatocellular carcinomas (52, 53, 54) .

A difference between the results obtained in mouse lung tumor models and available human lung tumor data is that molecular detection of p53 mutations occurs at a higher frequency in human lung whole tumors. This may be due to the fact that the human tumors were more advanced and more aggressive when analyzed. Alternatively, p53 mutations may occur relatively early in human lung tumori-genesis, resulting in a more homogeneous distribution throughout tumors (36) .

In a previous study by Donnelly et al. (30) using the same AFB₁-induced whole lung tumor DNA, specific mutations in the first exon of K-ras were found in all of the tumors, suggesting that K-ras activation is an early, critical event in AFB₁-induced pulmonary carcinogenesis in AC3F1 mice. The early K-ras mutations and other genetic events in the AFB₁-induced tumors may contribute to genomic instability, resulting in mutations in genes such as p53. p53 mutations were identified in both adenomas and carcinomas, indicating that these genetic alterations are present before tumors progress from adenoma to carcinoma.

Almost all of the p53 mutations were transition mutations. A high prevalence of transversion mutations has been interpreted as suggestive of the action of exogenous carcinogens, whereas a high frequency of transitions suggests endogenous mutagenic processes (55, 56, 57) . A majority of the p53 mutations were not identified at CpG dinucleo-tides, suggesting that endogenous mechanisms such as DNA polymerase infidelity rather than deamination of 5-methylcytosine may be responsible for the mutations. Furthermore, mutations did not occur predominantly at G:C bp. This clearly contrasts with our observation that the mutations found in K-ras were exclusively at G:C bp, consistent with AFBO attacking guanine residues. It also contrasts with the recent finding of Denissenko et al. (58) that AFB₁ adducts occur at numerous different guanine residues in human p53.

Overall, the high frequency and variability of specific point mutations in the AFB₁-induced mouse lung tumors compared to those found with other carcinogens, coupled with the relatively high frequency but heterogeneous distribution of positive immunostaining, suggest that p53 mutations are carcinogen specific but occur after direct carcinogen actions in AFB₁-induced mouse lung tumorigenesis. They also demonstrate that the analysis of microdissected DNA by LCM is useful for characterizing heterogeneously distributed p53 mutations, which cannot be analyzed accurately in whole lung tumor DNA.