Organ-specific profiles of genetic changes in cancers caused by activation-induced cytidine deaminase expression

Authors


Abstract

Various molecular changes characterizing organ-specific carcinogenesis have been identified in human tumors; however, the molecular mechanisms of the genomic changes specific for each cancer are not well defined. A transgenic (Tg) mouse model with constitutive expression of the nucleotide-editing enzyme, activation-induced cytidine deaminase (AID), develops tumors in various organs as a result of the mutagenic activities of AID. This phenotypic character of AID Tg mice allowed us to analyze the organ-specific genetic changes in tumor-related genes commonly triggered by AID-mediated mutagenesis. Among the 80 AID Tg mice analyzed, 11 mice developed hepatocellular carcinomas, and 7 developed lung cancers. In addition, 1 developed the gastric cancer and 3 developed gastric adenomas. Organ-specific preferences for nucleotide changes were observed in some of the tumor-related genes in each epithelial tissue of the AID Tg mice. Of note, the c-myc and K-ras genes were the preferential targets of the mutagenic activity of AID in lung and stomach cancers, respectively, whereas mutations in the p53 and β-catenin genes were commonly observed in all 3 organs. Quantitative RT-PCR analyses revealed that alpha-fetoprotein, insulin-like growth factor2 and cyclin D1 genes were specifically upregulated in HCC, whereas upregulation of the matrix metalloproteinase7 gene was more marked in lung cancer. Our findings suggest that AID, a DNA mutator that plays a critical role linking inflammation to human cancers, might be involved in the generation of organ-specific genetic diversity in oncogenic pathways during cancer development. © 2008 Wiley-Liss, Inc.

Cancer develops in various organs as a consequence of a series of genetic changes, including nucleotide mutations and chromosomal rearrangement.1–3 A growing number of mutations in oncogenes and tumor suppressor genes have been identified in human cancer tissues.4, 5 However, it is widely recognized that the prevalence of mutations and the patterns of nucleotide changes in tumor-related genes differ among individual cancers. For example, the c-myc gene is a frequent target for genetic changes in human lung cancers and lymphomas, whereas changes in this oncogene are rarely detectable in hepatocellular carcinoma (HCC).5–7 Similarly, almost all pancreatic cancers contain the K-ras gene mutations,8 whereas the mutation frequencies in the K-ras gene are usually low in other types of tumors. These findings suggest that organ-specific genetic changes might be induced or maintained in the pathway to cancer development in each tissue. However, the molecular mechanisms underlying the accumulation of tissue-specific genomic changes in oncogenic pathways are not well defined.

Activation-induced cytidine deaminase (AID), an enzyme with homology to members of the apolipoprotein B mRNA-editing enzyme catalytic polypeptide (APOBEC) family,9, 10 is required for germinal center B-cells to undergo somatic hypermutation and class switch recombination in immunoglobulin genes.11–13 However, recent studies have revealed that the inappropriate expression of AID could act as a DNA mutator that contributes to tumorigenesis via its mutagenic activity.14–16 We recently demonstrated aberrant AID expression in inflammatory tissues, as well as in tumor cells, during the development of both human gastric cancer and HCC.16, 17 Moreover, AID expression is induced in response to proinflammatory cytokine stimulation, leading to the accumulation of genetic changes in several tumor-related genes in both human gastric epithelial cells and hepatocytes.17, 18 These findings suggest that AID is involved in the development of human gastric cancers and HCC through the accumulation of genetic changes.16–18 Consistent with this hypothesis, the constitutive and ubiquitous expression of AID caused the development of neoplasia in vivo.19 It is noteworthy that AID transgenic (Tg) mice develop tumors in several organs, including the liver, lung and lymphoid tissues, through the accumulation of genetic changes induced by the genotoxic effects of AID.18–20 In fact, somatic mutations in the c-myc and T-cell receptor genes were frequently detectable in the lymphoma tissues of AID Tg mice.19, 20 However, it remains unclear whether the carcinogenesis pathway triggered by the accumulation of genetic changes is common to the different types of epithelial tissues with AID expression. Therefore, in this study, we analyzed the mutational spectra and expression profiles of various tumor-related genes in 3 types of epithelial tissues, including the stomach, liver and lung, as well as gastric, liver and lung cancers that developed in AID Tg mice in an identical genetic background.

Abbreviations

18s rRNA, 18s ribosomal RNA; Afp, alpha-fetoprotein; AID, activation-induced cytidine deaminase; APOBEC, apolipoprotein B mRNA-editing enzyme catalytic polypeptide; HCC, hepatocellular carcinoma; RT-PCR, reverse transcription polymerase chain reaction; Tg, transgenic.

Material and methods

Generation of AID Tg mice

The generation of Tg mice with constitutive and ubiquitous expression of AID has been described previously.19 Tissue samples from Tg mice were removed and fixed in 4% (w/v) formaldehyde, embedded in paraffin, stained with hematoxylin and eosin (H&E) and examined for histological abnormalities. Tissue samples were also frozen immediately in liquid nitrogen for nucleotide extraction. The mice received humane care according to the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health, USA (NIH publication 86-23).

Subcloning and sequence analysis of the p53, c-myc, K-ras and β-catenin genes

Genomic DNA was extracted with DNeasy purification (Qiagen, Hilden, Germany). The p53, c-myc, K-ras and β-catenin genes were amplified by genomic PCR with Phusion High-Fidelity DNA Polymerase (FINNZYMES, Espoo, Finland). The primer sets for the amplification of those genomic sequences are shown in Supplemental Table I. After 35 cycles of PCR, the PCR products of the p53, K-ras and β-catenin genes were subcloned into the EcoRI-XhoI sites of the pcDNA3 vector (Invitrogen, Carlsbad, CA). The PCR product of the c-myc gene was also subcloned into the HindIII-XhoI sites of the pcDNA3 vector, followed by sequence analyses.

Quantitative real-time reverse transcription (RT)-PCR

Total RNA was extracted from tissues using the RNeasy Mini Kit (Qiagen). Gene expression was quantified by real-time RT-PCR using the 7300 Real-Time PCR System (PE Applied Biosystems, Foster City, CA). The sequences of the oligonucleotide primers used in this experiment are shown in Supplemental Table II. cDNA was synthesized using the oligo-dT primer and SuperScript III (Invitrogen). Real-time PCRs were set up in 50 μl of Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen) with the RT product and the forward and reverse primers, as described previously.21 Standard curves were generated for every target using a 10-fold serial dilution series of 7 independent transcripts derived from the normal liver tissues of wild-type mice. To assess the quantity of isolated RNA and the efficiency of cDNA synthesis, target cDNAs were normalized to the endogenous mRNA levels of the housekeeping reference gene for 18S ribosomal RNA (18S rRNA).16 For simplicity, the ratios are presented as values relative to the expression levels in lysate from control specimens.

Immunohistochemistry

Immunostaining for alpha-fetoprotein (Afp) was carried out according to a previously described protocol.22 The polyclonal antibody directed against Afp was purchased from Dako Cytomation (Glostrup, Denmark).

Statistical analysis

Results were analyzed using Student's t test. A value of p < 0.05 was deemed significant. Statistical analysis was performed with StatView 5.0 (Abacus Concepts, Berkeley, CA).

Results

Development of gastric cancer in AID Tg mice

In our previous analyses, we observed that AID Tg mice developed 3 types of tumors, including lymphoma, HCC and lung cancer.18, 19 In this study, we focused on the development of nonlymphoid tumors and analyzed the overall phenotypes of the epithelial organs in AID Tg mice (B2 line).19 Among the 80 AID Tg mice analyzed, most developed T-cell lymphomas and 8 had sarcomas of the skin. Furthermore, 11 AID Tg mice developed HCC, 7 developed lung cancers and 4 developed gastric tumors (Fig. 1a). Histological examination revealed that the 1 gastric tumor had the morphological appearance of typical human gastric cancers (Fig. 1b), and the remaining 3 tumors showed the pathological characteristics of gastric adenomas. These findings suggest that constitutive expression of AID could cause the development of neoplasia in 3 types of epithelial organs: the liver, stomach and lung.

Figure 1.

Epithelial neoplasia developed in AID Tg mice. (a) Numbers of epithelial tumors developed in AID Tg mice are shown in this figure. (b) Representative histological findings for the gastric, liver, and lung tumors developed in AID Tg mice. Top left, macroscopic view of a representative gastric carcinoma in a 53-week-old AID Tg mouse. Top right, macroscopic view of nontumor gastric tissues in the AID Tg mouse with gastric carcinoma. Middle left, hematoxylin and eosin (H&E)-stained section of gastric carcinoma (magnification ×200). Middle right, H&E-stained section of gastric carcinoma (magnification ×400), showing a typical histological pattern of gastric adenocarcinoma. Bottom left, histological pattern (H&E stained; magnification ×400) of a hepatocellular carcinoma in a 53-week-old AID Tg mouse. Bottom right, histological pattern (H&E stained; magnification ×400) of a lung adenocarcinoma in a AID Tg mouse.

Organ-specific mutational profiles in liver, stomach and lung cancers that developed in AID Tg mice

The finding that the constitutive expression of AID caused cancer development in different types of epithelial organs prompted us to ask whether AID could induce tissue-specific genotoxic effects in oncogenic pathways during the process of each carcinogenesis. To determine the tissue-specific mutational profiles of tumor-related genes, we took the advantage of 1 AID Tg mouse that developed both HCC and gastric cancer simultaneously and its littermate with lung cancer, and investigate the occurrence of genetic mutations in several tumor-related genes in both nontumor and tumor tissues of the liver, stomach and lung. For this purpose, we analyzed the genetic changes in the c-myc gene, which is the common target for abnormal editing in the lymphoma cells of AID Tg mice.20 In addition, we also analyzed the nucleotide changes in the p53, β-catenin and K-ras genes, all of which are thought to be involved in human carcinogenesis.3 First, we confirmed by sequencing that no mutations were present in those tumor-related genes in 32 randomly selected clones amplified from normal liver, stomach and lung tissue specimens obtained from the non-Tg littermates of the same mouse line. In contrast, several nucleotide changes had accumulated in the nontumor tissues as well as in the tumor tissues of the liver, stomach and lung of AID Tg mice (Table I). The most marked mutations had accumulated in the c-myc gene in the lung cancer developed in the AID Tg mouse, with a mutation frequency of 4.48 substitutions per 1 × 104 nucleotides. Notably, nucleotide changes in the c-myc gene had also accumulated in the noncancerous lung tissues (2.25 substitutions per 1 × 104 nucleotides). In contrast to the lung, far fewer or no mutations were observed in the c-myc gene in the stomach and liver tissues. To examine whether AID preferentially induces somatic mutations in the c-myc gene of the lung tissues, we analyzed the mutation frequencies of the c-myc gene in the nontumor tissues of the lung, liver and stomach of several AID Tg mouse littermates. Interestingly, we found that the incidence of nucleotide alterations in the c-myc gene of the lung tissues was substantially higher than that of the liver or stomach tissues in all the AID Tg mice examined (Table II). These findings suggest that the c-myc gene might be one of the lung-specific target genes in AID-mediated genotoxicity. In contrast to the c-myc gene, several nucleotide changes in the K-ras gene were observed in the gastric cancer tissues of AID Tg mice, whereas no mutations had accumulated in the lung and liver cancers. On the other hand, the p53 and β-catenin genes had acquired mutations in both the tumor and the nontumor tissues of all 3 organs, although the mutation frequencies were low compared to those observed in the c-myc gene of the lung or the K-ras gene of the stomach in AID Tg mice. Interestingly, 4 of 6 (67%) nucleotide changes in the p53 coding region observed in the AID Tg mice tissues resulted in amino acid substitutions with potential functional consequences, whereas all the mutations occurred in the c-myc gene were synonymous. These findings suggest that the mutational profiles induced by aberrant AID expression could differ in the liver, stomach and lung tissues.

Table I. Mutational Profile of the p53, c-myc, K-ras and β-catenin Genes in the Three Types of Cancers Developed in Aid Tg Mouse
GeneOrganCloneMutation frequency per 1041
Cell sourceMutated/totalMutations
  • 1

    Mutational frequency is expressed as the number of mutated nucleotide per 1 × 104 nucleotide in the each target gene.

p53 (exon7 and exon8)LiverTumor2/8920.42
Nontumor1/8910.21
StomachTumor2/8620.44
Nontumor1/8210.23
LungTumor1/11310.16
Nontumor1/8410.22
c-myc (exon1)LiverTumor1/8910.22
Nontumor1/9010.22
StomachTumor3/7630.79
Nontumor1/9110.22
LungTumor11/67154.47
Nontumor7/8092.25
K-ras (exon2)LiverTumor0/820<0.33
Nontumor1/8910.30
StomachTumor3/8641.28
Nontumor0/850<0.32
LungTumor0/1100<0.21
Nontumor0/890<0.31
β-catenin (exon3)LiverTumor2/8820.58
Nontumor0/880<0.29
StomachTumor0/850<0.30
Nontumor1/8810.29
LungTumor1/11010.20
Nontumor2/8620.60
Table II. Mutations Accumulated in the c-myc Gene in Nontumor Tissues of Several Littermates of Aid Tg Mice
GeneOrganCloneMutationsMutation frequency per 104
Mutated/total
c-myc (exon1)Liver0/290<0.68
Stomach2/3821.05
Lung3/3842.10
4/4141.95
5/4552.22
5/4252.38

Organ-specific activation of signaling pathways in the cancer tissues of AID Tg mice

The tissue-specific target preference of AID-induced mutagenesis in 3 types of epithelial organs prompted us to speculate that there might be differences in the expression profiles of tumor-related genes in each epithelial organ. Therefore, we analyzed the gene expression profiles in the liver, stomach and lung cancer tissues of the AID Tg mice. In the current study, we selected the following 7 tumor-related genes and analyzed their expression by quantitative real-time RT-PCR assay using primers and probes specific for each gene (Supplemental Table II); the human HCC tumor marker alpha-fetoprotein (Afp), cell-proliferation related insulin-like growth factor2 (IGF2), the cell-cycle regulator cyclin D1, matrix metalloproteinase7 (MMP7), c-myc, K-ras and β-catenin gene. We found that the expression of Afp, IGF2 and cyclin D1, all of which are closely associated with the development of human HCC,7, 23, 24 which were markedly enhanced in all the HCC tissues examined, compared to their expression in the stomach and lung cancers that developed in AID Tg mice (p < 0.01) (Fig. 2). Immunohistochemical analysis confirmed that Afp was expressed only in HCC cells, and not in lung or gastric cancer cells (Fig. 3). This indicates that the upregulation of Afp expression occurs only in the process of hepatocarcinogenesis, even though the mutagenic impact was common to every epithelial organ. Interestingly, the expression of IGF2 and cyclin D1 genes was also upregulated in the nontumor liver tissues comprising the constitutive expression of AID (Fig. 2). In vitro analysis showed that the transient expression of AID in cultured hepatoma-derived cells did not cause any changes in the transcription levels of cyclin D1 and Afp genes (data not shown), suggesting that the changes in the expression of these genes in the livers of AID Tg mice might be triggered by the accumulation of genetic changes, not by their transcriptional upregulation caused by AID expression in hepatocytes. In contrast, the expression of the c-myc and MMP7 genes, implicated in various types of carcinogenesis in human lung cancer,25 was significantly upregulated in the lung cancer tissues compared with that in the liver and stomach tumors (p < 0.01) (Fig. 2). In gastric cancer tissue, the K-ras gene expression was relatively high compared to that in the liver and lung, although not significantly so. Taken together, these findings indicate that organ-specific genetic changes in oncogenic pathways occur during the processes of cancer development in AID Tg mice and suggest the possibility that the gene expression profiles differ strikingly between liver, stomach and lung cancers, even though the causative genotoxic effect and the genetic background of the cancers are completely identical.

Figure 2.

Relative expression levels of 7 tumor-related genes in 3 types of epithelial tissues of the representative AID Tg mice. Relative mRNA levels of each gene were measured by quantitative RT-PCR assay. Values shown in the graphs are normalized relative to the specimens of wild-type (Wt) mice (mean ± SD of 3 mice).

Figure 3.

Immunostaining for alpha-fetoprotein (Afp) in epithelial tumors from AID Tg mice. Top left, liver section immunostained for Afp in a 53-week-old AID Tg mouse (magnification ×40). The HCC (arrowheads) is clearly distinct from nontumor tissues. Top right, higher magnification (magnification ×200) of the same HCC. Afp-immunostained tumor cells are scattered in this view. Bottom left, lung cancer section immunostained for Afp (magnification ×200). Lung cancer cells are not immunostained for Afp. Bottom right, stomach cancer section immunostained for Afp (magnification ×200). Stomach cancer cells, like lung cancer cells, are not immunostained for Afp.

Discussion

It is well recognized that human cancer is caused by genetic changes including the accumulation of various mutations in tumor-related genes.1–3 In fact, a large number of nucleotide alterations have been identified in various types of human cancers.26–28 Although all cancer cells seem to share a common set of genetic changes required for carcinogenesis,29 organ-specific genetic changes and gene expression signatures of oncogenic pathways are present in each human malignancy.30, 31 A recent study defined the genetic landscape of 2 human cancer types, breast and colorectal cancers, with a systematic analysis of their mutational spectra.32 Moreover, Greenman et al.33 demonstrated that mutational signatures differed between cancer types by surveying the numbers and patterns of somatic mutations in a diverse set of human cancer genomes. These tissue-specific expression profiles and somatic mutations would provide clues to understand the cellular processes involved in tumorigenesis. However, the molecular mechanisms underlying these organ-specific genomic changes are unknown. In the current study, we took advantage of AID Tg mice, which develop various types of solid tumors through the accumulation of genetic alterations, and investigated the mutational profiles and changes in expression profiles in 3 types of epithelial neoplasia: liver, stomach and lung cancers. We demonstrated that the constitutive genotoxic effects of AID resulted in the appearance of tissue-specific mutational spectra and gene expression profiles, possibly leading to the activation of organ-specific oncogenic pathways in liver, stomach and lung tissues.

AID is a unique cellular enzyme that can trigger point mutations and chromosomal translocations, both of which potentially lead to the initiation and progression of human lymphoid malignancies.15, 34–37 A causal relationship between the ectopic expression of AID and epithelial tumorigenesis has been suggested from the results of analyses of clinical specimens and an in vitro study of AID-mediated mutagenesis in human hepatocytes and gastric epithelial cells.16–19 These findings suggested that the inappropriate expression of AID acts as a DNA mutator, which enhances the genetic susceptibility to mutagenesis in various human epithelial cells.16 Consistent with these findings, the analyses described here provides the first evidence that AID Tg mice, in which the inappropriate expression of AID was driven under the control of an ubiquitous promoter, caused carcinogenesis in the stomach as well as in the liver, with the accumulation of genomic mutations in these cancer tissues. These findings indicate that aberrant AID expression can lead to carcinogenesis in the liver and stomach, in which organs the expression of AID is absent under physiological conditions.

We have shown here that organ-specific changes in mutational and gene expression profiles are present in liver, stomach and lung tissues, caused by the common genotoxic effects of aberrant AID activation. Indeed, the c-myc gene was the preferential target of AID-induced mutagenesis in the lung cancers of AID Tg mice. In addition, mutational changes in the c-myc gene also accumulated in the nontumor lung tissues of AID Tg mice. In contrast, mutations in the K-ras gene were frequently observed in the gastric cancer. These findings suggest that the c-myc and K-ras genes could be the preferential targets of AID-mediated genotoxic effects in the lung and stomach, respectively, and that the target selection of AID-mediated mutagenesis might be associated with organ-specific genetic changes in oncogenic pathways. However, it is not clear why the c-myc and K-ras genes are more sensitive to AID-mediated mutagenesis in the lung and stomach tissues, respectively, compared to the other organs during tumorigenesis. One possible explanation for the selective mutations of the c-myc gene is that increased levels of the c-myc gene transcription in lung tissue might cause the accumulation of somatic hypermutations induced by AID activity, as AID-induced hypermutation depends on the transcription level of the target gene.38–41 It must be noted that cancer cells with the specific c-myc or K-ras gene mutations might survive selectively during tumorigenesis in AID Tg mice, although all the nucleotide changes detected in the current analyses were sporadic, not clonal, in the tumor tissues. Thus, further analysis would be required to determine whether AID-mediated genotoxic effects lead to clonal mutations in certain tumor-related genes and how AID-induced mutagenesis contributes to changes in the expression levels of the target genes.

Interestingly, the Afp transcript was specifically upregulated in HCC, but not in the stomach and lung cancers that developed in AID Tg mice. Afp is a glycoprotein synthesized by the fetal liver and is the best known and widely used tumor marker for human HCC. The significance of the specific aberrant upregulation of Afp in the cancers that developed in the liver of AID Tg mice cannot be overemphasized, as it suggests that the molecular process of hepatocarcinogenesis in AID Tg mice might be very similar to that in the development of human HCC.28 It remains unclear why Afp overexpression was specifically observed in HCC tissues. It might be tempting to assume that the regulatory pathway for Afp gene expression is a specific target of AID-mediated genotoxic effects in hepatocytes during hepatocarcinogenesis. Another possibility is that Afp-expressing cells survived selectively as the tumor cells during the process of liver cancer development, irrespective of AID-induced mutagenesis.

In conclusion, the present study clearly demonstrates that the genetic changes induced by the genotoxic activity of AID show organ-specific profiles and are quite different in liver, stomach and lung cancers. As AID expression is induced by proinflammatory cytokine stimulation in various human epithelial tissues, our findings suggest that the target preference of AID-induced mutagenesis might contribute to the diversity of tissue-specific oncogenic pathways in the various epithelial organs.

Further analyses would be required to determine whether organ specific mutations observed in the AID Tg mice were the result of the preferential target selection of the AID activity or the characteristics of the tumor cells selectively surviving in each organ during the process of tumorigenesis. Moreover, a recent study has demonstrated that AID is required for the translocation between the c-myc and IgH loci by a mechanism common to class-switch recombination in immunoglobulin genes.11 This suggests that aberrant AID expression contributes to the occurrence of chromosomal abnormalities in epithelial organs. Therefore, determining the role of AID in the generation of chromosomal abnormalities in epithelial tissues will be a future challenge.

Acknowledgements

The authors thank Dr. K. Kinoshita (Shiga Medical Center Research Institute) for helpful comments; Dr. Y. Endo (Kyoto University) for the AID-stable transfectant cells.

Ancillary