Gene Expression Analysis on the Dicyclanil-Induced Hepatocellular Tumors in Mice

Abstract

Our previous studies showed the possibility that oxidative stress, including oxidative DNA damage, is involved in the mechanism of dicyclanil (DC)-induced hepatocarcinogenesis at the preneoplastic stage in mice. In this study, the expression analyses of genes, including oxidative stress-related genes, were performed on the tissues of hepatocellular tumors in a two-stage liver carcinogenesis model in mice. After partial hepatectomy, male ICR mice were injected with N-diethylnitrosamine (DEN) and given a diet containing 0 or 1500 ppm of DC for 20 weeks. Histopathological examinations revealed that the incidence of hepatocellular tumors (adenomas and carcinomas) significantly increased in the DEN + DC group. Gene expression analysis on the microdissected liver tissues of the mice in the DEN + DC group showed the highest expression levels of oxidative stress-related genes, such as Cyp1a1 and Txnrd1, in the tumor areas. However, no remarkable up-regulation of Ogg1—an oxidative DNA damage repair gene—was observed in the tumor areas, but the expression of Trail—an apoptosis-signaling ligand gene—was significantly down-regulated in the tumor tissues. These results suggest the possibility that the inhibition of apoptosis and a failure in the ability to repair oxidative DNA damage occur in the hepatocellular DC-induced tumors in mice.

Introduction

Dicyclanil (DC)—4,6-diamino-2-cyclopropylamino-pyrimidine-5-carbonitrile—is a pyrimidine-derived insect growth regulator that inhibits the molting and development of insects and is used in the field of veterinary medicine to prevent myiasis (fly strike) in sheep. As a result, minimal amounts of the parent drug and its metabolites are occasionally detected as residues in the edible tissues of sheep, such as the muscle, liver, and fat (WHO, 2000). It has been reported that the incidence of hepatocellular carcinomas was increased in mice that were fed a diet containing 1500 ppm of DC for 18 months, but negative results were obtained from the in vivo and in vitro genotoxicity studies of DC (WHO, 2000). Based on these results, the 54th meeting of the Joint Food and Agriculture Organization (FAO)/World Health Organization (WHO) Expert Committee on Food Additives (JECFA) concluded that DC is a nongenotoxic rodent carcinogen (WHO, 2000).

Recently, to clarify the mechanism of DC-induced hepatocarcinogenesis, we performed 2 experiments—a 2-week feeding study of DC in mice and a short-term study conducted using a 2-stage hepatocarcinogenesis model of mice with partial hepatectomy that were administered DC for 7 weeks. The results of the molecular pathological analysis on the livers of the mice that were fed a diet containing 1500 ppm DC for 2 weeks showed an up-regulation of the expression of several metabolism- and/or oxidative stress-related genes such as cytochrome P450 1A1 and 1A2 (Cyp1a1 and Cyp1a2) as well as thioredoxin reductase 1 (Txnrd1). In addition to the genes belonging to the same category, fluctuations in the expression of DNA damage/repair genes, such as 8-oxoguanine DNA glycosylase (Ogg1) and growth arrest/DNA-damage-inducible alpha (Gadd45a), were observed in the liver of the 2-stage hepatocarcinogenesis model of mice that were administered DC at the same dose (1500 ppm) for 7 weeks after an initiation treatment with N-dimethylnitrosamine (Moto et al., 2005).

In our second study, we used a two-stage hepatocarcinogenesis model of mice that were fed a diet containing DC for 13 and 26 weeks. In the liver of these mice, significant increases in the number of altered foci positive for γ-glutamyltranspeptidase (GGT-positive foci)—a predictive marker of preneoplastic foci in mice livers (Carter et al., 1985)—and the content of liver DNA 8-hydroxydeoxyguanosine (8-OHdG)—a representative marker of oxidative DNA damage (Kasai, 1997; Nakae et al., 1997; Umemura et al., 1998; Yoshida et al., 1999; Kinoshita et al., 2002; Dybdahl et al., 2003; Fortini et al., 2003) were observed (Moto et al., 2006). Based on these results, it was suggested that oxidative stress is probably involved in the mechanism of DC-induced hepatocarcinogenesis in mice (Moto et al., 2006). However, these evidences including the observed gene expressions were obtained by examining whole liver tissues at a tumor promoting stage of hepatocarcinogenesis, and it is unclear how these genes were actually expressed in the hepatocellular tumor tissues.

The aim of this study is to investigate the expression of genes including oxidative stress-related genes in the DC-induced tumor areas in mice. Recently, the laser-capture microdissection (LCM) technique was developed; biochemical or molecular biological analyses in small tissue areas can be performed using this technique (Michael et al. 1996; Suarez-Quian et al., 1999). This technique is a useful tool that enables the collection of only target tissues (areas) under the microscope and prevents contamination with non-target tissues. In this study, by using the LCM technique, histopathological examinations and gene expression analyses were performed on the DC-induced liver tumors obtained from the two-stage hepatocarcinogenesis model of mice with partial hepatectomy.

Materials And Methods

Animals and Chemicals

Four-week-old male ICR mice that were purchased from Japan SLC, Inc. (Hamamatsu, Japan) were maintained on a powdered basal diet (MF; Oriental Yeast, Co., Ltd., Tokyo, Japan) and tap water until they were 5 weeks of age. The mice were housed in polycarbonate cages with paper bedding and were maintained under standard conditions (room temperature, 22°C ± 2°C; relative humidity, 55% ± 5%; light/dark cycle, 12 hr). Animal care and experiments were carried out in accordance with the Guide for Animal Experimentation of the Tokyo University of Agriculture and Technology.

DC (CAS No. 112636-83-6) was kindly provided by Novartis Animal Health Inc. (Basel, Switzerland) for the experiment. N-diethylnitrosamine (DEN) was purchased from Nacalai Tesque, Inc. (Kyoto, Japan).

Experimental Design

A two-stage liver carcinogenesis model of mice was employed using the modified protocol of Porta et al. (1987) and Lee et al. (1989) (Figure 1). To initiate hepatocarcinogenesis, an ip injection of DEN at a dose of 30 mg/kg body weight was administered to the animals (day 0). Twelve hours before the DEN injection, a two-third partial hepatectomy was performed on the mice to enhance the regeneration of the liver with DNA damage. One week after the DEN injection, mice were fed a powdered diet containing DC at a concentration of 0 or 1500 ppm for 20 weeks. For liver sampling, after 20 weeks, the survivors were sacrificed by exsanguination from the abdominal aorta under ether anesthesia.

At necropsy, tissue samples were collected from all the remaining lobes of the liver. In the mice with liver tumors, the tissues including tumors were sampled. One-third of these samples were fixed with natural-buffered formalin for the histopathological examinations, and a second one-third was embedded in the OCT compound (Tissue-Tek; Miles Inc., Elkhart, USA) to freeze them for the gene expression analysis of the tumor areas and the staining of GGT. The remaining liver samples were weighed, frozen in liquid nitrogen, and stored at −80°C until subsequent gene expression analyses were performed by using two types of low density cDNA microarrays and real-time reverse transcription (RT) polymerase chain reaction (PCR).

Histology

For light microscopy, formalin-fixed liver tissues were embedded in paraffin, and 4-μm-thick tissue slices were sectioned. Hematoxylin and eosin (H & E) staining for the sections was performed according to routine histopathological methods. Histochemical staining of GGT was performed by using frozen liver slices, as described previously (Moto et al., 2005), in order to determine the presence of preneoplastic foci, altered foci, and neoplasms in the frozen sections.

Preparation of RNA from the Liver

For gene selection in the liver tissues by using microarrays and RT-PCR, total RNA was isolated from approximately 30 mg of the frozen liver tissues of all the animals by using an RNeasy Mini Kit (QIAGEN Inc., CA, USA) according to the manufacturer’s instructions. For gene expression analysis in the tumor areas, total RNA was isolated from the livers of three mice in the DEN (nontumor areas) and the DEN + DC groups (tumor and nontumor areas in the same animal), respectively. The frozen tissues embedded in the OTC compound were sliced as 10–15 serial sections with a thickness of 10 μm. One of the frozen sections was used for the histochemical staining of GGT to clarify the area of proliferative lesions. Other sections were stained with 0.05% truidin blue, and the tumor areas (2 carcinomas of 2 mice and 1 adenoma of one mouse) and nontumor areas of 3 mice in the DEN + DC group were collected using the laser microdissection system (AS LMD, Leica Microsystems, Wetzlar, Germany) under light microscopy. LMD samples of each mouse were dissected from 9–14 serial sections and pooled (total >20 mm² per each mouse) in RNAlater RNA Stabilization Reagent (QIAGEN Inc., CA, USA). Nontumor areas in 3 mice of the DEN group also collected similarly. The total RNA isolated from the collected samples was treated with DNase and also isolated by the same method described above.

Gene Expression Analyses

The relative expression of genes involved in stress, toxicity, and signal transduction in cancer were analyzed using two kinds of low-density and pathway-specific cDNA microrrays (Stress & Toxicity PathwayFinder cDNA GEArray (MM-12) and the Signal Transduction in Cancer cDNA GEArray (MM-44); SuperArray Inc., Bethesda, MD) containing up to 192 cDNA (96 cDNA per array) fragments from genes associated with these specific biological pathway. Using total RNA from frozen tissues of the 3 mice in which the altered- foci and tumors were respectively observed in the DEN group and the DEN + DC group, cDNA microarrays were performed according to the manufacturer’s protocol. Total RNA (3 μg) was reverse transcribed and double-stranded cDNA probes were generated by biotin-16-dUTP incorporation using the AmpoLabeling-LPR Kit (SuperArray), according to the manufacturer’s instructions.

Labeled cDNA probes were hybridized overnight. Following repetitive washing, hybridized cDNA probes were detected by chemiluminescence. Membranes were blocked for non-specific binding with GEAblocking solution (SuperArray). Bound biotinylated cDNA probe was detected with alkaline phosphatase-conjugated streptavidin and CDP-Star chemiluminscent substrate (SuperArray). Images of the membranes were recorded on X-ray film and digitally recorded on Bio Imaging System (Lab Works 4.0: UVP Inc., CA, USA). Gene spots were converted into numerical data using ScanAlyze software <http://rana.lbl.gov/EisenSoftware.htm>. Data were further processed with GEArray Analyzer (www.superarray.com), correcting for background noise by subtraction of the minimum value and normalizing to the value of 2 housekeeping genes (β-actin and GAPDH) of each individual array. Genes were considered present if the expression level was 2 times greater than that of the blank negative control. Genes were considered to be differentially expressed in the DEN + DC group if the mean of fold-changes was less than 0.5 or greater than 2.0 and observed to be statistically significant compared to the DEN group.

Real-time RT-PCR was carried out using the SuperScript III First-Strand Synthesis System (Invitrogen Corp., Carlsbad, CA, USA), and the cDNA aliquots were used in the quantitative real-time RT-PCR with SYBR Green using an ABI Prism 7000 Sequence Detection System (Applied Biosystems, CA, USA). The RT-PCR primers for the genes of Cyp1a1 (accession No. NM 009992), heme oxygenase 1 (Hmo1) (accession No. NM 010442), progesterone receptor (Pgr) (accession No.; NM 008829), tumor necrosis factor (TNF)-related apoptosis-inducing ligand 10 (Trail) (accession No. NM 009425), Txnrd1 (accession No. NM 015762), and Ogg1 (Accession No.; NM 010957) in this study were prepared as reported previously (Moto et al., 2005). To obtain the relative quantitative values for gene expression, β-actin was used as an internal control. Each sample was replicated 3 times, and all reactions were independently repeated 2 times in order to ensure the reproducibility of the results.

Statistical Evaluation

The quantitative data in the DEN and DEN+DC groups were represented as the means ± SD. In gene expression analysis of the liver tissues by real-time RT-PCR, the data were represented as the mean (bar) with individual spots. The significance of the difference in the data of body weight, food consumption, and gene expression ratio between the DEN alone and DEN + DC groups were analyzed by Student’s t-test. Data from histopathological examinations were assessed by the Wilcoxon test. A p-value less than 0.05 was considered to be statistically significant. Statistical analyses were performed using a statistical software (JMP 4.0.5J; SAS Institute, Inc., NC, USA).

Results

General Observations and Histopathological Findings in the Liver

During the experimental period, neither death nor remarkable treatment-related clinical signs were observed in either the DEN or the DEN + DC groups. However, a significant reduction in body weight gain was observed in the DEN + DC group at week 3 and from weeks 16 to 20 (Figure 2, Table 1). Macroscopically, the livers of all the mice in the DEN + DC group showed discoloration, and nodules/masses were observed on the surface of the liver in 3 of the 14 mice in this group (Figure 3A).

On histopathological examination, altered hepatocellular foci of the basophilic cell type were observed in both groups, and the incidence of these foci was 43% in the DEN + DC group and 38% in the DEN group. In the DEN + DC group, hepatocellular adenomas (29%) and carcinomas (14%) were also observed, and the total incidence of hepatocellular tumors was 36%; this incidence was significantly higher in this group as compared to the DEN group (p < 0.01). In the histochemical staining of GGT performed on frozen sections, altered foci and tumors showed positive reactions (Figure 3, Table 2). Similar to other histopathological findings, centrilobular hypertrophy and necrotic foci of the hepatocytes were also observed in the DEN + DC group (data not shown).

Gene Expressions in Liver Tissues

The findings of the gene expressions by the 2 microarray analyses are shown in Table 3. A level of up- or down-regulation greater than 2-fold was observed in 11 genes (up-regulation, 8 genes; down-regulation, 3 genes) of 192 genes in two microarrays. In the stress and toxicity pathway array and the signal transduction array, significant or remarkable up-regulations were observed in the oxidative stress- and metabolism-related genes, such as Cyp1a1, Hmox1, Cyp1a2, and Pgr (sex hormone gene), in the DEN + DC group. On the other hand, Trail was significantly down-regulated in both arrays.

The validation of gene expression in all animals was performed by real-time RT-PCR for the genes that showed significant and remarkable fluctuations in expression in the microarray analyses (Figure 4). In addition to these genes, Ogg1 and Txnrd1, which we focused on in our previous studies and which are oxidative stress-related genes, were also examined. A significant up-regulation in mean gene expression was observed in Cyp1a1, Hmox1, Ogg1, and Txnrd1 in the DEN + DC group. In the same group, the mean expression of Trail was significantly down-regulated. In the expression analysis carried out on the individual mice in the DEN + DC group, Trail showed a tendency toward low levels of expression in the mice with tumors. Ogg1 also showed a tendency toward low levels of expression in mice with tumors in the DEN + DC group, although its mean expression was significantly up-regulated as compared with its expression in the DEN group. The expression of the other genes in the present examination did not show any remarkable tendency toward either up- or down-regulation in the mice with hepatocellular tumors in the DEN + DC group.

Gene Expression Analysis in the Liver Tumor Areas

In the microarray and real-time RT-PCR analysis, Ogg1 and Trail were selected. In addition to these 2 genes, Cyp1a1 and Txnrd1 were also examined as marker genes of oxidative stress. The results of the expressions analysis of these genes in the microdissected tumor areas of the livers of the 3 mice selected from each group are shown in Figure 5.

The mean expression of Cyp1a1 and Txnrd1 showed a tendency toward up-regulation in the tumor areas, and a significant up-regulation was observed in the nontumor areas in the mice in the DEN + DC group as compared to the non-tumor tissues in the mice in the DEN group. The expression of Ogg1 was not remarkably up-regulated in the tumor areas, although significant up-regulation was observed in the non-tumor areas in the mice in the DEN + DC group. On the other hand, a significant down-regulation of Trail was observed in the tumor areas in the mice in the DEN + DC group, although the expression of Trail in the nontumor areas in this group was similar to that observed in the DEN group.

Discussion

The histopathological examinations conducted in the present study demonstrated that DC enhanced the induction of hepatocellular tumors in mice, and these data support our previous reports that found that DC has hepatocarcinogenic potential in mice. In the special staining of GGT, the altered hepatocellular foci and tumors in the liver showed a positive reaction. These data provide supportive evidence indicating that staining of GGT in the liver is a useful tool for the prediction of hepatocellular tumors in mice.

In gene expression analyses carried out by using microarrays in the liver tissues at the tumor formation stage, significant (or remarkable) fluctuations were observed in the expression of certain genes in the DEN + DC group. In addition to the expression of oxidative stress- and metabolism-related genes, such as Cyp1a1, Homx1, and Cyp1a2, significant fluctuations were observed in the expression of Pgr and Trail. Similar changes in the mean expression of Cyp1a1, Homx1, and Trail in the DEN + DC group were also confirmed by real-time RT-PCR analysis. In our previous study, DC also enhanced the production of reactive oxygen species (ROS) in vitro and the expression of Cyp1a1 and Cyp1a2 in the livers of mice at the early stage of hepatocarcinogenesis (Moto et al., 2005, 2006). It has been reported that CYP1A isoforms, such as CYP1A1 and CYP1A2, indirectly result in the production of very large amounts of oxidative stress-inducible substances, such as ROS, in comparison to other CYPs (Puntarulo and Cederbaum, 1998; Canistro et al., 2002). Hmox1 plays an effective role in counteracting oxidative damage, and it is expected that the activation of this gene has a potential of therapeutic tool for cancer (Fabiana et al., 2004). In addition to these genes, the up-regulation of the mean expression ofTxnrd1 and Ogg1, which were focused on as oxidative stress-related genes in our previous studies, were also observed in the DEN + DC group. TXNRD1 plays an important role in the redox regulation of multiple intra-cellular processes, including DNA synthesis, transcriptional regulation, cell growth, and resistance to cytotoxic agents inducing oxidative stress (Becker et al., 2000; Nyuyen et al., 2005). Based on these results, it can be considered that oxidative stress occurs in the livers of the DC treated-mice at the tumor formation stage, and it is possible that the persistence of oxidative stress plays an important role in hepatocarcinogenesis induced by DC.

In the present gene expression analysis in the liver tissues, tendencies toward low levels of expression were observed for Ogg1 and Trail in mice with hepatocellular tumors in the DEN + DC group. In fact, in the DEN group, the expression of Ogg1 in the tumor areas was approximately equal to that in the nontumor areas of the DEN alone group; however, the expression level of this gene in the nontumor areas was significantly up-regulated in the DEN + DC group. Ogg1, a gene involved in the repair of 8-OHdG, is known as an indicator of oxidative DNA damage and has been shown to be potentially involved in the carcinogenesis in various experimental models (Nakae et al., 1997; Yoshida et al., 1999; Shinmura and Yokota, 2001; Kinoshita et al., 2002, 2003). Our previous study reported that the administration of DC for a period of 13 and 26 weeks increased the formation of 8-OHdG in the liver DNA of mice at the stage of preneoplastic foci formation (Moto et al., 2006). Additionally, Cyp1a1 and Txnrd1, which were used as markers of oxidative stress genes in this study, were remarkably up-regulated in the tumor areas in the livers of mice in the DEN + DC group. Therefore, the present results suggest a possibility that the ability of Ogg1 to repair the oxidative DNA damage induced by DC failed in the hepatocellular tumors in which high levels of oxidative stress occurred.

A significant down-regulation in the expression of Trail was observed in the tumor areas in the livers of mice in the DEN + DC group. TRAIL is a member of the TNF family of cytokines that can induce apoptotic cell death in a variety of tumor tissues, and preclinical studies in mice and non-human primates have shown the potential utility of recombinant TRAIL for cancer therapy (Ashkenazi et al., 1999; Walczak et al., 1999; Yagita et al., 2004). Therefore, the down-regulation of Trail expression indicates the possibility that apoptosis and the self-regulation of carcinogenesis are inhibited in the tumors induced by DC. It is well recognized that one possible mechanism of nongenotoxic carcinogens is an alteration of apoptosis regulation because dysregulation of apoptosis results in the decreasing ability of cells to undergo apoptosis. For example, phenobarbital (PB) and WY-14643, representative nongenotoxic carcinogens in mice and rats, gradually increase anti-apoptosis proteins during the hepatocarcinogenesis in mice (Christensten et al., 1999). Additionally, it has been reported that short-term treatments of PB showed no increase of apoptosis in the liver of mice (Bursch et al., 2002). In our previous study, the down-regulation of Trail expression was not observed in our 2- and 7-week studies of DC. Taking these data into account, it is speculated that Trail plays an important role at the later stages of carcinogenesis, such as in the stages of promotion or progression, in the DC-induced hepatocarcinogenicity.

In conclusion, the present investigation on the gene expressions in the hepatocellular tumors induced by DC in a 2-stage hepatocarcinogenesis model of mice have demonstrated the possibility that oxidative stress, failure of the ability to repair oxidative DNA damage, and the inhibition of apoptosis occur in the tumor areas. In addition, the results of our previous and present studies suggest that the continuous treatment of DC induces oxidative stress in the liver and hepatocellular tumors, and oxidative stress plays an important role in the DC-induced hepatocarcinogenesis in mice.