¹Laboratory of Molecular Carcinogenesis, NIEHS, Research Triangle Park, North Carolina, NC 27709, USA

²Laboratory of Experimental Pathology, NIEHS, Research Triangle Park, North Carolina, NC 27709, USA

Correspondence to: Theodora R Devereux, Molecular Toxicology Group, MD D4-04, NIEHS, PO Box 12233, Research Triangle Park, North Carolina, NC 27709, USA

-catenin activation, and subsequent upregulation of Wnt-signaling, is an important event in the development of certain human and rodent cancers. Recently, mutations in the -catenin gene in the region of the serine-threonine glycogen kinase (GSK)-3 phosphorylation target sites have been identified in hepatocellular neoplasms from humans and transgenic mice. In this study we examined 152 hepatocellular neoplasms from B6C3F1 mice included in five chemical treatment groups and controls for mutations in the -catenin gene. Twenty of 29 hepatocellular neoplasms from mice treated with methyleugenol had point mutations at codons 32, 33, 34 or 41, sites which are mutated in colon and other cancers. Likewise, nine of 24 methylene chloride-induced hepatocellular neoplasms and 18 of 42 oxazepam-induced neoplasms exhibited similar mutations. In contrast, only three of 18 vinyl carbamate-induced liver tumors, one of 18 TCDD-induced liver tumors, and two of 22 spontaneous liver neoplasms had mutations in -catenin. Thus, there appears to be a chemical specific involvement of -catenin activation in mouse hepatocellular carcinogenesis. Expression analyses using Western blot and immunohistochemistry indicate that -catenin protein accumulates along cell membranes following mutation. The finding of mutations in both adenomas and carcinomas from diverse chemical treatment groups and the immunostaining of -catenin protein in an altered hepatocellular focus suggest that these alterations are early events in mouse hepatocellular carcinogenesis.

-catenin mutations; chemical carcinogenesis; mouse liver tumors; -catenin expression; H-ras

Introduction

In several recent studies -catenin is implicated as an important component of the Wnt-signaling pathway (Gumbiner, 1997), dysregulation of which may be a major factor in colon cancer, melanoma and hepatocellular carcinoma (de La Coste et al., 1998; Morin et al., 1997; Rubinfeld et al., 1997). The protein encoded by the adenomatous polyposis coli (APC) gene and the serine-threonine glycogen kinase (GSK)-3 together bind -catenin, modulating its degradation and expression (Munemitsu et al., 1995; Rubinfeld et al., 1996). Mutations in either APC or -catenin cause -catenin accumulation and upregulation of Wnt-signaling, the consequence of which is stimulation of cell proliferation and inhibition of apoptosis (Morin et al., 1997).

Recently, it was reported that hepatocellular tumors from L-PK/HC-myc, WHW/c-myc and L-PK/H-ras transgenic mice have mutations in the -catenin gene in the region containing the potential GSK-3 phosphorylation site (de La Coste et al., 1998). This is the first genetic alteration identified in both mouse and human hepatocellular carcinogenesis. We are particularly interested in this finding because of the high liver tumor response in chemically treated B6C3F1 mice in 2 year carcinogenesis studies. The identification in mouse liver tumors of chemical specific genetic alterations in genes relevant to human cancer will be useful information in the risk assessment process. Except for chemical-specific H-ras activation in some murine hepatocellular neoplasms (Devereux et al., 1993; Wiseman et al., 1986), the molecular alterations involved in mouse liver carcinogenesis are mostly unknown.

In this study we examined 152 hepatocellular neoplasms from five groups of treated mice in addition to control mice in order to determine if mutations in -catenin and the Wnt-signaling pathway play a role in chemically induced hepatocellular carcinogenesis in the B6C3F1 mouse. The treatment groups chosen included mice with hepatocellular tumors induced by vinyl carbamate, which is a genotoxic carcinogen associated with a high frequency of H-ras mutations; 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and methylene chloride, which are tumor promoters of both mutant H-ras and non-ras proliferative liver lesions; and oxazepam and methyleugenol, which are compounds that caused a significant increase in both hepatocellular adenomas and carcinomas without detectable H-ras mutations. Oxazepam is a commonly prescribed tranquilizer, and methyleugenol is used as a flavoring agent in food products and is structurally similar to known carcinogens safrole and eugenol. The -catenin mutation data were analysed for chemical specificity, for frequency in adenomas versus carcinomas to determine timing of these alterations, and for potential associations with H-ras activation (previously determined for these neoplasms). Additionally, we examined a subset of these tumors for protein expression and cellular localization of -catenin.

Results

Mutation analysis

One-hundred-and-fifty-two hepatocellular neoplasms from five diverse treatment groups and control B6C3F1 mice were analysed first for molecular alterations in exon 2 of the -catenin gene, the region that contains potential phosphorylation sites for GSK-3. Sixty-seven per cent (20/29) of the methyleugenol-induced neoplasms, 38% (9/24) of methylene chloride-induced neoplasms, 43% (18/42) of oxazepam-induced neoplasms, 17% (3/18) of vinyl carbamate-induced neoplasms, and 6% (1/18) of TCDD-induced neoplasms exhibited mutations in the -catenin gene (Table 1). In addition, mutations were detected in 9% (2/22) of the spontaneous tumors analysed (Table 1, Figures 1 and 2). Thus, there appears to be a chemical specific association with -catenin gene mutations in these neoplasms. In all cases the wild type allele was present as well, indicating the dominant character of these mutations.

Among the -catenin mutations identified in the chemically induced liver neoplasms included several deletions of varying lengths (Table 1). Most of these began near the beginning of exon 2, although the exact break points varied in each tumor. All of the point mutations detected affected codons 32, 33, 34, 36 or 41 (Table 1). While only these five codons were found mutated in the tumors, any of the three bases of these codons could be mutated. For example, among the -catenin mutations in the methyleugenol-induced tumors, eleven different base substitutions were identified among the four mutant codons observed. Moreover, there was not a preponderance of G to T or G to A mutations in the liver tumors as was observed for some chemically induced rat colon tumors (Takahashi et al., 1998). -catenin mutations in most of the five target codons and most of the different base substitutions were represented among the methyleugenol-induced, oxazepam-induced, and methylene chloride-induced liver tumors, the three groups of tumors with the highest frequency of mutations (Table 1). Thus, while there is an apparent association between specific chemicals and frequency of -catenin mutation, there does not appear to be a predominant chemical specific -catenin mutation pattern for these mouse hepatocellular neoplasms.

Deletion mutations that encompass large parts of the -catenin gene have been detected in human and transgenic mouse hepatocellular neoplasms (de La Coste et al., 1998), and these would not be detected by SSCP of exon 2. In order to better estimate the involvement of -catenin mutation in chemically induced hepatocellular neoplasms in the B6C3F1 mouse, we screened a representative sample of tumors that lacked exon 2 mutations for larger deletions. Because the vinyl carbamate, TCDD and methylene chloride tumors had the lowest frequencies of mutations within exon 2, we screened 4 - 5 of the exon 2 mutation negative tumors from each of these groups in which frozen tissue was available. By RT - PCR of total RNA only one of five methylene chloride tumors examined exhibited transcripts smaller than normal, and the wild type allele was present also (data not shown). We did not detect aberrant -catenin transcripts in any of the five vinyl carbamate-induced or four TCDD-induced hepatocellular tumors assessed. These data suggest that the -catenin mutation frequencies in the tumors observed for each chemical are not significantly underestimated. We also used comparison of actin transcript amplification to look for semiquantitative differences in -catenin expression between these tumors and normal liver. -catenin RNA expression in some tumors without detectable -catenin mutation appeared slightly greater than in normal liver (data not shown), but we did not consider this quantitative. These tumors were examined further with analysis of -catenin protein expression as described below, and generally it was not greater than that observed in normal liver.

For the oxazepam and methyleugenol-induced tumor groups, in which more than one treatment dose was used to induce tumors, we examined the -catenin mutation data for a relationship between dose response and mutation frequency. There appears to be a dose response relationship with -catenin mutation frequency for the oxazepam-induced tumors (Table 2). Twenty per cent of the 20 tumors in the 125 p.p.m. oxazepam dose group exhibited B-catenin mutations, whereas 58 and 70% of the 2500 p.p.m. and 5000 p.p.m. dose groups, respectively, had mutations. This response is supported by the low (9%) frequency of -catenin mutations in the spontaneous liver tumors. On the other hand, the -catenin mutation frequency in the methyleugenol-induced tumors did not exhibit an association with dose.

-catenin mutations were detected almost equally in adenomas and carcinomas, indicating that this is an early event in liver carcinogenesis. Of the 152 neoplasms examined for -catenin mutations, 70 were diagnosed as hepatocellular carcinomas, while the rest were adenomas. Thirteen of the 20 mutations in the methyleugenol-induced neoplasms and eight of the 17 mutations in the oxazepam-induced neoplasms were identified in adenomas. However, in the vinyl carbamate-induced and TCDD-induced tumors, four of 18 carcinomas had mutations compared to none of 18 adenomas. Because hepatocellular proliferative lesions may represent a continuum from adenoma to carcinoma, this does not rule out a much earlier event in tumorigenesis.

The tumors examined in this study were analysed previously for H-ras codon 61 mutations (Devereux et al., 1993, 1994; Watson et al., 1995). Except for the methylene chloride-induced liver tumors (Table 3), -catenin mutations were found only in tumors that lacked H-ras mutations. H-ras mutations were not detected in any of the 29 methyleugenol-induced liver tumors (data not shown) or high dose oxazepam-induced liver tumors (Devereux et al., 1994), the groups with the highest frequency of -catenin gene mutations. Fifty per cent of the vinyl carbamate-induced neoplasms that were examined in this study exhibited H-ras mutations (Watson et al., 1995), but the three -catenin mutations identified in this group were in non-H-ras activated tumors. On the other hand, in the methylene chloride-induced tumors 7 of the 9 -catenin mutations were identified in tumors that also exhibited H-ras codon 61 mutations (Table 3). It is interesting to note that the bases targeted for mutation in H-ras were different from those in -catenin in each of these tumors.

Protein expression

In addition to mutation analysis, expression of -catenin protein by Western blot analysis was performed on a subset of samples including seven methyleugenol-induced hepatocellular tumors, nine methylene chloride-induced tumors, seven vinyl carbamate-induced tumors, and five TCDD-induced tumors. In general, in tumors with mutant -catenin, the protein was expressed at higher levels than in normal liver or in those tumors in which mutations were not detected (Figure 3). In two methylene chloride-induced tumors with deletion mutations, a truncated -catenin product was expressed. Most of the tumor samples expressed a small amount of the phosphorylated form of -catenin, represented in the Western blot by a band slightly larger than the major -catenin band (Figure 3). The five vinyl carbamate-induced and four TCDD-induced liver tumors without detectable -catenin mutations by PCR of exon 2 and RT - PCR of the entire coding region exhibited low protein expression similar to that of normal liver (data not shown).

In order to examine the expression and intracellular localization of -catenin in tumor and normal tissues, we also evaluated some of the tumors by immuno-histochemical methods (Figure 4). We first examined 16 methyleugenol hepatocellular neoplasms, 10 of which exhibited -catenin mutations. Eight of these 10 tumors with mutations demonstrated some positive staining for the -catenin protein. The distribution of staining was patchy and non-uniform and predominately localized to the periphery of the tumor. The staining intensity in these areas was variable ranging from 1+ (above background, light brown) to 3+ (intense, dark brown). Three methyleugenol-induced tumors without detectable mutations had some patchy areas of positive staining. The remaining three methyleugenol tumors without mutations, and six other tumors without mutations from the vinyl carbamate treatment, TCDD treatment or control groups of mice exhibited very low staining (<1% of section with positive staining). Positive immunostaining was localized primarily to the cell membranes, although cytoplasmic staining was present in some of the neoplasms (Figure 4c). Strong positive staining of -catenin was found in an altered hepatocellular focus observed in a liver tissue section from a methyleugenol treated mouse; strong membranous and some outer cytoplasmic staining was observed in almost all cells of the focus (Figure 4e). Positive staining was also detected in focal areas of biliary proliferation (data not shown). There was little staining in normal liver, and if present, it was localized to cell membranes and some weak cytoplasmic staining (Figure 4a). Background staining was minimal or absent in all sections stained for -catenin, and no nuclear staining was observed in any of the tumors. Moreover, during examination of hematoxylin and eosin stained tumors we did not observe any histopathologic differences that correlated with the presence or absence of -catenin mutations.

Discussion

In this study we have identified somatic mutations of -catenin in a large proportion of mouse hepatocellular neoplasms induced by different chemicals. In contrast, mutations were rarely detected in spontaneous tumors. This finding is important because identical mutations are found in human hepatocellular cancers, and it may lead to an understanding of mechanisms of chemical carcinogenesis. All of the point mutations detected affected codons 32, 33, 34, 36 or 41, sites also shown to be mutated in human and transgenic mouse hepatocellular carcinomas (de La Coste et al., 1998; Miyoshi et al., 1998) and chemically induced rat colon tumors (Dashwood et al., 1998; Takahashi et al., 1998). While mutations of codons 33 and 41 affect the serine and threonine residues targeted for phosphorylation, mutations of codons 32, 34 and 36 are thought to affect phosphorylation-dependent ubiquitination and subsequent degradation of the protein (Aberle et al., 1997; Orford et al., 1997). Thus, mutation at any of these codons affects residues that play a role in regulation of -catenin turnover.

Our data show an equal number of adenomas and carcinomas with -catenin mutations and demonstrate accumulation of -catenin protein in an altered hepatocellular focus, indicating that -catenin mutation and subsequent increased expression of -catenin protein occur early in hepatocellular carcinogenesis. As in colon cancer, mutation of -catenin in hepatocellular tumors may represent a marker of initiation rather than tumor progression. Inactivation of the APC tumor suppressor gene appears to initiate most human colorectal neoplasia (Jen et al., 1994; Kinzler and Vogelstein, 1996), although almost all colon cancers that lack APC mutations have mutations of -catenin (Morin et al., 1997). APC, an important member of the Wnt signaling pathway, is considered a `gatekeeper' of the normal colon epithelium which maintains an appropriate equilibrium between cell growth and cell death by keeping -catenin expression in check (Kinzler and Vogelstein, 1996; Korinek et al., 1997). Mutation in either APC or -catenin causes a sustained increase in -catenin expression and an irreversible imbalance of cell growth over cell death in the colon. Our data suggest that accumulation of -catenin protein following mutation and subsequent increased cell proliferation occur early and may make a significant contribution to chemically induced hepatocellular carcinogenesis in the B6C3F1 mouse.

Generally, there was a good correlation in the hepatocellular tumors between -catenin mutations and accumulation of the -catenin protein, although nuclear staining was not observed. Current hypotheses based on findings in colon tumors suggest that following mutation, free -catenin accumulates in the cytoplasm and forms a transcription complex with transcription factors of the T cell factor (TCF) or lymphoid enhancer factor (LEF) families. The -catenin-bound complex then translocates to the nucleus and activates transcription of genes such as the c-myc protooncogene (He et al., 1998; Peifer, 1997; Pennisi, 1998), which results ultimately in cell proliferation. Most studies have demonstrated cytoplasmic and nuclear staining of -catenin protein in tumors with -catenin mutations (Takahashi et al., 1998), although a study by Fukuchi et al. (1998) describes accumulation along membranes in an endometrial tumor with a mutation and in several tumors without mutations; our tumors followed the latter pattern. One explanation for lack of nuclear staining in our tumors is that only a small increase in the -catenin-Tcf4 transcription complex, less than detectable by immunohistochemistry, accounts for a large effect on Wnt-signaling. Or it is possible is that the binding of the transcription complex with -catenin blocks antibody recognition sites. Another possibility is that during liver carcinogenesis accumulation of mutant -catenin at the cell membrane affects other processes such as cell adhesion in ways that may enhance the carcinogenesis process. This is being investigated further. Nevertheless, both our Western protein expression data and immunohistochemical findings provide evidence that -catenin accumulates in these tumors following mutation, and this may result in upregulation of Wnt-signaling.

No clear -catenin mutation `signature' was identified for any chemically induced tumor set, suggesting an indirect action of chemicals on this pathway. The finding of different bases in the H-ras and -catenin genes targeted for mutation in some of the methylene chloride-induced tumors provides evidence that the mutations in both genes were caused by indirect DNA damage or were spontaneous events. In a previous study we hypothesized that methylene chloride promoted liver cells that had spontaneous H-ras mutations (Devereux et al., 1993). However, unlike ras mutations that alter activity but have little effect on ras expression, -catenin mutations alter its protein binding capacity to APC and the GSK-3 kinase, thus interfering with its own regulation. The nature of these mutations is to reduce its own turnover and begin an expanding cascade of signaling. For this reason it seems likely, at least for -catenin, that chemicals may be more involved in the process of DNA damage (direct or indirect) than in promotion of the already mutated cells.

It has been puzzling that few genetic alterations that are important for carcinogenesis in humans have been identified in mouse liver tumors. For example, mutations in p53, Rb and p16, frequenty identified in many human cancers, are rare in mouse liver tumors (Gresani et al., 1998). Moreover, many chemically induced hepatocellular neoplasms in mice lack oncogenic ras mutations (Maronpot et al., 1995), and loss of heterozygosity is detected infrequently in mouse liver tumors (Davis et al., 1994; Manenti et al., 1995). Mutation of APC may not occur frequently in hepatocellular carcinogenesis as evidenced by a lack of APC mutations in human hepatocellular cancers in one study (Horii et al., 1992), and loss of heterozygosity at the APC locus is not frequent in either human or mouse hepatocellular carcinomas (Fujimori et al., 1991; Manenti et al., 1995). On the other hand, mutations in -catenin have been identified in a large proportion of hepatocellular tumors from both humans and transgenic mice (de La Coste et al., 1998; Miyoshi et al., 1998), and now in this study from chemically treated mice. The dominant nature and high frequency of -catenin mutations identified in the mouse liver adenomas and carcinomas in this study suggest that alteration in the stability and regulation of -catenin expression is an important early event in chemically induced hepatocellular carcinogenesis in the B6C3F1 mouse. Moreover, the -catenin mutations at the GSK-3 phosphorylation sites observed in this study are the same as those found in human hepatocellular carcinomas, suggesting that similar carcinogenic pathways exist in the two species. Future studies will investigate the roles of proteins that interact with -catenin to further understand the mechanisms of chemical induction of hepatocellular carcinogenesis in the mouse.

Materials and methods

Tumor samples

Collections of hepatocellular adenomas and carcinomas from B6C3F1 mice treated with vinyl carbamate or TCDD, and H-ras mutation analysis were described previously (Watson et al., 1995), and half of the spontaneous liver tumors examined from untreated B6C3F1 mice were from that study. B6C3F1 mouse liver tumor collection, tumor incidence data, and ras oncogene mutation analysis for the oxazepam and methylene chloride treatment studies have also been reported previously (Devereux et al., 1993, 1994; Kari et al., 1993; National Toxicology Program, 1993) In addition, approximately equal numbers of hepatocellular neoplasms from three dose groups (37, 75, and 150 mg/kg) of methyleugenol-treated and control B6C3F1 mice were used for this study. DNA isolation from these tumors and normal liver from B6C3F1 mice has been described previously (Devereux et al., 1993; Marmur, 1961).

-catenin mutation screening and identification

Single strand conformation polymorphism (SSCP) analysis was carried out on PCR products corresponding to exon 2 of the mouse -catenin gene. This is the region containing the GSK-3 targeted phosphorylation residues. The sequences of the PCR primers flanking the intron exon borders of exon 2 were : BCAT-1F, 5'-TACAGGTAGCATTTTCAGTTCAC-3' and BCAT-2R, 5'-TAGCTTCCAAACACAAATGC-3' (de La Coste et al., 1998). [³³P]dATP was incorporated into the PCR reactions for detection. Because the amplicon was about 296 bp, and in order to make the SSCP more sensitive, the restriction enzyme MspI was used to cut the PCR product before denaturation and gel electrophoresis. MspI cut exon 2 of -catenin at codon 16, and all higher condons in exon 2 were in the larger fragment (about 219 bp) that was observed for band shifts. Two gel conditions were used to detect mutations: 6% acrylamide gels with 10% glycerol were electrophoresed at 40 W for 6 h at 5°C, and 0.5´MDE (AT Biochem, Malvern, PA, USA) gels were electrophoresed at 3 W for 17 h at room temperature.

Mutations in samples with band shifts detected by SSCP analysis were identified by cycle sequencing with a [³³P]Thermo-Sequenase kit (Amersham, Cleveland, OH, USA). A fresh amplification reaction was performed, agarose gels electrophoresed, and the amplified bands cut out and purified on Qiagen columns (Qiagen, Valencia, CA, USA) for PCR purification prior to sequencing. The amplification primers also served as sequencing primers.

In addition to point mutations and deletions within exon 2 of -catenin, larger deletions in the gene have also been identified in hepatocellular neoplasms (de La Coste et al., 1998). Therefore, we also screened a representative sample of liver tumors, in which mutations were not detected by SSCP of exon 2, for larger deletions of this gene. Total RNA was isolated from tumors and normal liver, and RT - PCR was performed with primers described in de La Coste et al. (1998), and PCR products were run on 0.1 - 2% agarose gels. Following preparation of cDNA with Stratagene Superscript II Reverse Transcriptase kit (Lifetechnologies, Gaithersburg, MD, USA), three PCR reactions were performed on each sample to amplify the -catenin transcript; one with primers F1´R2 that included exons 1 - 5 (1042 bp), one with primers (F3´R5) that over lapped reaction one and encompassed the 3'-end of the transcript (1423 bp), and a final one with primers F1´R5 that spanned the entire coding region (2354 bp). Primers to mouse actin were also used to control for cDNA quality. Elongase Taq polymerase (Lifetechnologies) was used for the amplification reactions.

Western analysis for protein expression of -catenin

Cellular protein was extracted from frozen tumor tissues in radio-immunoprecipitation (RIPA) buffer [50 mM Tris-Cl pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.25% Na deoxycholate, 4 ug/ml A-PMSF, 20 ug/ml aprotinin and leupeptin, 1 mM NaF, 1 mM Na₃VO₄]. Aliquots of 75 ug protein were denatured by boiling in Laemmli sample buffer, resolved on 4 - 12% SDS - PAGE gels, and transferred to polyvinylidene fluoride (immobilon-P PVDF) membranes (Millipore Corp., Marlborough, MA, USA). The membranes were blocked in PBS containing 5% nonfat dry milk and probed with goat anti--catenin (cat. no. C-18) or goat anti-actin (cat. no. 1 - 19) at 1 : 250 dilution (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Peroxidase-conjugated anti-goat IgG (Santa Cruz) was used as the secondary antibody, and the membranes were developed with a chemiluminescence detection system (ECL reagents from Amersham).

Immunohistochemistry

Hepatocellular tumors were fixed in 10% neutral buffered formalin, processed routinely, and embedded in paraffin. Localization of -catenin protein expression was investigated using a polyclonal goat anti--catenin antibody (Santa Cruz, Santa Cruz, CA, USA) at a dilution of 1 : 100 on serial 6 mm sections. Slides were deparaffinized in xylene and hydrated through a graded series of ethanols to 1´Automation buffer^TM (Biomeda Corp., Foster City, CA, USA). Antigen unmasking was accomplished by heating in 200 ml citrate buffer in a microwave oven at 50% power for 5 min. Following a 1 min break, the cycle was repeated, and the slides were then allowed to cool for 20 min. Endogenous peroxidase activity was blocked with 3% H₂O₂ for 15 min. After rinsing in 1´Automation buffer^TM, sections were blocked with 5% normal goat serum for 30 min. The primary -catenin antibody was then applied to sections for 1 h at room temperature. Nonimmune rabbit IgG (Jackson Immunoresearch Labs, West Grove, PA, USA) 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 goat Immunocruz^TM staining system (Santa Cruz) according to the manufacturer's instructions and with 3,3'-diaminobenzidine as the color-developing reagent. Slides were counterstained with Harris hematoxylin, dehydrated through a graded series of ethanol washes to xylene, and cover-slipped with Permount^TM (Fisher, Springfield, NJ, USA).

Statistical analysis

Chi-Square analysis was used to compare the frequency of -catenin mutations in tumors after chemical treatment to frequency in spontaneous tumors.

Aberle H, Bauer A, Stappert J, Kispert A and Kemler R. (1997). EMBO J. 16, 3797-3804. Article MEDLINE

Dashwood RH, Suzui M, Nakagama H, Sugimura T and Nagao M. (1998). Cancer Res. 58, 1127-1129. MEDLINE

Davis LM, Caspary WJ, Sakallah SA, Maronpot R, Wiseman R, Barrett JC, Elliott R and Hozier JC. (1994). Carcinogenesis 15, 1637-1645. MEDLINE

de La Coste A, Romagnolo B, Billuart P, Renard CA, Buendia MA, Soubrane O, Fabre M, Chelly J, Beldjord C, Kahn A and Perret C. (1998). Proc. Natl. Acad. Sci. USA 95, 8847-8851. Article MEDLINE

Devereux TR, Foley JF, Maronpot RR, Kari F and Anderson MW. (1993). Carcinogenesis 14, 795-801. MEDLINE

Devereux TR, White CM, Sills RC, Bucher JR, Maronpot RR and Anderson MW. (1994). Carcinogenesis 15, 1083-1087. MEDLINE

Fujimori M, Tokino T, Hino O, Kitagawa T, Imamura T, Okamoto E, Mitsunobu M, Ishikawa T, Nakagama H and Harada H, et al. (1991). Cancer Res. 51, 89-93. MEDLINE

Fukuchi T, Sakamoto M, Tsuda H, Maruyama K, Nozawa S and Hirohashi S. (1998). Cancer Res. 58, 3526-3528. MEDLINE

Gresani KM, Rollins LA, Leone-Kabler S, Cline JM and Miller MS. (1998). Carcinogenesis 19, 1045-1052. MEDLINE

Gumbiner BM. (1997). Curr. Biol. 7, R443-R446. MEDLINE

He T-C, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B and Kinzler KW. (1998). Science 281, 1509-1512. MEDLINE

Horii A, Nakatsuru S, Miyoshi Y, Ichii S, Nagase H, Ando H, Yanagisawa A, Tsuchiya E, Kato Y and Nakamura Y. (1992). Cancer Res. 52, 6696-6698. MEDLINE

Jen J, Powell SM, Papadopoulos N, Smith KJ, Hamilton SR, Vogelstein B and Kinzler KW. (1994). Cancer Res. 54, 5523-5526. MEDLINE

Kari F, Foley J, Seilkop S, Maronpot R and Anderson M. (1993). Carcinogenesis 14, 819-826. MEDLINE

Kinzler KW and Vogelstein B. (1996). Cell 87, 159-170. MEDLINE

Korinek V, Barker N, Morin PJ, van Wichen D, de Weger R, Kinzler KW, Vogelstein B and Clevers H. (1997). Science 275, 1784-1787. Article MEDLINE

Manenti G, De Gregorio L, Gariboldi M, Dragani TA and Pierotti MA. (1995). Mol. Carcin. 13, 191-200.

Marmur J. (1961). J. Mol. Biol. 3, 208-218.

Maronpot RR, Fox T, Malarkey DE and Goldsworthy TL. (1995). Toxicology 101, 125-156. MEDLINE

Miyoshi Y, Iwao K, Nagasawa Y, Aihara T, Sasaki Y, Imaoka S, Murata M, Shimano T and Nakamura Y. (1998). Cancer Res. 58, 2524-2527. MEDLINE

Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B and Kinzler KW. (1997). Science 275, 1787-1790. Article MEDLINE

Munemitsu S, Albert I, Souza B, Rubenfeld B and Polakis P. (1995). PNAS 92, 3046-3050. MEDLINE

Orford K, Crockett C, Jenson JP, Weissman AM and Byers SW. (1997). J. Biol. Chem. 272, 24735-24738. Article MEDLINE

Peifer M. (1997). Science 275, 1752-1753. Article MEDLINE

Pennisi E. (1998). Science 281, 1438-1441. MEDLINE

National Toxicology Program. (1993). NTP TR # 443 U.S. Dept. HHS, PHS, NIH, RTP, NC. .

Rubinfeld B, Albert I, Porfiri E, Munemitsu S and Polakis P. (1997). Cancer Res. 57, 4624-4630. MEDLINE

Rubinfeld B, Albert I, Porfiri EFC, Munemitsu S and Polakis P. (1996). Science 272, 1023-1026. MEDLINE

Takahashi M, Fukuda K, Sugimura T and Wakabayashi K. (1998). Cancer Res. 58, 42-46. MEDLINE

Watson MA, Devereux TR, Malarkey DE, Anderson MW and Maronpot RR. (1995). Carcinogenesis 16, 1705-1710. MEDLINE

Wiseman R, Stowers S, Miller E, Anderson M and Miller J. (1986). Proc. Natl. Acad. Sci. USA 83, 5825-5829. MEDLINE

Figure 1 -catenin mutation screening by SSCP analysis (MDE gel shown). (a) Lanes 1 - 18, DNA from vinyl carbamate-induced hepatocellular neoplasms; lanes 19 - 36, TCDD-induced hepatocellular neoplasms; lane 37, spontaneous hepatocellular neoplasm. (b) Lane 1, DNA from normal B6C3F1 liver; lanes 2 - 11, DNA from methyleugenol-induced hepatocellular neoplasms; lanes 12 - 19, DNA from methylene chloride-induced hepatocellular neoplasms; lanes 20 - 32, DNA from oxazepam-induced hepatocellular neoplasms. `+' signs at the top of some lanes represent samples with mutations of exon 2 identified by sequencing

Figure 2 Sequencing of -catenin exon 2 mutations in mouse hepatocellular neoplasms. Lanes 1 - 3 DNA from vinyl carbamate-induced hepatocellular carcinomas; lane 4, TCDD-induced hepatocellular carcinoma; lane 5, spontaneous hepatocellular carcinoma, no mutation detected; lane 6, DNA from normal liver tissue; lanes 7 - 8, DNA from methyleugenol-induced hepatocellular carcinomas. Arrows point to mutant bands in each sequence, although the wild type allele is visible also

Figure 3 Expression of -catenin protein in chemically induced mouse hepatocellular neoplasms by Western blot analysis. Equivalent amounts of protein from total homogenates from the indicated samples were subject to gel electrophoresis and immunoblotting as described in the Materials and methods section. The blot was cut horizontally and developed with anti--catenin (top) or anti-actin (bottom). Lanes 1 and 2 represent VC-induced tumors; lanes 3 - 8 show methyleugenol-induced tumors, lanes 9 - 13 show methylene chloride-induced tumors, and lane 14 represents normal B6C3F1 mouse liver. `+' signs show samples in which either point mutations or deletion mutations in the -catenin gene were identified

Figure 4 Immunohistochemical analysis of -catenin expression in B6C3F1 mouse normal liver, an altered hepatocellular focus and a hepatocellular carcinoma from a methyleugenol treated mouse. (a) Normal liver reacted with -catenin antibody (´250); (b) normal liver control with non-immune rabbit IgG instead of primary antibody, (´250); (c) liver carcinoma with -catenin antibody, (´250); (d) liver carcinoma, control with non-immune rabbit IgG instead of primary antibody, (´250); (e) altered hepatic focus stained with -catenin antibody, (´200)

Table 1 Summary of -catenin mutations in B6C3F1 mouse hepatocellular neoplasms

Table 2 Relationship of -catenin mutations to chemical dose in liver tumors induced by oxazepam or methyleugenol

Table 3 Relationship of -catenin mutations to H-ras mutations in methylene chloride-induced hepatocellular neoplasms