According to the WHO report, 7.5 million people in the world died from cancer in 2005. The main types of cancer leading to this mortality were lung (1.3 million), stomach (1 million), liver (0.662 million) and colon (0.655 million).1 Thus, liver cancer is one of the most common tumors worldwide. Liver cancer could be a primary cancer or a metastasized cancer from other organs. Primary liver cancers have been categorized as hepatocellular carcinoma (HCC), cholangiocellular carcinoma and hepatoblastoma. Cancer patients having HCC account for 90% of the total in Japan. Most HCC is caused by hepatitis B virus and hepatitis C virus (HBV and HCV) infections, and is one of the most prevalent liver diseases in the world. The mechanisms of primary liver cancer causation by HBV and HCV infections are activation of oncogenes and transcription factors (e.g., ras and NF-κB) in liver epithelial cells.2 The over-activation of ras (K-, Ha-, and N-) results in uncontrolled signaling pathway in the cells, and thereby the Ras protein-product plays a significant role on its downstream signal transduction [e.g., over-expression of MAP kinase and inactivation of gap junctional intercellular communication (GJIC)].3
Carcinogenesis has been conceptualized as a multi-step, multi-mechanism process, consisting of an initiation, promotion and progression phase. While the exact mechanisms underlying each of these phases are not yet known, the reversible inhibition of gap junctional intercellular communication (GJIC) has been hypothesized to be a part of the tumor promotion phase and is supported by considerable evidence.4–8
Oncogenes have long been known to be derived from normal cellular growth-related genes that have been mutationally activated or over expressed and contribute to the neoplastic transformation of a normal cell. The ras oncogene is common to many cancers and when this oncogene is transfected into normal, noncarcinogenic epithelial cells, GJIC significantly decreases.9–16
Using a zinc-inducible metallothionein ras-T24 gene construct, the inhibition of GJIC by ras was determined to be dose dependent.10 Although the transfection of myc alone had no affect on GJIC in an epithelial cell line, the cotransfection with ras inhibited GJIC significantly more than cells transfected with ras alone.14 Several clones of the ras/myc-transfected cells were selected and a negative correlation was seen between the inhibition of GJIC and an increase in tumorigenicity as determined by increased anchorage independent growth (AIG), and tumor formation in nude mice inoculation with these various clones.14 Similarly, there was an excellent negative correlation between the loss of GJIC and the ras gene mutation rate in 4 solid tumor cell lines derived from malignant tissue.17
One strategy for efficacious chemoprevention and chemotherapy could be to prevent the down-regulation of GJIC by tumor promoting chemicals and to restore GJIC in GJIC-deficient tumor cells.5–7 There are many chemicals that have been tested for anti-tumorigenic properties using GJIC assay systems. One of the earliest antitumorigenic-compound tested on GJIC activity was lovastatin. Lovastatin prevented 12-O-tetradecanoylphorbol-13-acetate (TPA)-induced inhibition of GJIC that correlated with the suppression of cell transformation.18, 19 Lovastatin also up-regulated GJIC in ras-transformed lung epithelial cells.20 Another early reported anticarcinogenic compound reported to increase GJIC was cAMP.18, 21–23 Carotenoids and retinoids have been shown to increase GJIC in normal and initiated cell types, and up-regulate heterologous GJIC between the 2 cell types.24–26 Green tea extract and epigallocatechin gallate (EGCG), a bioactive component of green tea, prevented the inhibition of GJIC in F344 rat liver epithelial cells (WB), treated with various tumor promoters.27 Mice fed green tea, ad libitum in their water, prevented the tumor promoting effects of pentachlorophenol (PCP), and also reversed the PCP-induced inhibition of GJIC in their liver tissue.28 Diallyl disulfide, a sulfur compound from garlic, and the flavonoids, apigenin and tangeretin, were compounds that enhanced GJIC and prevented inhibition of GJIC by tumor promoters in rat liver epithelial cells.29
Recently we demonstrated that the ethanol soluble extract of psyllium increased GJIC and decreased AIG in ras-transfected F344 rat liver epithelial cells (WBras) but not in ras-myc, src and neu-transfected F344 rat liver epithelial cells (WBras-myc, WBsrc and WBneu) indicating specificity to a single oncogene.3 These data in particular brings up an important consideration in diet and cancer prevention. As cells acquire more mutations and multiple oncogenes, they tend to be more malignant and potentially less responsive to anticarcinogenic compounds, thus chemoprevention at the earlier stages of cancer by diet would be more likely effective. Follow-up experiments to the psyllium study successfully purified and identified the bioactive components of psyllium extracts as phytosterols, β-sitotesterol and stigmasterol.30 Commercially attained β-sitotesterol and stigmasterol, at concentrations similar to those estimated in the extracts, had very similar effects on the ras-transfected cells as did the psyllium extract.30 Caffeic acid phenyl ester, an active ingredient of honeybee propolis, also restored GJIC and inhibited AIG in the same ras-transfected cell line.31
GJIC, either as a cause or consequence of tumorigenic events, can serve as a central-based marker of measuring tumorigenic potential of exogenous chemicals or the antitumorigenic activity of pharmaceutical or phytochemical and nutrition based compounds. The assessment of GJIC can be done by simple, semi-high throughput in vitro assay systems. Targeting a centralized signaling pathway that is phenotypic of most cancers offers a very robust system to begin scientific studies on identifying and determining the mechanisms of tumorigenic and antitumorigenic compounds. Considering the difficulties of mechanistic studies using in vivo systems, high costs, ethical and humane considerations; establishing mode of actions, time- and dose-dependent parameters, first with a robust-mechanistic based in vitro cell system, and then, use these data for a more effective and efficient design for in vivo validation offers a more cost effective, science-based approach in identifying antitumorigenic compounds.
The primary objective of this study was to develop an in vivo model system that can be used to validate and match the in vitro results of antitumorigenic studies. The F344 rat liver epithelial cell (WB) system has been extensively used to study the effects of oncogenes and tumor promoters. Thus, we used the ras-transfected F344 rat liver epithelial cell (WBras) to establish an in vivo liver tumor model system in F344 rats. Liver tumors were induced by injecting these cells into the intraportal vein of F344 rats. Using this in vivo tumor model system, we demonstrated that the in vitro observations of decreased GJIC and gap junction protein expression, and increased expression of ras in these cells were also observed in the in vivo model system. The similarities between the in vitro and in vivo system indicates this paired model system should be valuable in mechanistic studies of determining the anticarcinogenic properties of chemopreventive and chemotherapeutic compounds specific to ras-induced tumors. We further characterized this in vivo model system by determining changes in the proteome of livers with the ras oncogene. Using 2-dimensional gel electrophoresis (isoelectric and SDS) and N-terminal sequencing analysis, 6 proteins unique to the ras-induced liver tumors were identified and offers additional markers for in vivo assessments.
Material and methods
Lucifer yellow-CH was purchased from Sigma (St. Louis, MO). Ampholines (pH 3.5–10 and pH 5–7) were purchased from GE Healthcare Bio-Science Corp. (NJ, USA). Ultra pure grade of urea was obtained from MP Biomedicals, LLC (Eschwege, Germany). Acrylamide specially prepared for IEF, dithioerythritol (DTE), 2-mercaptoethanol (2-ME) and 3-[(3-cholamidopropyl)dimethylammonio]-2-hydroxy-1-propanesulfonate (CHAPSO) were obtained from Nakalai Tesque, Inc. (Kyoto, Japan).
Cell lines and culture
The WB-F344 rat liver epithelial cell line (WB), obtained from Drs. J.W. Grisham and M.S. Tsao of the University of North Carolina (Chapel Hill, NC), is a non-tumorigenic diploid cell line with classic liver oval cell markers derived from a F344 male rat.32 The WBras cell line was developed from the transfection of WB cell line with a retroviral vector containing the v-Ha-ras oncogene and a neomycin-resistant marker.9 The cells were previously characterized as GJIC deficient in vitro and as tumorigenic in vivo.9
Male F344 rats, 5 weeks of age, purchased from Shizuoka Laboratory Animal Center (Hamamatsu City, Japan), were quarantined for 1 week and then randomly separated into experimental and control groups. Each animal was housed in each cage at the holding room controlled at 23 ± 2°C, 50 ± 10% humidity, and with a 12-hr light/12-hr dark cycle. Ten male rats (6 weeks of age) were divided into 2 groups, and were given AIN-76 diet in groups A and B for 4 weeks. At Week 1, rats were given intraportal-vein injections of WBras cells [2.5 × 106 cells in 0.2 ml of phosphate buffered saline (PBS)] in Group B or 0.2 ml of PBS in Group A under the sodium pentbarbital (50 mg/kg body weight)-anesthesia that was performed according to the de Feijter's procedure with some modification.9 All rats were provided with the diets and tap water ad libitum, and were sacrificed at Week 4. Whole livers and excised tumors were weighed. The tumors were white and firm, and can be distinguished from normal tissue macroscopically, and were removed with scalpels. The whole liver (i.e., mixing normal liver and tumors) was homogenized after frozen with liquid nitrogen for 2-DE protein analysis. For histological analysis of liver tissue, the whole liver was fixed with Lillie solution (10% formaldehyde in PBS, pH 7.0), and a section (5 μm thickness) was prepared on the glass plate. The section was stained with hematoxylin and eosin for histological analysis.
Scrape loading/dye transfer assay
WB cells and WBras cells (5 × 104) were plated in 35-mm-diameter culture plates (Becton Dickinson Labware, Franklin Lakes, NJ) with 2 ml of modified Eagle's medium (DMEM, GIBCO Laboratories, Grand Island, NY) containing 5% fetal bovine serum (5% FBS-DMEM) and cultured for 4 days. The cell monolayer became confluent at this time. GJIC was measured using the scrape loading/dye transfer technique.6, 28 Dye migration was observed and digitally photographed at 200× using a Zeiss Axiovert 25 microscope illuminated with an Osram HBO 50 W lamp and equipped with a Fuji film CCD camera. The distance of dye migration perpendicular to the scrape (i.e., between adjacent cells linked only by gap junctions) represents the ability of cells to communicate via GJIC.
Immunofluorescence stain of Cx43
WB cells and WBras cells (2 × 103) were plated in a 4-well glass attached chamber slide (Nalge Nunc International, Naperville, IL) with 1 ml of 5% FBS-DMEM and cultured for 4 days. After the incubation period, cells were washed with PBS thrice, and then fixed with 3.5% formaldehyde/0.5% Triton-X100 (30 min) and washed with PBS; permeabilized the membrane with 0.05% saponin/PBS (30 min) and then washed with PBS. For an immunohistochemistry analysis of liver tissue, a section (5 μm thickness) on the glass plate was used. The cells and the tissue section were blocked with 10% goat serum (Sigma, St Louis, MO)/PBS for 1 h, and then treated with anti-Cx43 antibody diluted 1:100 in 1% goat serum/PBS, and incubated on a shaker at 4°C for 12 hr. The secondary antibody was a Cy3-conjugated rat or mouse antibody IgG (Jackson Immuno Research Laboratories, Inc., West Grove, PA), which was diluted 1/200 in 1% goat serum/PBS and incubated in the dark on a shaker at room temperature for 1 hr. The cells were then treated with DAPI (1 ng/ml) for 5 min, and washed with PBS and mounted with a cover slip using poly-aquamount (Polysciences, Inc., Washington, PA). Microscopic images were digitally obtained from an epifluorescence microscope equipped with a CCD camera (Nikon, Tokyo, Japan).
Western blot analysis of Cx43 and Ras
Proteins of liver were extracted with 20% SDS solution according to the reported method.33 The protein content was determined with the DC assay kit (Bio-Rad Corp., Richmond, CA). The proteins (3 μg) were separated on 7.5% (for Cx43) or 14% (for Ras) SDS-PAGE34 and electrophoretically transferred from the gel to PVDF membranes (Millipore Corp, Bedford, MA).35 Cx43 and Ras were detected with anti-Cx43 polyclonal antibody (Zymed, South San Francisco, CA), and anti-Ras monoclonal antibody (Upstate biotech, Lake Placid, NY) using horseradish peroxidase-conjugated secondary antibody (Bio-Rad Corp., Richmond, CA), and then observed with the ECL detection kit (Amersham Biosci. Corp., Piscataway, NJ).
Protein extraction and 2-DE analysis with N-terminal sequencing analysis
Frozen liver samples were dissolved in lysis buffer [8 M urea, 4% (w/v) CHAPSO], and lysed using spatulas and syringes. The lysates were centrifuged at 16,000 × g for 5 min to remove insoluble materials. The centrifugation was repeated to remove the remaining insoluble matter. Protein concentration was determined using the DC assay kit (Bio-Rad Corp., Richmond, CA) with bovine serum albumin in lysis buffer as the standard. DTE and ampholine (pH 3.5–10) were added to the protein samples in final concentrations of 65 mM and 2% (v/v), respectively.
The separation of proteins (150 μg) by 2-DE was performed by the O'Farrell's procedure, modified by Tsujimoto et al.36, 37 Briefly, polyacrylamide gels for IEF contained 8 M urea, 4% CHAPSO, 65 mM DTE, 2.5% ampholine (pH 3.5–10), and 2.5% ampholine (pH 5-7). Cathode and anode electrode solutions for IEF were 2% (v/v) TEMED and 0.02 M H3PO4. IEF was applied at 5°C using an NA-1313 system (Nihon Eido, Tokyo, Japan) (200 V for 1 hr, 400 V for 12–18 hr, and 800 V for 1 hr). The gels were washed thrice with distilled water, and then equilibrated with SDS-PAGE sample buffer [2.5% (w/v) SDS, 62.5 mM Tris-HCl, pH 6.8] containing 5% (v/v) 2-ME at room temperature for 30 min. The gels then were placed onto SDS-polyacrylamide and overlaid with agarose solution [1–2% (w/v) agarose, 0.01% (w/v) bromphenol blue in SDS-PAGE sample buffer]. SDS-PAGE was performed using NA-1113 units (Nihon Eido, Tokyo, Japan) by the method of Laemmli.34 To visualize the protein spots, the gels were stained with CBB Stain One (Nakalai Tesque, Inc.).
To identify proteins, the proteins on the SDS-PAGE gel were electrophoretically transferred to polyvinylidenfluoride membranes. N-Terminal sequencing of proteins was carried out by an automated Edman degradation and sequence analysis with a Shimadzu PPSQ-21 system (Kyoto, Japan).
Statistical comparisons were made using Student's t-test. The results were considered significantly different when p < 0.05.
Ras expression, localization of Cx43 and GJIC in rat liver cells (in vitro)
Immunological analysis was used to assess the levels of the tumorigenic protein marker, Ras (Western blots), and the anti-tumorigenic protein marker, Cx43 (immuno-localization), in the in vitro cell lines used for this study. As expected, the normal rat liver epithelial WB cell line had low levels of Ras and high levels of Cx43 in the plasma membranes, while the ras-transfected WBras cell line expressed high levels of the Ras protein and low levels of Cx43 in the plasma membranes (Fig. 1, panels a and b). Note that Cx43 staining in the WBras cells was primarily diffused through the cytoplasm. Cx43 in the non-tumorigenic WB cells clearly formed functional gap junction channels that allowed the passive diffusion of the fluorescent dye, Lucifer yellow, from the cells injected with this dye by the scalpel blade that traversed through several cell layers adjacent to the scrape (Fig. 1, panel c). Conversely, the WBras cells exhibited a very low rate of dye migration (Fig. 1, panel c), which is probably a function of the Cx43 being located primarily in the cytoplasm and not the plasma membrane (Fig. 1, panel b). Thus, the results of these 3 markers (Ras, Cx43, and GJIC) are consistent with the predicted phenotypes of normal and tumorigenic cell types.
Tumor incidence and amount of Ras and Cx43 in rat liver (in vivo)
The protocol for in vivo studies are indicated in Fig. 2, panel a, and the mean body weight, liver weight, and relative liver weight (g/100 g body weight) in each group are indicated in Table I. The mean body weight of rats receiving WBras cells (Group B) was not significantly lower than the PBS sham Group A (Table I). Macroscopic examination revealed that all 5 rats in Group B had liver tumors, while none of the 5 rats in Group A had detectable liver tumors (Fig. 2, panel b). Rats in Group B had no detectable tumors in other organs. The average liver weight, relative liver weight, tumor weight and tumor weight/liver weight among the 5 rats in Group B were significantly higher than those in Group A (Table I). The tumor weight/liver weight of Group A rats was 0%, and that of Group B rats was 64.1% range from 60.4-67.8% (Fig. 2, panel b). No macroscopic tumors were detected in other organs. Multiple small focal lesions of WBras cells and large aggressive neoplasm invading the normal liver tissue and lung metastasis at 1 and 3 weeks were observed after the injection of WBras cells (1.0 × 107 cells) in the hepatic portal vein of F344 rats.9 Our results indicate that an intraportal-vein injection of WBras cells (2.5 × 106 cells) in 6-week old male rats results in stably macroscopic tumor production in liver without macroscopic metastasis.
Table I. Weight of Mean Body, Liver, Relative Liver, Tumor and Relative Tumor for Group “A” and “B” Rats
Body weight (g)
Liver weight (g)
Relative liver weight (%)
Tumor weight (g)
Tumor weight/liver weight (%)
Relative liver weight is % of liver weight/body weight. Values in group B showing significant difference from corresponding group A (p < 0.05) by Student's t-test are indicated by an asterisk.
210.8 ± 7.0
8.74 ± 0.47
4.1 ± 0.1
204.7 ± 6.4
19.9 ± 4.5*
9.7 ± 2.2*
12.8 ± 3.5*
64.1 ± 3.3*
Considering that Ras and Cx43 were a positive and negative phenotypic proteins of the WBras cells, we also assessed the tumors of group-B rats for these 2 proteins and compared the results to the sham treated rats. As predicted, the rats injected with the WBras cells demonstrated high levels of Ras and low levels of Cx43 proteins in the liver as compared to the sham treated animals (Fig. 2, panel c).
Histochemical analysis in rat liver tissue (in vivo)
In macroscopic observations, hematoxylin-eosin stained the tumors blue indicating basophilic cells in the liver of rats injected with WBras cells (Group B) (Fig. 3, panel a). In microscopic observations, the tumor consisted high densities of basophilic cells, and a disordered arrangement of cells. In contrast, the normal liver tissue exhibited eosinophilic, cytoplasmic-rich cells, and had a radial cellular orientation from the central vein (Fig. 3, panel b). Cx43 protein in normal tissue was abundantly expressed in the cytoplasm and also formed distinct plaques on the plasma membrane of cells in normal tissue (Fig. 3, panel c). In contrast, Cx43 was rarely expressed in either the cytoplasm or the plasma membranes in the tumor tissue (Fig. 3, panel c). Cells in normal tissue clearly exhibited functional gap junction channels that allowed the passive diffusion of the fluorescent dye from the cells injected with Lucifer yellow by a scalpel blade through several cell layers adjacent to the scrape (Fig. 3, panel d). The tumor exhibited a very low rate of dye migration (Fig. 3, panel d), which is probably a function of the Cx43 being located on the plasma membrane.
2-DE analysis of rat liver protein (in vivo)
2-DE/N-terminal sequencing method for proteomic analysis extracted from all 10 rats of groups A and B was used to identify proteins uniquely expressed in the tumor-bearing animals. We chose to assess the entire liver tissue of the tumor-bearing rats considering that genomic changes are not only restricted to the oncogenic cells but also the cells neighboring the tumor tissue. A representative 2-DE image comparing the profiles of protein extracted from the livers of the sham and WBras-injected rats is shown in Fig. 4, panel a. The spots detected were between the molecular weight and pI ranges of 16–130 kDa, and 4.0–7.5, respectively. We limited our assessment of spots to this range because they were significantly clearer and more distinct than the spots observed out of this range, thus allowing for a more accurate determination of altered protein expression between the sham (Group A) vs WBras treated (Group B) rats. Using these ranges as a criterion for selecting spots, 10 spots were visually detected as significantly different between the 2 treatment groups (Fig. 4, panel a).
Four spots (Fig. 4, spots 1–4) decreased (Group A > Group B) and 2 spots (Fig. 4, spots 5 and 6) increased (Group A < Group B) in the animals bearing liver tumors. N-Terminal amino acid sequencing was used to identify the proteins in these spots. The identities of Spot 1 was glucose-regulated protein 78 (GRP78; 75 kDa, pI 4.9; P06761), Spots 2 and 3 were heat shock protein 60 (Hsp60; 50 kDa, pI 5.3 and 5.4; P63039), Spot 4 was the β-chain of ATP synthase (45 kDa, pI 5.0; P10719), Spot 5 was Vimentin with a 71 amino acid deletion from its N-terminus [Vimentin (Δ1–71); 41 kDa, pI 4.7; P31000] and Spot 6 was β-Actin with a 46 amino acid deletion from its N-terminus [β-Actin (Δ1–46); 35 kDa, pI 5.4; P60711]. Spot 5 was not detected in the normal liver (Group A), and therefore the amount of Vimentin (Δ1–71) in the liver of normal rats was estimated to be less than the detection limit (0.2%, 0.3 μg in 150 μg of liver protein) in the N-terminal sequencing system. Spots 1–4 and 6 were detected in the liver both Groups A and B. Three additional spots (7: 69 kDa, pI 5.4; 8: 47 kDa, pI 4.9; 9: 32 kDa, pI 5.1) decreased, and one additional spot (10: 40 kDa, pI 4.9) increased in the WBras injected rats as compared to the sham, but these proteins could not be sequenced, possibly blocking the N-terminal amino acids of these proteins were blocked (Fig. 5).
We used a 2-DE/N-terminal sequencing method to identify proteins that are uniquely up-regulated and down-regulated in the liver tissue of rats treated with a WBras cell line that induces hepatocellular tumors. The over-expression of ras in cells constitutively activates mitogen activated protein kinase signaling pathways, resulting in uncontrolled cell growth.14, 38, 39 In the specific cell lines used in this study, the cell doubling time of WB cells was 24 hr, while the doubling time for the WBras cells was only 20 hr (data not shown). In addition, WB cells sufficiently expressed Cx43 on the plasma membrane and exhibited significant intercellular communication, while the WBras cells that had an over-expression of the Ras protein resulted in the redistribution of Cx43 into the cytoplasm, resulting in deficient intracellular communication through gap junctions (Fig. 1, panels a and b).
Previous studies indicated that the injection of WBras cells in male F344 rats induced tumor formation, while male F344 rats injected with the normal epithelial rat liver cells (WB cells) exhibited no detectable tumors.14 Our data confirmed that the injection of WBras cells into male F344 rats resulted in the formation of visible tumors (Fig. 2, panel b), as well as several tumor formation parameters (liver weight, relative liver weight, tumor weight, and tumor weight/liver weight; Table I). Similar to previously published results using this same cell line,39 tumorigenicity in the liver was directly proportional to Ras expression and inversely proportional to Cx43 expression (Fig. 2, panel c). Thus, in addition to scoring tumorigenicity by macroscopic detection and tumor weight/liver weight, Western blot determinations of the Ras and Cx43 proteins could be prospective biomarkers that complement the traditional endpoints and also be potentially used for the early detection of tumors in rodent model systems. In addition, the nature of the tumors induced by WBras was identified as densely basophilic cells without Cx43 expression, and functional GJIC (Fig. 3).
To further identify other potential unique markers of hepatocellular carcinomas, 2-DE/N-terminal sequencing method for proteomic analysis was used to identify proteins that were uniquely expressed in the tumor-bearing animals that were treated with the WBras cells. This small scale proteomic approach detected 10 proteins that were either up-regulated or down-regulated in the tumor-bearing animals as compared to the control animals. N-Terminal amino acid analysis was able to identify six of these 10 proteins with the following 4 proteins being down-regulated: 2 isoforms of Hsp60, the β-chain of ATP synthase, and GRP78. The 2 other proteins that were up-regulated were identified as β-Actin (Δ1–46) and Vimentin (Δ1–71).
The mechanisms of GRP78 and Hsp60 are well studied and are known to act primarily as chaperones,40 and thus could negatively alter the function and location of proteins that play vital roles in growth and control. Future experiments will be done to determine if these proteins play a role in the expression and function of gap junction proteins. The induction of rat liver tumors by N-nitorosomorpholine41 also resulted in the decrease of Hsp60 and the β-chain of ATP synthase in rat liver, and a decrease of GRP78 induced mutation in human RC cells.42 Moreover the β-chain of ATP synthase also plays a regulatory role of apolipoprotein A-I receptor, which is involved with hepatic HDL endocytosis for cholesterol transportation to liver cells, its metabolism, and the secretion to bile.43 Therefore, the decrease of the β-chain of ATP synthase by the tumor-bearing animals with WBras cells might be involved with the comprehensive functions of the liver.
Vimentin is a growth-related gene that is frequently expressed in growth factor-induced proliferation of epithelial cells and up-regulated by tumor promoters.44 Although disruption of the Actin cytoskeleton is a common phenotype of cancer cells, the up-regulation of Actin in total protein extracts is difficult to interpret, but these 2 proteins with N-terminal deletions can serve as potential markers of liver tumor incidence similar to that of Ras and Cx43 proteins. In the literature, the intact form of Vimentin protein has already been used in pathological tissue diagnosis as a marker for mesenchymal malignant tumors,45 and as a liver oval cell marker.46
In summary, the use of a 2-DE/N-terminal sequencing method for proteomic analysis, allowed for the identification of 6 proteins that were either uniquely up-regulated or down regulated in the livers of tumor-bearing animals that were injected with WBras cells, which is a F344 derived rat liver epithelial cell line that was transfected with the ras oncogene. Further, the down-regulation of gap junctional intercellular communication by the oncogenic Ras protein is consistent with the hypothesis that the promotion of an initiated cell must be accompanied by a significant decrease in intercellular communication to prevent the suppression of growth by neighboring cells. In contrast, the up-regulation of GJIC has been demonstrated to restore a non-tumorigenic phenotype in oncogene containing cells47, 48 (i.e., normal epithelial morphologies, lack of anchorage-independent growth, lack of frank tumors in nude mice). Up-regulation of GJIC has also be linked to increased apoptosis,49 and an increase in the expression of Cx43 relocalization into the plasma membrane induced apoptosis in the WBras cell line,50 which was used in this study. Therefore, Cx43, Ras, Vimentin, β-Actin, the 2 isoforms of Hsp60, GRP78, and the β-chain of ATP synthase can serve as markers of tumorigenicity. Also, these results indicate that the injection of WBras cells into male F344 rats is a reliable model system of tumorigenicity, which can be used to assess drugs, nutritional agents, phytochemical agents for anticarcinogenic properties. We previously demonstrated that β-sitosterol in psyllium increased the level of Cx43 and restored GJIC in this WBras cell line30; thus, we are planning future experiments that will use this cancer model system and the markers established in this study to determine the efficacy of antitumorigenic compounds, such as orally administered β-sitosterol in preventing hepatocellular carcinomas. Having an in vitro/in vivo tumor model system allows for an effective, science-based method to screen for antitumorigenic compounds, first, by identifying and determining doses of antitumorigenic compounds using in vitro cell model system that uses GJIC as the primary indicator of anti-tumorigenicity, and then validate these results in vivo by treating the rats that had tumors induced by the same cell system used in vitro. Similarly, such a system would also be valuable in determining the underlying mechanisms of antitumorigenic chemicals.
In addition, the identification of novel new markers will also lead to new mechanistic-based experiments in understanding ras-induced liver cancers, such as variation of cleavage in maturation process of proteins or alternate splicing in the post-transcriptional phase of gene expression leading to N-terminal deleted proteins. The direct implication is to assist in reducing the risk to human liver cancers in large parts of the world.
This research was supported in part by Sumitomo Foundation grant to Yoshiyuki Tsujimoto; and by NIEHS grant R01 ES013268-01A2 to Brad L. Upham, and its findings are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS.