Omega-3 polyunsaturated fatty acids inhibit hepatocellular carcinoma cell growth through blocking β-catenin and cyclooxygenase-2

Hepatocellular carcinoma (HCC) is a common human cancer with high mortality, and currently, there is no effective chemoprevention or systematic treatment. Recent evidence suggests that cyclooxygenase-2 (COX-2)–derived PGE2 and Wnt/β-catenin signaling pathways are implicated in hepatocarcinogenesis. Here, we report that ω-3 polyunsaturated fatty acids (PUFA), docosahexaenoic acid (DHA), and eicosapentaenoic acid (EPA) inhibit HCC growth through simultaneously inhibition of COX-2 and β-catenin. DHA and EPA treatment resulted in a dose-dependent reduction of cell viability with cleavage of poly ADP ribose polymerase, caspase-3, and caspase-9 in three human HCC cell lines (Hep3B, Huh-7, HepG2). In contrast, AA, a ω-6 PUFA, exhibited no significant effect. DHA and EPA treatment caused dephosphorylation and thus activation of GSK-3β, leading to β-catenin degradation in Hep3B cells. The GSK-3β inhibitor, LiCl, partially prevented DHA-induced β-catenin protein degradation and apoptosis. Additionally, DHA induced the formation of β-catenin/Axin/GSK-3β binding complex, which serves as a parallel mechanism for β-catenin degradation. Furthermore, DHA inhibited PGE2 signaling through downregulation of COX-2 and upregulation of the COX-2 antagonist, 15-hydroxyprostaglandin dehydrogenase. Finally, the growth of HCC in vivo was significantly reduced when mouse HCCs (Hepa1-6) were inoculated into the Fat-1 transgenic mice, which express a Caenorhabditis elegans desaturase converting ω-6 to ω-3 PUFAs endogenously. These findings provide important preclinical evidence and molecular insight for utilization of ω-3 PUFAs for the chemoprevention and treatment of human HCC. [Mol Cancer Ther 2009;8(11):3046–55]


Introduction
Hepatocellular carcinoma (HCC) is the fifth most common human cancer with high mortality, and its incidence is increasing worldwide. The overall survival of patients with HCC is dismal, and currently, no efficient secondary prevention or systemic treatments are available. HCC usually develops in the presence of continuous inflammation and hepatocyte regeneration in the setting of chronic hepatitis and cirrhosis (1). Increased cellular turnover and regeneration within the context of a noxious chronically inflamed environment cause accumulation of chromosomal damages, which eventually affect the structure and expression of oncogenes and tumor suppressor genes leading to carcinogenesis. Recent studies have shown that mediators of inflammation, such as prostaglandins (PG), play an important role in hepatocarcinogenesis (2). For example, increased cyclooxygenase-2 (COX-2) expression has been found in human and animal HCCs and in dysplastic hepatocytes (3)(4)(5)(6)(7)(8)(9). Elevated levels of PGs, most notably PGE 2 , have also been detected in HCC (10). Overexpression of COX-2 or treatment with exogenous PGE 2 increases human HCC cell growth and invasiveness (8,11). The COX inhibitors, nonsteroidal anti-inflammatory drugs, inhibit the proliferation and induce apoptosis in cultured HCC cells and in animal models of hepatocarcinogenesis (2), although these inhibitors are known to mediate effects through both COX-dependent and COX-independent mechanisms.
In addition to upregulation of COX-2, Wnt/β-catenin activation has also been implicated in various stages of hepatic tumorigenesis, including the dysplastic foci, hepatic adenoma, hepatoblastoma, and HCC (12)(13)(14)(15)(16). Activation of the Wnt/β-catenin pathway occurs in approximately 30% to 40% of HCCs (17). Multiple mechanisms of β-catenin activation or stabilization have been reported in hepatic tumorigenesis, including mutations in the β-catenin gene (Ctnnb1), or components of its degradation machinery such as Axin and GSKβ inactivation (12)(13)(14)(15)(16). In mice, hepatic deletion of APC, another degradation component of β-catenin, leads to HCC (18). Recently, upregulation of a member of Wnt receptors, Frizzled-7, has been shown as another possible mechanism of β-catenin activation in HCC (19). In addition, Wnt/β-catenin also plays an important role in regulation of hepatocyte proliferation, survival, liver regeneration, and in the maintenance and self-renewal of pluripotent stem cells and progenitor cells (12); hence, they may play a role in the maintenance of the cancer stem cell compartment. Indeed, β-catenin activation has been identified in oval cells (liver stem cells), which might be precursors of a subset of HCC (13). Thus, there seems to be multiple mechanisms of β-catenin activation leading to liver neoplasia. Although PGE 2 has recently been shown to activate βcatenin in colon cancer cells (20,21), it remains unknown whether the COX-2/PG and Wnt/β-catenin signaling pathways converge during hepatocarcinogenesis.
In contrast to the documented carcinogenic effect of the PGs (PGE 2 in particular) derived from arachidonic acid [AA; an ω-6 polyunsaturated fatty acid (PUFA)], there is abundant experimental evidence that the ω-3 PUFAs rich in fish oil, such as docosahexaenoic acid (DHA) and eicosapentaenic acid (EPA), prevent carcinogenesis (22,23). However, the molecular mechanisms for the anticancer actions of ω-3 PUFAs remain incompletely understood. This study was designed to investigate the effect and mechanism of ω-3 PUFAs in HCC cells. Our results show that DHA and EPA inhibited the growth of three human HCC cells (Hep3B, Huh-7, HepG2), in vitro. The growth of HCC in vivo was also significantly reduced when mouse HCC cells (Hepa1-6) were inoculated into the syngeneic Fat-1 transgenic mice that carry a Caenorhabditis elegans desaturase converting ω-6 to ω-3 PUFAs. Moreover, our data reveal that COX-2-derived PGE 2 activates β-catenin signaling pathways in human HCC cells and that ω-3 PUFAs inhibit HCC growth by simultaneously blocking β-catenin and COX-2 signaling pathways. These findings provide important preclinical evidence and molecular framework for utilization of ω-3 PUFAs in the chemoprevention and treatment of HCC.

Materials and Methods
Materials α-MEM, DMEM, RPMI 1640, fetal bovine serum (FBS), glutamine, antibiotics, and Lipofectamine plus reagent were purchased from Life Technologies, Inc. PGE 2 was purchased from Calbiochem. The cell proliferation assay reagent WST-1 was purchased from Roche Molecular Biochemicals. The antibody for human COX-2, 15-hydroxyprostaglandin dehydrogenase (PGDH) were purchased from Cayman Chemical Company. The antibodies against human Axin, β-catenin, poly ADP ribose polymerase (PARP), caspase-3, caspase-9, and c-Met were purchased from Santa Cruz Biotechnology. The antihuman β-actin monoclonal antibody was purchased from Sigma. The horseradish peroxidaselinked streptavidin and chemiluminescence detection reagents were from Amersham Pharmacia Biotech, Inc. The rabbit antibodies for phospho-Akt (Thr308), Akt, phospho-GSK-3β (Ser9), and GSK-3β were purchased from Cell Signaling Technology. Mouse monoclonal anti-GSK-3β was purchased from Transduction Laboratories, and cytochrome c was purchased from BD Bioscience. The Bio-Rad protein assay system was obtained from Bio-Rad Laboratories. The Tris-glycine gels were obtained from Invitrogen Life Technologies, Inc. Dr. T. Hla at the University of Connecticut Health Center, Farmington, CT, provided the COX-2 expression plasmid (containing full length of human COX-2 cDNA in sense orientation cloned in mammalian expression vector PCDNA3).
Cell Culture The human HCC cell lines (Hep3B, HepG2, and Huh7) were obtained from American Type Culture Collection and cultured according to our previous described methods (8,11,24). Briefly, the cells were cultured in EMEM supplemented with 10%FBS, 2 mmol/L L-glutamine, and penicillin/ streptomycin. The cells were incubated at 37°C in a humidified CO 2 incubator. The experiments were done when cells reached ∼80% confluence and conducted in serumfree medium (with serum deprivation for 24 h before each experiment).
Cell Growth Assay Cell growth was determined using the cell proliferation reagent WST-1, a tetrazolium salt that is cleaved by mitochondrial dehydrogenases in viable cells. Briefly, 100 μL of cell suspension (containing 0.5-2 × 10 4 cells) were plated in each well of 96-well plates. After 24 h of culture to allow reattachment, the cells then were treated with specific reagents such as DHA, EPA, or Wnt3a-conditioned medium for indicated time points. At the end of each treatment, the cell proliferation reagent WST-1 (10 μL) was added to each well, and the cells were incubated at 37°C for 0.5 to 5 h. Absorbance at 450 nm was measured using an automatic ELISA plate reader.
Immunoprecipitations Equal amount of cellular protein from the treated cells was incubated with 10 μL of rabbit antihuman Axin polyclonal antibody at 4°C for overnight, followed by addition of 20 μL Protein A/G PLUS agarose (Santa Cruz Biotechnology). The mixture was incubated for 2 h and then washed thrice with the cell lysis buffer [50 mmol/L HEPES (pH 7.55), 1 mmol/L EDTA, 1 mmol/L DTT, and protease inhibitor cocktail tablets from Roche Diagnostics]. The final pellets were dissolved in 20 μL 2× protein loading buffer, and the samples were subjected to SDS-PAGE and Western blot analysis using 1:1,000 dilution mouse anti-human GSK-3β or β-catenin monoclonal antibodies and enhanced chemiluminescence Western blot detection system (Amersham Pharmacia Biotech, Inc.). Luciferase Reporter Activity Assay The cultured cells were seeded at a concentration achieving 80% confluence in 12-well plates for 18 h before transfection. The cells were transiently transfected with 0.2 μg per well translucent TCF/LEF-Luc reporter vector, which was designed to measure the β-catenin transcriptional activity of TCF/LEF responsive genes. After transfection, the cells were treated with specific reagents including DHA or Wnt3a conditioned medium in serum-free medium for 24 h. The cell lysates were then obtained with 1× reporter lysis buffer (Promega). The luciferase activity was assayed in a Berthold AutoLumat LB953 Luminometer by using the luciferase assay system from Promega. The relative luciferase activity was calculated after normalization of cellular proteins. All values are expressed as percentage of activity induction relative to control activity.

Immunoblotting
At the end of each indicated treatment, the cells were scraped off the plates and centrifuged, washed twice with cold PBS containing 0.5 mmol/L phenylmethylsulfonyl fluoride and 10 μg/mL leupeptin, and resuspended in 5fold volume of hypotonic buffer consisting of 50 mmol/L HEPES (pH 7.55), 1 mmol/L EDTA, 1 mmol/L DTT, and protease inhibitor cocktail tablets (Roche Diagnostics GmbH). After sonication, the whole-cell lysate was collected by centrifugation at the speed of 15,000 g at 4°C for 10 min to remove cell debris and stored in aliquots at −80°C until use. The protein concentration in the cell extracts were determined by the Bio-Rad protein assay (Bio-Rad). Thirty micrograms of cellular protein were subjected to SDS-PAGE on 4% to 20% Tris-glycine gels for PARP, β-catenin, GSK-3β, phosphor-GSK-3β, Akt, phosphor-Akt, c-Met, Axin cytochrome c, caspase-3, caspase-9, COX-2, 15-PGDH, or β-actin. The separated proteins were electrophoretically transferred onto the nitrocellulose membrane (Bio-Rad). Nonspecific binding was blocked with 0.5% Tween 20 in PBS (PBS-T) containing 5% nonfat milk for 1 h at room temperature. The membranes were then incubated overnight at 4°C with individual primary antibodies in PBS-T containing 1% nonfat milk at the dilutions specified by the manufactures. Following three washes with PBS-T, the membranes were then incubated with the horseradish peroxidaseconjugated secondary antibodies at 1:10,000 dilution in PBS-T containing 5% nonfat milk for 1 h at room temperature. The membranes were then washed thrice with PBS-T and the protein bands were visualized with the enhanced chemiluminescence Western blotting detection system.

Animals and Tumorigenicity Experiments
The Fat-1 transgenic mice (from Dr. J.X. Kang of Harvard University, Boston, MA; ref. 25). The animals were kept at 22°C under a 12-h light/dark cycle and fed standard mouse chow (Prolab IsoPro 5P75 RMH 3000) with water ad libitum. The fatty acid compositions in the Fat-1 transgenic mice and wild-type control mice housed at our animal facility are shown in Supplementary Table S1. The mice were kept under specific pathogen-free conditions in standard cages, and used 8-to 10-wk-old male for this experiment. Each mouse was injected s.c. into the area overlying the right flank with 1.5 × 10 6 mouse HCC cells (Hepa1-6) suspended in 100 μL of serum-free medium. After inoculation, the animals were closely monitored for the development of s.c. tumor. The tumor size was measured with a caliper every 2 d. Upon sacrifice, the tumor volume was calculated according to the following formula: tumor volume = L × W2 × 0.5. The animal experiments were carried out according to the protocol approved by the University of Pittsburgh Institutional Animal Care and Use Committee (#0201740B).

Results
Immunohistochemical Stains for COX-2 and β-Catenin in Human HCC Tissues Twenty paired human HCCs and their matched nonneoplastic/nondysplastic liver tissues were analyzed by immu-nohistochemistry for the expression of COX-2 and βcatenin. Increased cytoplasmic staining for COX-2 and nuclear staining for β-catenin was observed in HCC cells when compared with the nontumor liver tissue ( Supplementary  Fig. S1). The average staining intensity for COX-2 in HCC is 2.10 ± 0.78, which is significantly higher than that in nontumor liver tissue (0.20 ± 0.09; P < 0.01, Student's t test). Whereas COX-2 is expressed exclusively in the cytoplasm of both HCC cells and to a less degree in hepatocytes, the expression pattern of β-catenin between HCC cells and nonneoplastic hepatocytes are distinctly different. In nontumorous hepatocytes, β-catenin is expressed exclusively in the plasma membrane with no significant cytoplasmic staining and absence of nuclear staining in all 20 patients. In contrast, in HCC cells, nuclear staining for β-catenin was observed in 5 of 20 patients (25%), with focal cytoplasmic staining and decreased membrane staining, indicating βcatenin nuclear translocation and activation. Thus, COX-2 and β-catenin signaling pathways are active in a significant percentage of human HCCs.
ω-3 PUFAs Induce HCC Cell Apoptosis Human HCC cell lines were examined for their response to DHA, EPA, and AA treatment. As shown in Fig. 1A, treatment of Hep3B cells with two ω-3 PUFAs (30 μmol/L), DHA and EPA, induced a time-dependent reduction of cell viability; in contrast AA, a ω-6 PUFA, had no significant effect. Treatment with 30 μmol/L EPA for 12, 24, 48, and 72 hours induced approximately 45%, 60%, 70%, and 75% reduction of viable cells, respectively. DHA seems to have more effect, with ∼75% reduction of viable cells at 12 hours and >90% reduction at 24, 48 and 72 hours. The cells treated with DHA and EPA show morphologic features of cell death, characterized by shrinkage, roundness, and detachment. In contrast, AA treatment did not significantly alter the cell morphology. The effect of DHA and EPA is dose dependent in all three human HCC cell lines (Hep3B, Huh7, and HepG2; Fig. 1B). The observations that DHA induced the cleavage of PARP, caspase-3, and caspase-9, with concomitant release of cytochrome c from mitochondria to cytosol confirm the induction of apoptosis (Fig. 1C). Taken together, these results document induction of apoptosis by ω-3 PUFAs in HCC cells. We have also tested the effect of DHA and EPA in primary cultures of liver parenchymal cells and these compounds were found to have no cytotoxic effect in primary cells. 4

DHA Decreases the Level of β-Catenin and Inhibits TCF/LEF Transcription Activity in HCC Cells
Further experiments were done to assess the mechanisms by which ω-3 PUFAs induce HCC apoptosis. Because β-catenin activity is importantly involved in hepatocarcinogenesis, the potential effect of ω-3 PUFAs on β-catenin protein level and activity was examined. As shown in Fig. 2A and B, treatment with DHA or EPA reduced the level of β-catenin protein; this effect was time-dependent (observed 1-6 hours after treatment). As c-Met is a β-catenin controlled downstream gene, the potential effect of DHA and EPA on c-Met protein expression was also examined. Indeed, DHA and EPA treatment also reduced the expression of c-Met. In contrast, treatment with AA did not alter β-catenin or c-Met level (Fig. 2C).
Because β-catenin regulates gene expression via binding as a transcription factor in complex with the TCF/LEF transcription factor family to the promoter region of target genes, we further examined the effect of DHA on TCF/ LEF reporter activity. The TCF/LEF transcription activity was assayed after transient transfection of a luciferase reporter construct under the control of TCF/LEF response element. As shown in Fig. 2D, DHA treatment significantly inhibited the TCF/LEF reporter activity. This result further confirms suppression of β-catenin activity by DHA.
ω-3 PUFAs Induce β-Catenin Degradation through Inhibition of GSK-3β Phosphorylation The level of β-catenin in cells is tightly controlled by its degradation complex composed of Axin, APC, GSK-3β, and β-catenin, in which GSK-3β phosphorylates β-catenin and thus triggers its ubiquitination and subsequent proteosomal degradation. The activity of GSK-3β is regulated by its phosphorylation status, with GSK-3β phosphorylation at Ser-9 being functionally inactive. To determine whether ω-3 PUFAs might induce β-catenin degradation through inhibition of GSK-3β phosphorylation, we examined the phospho-Ser-9-GSK-3β and total GSK-3β protein levels in Hep3B cells treated with PUFAs. As shown in Fig. 3A, DHA treatment reduced GSK-3β phosphorylation, whereas it had no effect on the protein level of total GSK-3β. Similarly, Figure 3. ω-3 PUFAs inhibit GSK-3β phosphorylation in HCC cells. A, Hep3B cells were treated with 60 μmol/L of DHA, EPA, or AA in serum-free medium for indicated time periods. The cell lysates were obtained for Western blot analysis using antibodies against p-GSK-3β and GSK-3β as well as p-Akt as described in Materials and Methods. DHA or EPA treatment resulted in p-GSK-3β dephosphorylation, whereas AA had no effect. B, inhibition of GSK-3β by LiCl prevents DHA-induced βcatenin degradation. Top, Hep3B cells were pretreated with LiCl (20 mmol/L) for 1 h before DHA treatment (20 μmol/L) for 24 h; the cell lysates were obtained for Western blot analysis using the antibody against β-catenin. Bottom, inhibition of GSK-3β by LiCl prevents DHAmediated reduction of TCF/LEF reporter activity. Hep3B cells transfected with the TCF/ LEF reporter construct were treated with LiCl (20 mmol/L) for 1 h followed by DHA treatment (20 μmol/L) for 24 h. The cell lysates were obtained to determine the luciferase activity. Columns, mean of six independent experiments (*, P < 0.01 compared with control; **, P < 0.01 compared with DHA treatment); bars, SD.
C, LiCl prevents DHA-induced PARP cleavage and cell viability. Top, Hep3B cells were pretreated with LiCl (20 mmol/L) for 1 h followed by DHA treatment (20 μmol/L) for 24 h. The cell lysates were obtained for Western blot analysis using the antibody against PARP. Bottom, LiCl restores DHA-induced cell death. Hep3B cells pretreated with LiCl (20 mmol/L) for 1 h were incubated with DHA for 24 h. The cell growth was determined using WST-1 assay. Columns, mean of six independent experiments (*, P < 0.01 compared with control; **, P < 0.01 compared with DHA treatment); bars, SD. EPA was not used for LiCl experiments.

Inhibition of HCC by ω-3 PUFAs
Mol Cancer Ther 2009;8 (11). November 2009 EPA treatment also decreased the level of phosphor-GSK-3β, whereas AA had no effect. Because the phosphorylation of GSK-3β is controlled by Akt, we also examined the potential effect of DHA on Akt phosphorylation. Our data showed that DHA had no effect on Akt phosphorylation (Fig. 3A). Thus, DHA most likely inhibited GSK-3β phosphorylation through mechanism independent of Akt. Taken together, the above findings provide evidence for GSK-3β dephosphorylation (activation) in ω-3 PUFA-induced degradation of β-catenin in HCC cells.

Catenin Degradation and Cell Death
To further determine the role of GSK-3β in DHA-induced β-catenin degradation, Hep3B cells were pretreated for 1 hour with LiCl before DHA treatment to determine the level of β-catenin protein, TCF/LEF reporter activity, and cell growth. As shown in Fig. 3B, inhibition of GSK-3β by LiCl prevented DHA-induced reduction of β-catenin protein and TCF/LEF reporter activity. Accordingly, LiCl pretreatment also prevented DHA-induced PARP cleavage and restored DHA-induced cell death (Fig. 3C). These findings further support the role of GSK-3β activation (dephosphorylation) in DHA-induced β-catenin degradation in HCC cells.

DHA Induces the Association of Axin with GSK-3β and β-Catenin in HCC Cells
The degradation of β-catenin strictly depends upon β-catenin phosphorylation, which occurs in a multiprotein complex containing Axin and GSK-3β and β-catenin. It is believed that in this complex assembled by Axin, GSK-3β phosphorylates the β-catenin primarily when it is bound to Axin. To determine whether DHA alters the assembly of the Axin/GSK-3β/β-catenin complex, immunoprecipitation and Western blot experiments were done to detect the Axin/GSK-3β/β-catenin binding complex. As shown in Fig. 4A to C, treatment of Hep3B cells with DHA induced the association of Axin with GSK-3β as well as β-catenin.
This effect was observed within 1 hour and persisted at 5 hours. In contrast, AA did not affect the association between Axin and GSK-3β (Fig. 4D). These findings indicate that DHA induces the association of Axin with GSK-3β and βcatenin, thereby facilitating the formation of β-catenin destruction complex. Taken together, our data suggest that ω-3 PUFAs induce β-catenin degradation through dephosphorylation of GSK-3β and formation of β-catenin destruction complex in HCC cells.
Activation of β-Catenin by Wnt3a Partially Protects DHA-Induced HCC Cell Death Because Wnt3a is known to activate β-catenin signaling in cells, further experiments were carried out to determine whether Wnt3a might protect HCC cells from DHAinduced apoptosis. Indeed, treatment of Hep3B cells with Wnt3a conditioned medium partially prevented DHAinduced cell death ( Supplementary Fig. S2A). The effect of Wnt3a on β-catenin activation was confirmed by the observation that Wnt3a conditioned medium prevented DHA-induced reduction of TCF/LEF transcription activity ( Supplementary Fig. S2B). These results further show that DHA inhibits HCC growth at least in part through downregulation of Wnt/β-catenin signaling pathway.

DHA Inhibits COX-2 Expression in HCC Cells
We next examined whether DHA might also affect the expression of COX-2 in HCC cells. As shown in Supplementary Fig. S3, DHA significantly inhibited the COX-2 promoter activity and COX-2 protein expression in HCC cells. These findings suggest that DHA inhibits the expression of COX-2 through suppression of gene transcription.
DHA Induces 15-PGDH Expression in HCC Cells 15-PGDH catalyzes the rate-limiting step of PG catabolism and thus represents a physiologic antagonist of COX-2 (26,27). Recent emerging evidence suggests that elevated PGE 2 in cancers may be the result of enhanced COX-2-mediated PGE 2 synthesis as well as reduced 15-PGDH-mediated degradation of PGE 2 . Therefore, we sought to further determine whether DHA might affect 15-PGDH expression in HCC cells. As shown in Supplementary Fig. S4, DHA treatment enhanced the expression of 15-PGDH in a dosedependent manner in HCC cells (Hep3B, HepG2, and Huh7). These data are consistent with the observation that DHA and EPA inhibit PGE 2 production in Hep3B cells (Supplementary Fig. S5).
DHA Prevents PGE 2 Induced β-Catenin Activation Because DHA reduces PGE 2 level through concomitant inhibition of COX-2 and induction of 15-PGDH, we postulate that DHA might also inhibit β-catenin through inhibition of PGE 2 . To evaluate this hypothesis, further experiments were done to examine the direct effect of PGE 2 on β-catenin activation and to determine whether DHA might prevent PGE 2 effect. As shown in Fig. 5, PGE 2 treatment resulted in dissociation of Axin from GSK-3β and enhanced TCF/LEF reporter activity in Hep3B cells; these effects were significantly blocked by cotreatment with DHA. These findings suggest that suppression of PGE 2 by DHA represents another mechanism for β-catenin degradation.
ω-3 PUFAs Prevent HCC Growth In vivo After the in vitro effect of ω-3 PUFAs on HCC cell growth was documented, further experiments were carried out to evaluate the effect of ω-3 PUFAs on HCC growth in vivo. We implanted murine HCC cells (Hepa1-6) into the syngeneic Fat-1 transgenic and control mice (with C57BL/6 genetic background) and examined the growth of the inoculated tumor cells in these animals. The Fat-1 transgenic mice carry a Caenorhabditis elegans desaturase gene that adds a double bond into a saturated fatty acid hydrocarbon chain and converts ω-6 to ω-3 PUFAs, resulting in a significant increase in ω-3 PUFAs and reduction in ω-6 PUFAs in all the organs and tissues (25). The Hepa1-6 cell line was chosen because it was derived from HCC of C57BL/6 strain and can be grown to form tumors in mice with C57BL/6 genetic background. As shown in Fig. 6, there is a marked difference in the tumor size and tumor volume between Fat-1 transgenic (n = 10) and wild-type mice (n = 12). Over an observation period of 14 days, all wild-type mice developed a palpable tumor by day 4, whereas only 5 of 10 Fat-1 transgenic mice developed a minor tumor palpable by day 4 and the mass of all palpable tumor almost disappeared at day 12. Mice with homozygous mutation for the PG receptor EP 1 (in C57BL/6 background) was used as an additional control, which showed a similar degree of tumor growth as the wild-type mice. These findings show that ω-3 PUFAs inhibit HCC growth, in vivo.
The Effect of ω-3 PUFAs on Hepa1-6 Cell Growth

In vitro
Given the marked reduction of Hepa1-6 cell growth in the Fat-1 transgenic mice, we conducted subsequent experiments to evaluate ω-3 PUFA actions in Hepa1-6 cell growth, in vitro. Both DHA and EPA significantly reduced the viability of cultured Hepa1-6 cells (Supplementary Fig. S6). Treatment of Hepa1-6 with DHA led to reduction of β-catenin protein as well as TCF/LEF reporter activity (Supplementary Fig. S7). In parallel, DHA also inhibited the expression of COX-2 and induced the expression of 15-PGDH in dissociation. Hep3B cells were treated with PGE 2 (10 μmol/L) for different time points in the serum-free medium. The cell lysates were obtained for immunoprecipitation with anti-Axin antibody followed by immunoblotting with antibodies against GSK-3β and Axin. C, DHA prevents PGE 2 -induced Axin/GSK-3β dissociation. Hep3B cells were treated with PGE 2 (10 μmol/L) or vehicle in the absence or presence of DHA (60 μmol/L) in serum-free medium for 1 h. The cell lysates were obtained for immunoprecipitation with anti-Axin antibody followed by immunoblotting with antibodies against GSK-3β and Axin. D, DHA inhibits PGE 2 -induced TCF/LEF reporter activity. Hep3B cells were transiently transfected with TCF/LEF-Luc reporter vector. After transfection, the cells were treated with PGE 2 with or without DHA in the serum-free medium for 24 h. The cell lysates were obtained to determine the luciferase activity. Columns, mean of six independent experiments; bars, SD. DHA treatment significantly decreased TCF/LEF reporter activity (*, P < 0.01 compared with control; **, P < 0.01 compared with PGE 2 treatment). Fig. S8A). Consistent with the latter observations, DHA treatment also inhibited the production of PGE 2 in Hepa1-6 cells ( Supplementary  Fig. S8B). Therefore, the effects of ω-3 PUFAs in Hepa1-6 cells are similar to those in human HCC cells. These findings show that ω-3 PUFAs inhibit β-catenin and COX-2 signaling in both murine and human HCC cells.

Discussion
Both the COX-2/PGE 2 and Wnt/β-catenin signaling pathways are active in human HCCs. There is constitutively high expression and activation of COX-2 in human liver cancers and precancerous inflammatory liver diseases; COX-2 activation enhance the production of PGs from AA that subsequently promote hepatic inflammation and neoplasia (2). In parallel, the Wnt/β-catenin pathway is also activated in various stages of hepatic tumorigenesis (13)(14)(15)(16)(28)(29)(30)(31). Therefore, we postulate that therapies aimed at simultaneous disruption of the COX-2/PGE 2 and Wnt/βcatenin pathways may produce effective chemopreventive and antitumorigenic effects. This study provides important experimental evidence and mechanisms for inhibition of both Wnt/β-catenin and COX-2/PGE 2 signaling pathways by ω-3 PUFAs in HCC (Supplementary Fig. S9).
Compelling epidemiologic and experimental studies have indicated a relationship between PUFAs and the risk of cancer. For example, a high dietary intake of ω-6 PUFAs, such as linoleic acid (18:2ω-6), is associated with a high risk for colon cancer, whereas high intake of ω-3 PUFAs from fish oils, such as DHA (22:6ω-3) and EPA (20:5ω-3), decreases it (22,23). Experimental data have shown that the ω-6 fatty acids stimulate carcinogenesis, tumor growth, and metastasis, whereas the ω-3 fatty acids exert suppressive effects. In this study, we used both in vitro and in vivo models to evaluate the effect of ω-3 PUFAs on HCC growth. Treatment with DHA and EPA induced a dose-and time-dependent growth inhibition and apoptosis in three human HCC cell lines. The induction of apoptosis is confirmed by cleavages of PARP, caspase-3, and caspase-9 and release of cytochrome c. To evaluate the antitumor effect of ω-3 PUFAs on HCC growth in vivo, we implanted murine HCC cells (Hepa1-6) into the syngeneic Fat-1 transgenic and control mice. The Fat-1 transgenic mice ubiquitously express a Caenorhabditis elegans desaturase, leading to significant increase in ω-3 PUFAs and reduction in ω-6 PUFAs in all the organs and tissues (25,32). This model was selected because it provides a balanced ratio of ω-6 to ω-3 fatty acids in mouse tissues and eliminates the potential dietary variation associated with Figure 6. Tumorigenicity of Hepa1-6 mouse HCC cells in Fat-1 transgenic and wild-type mice. A and B, tumor size and tumor volume in the Fat-1 transgenic and wild-type control mice. Hepa1-6 murine HCC cells were obtained from the American Type Culture Collection and cultured in DMEM containing 4 mmol/L L-glutamine and 10% FBS. Cells (1.5 × 10 6 ) were injected s.c. into the right flank of each of Fat-1 transgenic and wildtype mice (8-wk-old male). Mice with homozygous mutation for the PG receptor EP 1 (in C57BL/6 background) were used as an additional control. Tumor size was monitored at the indicated time with a caliper. Tumor volume was calculated on the basis of the following formula: Tumor volume = (length × width × width)/2. Points, mean of tumor size or volume of wild-type group (n = 12), EP 1 knockout (n = 10), or Fat-1 transgenic (n = 10) mice; bars, SD. C, representative photographs showing tumor formation at three different time points after cell implantation.
long-term feeding of PUFAs. A significant reduction of HCC tumor size and tumor volume was observed in the Fat-1 transgenic mice. These findings provide important in vivo evidence for inhibition of HCC by ω-3 PUFAs. The effect of dietary DHA and EPA on hepatocellular cancer growth remains to be further evaluated.
A prominent mechanism for the chemopreventive action of ω-3 PUFAs is their suppressive effect on the production of AA-derived prostanoids, particularly PGE 2 (23,33). This is important because PGE 2 is implicated in multistages of tumorigenesis, including modulation of inflammation, cancer cell proliferation, differentiation, apoptosis, angiogenesis, metastasis, and host immune response to cancer cells (2). Our data in this study show that ω-3 PUFAs inhibit COX-2 expression in HCC cells, which is consistent with recently reported downregulation of COX-2 by ω-3 PUFAs in colon cancer cells (34). Moreover, our findings provide novel evidence for induction of 15-PGDH, a rate-limiting key enzyme in PG catabolism, by ω-3 PUFAs in human cancer cells. The latter observation is noteworthy, because 15-PGDH is a PG-degrading enzyme that physiologically antagonizes COX-2 and suppresses tumor growth.
In addition to modulation of COX-2 and 15-PGDH by ω-3 PUFAs, our results reveal that degradation of β-catenin is a novel parallel mechanism for ω-3 PUFA-mediated antitumor effect. β-Catenin is a key molecule in the canonical Wnt pathway that regulates multiple biological functions, including embryogenesis and tumorigenesis (35)(36)(37)(38). In the absence of Wnt ligands, cytoplasmic β-catenin associates in a complex with GSK-3β, Axin, and APC, where it is phosphorylated and targeted for proteosomal degradation. Activation of Wnt signaling causes dissociation of the β-catenin degradation complex, leading to β-catenin accumulation in the cytoplasm and translocation into the cell nucleus. In the nucleus, β-catenin binds the transcription factor TCF/LEF that induce transcription of important downstream target genes implicated in cell proliferation, differentiation, and apoptosis (35)(36)(37)(38). Recent evidence has shown that PGE 2 induces the cytoplasmic and nuclear accumulation of β-catenin in human colon cancer cells. Castellone and colleagues (20) reported that PGE 2 activates its G protein-coupled receptor, EP 2 , resulting in direct association of the G protein α subunit with the regulator of G protein signaling domain of axin; this causes release of GSK-3β from its complex with axin, thus leading to β-catenin accumulation. A separate study by Shao et al. (21) showed the involvement of cyclic AMP/protein kinase A pathway in PGE 2 -induced β-catenin accumulation in colon cancer cells. The current study provides evidence that PGE 2 induces dissociation of GSK-3β from Axin, thus preventing β-catenin reduction in HCC cells.
Our data suggest that ω-3 PUFAs induce β-catenin degradation through three interrelated mechanisms. First, we show that DHA and EPA induce a rapid dephosphorylation of GSK-3β in HCC cells, suggesting that GSK-3β activation is involved in ω-3 PUFA-induced β-catenin degradation. This assertion is further supported by the observations that the GSK-3β inhibitor, LiCl, prevents DHA-induced reduc-tion of β-catenin protein and transcription activity and restored DHA-induced cell death. Second, DHA treatment induces the association of Axin with GSK-3β forming βcatenin destruction complex. Third, ω-3 PUFAs suppress PGE 2 signaling through concomitant inhibition of COX-2 and induction of 15-PGDH, thus preventing PGE 2 -induced β-catenin accumulation. The involvement of β-catenin degradation in ω-3 PUFA-induced inhibition of tumor growth is further supported by the observation that Wnt3a conditioned medium partially protects HCC cells from DHA-induced apoptosis.
In summary, this study provides encouraging preclinical evidence and important mechanism for utilization of ω-3 PUFAs in the chemoprevention and treatment of HCC, although the data should be interpreted with caution because the concentration of EPA and DHA used in the cultured cells is relatively high and most likely will not be achieved in vivo. Our findings have significant clinical implications, given that HCC is a common and highly malignant human cancer. It is conceivable that ω-3 PUFAs, either applied alone or in conjunction with current modalities, may represent an effective, nontoxic, and safe chemopreventive and therapeutic agent for patients with HCC or at high risks for development of this devastating tumor.

Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.