Retinoic acid receptors (RAR; α, β, and γ), members of the nuclear receptor superfamily, mediate the pleiotropic effects of the vitamin A metabolite retinoic acid (RA) and derivatives (retinoids) in normal and cancer cells. Abnormal expression and function of RARs are often involved in the growth and development of cancer. However, the underlying molecular mechanisms remain largely elusive. Here, we report that levels of RARγ were significantly elevated in tumor tissues from a majority of human hepatocellular carcinoma (HCC) and in HCC cell lines. Overexpression of RARγ promoted colony formation by HCC cells in vitro and the growth of HCC xenografts in animals. In HepG2 cells, transfection of RARγ enhanced, whereas downregulation of RARγ expression by siRNA approach impaired, the effect of RA on inducing the expression of α-fetoprotein, a protein marker of hepatocarcinogenesis. In studying the possible mechanism by which overexpression of RARγ contributed to liver cancer cell growth and transformation, we observed that RARγ resided mainly in the cytoplasm of HCC cells, interacting with the p85α regulatory subunit of phosphatidylinositol 3-kinase (PI3K). The interaction between RARγ and p85α resulted in activation of Akt and NF-κB, critical regulators of the growth and survival of cancer cells. Together, our results show that overexpression of RARγ plays a role in the growth of HCC cells through nongenomic activation of the PI3K/Akt and NF-κB signaling pathways. Cancer Res; 70(6); 2285–95
- retinoid receptor
- nongenomic action
- nuclear factor κB
- hepatocellular carcinoma
Retinoic acid receptors (RAR) and retinoid X receptors (RXR) are members of the nuclear receptor superfamily, which mediate the pleiotropic effects of vitamin A metabolite retinoic acid (RA) and its natural and synthetic derivatives (retinoids; refs. 1, 2). Both RAR and RXR are encoded by three distinct genes, α, β, and γ, and they regulate a variety of important cellular processes in normal and cancer cells (1, 2). The distinct spatial and temporal expression patterns of retinoid receptors during development and in adult tissues suggest that each of the subtypes has discrete functions (1, 2). Altered expression and function of retinoid receptors are associated with the development of certain malignancies (3–5). Integration of HBV gene at FRA3A within the RARβ gene (6, 7), altered phosphorylation of RXRα (8), and limited proteolysis of the RXRα protein (9) have been implicated in the development and progression of hepatocellular carcinoma (HCC).
RARs, like other nuclear receptors, are known to act as transcriptional factors to regulate gene expression by binding to cis-acting response elements (retinoic acid response elements, RARE) in the promoter/enhancer region of target genes as heterodimer with RXRs (1, 10). However, only a limited number of their target genes are described (11). A considerable number of studies have now shown that RARs can nongenomically regulate some rapid biological responses to retinoids (11, 12). Thus, a membrane-associated RARα can mediate the effect of RA on spine formation of hippocampal neurons through phosphorylation of extracellular signal-regulated kinase 1/2 and p70s6k (13), whereas RARγ can bind to and catalytically activate c-Src kinase during neuritogenesis (14). RARs can also interact with the p85α regulatory subunit of phosphoinositide 3-kinase (PI3K), leading to its activation in various cell types (15–17). Consistently, retinoid receptors are found to reside in the cytoplasm at certain stages during development (18, 19) and in response to differentiation (14, 20), apoptosis (21), and inflammation (22–26). Altered subcellular localization of retinoid receptors likely reflects changes in the cellular environment but may also play an active role in the regulation of cellular activities. In a recent study, we showed that RARγ is unique among three RAR subtypes in that it often resides in the cytoplasm of cancer cells (9, 27). Whether and how the cytoplasmic RARγ regulates the growth of cancer cells remained undefined.
In the present study, we showed that RARγ was overexpressed in a majority of tumor tissues from HCC patients and in HCC cell lines. By transfection and siRNA approaches, we showed that RARγ played a critical role in promoting colony formation of HCC cells in vitro and the growth of HCC xenografts in animals and in mediating the effect of RA on inducing the expression of α-fetoprotein (AFP), a protein marker of hepatocarcinogenesis (28, 29). We also found that RARγ predominantly resided in the cytoplasm where it interacted with p85α, leading to activation of Akt and NF-κB in HCC cells. Together, our results show that RARγ overexpression contributes to the growth and development of HCC through its interaction with p85α, which activates both Akt and NF-κB pathways.
Materials and Methods
Lipofectamine 2000 and Trizol LS (Invitrogen); enhanced chemilumienescence (ECL) reagents, goat anti-rabbit and anti-mouse secondary antibodies conjugated to horseradish peroxidase (Thermo); polyclonal antibodies against RARγ, RARα (C-20), RARβ (C-19), IκBα, Akt1/2/3, pAkt1/2/3, p65, OctA-probe (Flag, D-8), and proliferating cell nuclear antigen (PCNA; FL-261) and monoclonal antibodies against pIκBα, Myc (9E10), Hsp60 (H-1), p85α (B-9), AFP (1B10), and FITC-labeled antirabbit IgG (Santa Cruz Biotechnology); monoclonal anti-poly(ADP-ribose) polymerase (PARP) antibody (556494; BD Biosciences); antimouse IgG conjugated with Cy3 (Chemicon international); monoclonal antibodies against β-actin, α-tubulin, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and chemicals including all-trans-RA, cyclophosphamide, aflatoxin B1, tumor necrosis factor α (TNFα), MG132, LY-294002, and BMS-345541 (Sigma); protein A beads (GE Healthcare); polyvinylidene difluoride (PVDF) membranes (Millipore); and cocktail of proteinase inhibitors (80-6501-23; Amersham) were used in this study.
Seventeen Chinese patients with primary HCC without prior chemotherapy were recruited. Tumor and the adjacent nontumorous tissues were collected after surgical resection.
Construction of lentivirus-based siRNA expression vector
RARγ siRNA oligonucleotide (5′-gctaccaagtgcatcatca-3′) or nonsense siRNA oligonucleotide (5′-ttctccgaacgtgtcacgt-3′) with hairpin-containing sequence created as described (30) was ligated into pLL3.7 vector via HpaI/XhoI site. Lentivirus packaging and infection were performed as described (30).
HepG2, QGY-7703, and QSG-7701 cells (31) and stable clones expressing Myc-RARγ (HepG2/γ and QSG-7701/γ), empty vector (HepG2/v and QSG-7701/v), RARγ siRNA lentiviral vector (HepG2/γi and QGY-7703/γi), and control siRNA (HepG2/ci and QGY-7703/ci) were used. Cells cultured in six-well plates for 14 d were fixed and stained with Giemsa. The number of foci containing >50 cells was scored.
Nude mice (BALB/c, SPF grade, 16–18 g, 4–5 weeks old) were injected s.c. with 200 μL of wild-type HepG2 or stable clone (HepG2/γ, HepG2/γi, HepG2/v, or HepG2/ci) cells (5 × 106 per mouse). Mice with HepG2/wt xenografts were treated i.p. at day 8 of posttransplantation with saline, cyclophosphamide (CP, 40 mg/kg), RA (10 mg/kg), or aflatoxin B1 (AFB1, 20 μg/kg) once every other day. Tumor size was measured every 4 or 5 d. Three weeks later, mice were sacrificed and tumors were removed for assessments (32).
Total proteins were prepared with a modified radioimmunoprecipitation assay buffer (33). Cytoplasmic proteins were purified with Buffer A [10 mmol/L HEPES-KOH (pH 7.9), 1.5 mmol/L MgCl2, 10 mmol/L KCl, and 0.5 mmol/L DTT], whereas nuclear proteins were prepared by resuspending the pellets in high-salt Buffer C [20 mmol/L HEPES-KOH (pH 7.9), 25% glycerol, 420 mmol/L NaCl, 1.5 mmol/L MgCl2, 0.2 mmol/L EDTA, and 0.5 mmol/L DTT; ref. 33].
Cells were lysed in a buffer containing 2 mmol/L Tris-HCl (pH 7.4), 10 mmol/L EDTA, 100 mmol/L NaCl, and 1% IGEPAL. After preclearing with protein A beads, lysates were incubated with 8 μL of primary antibody and 40 μL of protein A beads at 4°C. The beads were washed thrice with 1 mL of cold lysis buffer with exception of the NaCl concentration (150 mmol/L).
A cocktail of proteinase inhibitors was included in all protein purification. Equal amount of proteins was electrophoresed on an 8% SDS-PAGE gel and transferred onto PVDF membranes, which were then incubated with primary and secondary antibodies and detected using ECL system.
Tumor sections were immunostained with antibody against RARγ (1:500) or PCNA (1:500) and detected with corresponding secondary antibody. The slides were counterstained with hematoxylin. The proliferation indices were calculated by counting PCNA-positive cells versus total cells from three randomly selected regions. For fluorescent confocal microscopy, HepG2 cells were transfected with Myc-RARγ alone or in combination of Flag-p85α and stained with anti-Myc (1:100) and anti-Flag (1:100) antibodies. The slides were then stained with anti-Cy3 and anti-FITC antibodies. Cells were also costained with 4′,6-diamidino-2-phenylindole to visualize nuclei.
Reverse transcription–PCR analysis
Total RNAs were isolated by Trizol LS. The first strand was synthesized with RevertAid First Strand cDNA synthesis kits (Fermentas; ref. 33). The primers for PCR reactions are as follows: for human AFP, sense: 5′-tcagtgaggacaaactattgg-3′ and antisense: 5′-ctcttcagcaaagcagacttc-3′; for β-actin, sense: 5′-ctggagaagagctacgag-3′ and antisense: 5′-tgatggagttgaaggtag-3′.
Luciferase reporter assay
Expression of βRARE-luciferase reporter gene was determined in HepG2 and Hep3B cells after treatment with vehicle or 1 μmol/L RA for 20 h. NF-κB luciferase activity was assayed in HepG2 cells treated with 10 ng/mL TNFα, 1 μmol/L RA, or vehicle for 6 h. Each transfection also included β-galactosidase for normalization of the luciferase activities.
Data were expressed as mean ± SD. Each assay was repeated in triplicate in three independent experiments. Statistics was analyzed using Student's t test or ANOVA analysis. Values of P < 0.05 were considered significant.
Overexpression of RARγ in HCC tissues and cells
We recently reported that RARγ often resides in the cytoplasm of cancer cells depending on cell culture conditions (27). In this study, we wanted to explore RARγ expression and its cytoplasmic action in HCC cells. Immunoblotting studies showed that RARγ was readily detectable in all tumor and surrounding liver tissues prepared from HCC surgical specimens though at various levels (Fig. 1A). However, levels of RARγ were significantly higher in tumor tissues than the nontumorous tissues from a majority of patients (82%). Immunostaining studies showed that overexpression of RARγ in HCC tissues was predominantly cytoplasmic (Fig. 1B, left), which was consistent with our previous observation that overexpressed RARγ could translocate to the cytoplasm (27). The cytoplasmic localization of RARγ in HCC tissues were further confirmed by subcellular fractionation analysis, whereas in the surrounding liver tissues RARγ was mainly expressed in the nucleus (Fig. 1B, right). Thus, our results clearly show that RARγ is abnormally overexpressed in the cytoplasm of HCC cells.
Our results also showed that RARγ was highly expressed in the aggressive QGY-7703 HCC cells (Fig. 1C); however, it was only slightly expressed in QSG-7701 hepatocytes, a precancerous cell line derived from paracancerous tissue of a HCC patient (31). The expression levels of RARγ in QGY-7703 and QSG-7701 cells were comparable with those in HCC tissues and the paired nontumorous tissues. Compared with QSG-7701 hepatocytes, HepG2 and SMMC7721 HCC cells also expressed elevated levels of RARγ. However, no detectable level of RARγ was seen in Hep3B cells (Fig. 1C).
RARγ confers growth stimulatory effect of RA
We next studied the effect of RA on the growth of RARγ-positive HepG2 and RARγ-negative Hep3B cells. Treatment of Hep3B cells with RA resulted in growth inhibition in a dose-dependent manner (Fig. 1D, left). In contrast, RA treatment enhanced the growth of HepG2 cells. To determine whether RARγ expression levels were responsible for the distinct growth responses of RA, HepG2/γi cell line was then treated with RA. Our result showed that inhibition of RARγ by siRNA was able to render HepG2 cells sensitive to growth inhibition by RA (Fig. 1D, left), thus showing that RARγ mediates the growth stimulatory effect of RA in HepG2 cells. This was consistent with our observation that RARα and RARβ were similarly expressed in HepG2 and Hep3B cells (Fig. 1D, middle). Interestingly, when the effect of RA on activating βRARE reporter gene was examined in HepG2 and Hep3B cells, we observed that RA induced similar luciferase activity in both cell lines (Fig. 1D, right). This result suggests that transcriptional regulation of RA target genes is unlikely responsible for the different effects of RA on the growth of HepG2 and Hep3B cells.
RARγ enhances colony formation of HCC cells
Colony formation assays were then used to investigate the role of RARγ in cell transformation. Figure 2A showed that the ability of HepG2 cells in forming foci was significantly enhanced when the cells were transfected with RARγ expression vector. Conversely, inhibition of RARγ by siRNA strikingly impaired HepG2 cells to form colonies. When colony formation was compared between high-RARγ expressing QGY-7703 HCC cells and low-RARγ expressing QSG-7701 hepatocytes, we observed a significant number of foci formed by QGY-7703. In contrast, QSG-7701 cells seldom formed foci (Fig. 2A). When RARγ siRNA was introduced into QGY-7703 cells, it potently inhibited their colony formation, whereas overexpression of RARγ in QSG-7701 cells greatly increased their ability to form colonies (Fig. 2A). Together, these results convincingly show that RARγ plays an active role in promoting the clonogenic survival of HCC cells.
RARγ promotes the growth of HCC xenografts in animals
We next studied the effect of RARγ on the growth of HepG2 cells in nude mice. Figure 2B showed that RARγ transfection greatly enhanced the growth of HepG2 xenografts in mice, whereas inhibition of RARγ by siRNA significantly inhibited their growth. Consistent with our in vitro observation showing that RA was growth stimulatory in HepG2 cells, treatment of mice with RA stimulated rather inhibited the growth of HepG2 xenografts (112% of control, data not shown). When PCNA expression, an index of cell proliferation, were examined, we observed that ectopic overexpression of RARγ and RA treatment could strongly enhance immunoreactivity of PCNA in HepG2 xenograft (Fig. 2C). Interestingly, treatment of mice with cyclophosphamide, a potent cancer chemotherapeutic agent, greatly inhibited the growth of HepG2 xenografts (43% of control, data not shown), which was accompanied with reduction of RARγ expression, whereas AFB1, a key factor for promoting the development of HCC, increased RARγ expression (Fig. 2D). These results further show the role of RARγ in regulating the growth of HCC cells and suggest that it may also serve as a target of anticancer drugs.
Induction of AFP by RARγ
As AFP expression is a sensitive biomarker for hepatocarcinogenesis (28, 29), we then examined the role of RARγ in regulation of AFP expression in HepG2 cells. Figure 3A showed that RARγ transfection could strongly induce AFP transcription, which was further enhanced upon RA treatment. In contrast, inhibition of RARγ expression by siRNA abolished the effect of RA on inducing AFP expression (Fig. 3B). In HepG2 xenografts, AFP levels were reduced by cyclophosphamide but significantly increased by RA treatment (Fig. 3C). Together, these results suggest that abnormally increased expression of RARγ may play a role in hepatocarcinogenesis.
RARγ activates NF-κB in HCC cells
Because HCC is pathologically associated with chronic hepatitis and activation of NF-κB is a crucial for HCC growth and progression (34, 35), we next studied whether RARγ expression was involved in the regulation of NF-κB signaling pathway. We first determined the effect of RA on expression of IκBα, a key inhibitor of NF-κB nuclear translocation (36), in HepG2 and Hep3B cells. Figure 4A showed that RA treatment could strongly reduce IκBα expression in HepG2 cells, but it was ineffective in Hep3B cells. Both cell lines were sensitive to TNFα treatment. In the next studies, we showed that transfection of RARγ into HepG2 cells could dose-dependently reduce IκBα expression, which was further potentiated by RA treatment (Fig. 4B, left). Our results showed that RA/RARγ-induced IκBα reduction was due to activation of proteasome degradative pathway, as pretreatment of MG132, a potent inhibitor of the 26S proteasome, strongly inhibited this process (Fig. 4B, right). The ability of RARγ to activate the NF-κB pathway was also illustrated by our observation that RA treatment and RARγ transfection significantly induced nuclear translocation of p65 (Fig. 4C) and stimulated NF-κB luciferase activity (Fig. 4D). Together, our results show that RARγ can activate NF-κB pathway in HepG2 cells.
Interaction of RARγ with p85α and the activation of PI3K/Akt in HCC cells
As one of the major upstream activators of NF-κB is Akt (36, 37), we then examined whether RARγ activated Akt in HCC cells. Treatment of HepG2 cells with RA for 20 minutes increased levels of phosphorylated Akt protein in a dose-dependent manner (Fig. 5A). Such an effect of RA was not due to its enhancement of total levels of Akt protein, suggesting that RA could rapidly activate Akt. Transfection of HepG2 cells with RARγ also resulted in Akt activation, which was further enhanced when cells were treated with RA (Fig. 5A). Thus, RARγ might mediate the effect of RA on inducing Akt activation in HCC cells.
Recent studies showed that certain retinoid receptors could activate the PI3K/Akt pathway through interaction with the p85α regulatory subunit of PI3K (15, 17). We therefore studied the interaction of RARγ with p85α in HepG2 cells. Coimmunoprecipitation assays showed that immunoprecipitation of endogenous RARγ protein in HepG2 cells by anti-RARγ antibody but not by control IgG resulted in coimmunoprecipitation of a significant amount of p85α protein (Fig. 5B). The coimmunoprecipitation of p85α protein by anti-RARγ antibody was enhanced when cells were treated with RA. Similarly, RARγ was coimmunoprecitated by anti-p85α antibody in HepG2 cells. Furthermore, when HepG2 cells were cotransfected with Myc-RARγ and Flag-p85α, immunoprecipitation of Flag-p85α by anti-Flag antibody but not by control IgG resulted in coimmunoprecipitation of Myc-RARγ (Fig. 5C). Similarly, immunoprecipitation of Myc-RARγ by anti-Myc antibody led to coimmunoprecipitation of Flag-p85α. The interaction between Myc-RARγ and Flag-p85α occurred in the absence of RA, and the interaction could be enhanced by RA treatment. The interaction of Myc-RARγ with Flag-p85α was also illustrated by their colocalization in HepG2 cells (Fig. 5D). In the absence of Flag-p85α, Myc-RARγ was diffusely distributed in both the nucleus and cytoplasm of cells. When cotransfected with Flag-p85α that was primarily cytoplasmic, Myc-RARγ was mainly found in the cytoplasm, displaying distribution patterns that were extensively overlapped with those of Flag-p85α (Fig. 5D). Together, these results show that RARγ can interact with p85α, which may lead to activation of PI3K/Akt pathway.
RARγ activation of Akt precedes NF-κB activation
We next studied whether activation of Akt by RARγ was required for its activation of NF-κB. Time course analysis of the effect of RA on activation of Akt and NF-κB in HepG2 cells showed that Akt activation occurred 10 minutes after cells were treated with RA, with the optimal effect at 20 minutes after RA treatment (Fig. 6A). Such RA-induced Akt activation was accompanied with IκBα phosphorylation and its degradation. When the cells were pretreated with LY294002, a potent PI3K kinase inhibitor, RA-induced Akt phosphorylation was reduced. Consequently, IκBα phosphorylation and its degradation induced by RA were inhibited (Fig. 6A). Furthermore, transfection of dn-Akt, a dominant-negative Akt mutant (38), could dose-dependently reverse the effect of RA on IκBα degradation (Fig. 6B, left) and abrogate the effect of RARγ on stimulating NF-κB activity in HepG2 cells (Fig. 6B, right). These findings suggest that RARγ-induced IκBα degradation and NF-κB transactivation are mediated by its activation of Akt.
The role of RARγ in activating Akt and NF-κB pathways was further shown by our observation that inhibition of RARγ in HepG2 cells by siRNA reduced Akt activation and impaired the effect of RA on IκBα degradation (Fig. 6C). Interestingly, pretreatment of HepG2 cells with PI3K inhibitor LY294002 or IKK inhibitor BMS-345541 could strongly inhibit the effect of RA on inducing AFP expression (Fig. 6D), suggesting that RA induces AFP expression through its activation of PI3K/Akt and NF-κB pathways.
Retinoids and retinoid receptors are critical regulators of the growth, differentiation, and apoptosis of normal and malignant cells. Here, we report that RARγ was overexpressed in HCC tissues and cells and that overexpression of RARγ facilitated the growth of HCC cells through a RARγ/PI3K/Akt/NF-κB signaling axis.
Our results showed that RARγ was overexpressed in a majority of HCC tissues (Fig. 1A) and in several HCC cell lines (Fig. 1C). We showed that elevated RARγ expression was associated with the resistance of HCC cells to the growth inhibitory effect of RA in in vitro (Fig. 1D) and in animal, which might explain a previous clinical observation showing that RA treatment of some HCC patients resulted in the development of more aggressive phenotypes (39). Our results further showed that RARγ played an active role in promoting the growth of HCC cells as overexpression of RARγ greatly enhanced the colony formation of HCC cells (Fig. 2A) and stimulated tumor growth in mice (Fig. 2B). In addition, overexpression of RARγ strongly increased PCNA expression (Fig. 2C), an index of proliferation, and reactivated AFP transcription (Fig. 3), a sensitive marker for evaluating tumor activity during heptocarcinogenesis (28, 29). Consistently, the oncogenic activity of RARγ has been suggested in other tissues. In skin, high level of RARγ was believed to mediate the effect of RA on inducing hyperplasia (40). Overexpression of RARγ could significantly inhibit RA-induced expression of RARβ, a tumor suppressor frequently downregulated in many malignant diseases (41). The role of RARγ in maintaining the growth and survival of HCC cells was also illustrated by our observation that inhibition of the growth of HepG2 xenografts by cyclophosphamide was associated with a decrease of RARγ expression (Fig. 2D). Although how cyclophosphamide inhibited RARγ expression remains to be determined, such an observation suggests that RARγ may be targeted for drug development.
Data presented here show that overexpression of RARγ in HCC cells plays a critical role in the activation of PI3K/Akt pathway. We recently reported that unlike other retinoid receptors, RARγ often resided in the cytoplasm of cancer cells (27). Consistently, our present study revealed a strong cytoplasmic accumulation of RARγ in HCC cells but not in the surrounding liver cells (Fig. 1B). Altered subcellular localization of retinoid receptors likely reflects changes in the cellular environment, such as abnormal survival signalings, due to the effects of growth factors during carcinogenesis. Alternatively, the cytoplasmic expression of retinoid receptors may play an active role in the modulation of signal pathways through their direct interaction with protein kinase cascades (11, 12). These actions often occur rapidly and are recognized as nongenomic effects contrasting to their classic genomic RARE-based transcriptional regulation (1, 11, 12). In this study, we showed that RARγ could interact with p85α, which was further enhanced by RA treatment (Fig. 5B and C). As p85α is often highly expressed in human cancers including HCC (42, 43), p85α may serve to retain RARγ in the cytoplasm through their physical interaction in HCC. Indeed, transfection of Myc-RARγ alone showed a diffused distribution of Myc-RARγ in both nucleus and cytoplasm. However cotransfection of Flag-p85α and Myc-RARγ resulted in increased cytoplasmic localization of Myc-RARγ, displaying a distribution pattern that was extensively colocalized with Flag-p85α (Fig. 5D). The interaction of RARγ with p85α may lead to constitutive activation of Akt often detected in HCC (43, 44) and may also be responsible for RA-induced activation of Akt in RARγ overexpressing HCC cells (Fig. 5A). We showed that RARγ was essential for Akt activation as transfection of RARγ strongly enhanced the effect of RA on Akt activation (Fig. 5A), whereas downregulation of RARγ by siRNA impaired its effect (Fig. 6C). Because various RAR subtypes are able to activate PI3K/Akt pathway (15, 17), this pathway seems to be a general downstream target of RARs. However, its biological outcomes seem to be different, depending on different cell types. RARα-dependent PI3K/Akt signaling was shown to be required for RA-induced differentiation in SH-SY5Y neuroblastoma (15), whereas RARβ-dependent PI3K/Akt signaling in human breast cancer facilitated iodide uptake (17). Our results showed that RARγ-dependent activation of PI3K/Akt signaling led to NF-κB activation (Fig. 4C and D) and AFP induction (Figs. 3A and 6D), which was associated with the growth of HCC cells. The PI3K/Akt pathway is one of the major survival pathways in cancer cells (42), whereas activation of NF-κB provides an additional mechanism for cell growth and survival (44). The regulation of NF-κB activity by RARγ through PI3K/Akt pathway described here is distinct from the established mutually antagonistic mode between NF-κB and retinoid acid receptors, which mainly occurs within the nucleus (45).
In summary, we report here that RARγ is often overexpressed in HCC and in liver cancer cells, conferring their survival in vitro and in vivo. RARγ predominantly resides in the cytoplasm of HCC cells and is able to interact with p85α, leading to activation of the PI3K/Akt and IκBα/NFκB pathways and induction of AFP. Our characterization of this mechanism may provide a basis for developing novel strategies and agents for treatment of HCC by targeting the RARγ/PI3K/NF-κB signaling axis.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
We thank Dr. Xin Zhang from Chinese National Human Genome Center at Shanghai for providing Hep3B cells.
Grant Support: 863 Program grant 2007AA09Z404, Natural Science Foundation of China grant 30971445, NSFC/RGC Joint Research Project grant 30931100431, Key Science and Technology Planning Project grants 2007I0023 and 2008Y0062, Natural Science Foundation grant 2009J01198 of Fujian Province, and 985 Project (J-Z. Zeng) and NIH grant CA109345 and U.S. Army Medical Research and Material Command grant PCRPW81XWH-08-1-0478 (X-k. Zhang).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received August 11, 2009.
- Revision received December 4, 2009.
- Accepted January 14, 2010.
- ©2010 American Association for Cancer Research.