Ciliary neurotrophic factor receptor α subunit–modulated multiple downstream signaling pathways in hepatic cancer cell lines and their biological implications

Authors

  • Xin Hu,

    1. Shanghai Medical College, Fudan University, Shanghai, China
    2. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yingjun Zhao,

    1. Shanghai Medical College, Fudan University, Shanghai, China
    2. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Xianghuo He,

    1. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    Search for more papers by this author
  • Jinjun Li,

    1. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    Search for more papers by this author
  • Tao Wang,

    1. Shanghai Medical College, Fudan University, Shanghai, China
    2. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    Search for more papers by this author
  • Weiping Zhou,

    1. Department of Surgery (III), Eastern Hepatobiliary Surgery Institute, Shanghai, China
    Search for more papers by this author
  • Dafang Wan,

    1. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    Search for more papers by this author
  • Hongyang Wang,

    1. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    Search for more papers by this author
  • Jianren Gu

    Corresponding author
    1. National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, Shanghai, China
    • National Laboratory for Oncogenes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong University School of Medicine, No.25/Ln2200, Xietu Road, Shanghai 200032, China
    Search for more papers by this author
    • fax: 86-21-64177401.


  • Potential conflict of interest: Nothing to report.

Abstract

Ciliary neurotrophic factor (CNTF) plays important roles in a variety of tissues including neural and non-neural systems, but the function of CNTF and its receptor (CNTFR) in liver remains unclear. In this study, we demonstrate that CNTFRα is expressed heterogeneously in normal human liver and hepatocellular carcinoma (HCC) specimens but not in hepatoblastoma specimens. We choose the CNTFRα+/CNTFRα (CNTFRα positive/ CNTFRα negative) cell models of hepatic origin to study multiple downstream pathways of CNTFRα. We show that the presence of CNTFRα determines the temporal activation patterns of downstream signaling molecules and serves as a key modulator in regulating PI3K and AMP–activated protein kinase (AMPK) dynamically under CNTF stimulation, thus resulting in the increase of glucose uptake and translocation of glucose transporter 4 (GLUT4). Furthermore, CNTF-induced mitogen-activated protein kinase (MAPK) activation suppresses AMPK activity in the early phase of CNTF stimulation. Moreover, the protective role of CNTF against cell-cycle arrest is dependent on the presence of CNTFRα and is modulated by the glucose concentration of the culture medium. Conclusion: Our results demonstrate the importance of CNTFRα-mediated downstream signaling pathways and their functional implications in hepatic cancer cells, thus highlighting a better understanding of the biological roles of CNTFRα in human liver abnormalities, including metabolic diseases and hepatocarcinogenesis. (HEPATOLOGY 2008.)

Ciliary neurotrophic factor was first denoted as a neuron-nurturing factor in chick ciliary ganglion neurons more than 2 decades ago and was subsequently defined as a member of the interleukin-6 cytokine family.1 In past decades, ciliary neurotrophic factor (CNTF) has been shown to promote the differentiation of sympathetic neurons and glial progenitor cells into astrocytes, and to promote the survival of a variety of neuronal cells, such as sensory, motor, hippocampal, and cerebral neurons.2 A recent study revealed that CNTF can suppress food intake and induce weight loss through a leptin-like mechanism in ob/ob mice.3 Moreover, CNTF increases metabolic rate and energy expenditure of peripheral metabolic organs, independent of signaling in the brain.4 It was further demonstrated that CNTF enhances fatty acid oxidation in muscle and reduced insulin resistance in obese, diabetic mice.5

CNTF exerts its biological functions through its receptor CNTFR to activate multiple downstream signaling pathways. CNTFR is a tripartite complex composed of 3 subunits: CNTF receptor α subunit (CNTFRα), gp130, and leukemia inhibitory factor receptor (LIFR).6 CNTFRα, the specific α subunit for CNTF, is anchored to the membrane by a glycosyl-phosphatidylinositol linkage without a cytoplasmic domain.7 Binding of CNTF to CNTFRα induces heterodimerization of the transmembrane subunits (gp130 and LIFR) to transduce cell signals via the Janus kinase (JAK)/signal transducer and activator of transcription (STAT) and mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) signaling pathways.8 In addition, phosphotidyl inositol-3′-phosphate-kinase (PI3K)/Akt participates in the survival response of neurons to CNTF.9 CNTF plays dual roles on the activation of AMP–activated protein kinase (AMPK). In the central nervous system, CNTF inhibits AMPK activity in the hypothalamus to promote leptin-sensitive neurogenesis in the arcuate nuclei and reduces neuropeptide Y gene expression; in the periphery, CNTF accelerates fat oxidation through an AMPK-dependent mechanism in skeletal muscle.5, 10 However, the function of CNTF/CNTFRα in human liver cancers and the underlying molecular mechanisms have not been elucidated.

Serendipitously, our previous study of a large-scale complementary DNA transfection screening showed that CNTFRα might have some positive effect on hepatocellular carcinoma cell growth.11 Here, we examined the expression patterns of CNTFRα in normal human liver, hepatocellular carcinoma (HCC), and hepatoblastoma tissues, as well as cancer cell lines. We chose the CNTFRα+/ CNTFRα cell models of hepatic origin to study multiple downstream pathways in hepatic cancer cells. After examining CNTFRα-dependent/CNTFRα-independent pathways, we found that CNTF could contribute to the process of glucose uptake and glucose transporter 4 (GLUT4) translocation by PI3K and AMPK pathways. Furthermore, we wanted to learn whether there would be some cross-talk between these CNTF-responding pathways, using specific inhibitors, RNA interference, and constitutively active plasmids, respectively. Moreover, we extended our study to the roles of CNTF on cell cycle progression, with the additional insight of the presence of CNTFRα and the glucose content in the culture medium.

Abbreviations

CNTF, ciliary neurotrophic factor; CNTFR, ciliary neurotrophic factor receptor; CNTFRα, ciliary neurotrophic factor receptor α subunit; ERK, extracellular signal-regulated kinase; GFP, green fluorescent protein; GLUT4, glucose transporter 4; HCC, hepatocellular carcinoma; JAK, janus kinase; LIFR, leukemia inhibitory factor receptor; MAPK, mitogen-activated protein kinase; MEK, mitogen-activated protein kinase kinase; PCR, polymerase chain reaction; PI3K, phosphatidyl inositol-3′-phosphate-kinase; STAT, signal transducer and activator of transcription; AMPK, AMP-activated protein kinase.

Materials and Methods

Cell Culture, Reagents, and Constructs.

Human HCC cell lines including Hep3B, SMMC-7721, BEL-7402; human hepatoblastoma cell line HepG2; immortalized human hepatocyte cell line LO-2; Cercopithecus aethiops cell line Cos 7; and human astrocytoma cell line U87 were cultured in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA), supplemented with 10% fetal bovine serum, 100 IU/mL penicillin G, and 100 mg/mL streptomycin sulfate. To regulate glucose concentrations in the culture medium, cells were cultured in Dulbecco's modified Eagle's medium without D-glucose and sodium pyruvate (Invitrogen) supplemented with different glucose concentrations. Experiments were replicated in the same cell lines with the same passage. Chemical reagents including JAK inhibitor AG490, mitogen-activated protein kinase kinase (MEK) inhibitors U0126 and PD98059, PI3K inhibitor LY294002, and AMPK inhibitor Compound C (Calbiochem, San Diego, CA) were dissolved in dimethylsulfoxide (DMSO) according to the manufacturer's protocol. CNTF was provided by Dr. Y. Gong.12 The plasmid Myc-Glut4-EGFP was provided by Dr. Shuichi Okada.13 Constitutively active MEK1 (S218, 222E) and dominant-negative MEK1 (K97M) vectors were provided by Dr. Koichi Takimoto.14

Reverse Transcription Polymerase Chain Reaction and Quantitative Polymerase Chain Reaction.

Total RNA was isolated from cells and tissues with Trizol reagent (Invitrogen) according to the manufacturer's protocol. The primers used were previously described,15 resulting in the following products: CNTFRα (638 bp); LIFR (478 bp); gp130 (473 bp); glyceraldehyde 3-phosphate dehydrogenase (515 bp). These results were obtained by 42 cycles of amplification. Quantitative polymerase chain reaction (PCR) was performed to detect CNTFRα and housekeeping gene glyceraldehyde 3-phosphate dehydrogenase using ABI PRISM 7300 Sequence Detection System (Applied Biosystems, Foster City, CA). The primers for TaqMan were as follows: CNTFRα, forward, 5′-ACA GCA CAC ACC ATC ACA GAT-3′; reverse, 5′-GCT ACG CTC CAG TCA CTC CAT-3′; probe FAM-5′-ACG CCG GGA AGG AGT ACA TTA TCC-3′-TAMRA. All samples were run in triplicate.

Immunoblot Analysis and Immunohistochemistry.

Immunoblotting was done essentially as described,8 with the following antibodies: CNTFRα (Santa Cruz Biotechnology, Santa Cruz, CA); phospho-STAT3 (Tyr705) (Cell Signaling Technology, Beverly, MA), STAT3 (Cell Signaling Technology); phospho-p44/42MAPK (Thr202/Tyr204) (Cell Signaling Technology), p44/42 MAPK(Cell Signaling Technology); phospho-Akt (Ser473) (Cell Signaling Technology), Akt (Cell Signaling Technology); phospho-AMPKα (Thr172) (Cell Signaling Technology), AMPKα (Cell Signaling Technology); β-actin (Sigma-Aldrich, St. Louis, MO), and appropriate secondary antibodies (Pierce, Rockford, IL). Membranes were stripped of antibodies using Restore Stripping Buffer (Pierce). The densitometry analysis was measured by Image-Pro Plus 5.0 (Media Cybernetics, Bethesda, MD).

Immunohistochemistry was carried out with a section of hepatic tissues as described.16 Hepatoblastoma specimens were from Eastern Hepatobiliary Surgery Institute.

Immunofluorescence Confocal Imaging.

Cells on the slides were fixed with 4% paraformaldehyde and incubated with CNTFRα antibodies (Santa Cruz Biotechnology) for 2 hours. The slides were incubated with fluorescein isothiocyanate (FITC)-conjugated secondary antibodies (Santa Cruz Biotechnology) and 4′,6-diamino-2-phenylindole (Sigma-Aldrich) for 1 hour. Pictures were taken with a confocal laser microscope (IX-70, Olympus 60M, Olympus, London, UK).

FITC-CNTF Binding Assay.

After starving with serum-free culture medium for 16 hours, cells were incubated with FITC-CNTF (50 ng/mL) at 4°C for 15 minutes. As a competition control, unlabeled CNTF was added to 50 μg/mL before the addition of FITC-CNTF. Cells were washed with cold Dulbecco's modified Eagle's medium and fixed in 4% paraformaldehyde for 15 minutes. The images were collected by a confocal laser microscope (IX-70, Olympus 60M, Olympus).

2-Deoxyglucose Uptake.

Cells were seeded in 6-well plates, and [3H]-2-deoxyglucose uptake was determined as described.17 The brief detail is described in Supplementary Methods.

Myc-Glut4-EGFP Translocation Assays.

GLUT4 translocation assays were done essentially as described.18 The brief detail is described in Supplementary Methods.

Cell-Cycle Distribution.

Cells were treated with 0.17 μM nocodazole (Sigma-Aldrich) for 16 hours, then replated in nocodazole-free medium with different glucose concentrations for 12 hours, additionally with or without CNTF (100 ng/mL). Then, cells were collected and fixed in ice-cold 70% ethanol overnight. Fixed cells were washed with phosphate-buffered saline and stained with a solution containing 25 μg/mL propidium iodide (Sigma-Aldrich), 10 μg/mL RNase A, 0.2% Triton X-100, and 0.05 mM ethylene diamine tetraacetic acid in phosphate-buffered saline for 30 minutes in the dark. For each sample, at least 10,000 cells were analyzed by FACSC cytometry (Beckman Coulter Epics Altra, Miami, FL) and MultiCycle AV for Windows 5.0 (Phoenix Flow Systems, San Diego, CA).

Statistical Analysis.

All data were expressed as mean and standard error. Statistical significance was tested using the Student t test. Statistical differences were considered to be significant if P < 0.05.

Results

Expression of CNTFRα in Normal Liver, HCC, Hepatoblastoma Tissues, and Hepatic Cancer Cell Lines.

We first examined the expression status of CNTFRα in normal human liver, HCC, corresponding noncancerous liver tissues from the same HCC patient, and hepatoblastoma tissues by immunohistochemistry assay (Fig. 1A). It was shown that CNTFRα+ cells were distributed in a heterogeneous manner. Positively staining hepatocytes were scattered in normal liver, HCC, and corresponding noncancerous liver tissues, but no significant difference was observed between them (data not shown). In hepatoblastoma tissues, most hepatocytes showed a negative signal. Based on these observations, we analyzed the messenger RNA expression of all the subunits of CNTFR, including CNTFRα, gp130, and LIFR. Our reverse transcription PCR results demonstrated that all the subunits of CNTFR were expressed in normal liver, HCC, corresponding noncancerous liver tissues, three HCC cell lines (SMMC-772, BEL-7402, Hep3B), and human hepatocyte cell line (LO-2); however, CNTFRα was not expressed in human hepatoblastoma tissues and hepatoblastoma cell line (HepG2) (Fig. 1B), with Cercopithecus aethiops cell line (Cos7) and human astrocytoma cell line (U87) used as controls.19 The results of the immunoblot analysis also confirmed the absence of CNTFRα in HepG2 cells (Fig. 1C). For further study of the biological roles and downstream signaling pathways of CNTFRα, we used SMMC-7721 and HepG2 as CNTFRα+ and CNTFRα cells, respectively.

Figure 1.

CNTFR expression levels in human liver tissues and hepatic cancer cell lines. (A) Immunohistochemistry analysis of liver sections revealed the heterogeneous distribution of CNTFRα in human hepatic tissues. (CNTFRα+, black arrow; CNTFRα, white arrow); (B) CNTFR heterotripartite (CNTFRα, gp130 and LIFR) expression was examined by reverse transcription PCR in human hepatic tissues and cancer cell lines, using U87 cells as positive control, Cos7 cells as negative control (H, HCC; N, corresponding noncancerous liver tissues from the same HCC patient); (C) Immunoblots of CNTFRα expression in hepatic cancer cell lines, U87 and Cos7 cells.

Different Temporal-Activation Patterns of JAK/STAT3, MAPK/ERK, PI3K/Akt and AMPK Pathways in CNTFRα+/ CNTFRα Cells.

To elucidate the molecular function of CNTF/CNTFR in hepatic cancer cells, we first examined intracellular signaling transduction, including JAK/STAT3, MAPK/ERK, PI3K/Akt and AMPK signaling pathways, in CNTFRα+ (SMMC-7721) and CNTFRα cells (HepG2), using the phosphorylation of STAT3, ERK1/2, Akt, and AMPK molecules as indicators for relevant activated pathways.

We first studied the temporal-activation courses of these 4 signaling pathways in SMMC-7721 and HepG2 cells from 0 to 240 minutes after CNTF treatment. In SMMC-7721 cells, all these pathways were activated throughout the analyzed period (Fig. 2A). The temporal-activation patterns of phospho-STAT3 and phospho-Akt were similar as a persistent activation mode at all the intervals. However, the phosphorylation of ERK1/2 was observed to reach a peak at 15 minutes and decline abruptly by 30 minutes; whereas phosphorylation of AMPK had a lag phase of 30 minutes, then increased immediately and remained persistent. Surprisingly, the temporal patterns of these 4 signaling pathways were quite different in HepG2 cells (Fig. 2B). In HepG2 cells, no phospho-ERK1/2 and phospho-Akt were detected within all analyzed intervals (0-240 minutes), but CNTF did increase the phosphorylation of STAT3 and AMPK. The CNTF-induced increase of phospho-STAT3 was observed 5 to 30 minutes after the treatment and descended later, whereas phosphorylation of AMPK was an immediate and persistent mode (Fig. 2B). These results showed that the presence or absence of CNTFRα determined the characteristic temporal-activation patterns of signaling pathways in hepatic cancer cell lines.

Figure 2.

CNTFRα determined the characteristic temporal-activation patterns of downstream signaling pathways in hepatic cancer cell lines. (A and B) Immunoblots of SMMC-7721 (CNTFRα+) and HepG2 (CNTFRα) extracts after CNTF treatment (100 ng/mL) at the indicated times, using antibodies against phospho-STAT3, phospho-ERK1/2, phospho-Akt, and phospho-AMPK. Insulin treatment was used as positive control. The amount of STAT3, ERK1/2, Akt, AMPK, and β-actin was detected as a loading control; (C) CNTFRα was ablated by targeting small interfering RNA in SMMC-7721 cells. The extracts from cells were analyzed by western blot using the indicated antibodies at 15 minutes after CNTF treatment (100 ng/mL).

To further confirm these results, 3 CNTFRα-specific small interfering RNA duplexes were introduced into SMMC-7721 cells, causing an 80% to 95% decrease of CNTFRα messenger RNA, as verified by quantitative PCR (Supplementary Fig. 1). As shown in Fig. 2C, silenced CNTFRα reduced the phosphorylation of STAT3, ERK1/2, and Akt after CNTF treatment (15 minutes). Surprisingly, a dramatic increase of phospho-AMPK was observed. Taken together, CNTF-activated MAPK/ERK and PI3K/Akt pathways depended on the presence of CNTFRα in HCC cells. Even though the JAK/STAT3 pathway could be activated by CNTF in both the cell lines, silenced CNTFRα did impair STAT3 phosphorylation, indicating its partial dependence of CNTFRα. Moreover, the activation of AMPK was independent of the presence of CNTFRα.

CNTF-Induced Enhancement of Glucose Uptake in a CNTFRα-Independent Manner in Human Hepatic Cancer Cell Lines.

Because CNTF has been known to be an important cytokine in regulating metabolism from the hypothalamus to peripheral organs,4 we attempted to examine the effects of CNTF on glucose metabolism, particularly the process of glucose uptake in human hepatic cancer cell lines. As shown in Fig. 3A, both insulin and CNTF treatment could enhance the [3H]-2-deoxyglucose uptake in SMMC-7721 cells. A similar result was obtained in CNTFRα+ hepatocytes (LO-2). The enhancement of glucose uptake by CNTF was also found in CNTFRα cells (HepG2) and CNTFRα-silenced SMMC-7721 cells (Fig. 3B). This unexpected discovery implied that some alternative CNTFRα-independent mechanisms might be involved.

Figure 3.

The process of CNTF-induced glucose uptake did not rely on the presence of CNTFRα. (A) Stimulation of CNTF (100 ng/mL) and insulin (150 nM) for 30 minutes enhanced [3H]-2-Deoxyglucose uptake in SMMC-7721 cells. *P < .05 versus control; (B) Stimulation of CNTF (100 ng/mL) for 30 minutes enhanced [3H]-2-deoxyglucose uptake in SMMC-7721 and LO-2 (CNTFRα+) cells; HepG2 (CNTFRα) and CNTFRα silenced SMMC-7721 cells. *P < 0.05 versus control.

To further investigate the functional subunit CNTFRα in response to CNTF, the distribution of CNTFRα protein and its ability to complex with FITC-CNTF were examined. Consistent with the previous results (Fig. 1B,C), CNTFRα protein was distributed on the cellular membrane of SMMC-7721, LO-2, and U87 cells, but not on HepG2 cells (Fig. 4A). Nevertheless, the FITC-CNTF binding assay revealed that the binding between CNTF and the cellular membrane occurred in both CNTFRα+ cells (SMMC-7721, LO-2, and U87) and CNTFRα cells (HepG2) (Fig. 4B, top). In the parallel competitive experiments, fluorescence of FITC-CNTF was impaired by pretreatment of excessive unlabeled CNTF (Fig. 4B, bottom). These results suggested that the binding between CNTF and the cellular membrane was not exclusively mediated through CNTFRα. This observation provided additional support to the previous notion that the CNTF-enhanced glucose uptake might be attributed to a CNTFRα-independent mechanism in hepatic cancer cells.

Figure 4.

CNTF could conjugate to CNTFRα+ and CNTFRαhepatic cancer cell lines. (A) Representative confocal images of CNTFRα protein were stained by indirect-immunofluorescence in SMMC 7721, HepG2, Lo2, and U87 cells. [CNTFRα (green), nuclei (blue)]; (B) The cells incubated with FITC-CNTF (50 ng/mL) at 4°C for 15 minutes were examined by confocal microscopy. The conjugation between FITC-CNTF and the cellular membrane existed in both CNTFRα+ and CNTFRα cells (top). Pretreatment with excessive unlabeled CNTF (50 μg/mL) reduced the fluorescence of FITC-CNTF (50 ng/mL) at the periphery of the cells (bottom).

CNTF-enhanced GLUT4 Translocation and Glucose Uptake Is Attributed to the Activation of PI3K/Akt and AMPK Pathways.

For a better understanding of the molecular mechanisms of CNTF-enhanced glucose uptake, we further explored the possible roles of glucose transporter GLUT4, which has been described to be involved in insulin-induced glucose uptake.18, 20 GLUT4 translocation was monitored using a Myc-GLUT4–green fluorescent protein (GFP) plasmid. This protein was fused with GFP at its C terminus to visualize the total transporter and contained 7 copies of c-myc in its first extracellular loop to reveal the transporters on the cell surface. As shown in Fig. 5A, in SMMC-7721 cells, CNTF enhanced GLUT4 translocation from the cytosol-pool to the cell surface, similarly to insulin.

Figure 5.

Activation of PI3K/Akt and AMPK contributed to the process of CNTF-induced GLUT4 translocation and glucose uptake in HCC cell lines. (A) SMMC-7721 cells were transfected with a GLUT4-GFP reporter containing c-myc epitope tags for 48 hours. After stimulation with CNTF (100 ng/mL) and insulin (150 nM) for 15 minutes, cells were incubated with anti-myc. (Staining of myc on cell surface, shown in red; total GLUT4 fused GFP, shown in green; merge, shown in yellow); (B) The cells transfected with the Myc-GLUT4-GFP reporter were pretreated with MEK1/2 inhibitor U0126 (10 μM) and PI3K inhibitor LY294002 (10 μM) for 2 hours, and AMPK inhibitor compound C (10 μM) for 15 minutes. Then the cells were treated as in (A). (C) After treatment with pathway inhibitors (LY294002 and compound C), the amount of glucose uptake by CNTF or insulin in SMMC-7721 was suppressed (*P < 0.05, **P < 0.01). (D) The densitometry analysis of PI3K/Akt and AMPK pathways (Fig. 2A,B) showed that their temporal-activation patterns were different in SMMC 7721 (CNTFRα+) and HepG2 (CNTFRα) cells. (E) Immunoblots of SMMC-7721 (CNTFRα+), LO-2 (CNTFRα+) and HepG2 (CNTFRα) cell extracts after CNTF treatment of 15 minutes. Immunoblots were incubated with antibodies against phospho-Akt and phospho-AMPK followed by total Akt, AMPK, and β-actin.

We investigated the related intracellular signaling pathways using corresponding kinase inhibitors. U0126 could inhibit MEK1/2 (ERK1/2 upstream kinase) and reduce ERK1/2 phosphorylation (Supplementary Fig. 2A). LY294002 could inhibit PI3K and ablate phospho-Akt completely (Supplementary Fig. 2B). AG490 failed to halt the phosphorylation of STAT3 caused by CNTF by its specific blockage of JAK2/3 in SMMC-7721 cells (Supplementary Fig. 2C).8, 21 Thus, we applied U0126 (MEK1/2 inhibitor), LY294002 (PI3K inhibitor), and compound C (AMPK inhibitor) in the subsequent experiments to block relevant kinases before CNTF stimulation. As shown in Fig. 5B, CNTF-triggered GLUT4 translocation in SMMC-7721 cells was strikingly blocked by pretreatment with LY294002 or compound C but not by U0126. Moreover, [3H]-2-deoxyglucose uptake experiments provided additional evidence that pretreatment of LY294002 and compound C reduced CNTF-triggered glucose uptake by approximately 0.5-fold (Fig. 5C). All of these data indicated that PI3K and AMPK pathways were involved in CNTF-induced glucose uptake and GLUT4 translocation.

To elucidate the roles of the PI3K/Akt and AMPK pathways, we established temporal-activation patterns according to the phosphorylation status of Akt and AMPK at intervals from 0 to 240 minutes in both SMMC-7721 and HepG2 cells (Fig. 5D). Quantitative analysis revealed that phosphorylation of Akt had a quick and persistent increase from 5 to 240 minutes, but phosphorylation of AMPK had a lag phase of 30 minutes, then increased sharply and reached the plateau from 30 to 240 minutes in CNTFRα+ cells (SMMC-7721). However, in CNTFRα cells (HepG2), CNTF could not activate Akt phosphorylation but induced an immediate and persistent phosphorylation of AMPK. These results indicated the regulatory role of CNTFRα in downstream signaling pathways. We chose 15 minutes after CNTF treatment as the report time to validate the significance of these results. As anticipated, Akt activation played the major role in CNTFRα+ cells (SMMC-7721 and LO-2). In contrast, AMPK activation was dominant in CNTFRα cells (HepG2) at the same time point (Fig. 5E). The results in CNTFRα-silenced SMMC-7721 cells provided further evidence that CNTFRα knockdown could reduce Akt activation but increase AMPK activation at 15 minutes after treatment (Fig. 2C). Taken together, the existence of CNTFRα determined different temporal-activation patterns of Akt and AMPK, which contributed to glucose uptake and GLUT4 translocation. Although the PI3K/Akt pathway was suppressed in CNTFRα or CNTFRα-silenced cells, the AMPK pathway could be activated immediately as a compensatory pathway to ensure glucose uptake.

CNTF-induced AMPK Activation Is Suppressed by MAPK Activation in the Early Phase of Stimulation.

As described previously, we unexpectedly observed 2 phenomena: activation of AMPK had a special lag phase in SMMC-7721 compared with that in HepG2 cells; and the increasing phospho-AMPK level accompanied by a decreasing MAPK activation in CNTFRα-silenced cells. To further explore the mechanism regulating AMPK activation, we analyzed the temporal-activation pattern of the four CNTF-induced signaling pathways in SMMC-7721 cells. As shown in Fig. 6A, the peak of MAPK activation and the suppression of AMPK occurred at the same time, thus strongly implicating the intrinsic correlation of the 2 pathways. Moreover, we applied inhibitors to determine whether the other pathways affected AMPK activation in the early phase (15 minutes) of CNTF stimulation. The AMPK activation was enhanced by pretreatment with MEK 1/2 inhibitor U0126 and PD98059, but not by PI3K inhibitor LY294002 (Fig. 6B). To further clarify the relationship between MAPK activation and AMPK suppression, we transfected a constitutively active MEK1 (S218, 222E) plasmid into CNTFRα cells (HepG2) to activate ERK1/2. This transfection could mimic CNTF-activated MAPK pathway in SMMC-7721 cells and did cause the suppression of AMPK in HepG2 cells (Fig. 6C). These data strongly suggested that MAPK activation could temporarily suppress AMPK activation at the early time points of CNTF stimulation.

Figure 6.

Activation of MAPK suppressed AMPK phosphorylation in the early phase of CNTF treatment. (A) The densitometry analysis of CNTF-induced signaling pathways (JAK/STAT3, MAPK/ERK, PI3K/Akt, and AMPK pathways, as shown in Fig. 2A) was integrated to build a temporal-activation mode in SMMC-7721 cells. (B) SMMC-7721 cells were pretreated with MEK1/2 inhibitor U0126 (10 μM), PD98059 (10 μM), and PI3K inhibitor LY294002 (10 μM) for 2 hours. After treatment with CNTF (100 ng/mL) for 15 minutes, the cell extracts were analyzed by western blot using the indicated antibodies. Phospho-AMPK/AMPK was quantified by densitometry analysis. (C) HepG2 cells were transfected with constitutively active MEK1 (S218, 222E) and dominant-negative MEK1 (K97M) and pcDNA3 plasmids for 48 hours. After 15 minutes of CNTF stimulation, the cell extracts were analyzed by western blot using the indicated antibodies, and phospho-AMPK/AMPK was quantified.

CNTF-Induced Recovery from G2/M Cell Cycle Arrest Is Related to the Presence of CNTFRα and the Extracellular Glucose Concentration.

Our previous work has implicated CNTFRα with HCC cell growth,11 and preliminary studies have shown that the proliferation of SMMC-7721 cells was impaired after RNA interference–mediated knockdown of CNTFRα (Supplementary Fig. 3). Here, we further analyzed the effect of CNTF on releasing cells from nocodazole-induced G2/M arrest. Considering the importance of CNTF on glucose metabolism, we extended our analysis to the effect of the glucose concentration in culture medium. The data from a representative experiment demonstrated that the proportion of cells releasing from G2/M arrest increased after CNTF treatment under different glucose concentrations (Fig. 7A). The statistical analysis revealed that the CNTF-induced cells releasing from G2/M arrest were 6.48% and 4.45% at physiological glucose concentrations, 3 mM and 6 mM, respectively; whereas the increment was 1% to 3% at the rest points, including the glucose concentration in conventional culture medium, 10 mM (Fig. 7B, left). Thus, we demonstrated that CNTF could promote cells to release from G2/M arrest, and this effect could be modulated by the extracellular glucose concentration. In parallel experiments, no similar trend was detected in CNTFRα cells (HepG2) after CNTF treatment (Fig. 7B, right). These data suggested that the CNTF-induced protective effect on cell cycle arrest was principally dependent on the presence of CNTFRα, and more importantly, this effect could be unveiled at a physiological glucose concentration.

Figure 7.

CNTF induced cells to release from G2/M arrest in different glucose concentrations. (A) Treatment with CNTF could enhance the number of SMMC-7721 cells releasing from G2/M arrest preferring to physiological glucose concentrations (3 mM and 6 mM). After nocodazole treatment for 16 hours, SMMC-7721 cells were released into the serum-free medium with a series of glucose concentration from 1 mM to 20 mM, additionally with or without CNTF for 12 hours, and cell cycle distribution was analyzed with flow cytometry (representative data). (B) The statistical analysis of the percentage of CNTF-induced cells released from G2/M arrest under different glucose concentrations in SMMC-7721(CNTFRα+) and HepG2 (CNTFRα) cells. The data were reproduced 3 times. *P < 0.05, **P < 0.01

Discussion

Despite vigorous studies of CNTF in CNS and peripheral organs,2 little is known about its functional role in hepatic cancer cell lines and its relationship with CNTFRα. Here, we demonstrated that the CNTFRα+ cells are unevenly distributed, exhibiting a heterogeneous expression pattern in normal liver, HCC, and corresponding noncancerous liver tissues. For example, in normal liver, CNTFRα was preferentially expressed in hepatocytes at the peripheral area of hepatic lobules, whereas negative or weak staining was shown in most hepatocytes at the intermediate area of hepatic lobules and the zone around the central vein. These features implied that CNTFRα expression might be related to the status of hepatocyte metabolism, as described in other organs.5 The heterogeneous expression of CNTFRα has also been observed in hepatic cancer cell lines. Since hepatoblastoma has a specific identity based on its different clinical and pathological characteristics,22 it is not surprising to find negative CNTFRα expression in hepatoblastoma tissues and HepG2 cells, a cell line derived from hepatoblastoma. Thus, the impact of CNTFRα on hepatic tissues needs further clarification.

We focused on two functions of CNTF and its receptor in hepatic cancer cells: the enhancement of glucose uptake and the effect on recovering from drug-induced cell cycle arrest. Because cancer cells, including HCC, are well known for their high consumption of glucose,23 F18-labeled 2-deoxyglucose is currently used in clinical positron emission tomography imaging diagnosis of cancer.24 Previous studies examining the effects of insulin showed that insulin can activate PI3K and AMPK signaling pathways to promote glucose uptake,18, 20 but little is known about the roles of CNTF for glucose uptake and its molecular mechanism in hepatic cancer cells. Here, it was shown that CNTF-triggered PI3K/Akt and AMPK signaling pathways also contribute to glucose uptake in hepatic cancer cells, and induce the translocation of GLUT4 from intracellular storage to the plasma membrane. In CNTFRα+ cells, the activation of PI3K/Akt was an immediate and persistent response, whereas AMPK activation had a lag phase of 30 minutes. In the CNTFRα or CNTFRα-silenced cells, AMPK activation was immediate, whereas no activation of PI3K/Akt was observed. These results highlight that the responses of PI3K/Akt and AMPK pathways triggered by CNTF are dynamic processes, and CNTFRα serves as a key modulator to transduce multiple signals in hepatic cancer cell lines.

Numerous studies have reported that CNTF-induced activation of MAPK/ERK, JAK/STAT3, and PI3K/Akt pathways play important roles in carcinogenesis. The suppression of ERK activation can inhibit human tumor cell proliferation in vitro and control the growth of human tumor xenografts in mice.25 The constitutively active PI3K/Akt pathway helps tumor cells prolong survival and lower the mitogenic threshold required for cell division.26 STAT3 displays its roles in oncogenesis by regulating cell growth, survival and differentiation.27 Therefore, the effect of CNTF-CNTFR on cell cycle progression in hepatic cancer cells was not beyond our expectations. Interestingly, the protective effect on cell cycle arrest is more sensitive at physiological glucose concentrations (3-6 mM), compared with that at 10 mM glucose, which is the concentration currently used in conventional culture medium. The reason for this phenomenon is presently unknown, but these interesting results remind us of the importance of the glucose concentration in culture medium, which has been previously ignored for other in vitro studies of cytokines or growth factors involved in glucose metabolism in cancer biology.

The data presented here show that exogenous CNTF can activate multiple signaling pathways in CNTFRα+or CNTFRα hepatic cancer cell lines, and we proposed a model to differentiate the four CNTF-induced signaling pathways into CNTFRα-dependent or CNTFRα-independent processes, including JAK/STAT3, MAPK/ERK, PI3K/Akt, and AMPK signaling pathways (Fig. 8). Based on our observations, the activation of MAPK/ERK, PI3K/Akt, and JAK/STAT3 pathways are dependent or partially dependent on the presence of CNTFRα, and the AMPK pathway is activated in an independent manner. In our experiments, CNTF could activate JAK/STAT3 and AMPK pathways in both CNTFRα+ and CNTFRα hepatic cancer cell lines. Similar results have been reported in skeletal muscle cells.5 In addition, the results of the FITC-CNTF binding assay and glucose uptake also provided evidence that not only does CNTF function in CNTFRα+ hepatic cancer cells, but there is at least one other candidate receptor for CNTF. Some studies have presumed that interleukin-6 Rα might serve as a potential α receptor for CNTF to induce downstream STAT3 activation,28 but whether the downstream AMPK pathway can also be activated remains unclear.

Figure 8.

A model of CNTF-induced multiple signaling pathways in hepatic cancer cell lines. Responding to CNTF stimulation, MAPK/ERK and PI3K/Akt activation is dependent on the presence of CNTFRα, and JAK/STAT3 activation is partially dependent on CNTFRα, whereas AMPK activation is independent of CNTFRα. In hepatic cancer cells, PI3K/Akt and AMPK activation contributes to CNTF-induced glucose uptake and GLUT4 translocation. Furthermore, activated MAPK suppresses AMPK activation in the early phase of CNTF stimulation, showing the cross-talk between these signaling pathways.

Based on the difference in AMPK temporal-activation between CNTFRα+ and CNTFRα cells, we further investigated the regulation of the AMPK pathway under special conditions and found that the activation of the MAPK/ERK pathway can suppress AMPK activation in the early phase of CNTF stimulation. Here, we revealed that: (1) The concurring MAPK activation and AMPK suppression appeared in the same phase (15 minutes); (2) in CNTFRα-silenced SMMC-7721 cells, a decrease of MAPK phosphorylation was accompanied by an increase of AMPK phosphorylation; (3) MEK1 inhibitors could enhance AMPK activation in early phase; and (4) the constitutive activation of MEK1 could partially suppress the AMPK activation induced by CNTF. For the first time, our results described the interaction between CNTF-induced MAPK and AMPK pathways, in which CNTFRα might act as a modulator to regulate the cross-talk between them.

In summary, CNTFRα was expressed in a heterogeneous manner and modulated the activation of PI3K and AMPK pathways, which contributed to CNTF-induced glucose uptake and GLUT4 translocation in hepatic cancer cell lines. The presence of CNTFRα was also involved in the cross-talk between MAPK and AMPK pathways as follows: CNTF-induced AMPK activation was suppressed by CNTFRα-dependent MAPK activation in the early phase. CNTFRα also played an important role in CNTF-induced recovery from G2/M cell cycle arrest. The extending investigation revealed that the extracellular glucose concentration could make HCC cells more sensitive to the protective effects of CNTF-CNTFR against cell cycle arrest. The significance in the context of HCC biology for this special cytokine is still to be fully ascertained. These findings will further the understanding of the biological roles of CNTF-CNTFR and their relevant signaling pathways in human liver abnormalities, including metabolic diseases and the progression of liver cancer.

Acknowledgements

We thank Dr. Y. Gong from Shanghai Institute of Biochemistry for providing CNTF protein, Dr. S. Okada from Gunma University for providing Myc-GLUT4-eGFP plasmid, and Dr. K. Takimoto from University of Pittsburgh for providing constitutively active MEK1 (S218, 222E) and dominant-negative MEK1 (K97M) plasmids.

Ancillary

Advertisement