Grb2-associated binder-1 plays a central role in the hepatocyte growth factor enhancement of hepatoma growth inhibition by K vitamin analog compound 5

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Compound 5 (Cpd 5), a K vitamin analog, has been shown to inhibit Hep3B human hepatoma cell growth in cultures and rat hepatoma growth in vivo through prolonged epidermal growth factor receptor (EGFR)–extracellular response kinase (ERK) phosphorylation, and hepatocyte growth factor (HGF) synergizes with Cpd 5 to enhance the inhibition of Hep3B cell and rat hepatoma growth. To explore the mechanisms mediating the HGF/Cpd 5 synergy, we examined the possible involvement of the Grb2-associated binder-1 (Gab1) docking protein because it interacts with both EGFR and HGF receptor c-Met pathways. We found that HGF enhanced Cpd 5–induced c-Met phosphorylation at Tyr-1349, a binding site for Gab1, resulting in increased c-Met binding to Gab1, and induced strong and prolonged Gab1 tyrosine phosphorylation. Prolonged Gab1 phosphorylation by HGF/Cpd 5 in turn enhanced the ability of Gab1 to bind to protein tyrosine phosphatase SHP2 and enhanced the activation of its downstream mitogen-activated protein kinase pathway. In contrast, this same HGF/Cpd 5 treatment inhibited Gab1 binding to phosphatidylinositol 3-kinase (PI3K), leading to the inactivation of the PI3K-Akt pathway. The inhibition of Akt phosphorylation by HGF/Cpd 5 further activated the Raf-MEK-ERK signaling cascade via an Akt-Raf1 interaction, leading to strong and prolonged ERK phosphorylation. The transfection of Hep3B cells with mutated Gab1 (Gab1 Y627F), which had lost its ability to bind SHP2, antagonized HGF/Cpd 5–induced ERK phosphorylation, whereas the transfection of Hep3B cells with mutated Gab1 3YF, which lost its ability to bind PI3K, further enhanced HGF/Cpd 5–induced ERK phosphorylation and cell growth inhibition. Conclusion: Gab1 plays a central role in regulating HGF/Cpd 5 synergy in their actions on Hep3B cell growth inhibition. (HEPATOLOGY 2007;46:2003–2013.)

Hepatocyte growth factor (HGF) was originally identified as a potent mitogen for primary hepatocytes and other cell types1, 2 but is now widely recognized as a multifunctional protein.3 Biological responses of the cells to HGF are not confined to the stimulation of cell growth but include the enhancement of cell motility, the induction of morphogenesis, and the inhibition of tumor cell growth.3 Although HGF is well known as a strong promoter of cell survival, it has been found to induce apoptosis and to act as a cytotoxic factor that suppresses the growth of some tumor cell lines, including HepG2 human hepatoblastoma cells.4, 5 In addition, HGF has been shown to inhibit the growth of some hepatocellular carcinomas (HCCs) in vivo and c-myc–induced hepatocarcinogenesis in a transgenic mouse model coexpressing c-myc and hgf.6, 7 However, the molecular mechanisms by which HGF elicits opposing effects on cell survival and death remain poorly understood.

The multiple biological responses elicited by HGF are transmitted through the activation of its high-affinity receptor c-Met.8 The binding of HGF to c-Met induces its tyrosine phosphorylation in cells in which HGF acts as a mitogen, motogen, or growth inhibitor.9, 10 c-Met activation by HGF results in its autophosphorylation on specific tyrosine residues. The phosphorylation of 2 tyrosine residues (Tyr-1234 and Tyr-1235) located within the activation loop of the tyrosine kinase domain activates the intrinsic receptor kinase activity,10 whereas the phosphorylation of 2 other tyrosine residues in the c-Met C-terminal (Tyr-1349 and Tyr-1356) is required for all biological activities of the receptor.11 Phosphorylation at Tyr-1349 provides a direct binding site for Grb2-associated binder-1 (Gab1),12 which couples with multiple signaling proteins, including phosphatidylinositol 3-kinase (PI3K), phospholipase Cγ, and the tyrosine-specific phosphatase SHP2.13

Gab1 belongs to a newly identified group of docking proteins that function as specific substrates of tyrosine kinases. Following the binding of ligands, such as epidermal growth factor (EGF) or HGF, to their respective receptors, Gab1 associates with the activated growth factor receptor and is itself phosphorylated on specific tyrosine residues, in turn recruiting a series of SH2 domain–containing proteins that initiate intracellular signaling cascades. A number of growth factors, such as HGF and EGF, can induce Gab1 tyrosine phosphorylation, and it then recruits distinct sets of signal transducers to induce different responses. For example, HGF induces cell branching morphogenesis after binding to c-Met, whereas EGF induces cell growth stimulation after binding to epidermal growth factor receptor (EGFR).14, 15 However, the mechanisms by which HGF-induced or EGF-induced Gab1 phosphorylation can induce different downstream cellular effects, such as cell growth and morphogenesis, have not been elucidated. Because phosphopeptide maps of Gab1 following stimulation by either EGF or HGF are identical, the possible mechanisms by which Gab1 could modulate distinct biological signals are unclear. One possibility is that Gab1 is phosphorylated with distinct kinetics; that is, EGF induces transient Gab1 phosphorylation, which is related to cell growth, whereas HGF induces sustained Gab1 phosphorylation, which is responsible for cell branching morphogenesis.16, 17 Furthermore, EGF induces transient activation of extracellular response kinase (ERK), whereas HGF induces sustained ERK phosphorylation paralleling Gab1 phosphorylation; this suggests that Gab1 provides a link from receptor tyrosine kinase to the mitogen-activated protein kinase (MAPK) pathway.18

We have previously shown that vitamin K analog compound 5 (Cpd 5), or 2-(2-mercaptoethanol)-3-methyl-1,4-naphthoquinone, can inhibit Hep3B human hepatoma cell growth and that HGF enhances this inhibitory effect, which is related to strong and persistent activation of the MAPK pathway in comparison with cells treated with either HGF or Cpd 5 separately.19, 20 We found that the enhancement by HGF of Cpd 5–induced Hep3B cell growth inhibition was regulated by an interaction between the MAPK and PI3K-Akt pathways because a combined HGF/Cpd 5 treatment of Hep3B cells caused the inhibition of Akt phosphorylation, resulting in Raf-1 kinase activation.20 In this study, we explored the mechanisms by which HGF regulates the interaction between MAPK and PI3K-Akt pathways. We found that HGF enhanced Cpd 5–induced c-Met phosphorylation at Tyr-1349, leading to increased c-Met–Gab1 interaction and prolonged Gab1 phosphorylation at Tyr-627, which consequently increased its binding to SHP2 phosphatase and led to activation of the MAPK pathway. Moreover, Gab1 phosphorylation at Tyr-627 inhibited the PI3K-Akt pathway and further activated the MAPK pathway by removing the inhibitory effect of Akt on Raf-1 kinase activity.

Abbreviations

Cdc25A, cell division cycle 25 homolog A; cDNA, complementary DNA; Cpd 5, compound 5; DEN, N-nitrosodiethylamine; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; ERK, extracellular response kinase; Gab1, Grb2-associated binder-1; Gab1-WT, wild-type Grb2-associated binder-1; GST-pi, glutathione S-transferase-pi; HCC, hepatocellular carcinoma; HGF, hepatocyte growth factor; MAPK, mitogen-activated protein kinase; MEK, ERK kinase; MEM, minimal essential medium; pCMV, cytomegalovirus plasmid; PI3K, phosphatidylinositol 3-kinase; Shc, Src homology 2 domain–containing; siRNA, small interfering RNA; Tris, trishydroxymethylaminomethane.

Materials and Methods

Cell Culture.

The effects of HGF and Cpd 5 on cell growth inhibition in human hepatoma cell lines Hep3B, HepG2, and PLC/PRF/5 and in rat hepatoma JM-1 cells (a gift from Dr. G. Michalopoulos, University of Pittsburgh)21 were evaluated by the determination of the cell numbers. Cells were plated at a concentration of approximately 2 × 104 cells per well in 24-well plates. Twenty-four hours after the plating, the medium was replaced with fresh minimal essential medium (MEM) containing HGF (50 ng/mL) with Cpd 5 at the indicated concentrations. Three days after the treatment, the medium was removed, and the plates were stored at −80°C until the day of the assay. The cell number was estimated by a DNA fluorometric assay with the fluorochrome Hoechst 33258.22

In Vivo Transplantable Rat Hepatoma Growth and Treatment.

Under direct vision, 1 × 106 JM-1 cells per rat were injected into a mesenteric vein feeding into the portal vein. Rat liver HGF gene transfection was performed as described.23 Briefly, the recombinant human HGF expression plasmid (pCMV-HGF; a gift from Dr. Y. Liu, University of Pittsburgh) was diluted in a 10-mL saline solution at a concentration of 10 μg/mL and then injected through the tail vein in 8-10 seconds. The HGF plasmid was injected 1 day prior to JM-1 cell injection. The Cpd 5 treatment was started 1 day after the implantation of JM-1 cells. The in vivo JM-1 transplantable HCC experiment was divided into 4 groups, with 4 rats in each group: control (vehicle only), Cpd 5 (5 mg/kg of body weight), HGF plasmid (100 μg per rat), and combined HGF/Cpd 5 treatment. Two weeks after the first injection of Cpd 5, the animals were sacrificed, and the tumors were excised and weighed. The total weight of the tumors from each rat liver was used as the tumor volume.

In Vivo N-Nitrosodiethylamine (DEN)–Induced Preneoplastic Focus Formation and Treatment.

A stock solution of DEN at a concentration of 50 mg/mL was made in dimethyl sulfoxide. Rats were injected intraperitoneally with a single injection at a dose of 200 mg/kg of rat body weight. After 2 weeks of DEN injection, the rats were divided into 4 groups, with 4 rats in each group: control (vehicle only), HGF plasmid injection (100 μg per rat), Cpd 5 (5 mg/kg of body weight), and combined HGF/Cpd 5 treatment. Two weeks after the first injection of Cpd 5, the animals were sacrificed, and the liver sections were fixed in 10% formalin and immunostained with the glutathione S-transferase-pi (GST-pi) antibody with an ABC kit (Vector Labs, Burlingame, CA).

Immunoprecipitation and Western Blot Assay.

Hep3B cells were grown in MEM with 10% fetal bovine serum. After about 70% of confluence, the cells were serum-starved for 24 hours and treated with the vehicle, HGF (50 ng/mL), Cpd 5 (10 μM), or HGF/Cpd 5 for various times. After the treatment, the cells were harvested and then lysed in 100 μL of RIPA immunoprecipitation assay buffer [150 mM NaCl, 50 mM trishydroxymethylaminomethane (Tris)-HCl (pH 8.0), 0.1% sodium dodecyl sulfate, 1% Triton X-100, 1 mM orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 10 mg/mL leupeptin, and 10 mg/mL aprotinin]. Whole cell extracts were immunoprecipitated with the indicated antibody with protein A-agarose (Sigma, St. Louis, MO) overnight. The protein A-agarose pellets were washed 3 times with RIPA immunoprecipitation assay buffer and boiled 2 times in 40 μL of a sample buffer for 5 minutes, and the proteins were resolved through 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, western-blotted with the indicated antibody, and detected with enhanced chemiluminescence (Amersham, Arlington Heights, IL).

Preparation of the Gab1 Expression Plasmids and Gab1 Mutants.

The wild-type Grb2-associated binder-1 (Gab1-WT) complementary DNA (cDNA) was amplified by a polymerase chain reaction using cDNA from human tissues (Panomics, Fremont, CA) and oligonucleotide primers (5′-primer ATG AGC GGT GGT GAA GTG GTC and 3′-primer TCA TTT CAC ACT CTT CGC TGG CGT).22 The Gab1-WT cDNA was then subcloned into the pcDNA3 plasmid to generate pcDNA-Gab1-WT. Site-specific mutagenesis of Gab1 was performed with a Quick-Change II site-directed mutagenesis kit (Stratagene, La Jolla, CA). The mutagenic primer, 5′-CAA ACA GGT GGA ATT CTT AGA TCT CGA C-3′, was used to replace Tyr-627 with phenylalanine in the mutation of the SHP2 binding site (Gab1 Y627F). Three oligonucleotides—5′-CTG GAT GAA AAT TTC GTC CCA ATG AAT C-3′, 5′-CAG GAA GCA AAT TTT GTG CCA ATG ACT C-3′, and 5′-CAG TGA AGA GAA TTT TGT TCC CAT GAA CC-3′—were used to generate the triple Y447F/Y472F/Y589F mutations of the Gab1 PI3K binding sites (Gab1 3YF). All mutations were confirmed with DNA sequencing.

Transient Gab1 cDNA Transfection.

Transfections of Gab1-WT, Gab 1 Y627F, and Gab 1 3YF to Hep3B cells were carried out with the Lipofectamine method according to the manufacturer's instructions (Invitrogen Co., Grand Island, NY). Briefly, Hep3B cells (100,000 per well) were plated in 6-well plates and transfected with 2.0 μg per well of plasmid DNA in the Opti-MEM medium with the Lipofectamine 2000 reagent (Invitrogen). After 5 hours of transfection, the medium was replaced with complete growth medium, and the cells were allowed to recover for 48 hours. The cells were treated with the medium alone, HGF (50 ng/mL), Cpd 5 (10 μM), or HGF/Cpd 5 for 60 minutes. The cell lysates were then analyzed with a western blot or immunoprecipitation assay.

Gab1 Small Interfering RNA (siRNA) Transfection.

The Gab1 siRNA was purchased from Upstate (Waltham, NY). The Gab1 siRNA transfection of Hep3B cells was carried out by the Lipofectamine method according to the manufacturer's instructions (Invitrogen). Briefly, Hep3B cells (100,000 per well) were plated in 6-well plates and transfected with 200 pmol per well of Gab1 siRNA (the final concentration was 100 nM) in the Opti-MEM medium with the Lipofectamine 2000 reagent (Invitrogen). After 48 hours of transfection, the medium was replaced with serum-free MEM, and the cells were treated with the medium alone, HGF, Cpd 5, or HGF/Cpd 5 for an additional hour. The cell lysates were analyzed with a western blot using the anti–phospho-Gab1 (Tyr-627), anti–phospho-ERK, or anti-Gab1 antibody. The mock transfection was performed with nonspecific siRNA (Upstate).

PI3K Activity Assay.

The PI3K activity assay was performed with an enzyme-linked immunosorbent assay kit according to the manufacturer's instructions (Echelon Biosciences, Inc., Salt Lake City, UT) with some modifications. Briefly, Hep3B cells were grown to 70%-80% of confluence in MEM supplemented with 10% fetal bovine serum, which was then changed to serum-free MEM for 24 hours. The serum-starved cells were treated with the medium alone, HGF (50 ng/mL), Cpd 5 (10 μM), or HGF/Cpd 5 for 1 hour. After the treatment, the cells were washed with 10 mL of ice-cold buffer A [137 mM NaCl, 20 mM Tris-HCl (pH 7.4), 1 mM CaCl2, 1 mM MgCl2, and 0.1 mM sodium orthovanadate] 3 times and lysed in 1 mL of an ice-cold lysis buffer (buffer A plus 1% Nonidet P-40 and 1 mM phenylmethanesulfonylfluoride). The cell extracts were immunoprecipitated with 5 μL of the anti-PI3K antibody (Upstate Biotechnology) and 60 μL of a 50% slurry of protein A-agarose beads at 4°C overnight. The immunoprecipitated enzyme was then washed with buffer A plus 1% Nonidet P-40 3 times and with an LiCl buffer [0.1 M Tris-HCl (pH 7.4), 5 mM LiCl, and 0.1 mM sodium orthovanadate] 3 times, and this was followed by a TNE (Tris-NaCl-EDTA) buffer [10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM ethylene diamine tetraacetic acid, and 0.1 mM sodium orthovanadate] twice. The immunoprecipitated PI3K beads were resuspended in 5 μL of a 10× reaction buffer [25 mM MgCl2, 50 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (pH 7.0), and 250 μM adenosine triphosphate], 10 μL of a phosphatidylinositol (4,5)-bisphosphate (20 μM) substrate solution, and 35 μL of H2O for a total volume of 50 μL. This kinase reaction was maintained for 1-3 hours at room temperature, and then the enzyme activity was measured with the phosphatidylinositol (3,4,5)-triphosphate detector provided by the manufacturer.

Results

HGF Enhances Cpd 5–Induced Hepatoma Growth Inhibition in a Cell Culture.

We have previously shown that HGF enhances Cpd 5–induced Hep3B human hepatoma cell growth inhibition, which involves strong and prolonged ERK phosphorylation.20 To confirm that this HGF/Cpd 5 synergy is not cell line–specific, we examined the HGF/Cpd 5 cell growth inhibition effects on 2 other human hepatoma cell lines, HepG2 and PLC/PRF/5. We found that HGF enhanced Cpd 5–induced cell growth inhibition in HepG2 (Fig. 1B) and PLC/PRF/5 (Fig. 1C) cells in a pattern similar to that for Hep3B cells (Fig. 1A). Furthermore, HGF also enhanced Cpd 5–induced ERK phosphorylation in these 2 cell lines (Fig. 1D).

Figure 1.

HGF enhances Cpd 5–induced hepatoma cell growth inhibition in culture. (A) Hep3B, (B) HepG2, and (C) PLC/PRF/5 human hepatoma cells were plated at a concentration of 2 × 104 cells per well in 24-well plates. Twenty-four hours after the plating, the medium was replaced with fresh MEM containing Cpd 5 at the indicated concentrations (0-10 μM) in the presence or absence of HGF (50 ng/mL). After 3 days of culture, the cell number was estimated by a DNA fluorometric assay using the fluorochrome Hoechst 33258. The data represent the mean ± the standard deviation of 3 experiments, and each experiment was performed with triplicate dishes. (D) Hep3B, HepG2, and PLC/PRF/5 cells were treated with the control, HGF (50 ng/mL), Cpd 5 (10 μM), or HGF/Cpd 5 for 1 hour, and the cell lysates were immunoblotted with an anti–phospho-ERK antibody. ERK2 was used as the loading control.

HGF Enhances Cpd 5–Induced Hepatoma Growth Inhibition In Vivo.

To investigate whether HGF could synergize with Cpd 5 to inhibit hepatoma growth in vivo, we used a rat transplantable hepatoma model24 with JM-1 rat hepatoma cells, which were established from a transplantable hepatoma grown in male Fischer rats by a treatment with DEN. This model was particularly attractive because the JM-1 cells were also susceptible to growth inhibition by HGF and Cpd 5 in cultures (data not shown). We also used a carcinogen DEN-induced GST-pi foci model because GST-pi–expressing hepatocytes are thought to be possible precursors to preneoplastic foci and eventually liver tumors.25 We first examined the effects of the combined HGF/Cpd 5 treatment on rat transplantable hepatoma growth. Fischer F344 rats were injected with JM-1 cells (1 × 106 cells per rat), as described in the Materials and Methods section, and then the rats were divided into 4 groups, with 4 rats in each group: (1) vehicle only, (2) Cpd 5 only (5 mg/kg of body weight intraperitoneally every other day for 10 days), (3) HGF only (100 μg of HGF plasmids in 10 mL of saline was injected into the rat tail vein in less than 10 seconds 1 day prior to the JM-1 cell injection and the injection was repeated 1 week thereafter), and (4) combined HGF/Cpd 5 treatment (the same as the separate HGF and Cpd 5 treatments). Two weeks after the first injection of Cpd 5, the animals were sacrificed, and the tumors were excised and weighed. The total weight of the tumors from each rat liver was measured. We found that Cpd 5 at lower doses (5 mg/kg of body weight) slightly inhibited tumor growth in vivo (no statistical significance), and this was consistent with our previous report24; however, the combined HGF/Cpd 5 treatment had a synergistic effect on tumor growth inhibition, reducing the tumor volume by about 75% (Fig. 2A), and showed statistical significance in comparison with the Cpd 5 treatment alone (P < 0.05). We then examined the effects of HGF/Cpd5 on the expression of GST-pi hepatocytes induced by DEN as a possible marker of preneoplasia. Rats were injected intraperitoneally with a single dose of DEN at 200 mg/kg of rat body weight. After 2 weeks of recovery, the rats were divided into 4 groups, with 4 rats in each group, and the treatment was the same as that used for the transplantable hepatoma model. We found that GST-pi focus formation was more sensitive to the Cpd 5 treatment than a JM-1 transplantable hepatoma. Figure 2B shows that Cpd 5 alone significantly inhibited the number of GST-pi–expressing hepatocyte foci (P < 0.01), and HGF further enhanced the Cpd 5 inhibitory effects (P < 0.05). The data provide evidence that HGF can synergize with Cpd 5 to inhibit hepatoma growth both in vitro and in vivo.

Figure 2.

(A) HGF enhances Cpd 5–induced rat transplantable hepatoma growth inhibition in vivo. The in vivo JM-1 cell transplantable hepatoma experiment was divided into 4 groups, with 4 rats in each group: (1) vehicle only, (2) Cpd 5 only (5 mg/kg of body weight intraperitoneally every other day for 10 days), (3) HGF only (100 μg of HGF plasmids in 10 mL of saline was injected into the rat tail vein 1 day prior to the JM-1 cell injection, and the injection was repeated 1 week thereafter), and (4) combined HGF/Cpd 5 treatment (exactly the same as the separate HGF and Cpd 5 treatments). Two weeks after the first injection of Cpd 5, the animals were sacrificed, and the tumors were excised and weighed. The total weight of the tumors from each rat liver was used as the tumor volume. The results are expressed as the mean ± the standard deviation. The P value is for the comparison of HGF/Cpd 5–treated and Cpd 5–treated rats. (B) HGF enhances Cpd 5–induced inhibition of GST-pi focus formation. Rats were injected intraperitoneally with a single dose of DEN (200 mg/kg of body weight). After 2 weeks of recovery, the rats were divided into 4 groups, with 4 rats in each group, and treated as described in the Materials and Methods section. The animals were sacrificed 2 weeks after the start of the Cpd 5 treatment, and liver sections were immunostained for GST-pi. The number of stained foci per square centimeter was counted. The data are expressed as the mean ± standard deviation. The P value is for the comparison of Cpd 5–treated and control rats and HGF/Cpd 5–treated and Cpd 5–treated rats.

Gab1 Is Important in Regulating HGF/Cpd 5–Induced MAPK Activation.

Because we previously showed that the synergy between HGF and Cpd 5 in Hep3B cell growth inhibition is mediated by strong and prolonged ERK phosphorylation through an interaction between the PI3K-Akt and MAPK pathways,20 and because Gab1 has been implicated as a major link between a variety of growth factor receptors and both the MAPK and PI3K-Akt pathways,26 we examined the possibility of Gab1 involvement in this synergy. Gab1 siRNA was transfected into Hep3B cells to examine whether the suppression of Gab1 expression had any effects on HGF/Cpd 5-regulated ERK phosphorylation and cell growth inhibition. As shown in Fig. 3A,B, Gab1 siRNA transfection significantly suppressed Gab1 expression and almost completely abolished the effects of HGF, Cpd 5, and HGF/Cpd 5 on both Gab1 phosphorylation and ERK phosphorylation. Furthermore, HGF/Cpd 5–induced cell growth inhibition was also attenuated by Gab1 siRNA transfection (Fig. 3C). These results suggest that Gab1 plays an important role in regulating both HGF/Cpd 5–induced ERK phosphorylation and cell growth inhibition.

Figure 3.

The suppression of Gab1 expression antagonizes the HGF enhancement of Cpd 5–induced Gab1 and ERK phosphorylation and also antagonizes cell growth inhibition. (A) Hep3B cells were transfected with Gab1 or mock (unrelated) siRNA, and the cell lysates were immunoblotted with the anti-Gab1 antibody. β-Actin was used as the loading control. (B) Mock or Gab1 siRNA–transfected Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 for 1 hour, and the cell lysates were immunoblotted with the anti–phospho-Gab1 (Tyr-627) or anti–phospho-ERK antibody. ERK2 was used as the loading control. (C) Mock or Gab1 siRNA–transfected Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 for 3 days, and the cell number was estimated by a DNA fluorometric assay, as described in Fig. 1A. The data are expressed as the mean ± the standard deviation of 3 experiments, and each experiment was performed with triplicate 6-well dishes. The P values are for the comparison of Cpd 5–treated cells with the control and of HGF/Cpd 5–treated cells with Cpd 5–treated cells.

HGF Enhances Cpd 5–Induced Gab1 Tyrosine Phosphorylation.

It has been reported that Gab1 is tyrosine-phosphorylated after growth factor stimulation and that its tyrosine phosphorylation status regulates cell growth and morphology.27 We hypothesized that the involvement of Gab1 in HGF/Cpd 5–mediated Hep3B cell growth inhibition is related to its phosphorylation status. To study the effects of HGF and Cpd 5 on Gab1 tyrosine phosphorylation, Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 from 15 minutes to 3 hours. The cell lysates were immunoprecipitated with the anti-Gab1 antibody and then western-blotted with the anti–phospho-Gab1 (Tyr-627) antibody, which reacted with the Tyr-627 residue, the binding site of Gab1 for protein tyrosine phosphatase SHP2.28 As shown in Fig. 4, both HGF and Cpd 5 induced Gab1 phosphorylation at Tyr-627 in a time-dependent manner. However, the treatment of Hep3B cells with combined HGF/Cpd 5 resulted in very strong and prolonged Gab1 Tyr-627 phosphorylation. The total Gab1 tyrosine phosphorylation with the anti-tyrosine antibody showed similar results (data not shown).

Figure 4.

HGF enhances Cpd 5–induced Gab1 tyrosine phosphorylation. Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 from 15 minutes to 3 hours. The cell lysates were immunoprecipitated with the anti-Gab1 antibody and then western-blotted with the anti–phospho-Gab1 (Tyr-627) antibody. Gab1 was used as the loading control.

Our previously published study showed that Cpd 5 can induce EGFR tyrosine phosphorylation by inhibiting cell division cycle 25 homolog A (Cdc25A) activity,29 and because HGF is a c-Met kinase stimulant, we reasoned that the effects of a combined HGF/Cpd 5 treatment on Hep3B cells might be related to EGFR and/or c-Met phosphorylation and their interaction with Gab1. We first examined EGFR phosphorylation at Tyr-845 (in the activation loop of the kinase domain) and Tyr-1068 (the Gab1 and Grb2 binding site) and the EGFR-Gab1 interaction. We found that Cpd 5 induced EGFR phosphorylation at both tyrosine sites, but HGF had no effect on either site. Combined HGF/Cpd 5 did not show an HGF enhancement of Cpd 5–induced EGFR phosphorylation. A coimmunoprecipitation assay for EGFR and Gab1 showed that Cpd 5 induced the EGFR-Gab1 physical interaction, whereas HGF attenuated the Cpd 5–induced EGFR-Gab1 interaction (Fig. 5A). We then examined the effects of HGF, Cpd 5, and HGF/Cpd 5 on c-Met phosphorylation at Tyr-1234 (in the kinase activation site) and Tyr-1349 (in the Gab1 binding site) and the c-Met–Gab1 interaction. Figure 5B shows that only HGF induced c-Met Tyr-1234 phosphorylation, whereas both HGF and Cpd 5 induced c-Met Tyr-1349 phosphorylation. Moreover, the combined HGF/Cpd 5 treatment showed a synergistic effect on both c-Met Tyr-1349 phosphorylation and on the c-Met–Gab1 interaction in comparison with Cpd 5 alone (Fig. 5A,B). The data suggest that the combined HGF/Cpd 5 treatment enhanced the c-Met–Gab1 interaction, leading to Gab1 Tyr-627 phosphorylation.

Figure 5.

(A) HGF does not enhance Cpd 5–induced EGFR tyrosine phosphorylation and EGFR-Gab1 coimmunoprecipitation. Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 for 60 minutes. The cell lysates were immunoprecipitated with the anti-EGFR antibody and then immunoblotted with the anti–phospho-EGFR (Tyr-845 or Tyr-1068) or anti-Gab1 antibody. EGFR was used as the loading control. (B) HGF enhances Cpd 5–induced c-Met phosphorylation at Tyr-1349 and the c-Met–Gab1 interaction. Hep3B cells were treated as in part A, and the cell lysates were immunoprecipitated with the anti–c-Met antibody and immunoblotted with the anti–phospho-c-Met (Tyr-1234 or Tyr-1349) or anti-Gab1 antibody. c-Met was used as the loading control.

Gab1 Tyrosine Phosphorylation by HGF/Cpd 5 Activates the SHP2-MAPK Pathway.

It has been reported that protein tyrosine phosphatase SHP2 can be recruited to the c-Met receptor through the Gab1 docking protein, and this recruitment of SHP2 is related to HGF-stimulated sustained MAPK activation.30 To determine the relationship between HGF/Cpd 5–induced prolonged ERK activation and Gab1 phosphorylation, we examined the coimmunoprecipitation of Gab1 with SHP2 and its downstream target, the Src homology 2 domain–containing (Shc) protein. Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 for 1 hour, and the cell lysates were immunoprecipitated with the anti-Gab1 antibody and immunoblotted with anti-SHP2 or anti-Shc antibodies. Figure 6A shows that Cpd 5 induced Gab1 binding to either SHP2 or Shc, and HGF strongly enhanced the Cpd 5 effects. Furthermore, HGF also enhanced Cpd 5–induced SHP2, Shc, and ERK phosphorylation in regular western blot experiments (Fig. 6B).

Figure 6.

HGF/Cpd 5 induces the Gab1-SHP2 interaction and MAPK activation. (A) Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 for 60 minutes, and the cell lysates were immunoprecipitated with the anti-Gab1 antibody and immunoblotted with the anti-SHP2 or anti-Shc antibody. Gab1 was used as the loading control. (B) Hep3B cells were treated as in Fig. 5 (A), and the cell lysates were immunoblotted with the anti–phospho-SHP2, anti–phospho-Shc, or anti–phospho-ERK antibody. ERK2 was used as the loading control.

Gab1 Tyrosine Phosphorylation by HGF/Cpd 5 Inhibits PI3K Activity.

PI3K-Akt is another important pathway that is triggered by HGF via c-Met phosphorylation, and this has been reported to be involved in HGF-induced cell survival and antiapoptosis activity.31 Although PI3K can associate directly with c-Met via the multisubstrate binding site, more PI3K activity coimmunoprecipitates with Gab1 than with c-Met, and this suggests that PI3K associates primarily with Gab1.8 We therefore examined whether HGF/Cpd 5–induced Gab1 tyrosine phosphorylation might alter the Gab1-PI3K association and PI3K activity. We found that although both HGF and Cpd 5 individually inhibited PI3K binding to Gab1, the combined HGF/Cpd 5 treatment further inhibited the binding of PI3K to Gab1 (Fig. 7A). The measurement of PI3K activity with a PI3K enzymatic assay showed that the combined HGF/Cpd 5 treatment inhibited PI3K activity (Fig. 7B). The inhibition of Akt phosphorylation (Ser-473) by HGF/Cpd 5 (Fig. 7C) provided more evidence that HGF/Cpd 5–induced Gab1 tyrosine phosphorylation had an inhibitory effect on Gab1-PI3K interaction and on the PI3K downstream Akt kinase activity.

Figure 7.

HGF/Cpd 5 inhibits Gab1-PI3K interaction and PI3K-Akt activity. (A) Hep3B cells were treated with HGF, Cpd 5, or HGF/Cpd 5 for 60 minutes, and the cell lysates were immunoprecipitated with the anti-Gab1 antibody and immunoblotted with the anti-PI3K p85 antibody. Gab1 was used as the loading control. (B) Hep3B cells were treated as in part A, and the cell lysates were immunoblotted with the anti–phospho-Akt (Ser-473) or anti-Akt antibody. (C) Hep3B cells were treated as in part A, and a PI3K activity assay was performed with an enzyme-linked immunosorbent assay kit according to the manufacturer's instructions with some modifications (for details, see the Materials and Methods section). The results are expressed as the mean ± the standard deviation of 3 experiments. The P values are for the comparison of HGF/Cpd 5–treated cells with control and Cpd 5–treated cells.

Gab1 Tyr-627 Phosphorylation Plays a Central Role in Regulating HGF/Cpd 5–Induced MAPK Activation and Cell Growth Inhibition.

To further demonstrate the importance of Gab1 Tyr-627 phosphorylation, we transfected Gab1 Tyr-627 mutant (Y627F), wild-type Gab1, or empty pcDNA3 plasmids into Hep3B cells and then treated these cells with HGF, Cpd 5, or HGF/Cpd 5 for 1 hour (western blot) or 3 days (cell growth inhibition). Figure 8A shows that although wild-type Gab1 transfection enhanced the HGF/Cpd 5–induced Gab1-SHP2 and Shc interaction in comparison with pcDNA3 transfection, transfection with the Gab1 Tyr-627 mutant significantly attenuated this interaction. Furthermore, the Gab1 Tyr-627 mutant also antagonized the HGF/Cpd 5–induced inhibition of the Gab1-PI3K interaction by pcDNA3 and wild-type Gab1 transfection (Fig. 8A). Consequently, the Gab1 Tyr-627 mutation abolished HGF/Cpd 5–induced ERK phosphorylation and cell growth inhibition (Fig. 8B,C). We also generated Gab1 PI3K binding site mutations of Gab1 3YF (Y447F, Y472F, and Y589F), transfected them into Hep3B cells, and treated the cells with HGF, Cpd 5, or HGF/Cpd 5. We found that Gab1 3YF completely inhibited PI3K binding to Gab1 when a coimmunoprecipitation assay was used, but there were no effects on the Gab1-SHP2 interaction (data not shown). However, Gab1 3YF transfection strongly enhanced HGF/Cpd 5–induced ERK phosphorylation (Fig. 8B). Taken together, the data suggest that Gab1 Tyr-627 phosphorylation plays a central role in regulating HGF/Cpd 5–induced ERK phosphorylation and cell growth inhibition.

Figure 8.

(A) The Gab1 Tyr-627 mutation antagonizes the HGF/Cpd 5–altered Gab1-SHP2 and Gab1-PI3K interaction. Hep3B cells were transiently transfected with pcDNA3 empty plasmids or wild-type or mutated Gab1 Y627F cDNA and then treated with HGF, Cpd 5, or HGF/Cpd 5 for 60 minutes. The cell lysates were immunoprecipitated with the anti-Gab1 antibody and immunoblotted with the anti-SHP2, anti-Shc, or anti-PI3K p85 antibody. Gab1 was used as the loading control. (B) The Gab1 Tyr-627 mutation antagonizes HGF/Cpd 5–induced ERK phosphorylation, whereas the Gab1 PI3K binding site mutation (Gab1 3YF) enhances HGF/Cpd 5–induced ERK phosphorylation. Hep3B cells were transiently transfected with pcDNA3 or wild-type or mutated Gab1 (Gab1 Y627F or Gab1 3YF) and treated with HGF, Cpd 5, or HGF/Cpd 5 as described in part A. The cell lysates were immunoblotted with the anti–phospho-ERK antibody. ERK 2 was used as the control. (C) The Gab1 Tyr-627 mutation antagonizes HGF/Cpd 5–induced cell growth inhibition. Hep3B cells were transfected with pcDNA3 or wild-type or Gab1 Y627F as described in part A and treated with HGF (50 ng/mL), Cpd 5 (10 μM), or HGF/Cpd 5 for 3 days. The cell number was estimated as described in Fig. 1. The results represent the mean ± the standard deviation of 3 experiments. Each experiment was performed with triplicate 6-well dishes. The P values are for the comparison of Cpd 5–treated cells with HGF/Cpd 5–treated cells in mock and wild-type Gab1 transfections.

Discussion

Patients with HCC, a common solid tumor of humans, tolerate known chemotherapy poorly because the tumor usually arises on the basis of a liver damaged by chronic hepatitis. Therefore, nontoxic chemotherapy combined with a protective treatment for liver function is badly needed for HCC patients. HGF, a hepatocyte growth stimulant, seems to be an excellent candidate for combined therapeutic use in the treatment of HCC with chronic liver damage. We previously reported that Cpd 5, a vitamin K analog and protein phosphatase Cdc25A inhibitor, inhibits human hepatoma Hep3B cell growth, and HGF enhances this growth inhibitory effect. In this study, we applied this HGF/Cpd 5 combination therapeutic strategy to a transplantable rat hepatoma model and to DEN-induced preneoplastic GST-pi–positive foci, and we found that HGF also significantly enhanced the Cpd 5 inhibitory effect on tumor growth and GST-pi focus formation in vivo (Fig. 2). These data are encouraging because the combined HGF/Cpd 5 treatment not only enhanced Cpd 5–mediated hepatoma cell growth inhibition but also protected normal hepatocyte functions as well. Therefore, the HGF/Cpd 5 synergy may provide a better strategy for battling this deadly tumor.

Our current research focused on the mechanisms by which the HGF/Cpd 5 combination enhanced cell growth inhibition. We previously showed that the synergism of HGF/Cpd 5 in Hep3B cell growth inhibition was mediated by an interaction between MAPK and PI3K-Akt signaling pathways,20 but the upstream regulator had not yet been elucidated. Because Gab1 is a multisubstrate signaling adapter protein that links activated growth factor receptors to specific downstream signaling pathways, we surmised that it might be involved in regulating the HGF/Cpd 5 synergism. To examine this hypothesis, we suppressed Gab1 expression in Hep3B cells with a Gab1 siRNA transfection assay and then treated these cells with HGF, Cpd 5, or HGF/Cpd 5. We found that although HGF/Cpd 5 induced both Gab1 Tyr-627 and ERK phosphorylation in mock (unrelated siRNA) transfected cells, Gab1 siRNA transfection almost completely abolished HGF/Cpd 5 effects. Furthermore, HGF/Cpd 5–induced cell growth inhibition in mock transfected cells was also antagonized by Gab1 siRNA transfection (Fig. 3). These results provide evidence that Gab1 may play an important role in regulating the HGF enhancement of Cpd 5–induced cell growth inhibition.

A number of growth factors, including EGF and HGF, can induce Gab1 tyrosine phosphorylation and its association with several signaling transducers, such as SHP2 phosphatase, phospholipase Cγ, and PI3K. It has been previously reported that Gab1 becomes uniquely phosphorylated by different growth factors or cytokines, but phosphorylated Gab1 can recruit distinct sets of signal transducers to induce different cellular responses.8 One explanation for this paradox of a single phosphorylation causing different downstream effects is the duration of the Gab1 phosphorylation status. For instance, EGF induces transient Gab1 phosphorylation, which stimulates cell growth, whereas HGF induces prolonged Gab1 phosphorylation, which stimulates cell morphological changes.16 In our current study, we found that HGF/Cpd 5 induced very strong and prolonged Gab1 phosphorylation, which resulted in cell growth inhibition. It seems that enhanced c-Met phosphorylation at Tyr-1349 by HGF/Cpd 5 was most likely responsible for the prolonged Gab1 phosphorylation. This is reasonable because c-Met Tyr-1349 is a binding site for Gab1. Cpd 5 is a Cdc25A protein phosphatase inhibitor, and the inhibition of Cdc25A by Cpd 5 can induce strong and prolonged EGFR phosphorylation because EGFR has been shown to be a Cdc25A substrate.28 However, Cpd 5 also induced c-Met Tyr-1349 phosphorylation, and this suggested to us that Cdc25A might be a c-Met phosphatase as well. However, although both EGFR and c-Met can induce Gab1 phosphorylation, the fact that HGF did not activate EGFR and had no effect on Cpd 5–induced EGFR phosphorylation suggested that c-Met, not EGFR, regulated HGF/Cpd 5–induced Gab1 phosphorylation.

Gab1 has multiple tyrosine residues, some of which, if phosphorylated, provide potential binding sites for SH2 domain–containing proteins, such as SHP2 and PI3K. We found that HGF/Cpd 5 induced Gab1 phosphorylation at Tyr-627, a binding site for the SHP2 protein phosphatase. This resulted in increased Gab1 binding to SHP2 and Shc, leading to enhanced SHP2 and Shc phosphorylation and activation of the downstream ERK kinase. When Gab1 Tyr-627 was mutated to Gab1 Y627F, the induction of Gab1 binding to SHP2 and Shc by HGF/Cpd 5 was abolished, and ERK phosphorylation and cell growth inhibition were also antagonized (Fig. 8). Our results are consistent with the report that the recruitment of SHP2 to Gab1 is required for the sustained activation of the ERK pathway in HGF-stimulated Madin-Darby canine kidney cells.32 Interestingly, we also observed that HGF/Cpd 5–induced Gab1 Tyr-627 phosphorylation attenuated Gab1 binding to PI3K, resulting in the inhibition of PI3K activity and the subsequent inhibition of phosphorylation of its target protein Akt (Fig. 7). Gab1 Y627F transfection reversed this effect, and this implied that Gab1 Tyr-627 phosphorylation could inhibit PI3K activity. One possible mechanism is that Gab1 Tyr-627 phosphorylation activates protein phosphatase SHP2, which then dephosphorylates PI3K binding sites on Gab1, resulting in the inhibition of the Gab1-PI3K interaction.33 To confirm this hypothesis, we mutated 3 PI3K p85 binding sites on Gab1 (Gab1 3YF) and transfected this mutant into Hep3B cells. We found that Gab1 3YF transfection had no effects on HGF/Cpd 5–induced Gab1 Tyr-627 phosphorylation and its binding to SHP2, but it completely inhibited Gab1 binding to PI3K in HGF-treated, Cpd 5–treated, or HGF/Cpd 5–treated cells (data not shown). We found that HGF/Cpd 5 induced super-ERK phosphorylation in Gab1 3YF-transfected cells, which was comparable to that of Gab1 wild-type–transfected cells (Fig. 8B). Taken together, these data suggest that HGF/Cpd 5–induced Gab1 Tyr-627 phosphorylation has dual actions: one is to activate the SHP2-MAPK pathway, whereas the other is to inhibit the PI3K-Akt pathway. The inhibition of Akt kinase activity blocked Akt phosphorylation on Raf1 Ser-259, an inhibitory phosphorylation site for Raf1 activity, resulting in further MAPK activation and enhanced cell growth inhibition.20 These results are most easily explained and summarized in Fig. 9.

Figure 9.

Hypothesis of the mechanisms by which Gab1 mediates the HGF enhancement of Cpd 5–induced prolonged ERK phosphorylation and cell growth inhibition. (A) Cpd 5, an inhibitor of EGFR phosphatase Cdc25A, induces prolonged EGFR tyrosine phosphorylation and the EGFR-Gab1 interaction, resulting in prolonged Gab1 Tyr-627 phosphorylation and activation of the SHP2-MAPK pathway, which leads to cell growth inhibition. (B) HGF stimulates c-Met phosphorylation and the c-Met–Gab1 interaction, resulting in prolonged but mild Gab1 Tyr-627 phosphorylation and activation of the SHP2-MAPK pathway, which causes cell growth inhibition only detectable in a colony-forming assay.20 (C) The combined HGF/Cpd 5 treatment of Hep3B cells synergistically induces c-Met Tyr-1349 phosphorylation and the c-Met–Gab1 interaction, resulting in strong and prolonged Gab1 Tyr-627 phosphorylation and activation of the SHP2-MAPK pathway. Furthermore, strong and prolonged Gab1 Tyr-627 phosphorylation inhibits Gab1 binding to PI3K, leading to the inhibition of PI3K-Akt activity. The inhibition of Akt phosphorylation at Ser-473 removes its inhibitory effect on Raf-1 activation, further inducing MAPK activation and cell growth inhibition.

Ancillary