Hepatocyte growth factor induces Mcl-1 in primary human hepatocytes and inhibits CD95-mediated apoptosis via Akt

Authors


Abstract

CD95 (APO-1/Fas)-mediated apoptosis of hepatocytes plays a central role in the pathophysiology of various human liver diseases. Hepatocyte growth factor (HGF) was shown to exert antiapoptotic functions in rodent hepatocytes. We previously showed that primary human hepatocytes (PHH) are a valuable tool for the investigation of apoptotic processes in liver cells. In this study, we analyzed the influence of HGF on CD95-mediated apoptosis of PHH and its molecular determinants. HGF significantly inhibited CD95-mediated apoptosis of PHH as well as cleavage of caspase-8 and poly (ADP-ribose)polymerase. HGF transcriptionally induced the expression of the anti-apoptotic Bcl-2 family member myeloid cell leukemia-1 (Mcl-1). In contrary, HGF did not alter the expression levels of Bcl-2 or Bcl-xL. HGF activated survival pathways such as the phosphatidylinositol-3 kinase (PI3K)/Akt pathway, the mitogen-activated protein kinase/extracellular signal-regulated kinase (ERK) kinase/ERK and the signal transducer and activator of transcription 3 (STAT3) pathway. Notably, HGF triggered serine727—but not tyrosine705—phosphorylation of STAT3. Pretreatment of PHH with the PI3K inhibitor LY294002 as well as adenoviral transduction of dominant negative Akt1 prevented HGF-mediated Mcl-1 induction and reversed the antiapoptotic effects of HGF. In conclusion, HGF confers survival of PHH by activation of the PI3K/Akt pathway. PI3K/Akt activation by HGF results in the induction of antiapoptotic proteins such as Mcl-1. Thus, application of HGF may be a therapeutic approach to prevent CD95-mediated hepatocellular damage in human liver diseases. (HEPATOLOGY 2004;39:645–654.)

CD95 (APO-1/Fas) belongs to the subfamily of death receptors among the tumor necrosis factor/nerve growth factor receptor superfamily.1 CD95 plays an important role in liver homeostasis.2 Hepatocytes express high amounts of CD95 and are very sensitive toward CD95 triggering.3, 4 Mice injected with agonistic anti-CD95 antibody rapidly die of liver failure.5 CD95-mediated apoptosis is involved in a broad spectrum of human liver diseases, including acute liver failure.6In vivo silencing of the CD95 gene via small, interfering RNA protects mice from liver failure as well as from fibrosis in a model of autoimmune hepatitis.7 A key event of CD95 signaling is the formation of a multimeric complex of proteins called death-inducing signaling complex (DISC). Two different pathways are described downstream of CD95. In type I cells, the death signal is propagated by a cascade that is initiated by the activation of large amounts of caspase-8 at the DISC and subsequent activation of downstream caspases.8 In type II cells, including hepatocytes, DISC formation is weak, and the propagation of the apoptotic signal depends on its amplification via mitochondria.8, 9

Members of the Bcl-2 family of proteins play a pivotal role in mitochondrial integrity. Bcl-2 family members can be subdivided into antiapoptotic and proapoptotic proteins. Various antiapoptotic family members, such as Bcl-2, Bcl-xL, and myeloid cell leukemia-1 (Mcl-1), contribute to the stabilization of mitochondria, thus counteracting mitochondrial activation. Proapoptotic family members, such as Bid, Bax, or Bak, actively can initiate permeabilization of the mitochondrial outer membrane and can trigger release of effector molecules such as cytochrome c. As soon as cytochrome c is released into the cytosol, a multiprotein complex called the apoptosome is formed, inducing caspase-9 activation. In vivo studies comparing wild-type and Bid-deficient mice indicated that mitochondrial activation is a central event for apoptosis of hepatocytes.10

Hepatocyte growth factor (HGF) is a multifunctional polypeptide and has been shown to contribute to liver cell survival in previous studies.11, 12 The receptor for HGF, Met, is a 190kDa transmembrane tyrosine kinase. Met is highly expressed in hepatocytes and can modulate signaling events leading to apoptosis.13 Production of HGF is markedly increased on liver injury12 and has been described as stimulating liver regeneration by its direct mitogenic activity.14 Several studies have demonstrated an inhibitory effect of HGF on experimental hepatocellular injury in rodents both in vitro and in vivo.15–18 It has been reported that intraperitoneal administration of HGF abrogates CD95-induced hepatic failure in mice.15 The antiapoptotic mechanisms of HGF have not been well characterized. It has been suggested that the expression of Met in the liver leads to sequestration of CD95, and thus to protection from CD95-mediated apoptosis in transgenic mice.19 Studies of HGF and other growth factors indicate a role for phosphatidylinositol-3 kinase (PI3K)/Akt and mitogen-activated protein kinase/extracellular signal-regulated kinase (ERK) kinase (MEK)/ERK signaling in growth-dependent cell survival.20 The central downstream mediator of PI3K is the serine/threonine kinase Akt/protein kinase B.21 PI3K-dependent activation of Akt leads to phosphorylation of several proteins that are involved in the control of cell survival, such as Bad22 and Gsk-3β.23

However, HGF also exerts proapoptotic effects on hepatocytes. If exposed in high concentrations, HGF can result in dissociation of Met from death receptors, leading to a sensitization of hepatocytes to death receptor-mediated apoptosis.19 In addition, in transformed hepatic cells, HGF has been shown to trigger activation of JNK1 (c-jun N-terminal kinase), resulting in rapid DISC formation and activation of caspase-8 and caspase-3.24

We previously showed that primary human hepatocytes are a valuable tool for the investigation of apoptosis signaling events in hepatocytes.4 In this study, we analyzed the effect of HGF on CD95-induced apoptosis pathways in primary human hepatocytes. We demonstrate that HGF inhibits CD95-induced apoptosis of human hepatocytes via activation of the PI3K/Akt pathway, resulting in induction of antiapoptotic proteins such as Mcl-1.

Abbreviations

DISC, death-inducing signaling complex; HGF, hepatocyte growth factor; PI3K, phosphatidylinositol-3 kinase; MEK, mitogen-activated protein kinase/extracellular signal-regulated kinase (ERK) kinase; Mcl-1, myeloid cell leukemia-1; PHH, primary human hepatocytes; STAT, signal transducer and activator of transcription; PARP, poly (ADP-ribose)polymerase; ERK, extracellular regulated kinase; c-FLIP, cellular FLICE inhibitory protein.

Materials and Methods

Reagents and Cell Lines.

Anti-APO-1 is an agonistic monoclonal antibody (IgG3, κ), recognizing an epitope in the extracellular portion of CD95 (APO-1/Fas), and was cross-linked by addition of protein A (10 ng/mL; Sigma, Deisenhofen, Germany).25 HGF, LY294002, and PD98059 (solubilized in DMSO (dimethyl sulfoxide)) were purchased from Sigma, and the signal transducer and activator of transcription 3 (STAT3) inhibitor peptide (PYpLKTK-mts) was purchased from Calbiochem (Schwalbach, Germany). The human T cell leukemia line Jurkat (clone J16) was maintained in RPMI 1640 (Invitrogen Corporation, Karlsruhe, Germany) supplemented with 10% FCS (foetal bovine (Invitrogen)).

Isolation and Culture of Primary Human Hepatocytes (PHHs).

PHHs were isolated from fresh surgical specimens of patients undergoing partial hepatectomy. Informed consent was obtained from each patient, and the procedure was approved by the Ethics Committee, University of Heidelberg. To isolate PHH, a modified two-step collagenase perfusion was performed as described.4 Isolated PHH were seeded at a density of 1.0 to 1.5 × 105 viable cells/cm2 onto collagen-coated culture plates (collagen R, type I; Serva Biochemicals, Heidelberg, Germany). Cells were maintained in Williams medium E (Invitrogen) supplemented with L-glutamine (5 mM), glucose (0.06%), HEPES (23mM, pH 7.4), gentamicin (50 μg/mL), penicillin (50 IU/mL), streptomycin (50 μg/mL), DMSO (1.5%), inosine (37 μM), hydrocortisone (4.8 μg/mL), and insulin (1 μg/mL).

Adenovirus Infection of PHH.

On day 2 after seeding, PHH cultures were infected with replication-defective adenoviruses expressing dominant negative (AdAkt1AAA) and constitutively active (Ad-myrAkt1) forms of murine Akt1,26 (kindly provided by Kenneth Walsh, Boston, MA). Adenoviral constructs were grown and titered in Hek293-cells. The dominant negative mutant of Akt1 (Akt1AAA) cannot be phosphorylated because of alanine substitutions.26 The constitutively active Akt1 (myrAkt1) encodes wild-type Akt1 protein with a myristoylation signal sequence.

Viability Test.

Cell viability was determined by a colorimetric MTT ([3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay in flat-bottomed 96-well microtiter plates as described.27 PHHs were treated on day 3 after seeding as indicated. The viability of 100% was defined by the absorbance obtained from untreated cells.

Detection of Apoptosis.

PHH were seeded onto 24-well plates (2.5 × 105 cells/well) and treated as indicated. Cells were collected, washed, and resuspended in lysis buffer containing 0.1% (w/v) sodium citrate, 0.1% (v/v) Triton X-100, and 50 μg/mL propidium iodide (Sigma). After overnight incubation at 4°C, nuclei from apoptotic cells were quantified by flow cytometry according to the method of Nicoletti et al.28 Specific apoptosis is indicated and was calculated as follows: 100 × (experimental apoptosis − spontaneous apoptosis) / (100 − spontaneous apoptosis).

SybrGreen Real-Time Quantitative Polymerase Chain Reaction (PCR).

Total RNA from 1 × 107 PHH was extracted using TRIzol Reagent (Invitrogen). Four micrograms of total RNA were reverse transcribed and analyzed by real-time quantitative PCR using the SybrGreen PCR Core Kit (Perkin-Elmer Applied Biosystem, Foster City, CA), the GeneAmp 5700 sequence detection system (Perkin-Elmer Applied Biosystem), and the following primers: GAPDH: (glyceraldehyde-3-phosphate dehydrogenase, house keeping gene) FP, 5′-TCCACTGGCGTCTTCACCA-3′; RP, 5′-CCCCTGCAAATGAGCCC-3′; Mcl-1: FP, 5′-ACGGGACTGGCTAGTTAAACA AAG-3′; RP, 5′-AAGAACTCCACAAACC CATCC-3′. The relative increase in reporter fluorescent dye emission was monitored. The level of Mcl-1 mRNA, relative to GAPDH, was calculated using the formula: Relative mRNA expression = 2 − (CtMcl-1CtGAPDH), where Ct is defined as the number of the cycle in which SybrGreen emission exceeds an arbitrarily defined threshold.

Cell Lysis and Western Blot.

For cell lysis, 0.5 to 1 × 106 PHH were collected and washed with phosphate buffered saline. The cell pellet was resuspended in 100 μL RIPA lysis buffer (25 mM Tris, pH 7.4, 150 mM KCl, 5 mM ethylenediaminetetraacetic acid, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, protease, and phosphatase inhibitors). Lysis was performed on ice for 15 minutes, and cell debris was removed by centrifugation (14,000g; 4°C). Proteins were separated by 10% to 15% SDS (sodium dodecyl sulfate)-polyacrylamide gel electrophoresis and transferred to a Hybond ECL nitrocellulose membrane (Amersham Pharmacia Biotech, Freiburg, Germany). Immunodetection was performed using the indicated primary antibodies: anti-phospho-Tyr705 and anti-phospho-Ser727 STAT3, anti-phospho-Thr308 and anti-phospho-Ser473 Akt, anti-Akt1 (all rabbit; New England Biolabs, Frankfurt, Germany), mouse anti-phospho-ERK1/2 (New England Biolabs), mouse anti-STAT3, anti-poly (ADP-ribose)polymerase (PARP) and anti-ERK-1 (Becton Dickinson Biosciences, Heidelberg, Germany), rabbit anti-Bcl-xL (Transduction Laboratories, Lexington, KY), mouse antihemagglutinin (Roche, Mannheim, Germany), rabbit anti-Mcl-1 (Santa Cruz, Heidelberg, Germany), mouse anti-CD95 (3D5; Qbiogene-Alexis, Grünberg, Germany), mouse anti-α-Tubulin (clone B-5-1-2; Sigma-Aldrich), C15 (mouse IgG2b) recognizing the p18 subunit of caspase-829 and NF-6 (mouse IgG1) recognizing cellular FLICE inhibitory protein (c-FLIP).30

Data Analysis.

All results obtained by viability assay and apoptosis assay are expressed as mean (+ standard deviation). All assays (including triplicates and quadruplicates of viability and apoptosis assays) were performed with PHH cultures derived from one individual donor, respectively, and are representative for at least three independent liver cell preparations from three donors. Data were analyzed by Student's t test (paired, two sided) with P < .05 was considered significant.

Results

Influence of HGF on CD95-Mediated Apoptosis of Human Hepatocytes.

PHHs have been shown to be highly sensitive toward CD95-induced apoptosis.3, 4 First, we analyzed whether HGF influences CD95-mediated cell death of PHH (Fig. 1A). After 6 hours of treatment with the agonistic CD95-antibody anti-APO-1, cell viability of PHH was reduced significantly (Fig. 1A). Concomitant treatment with HGF (10 ng/mL) significantly inhibited cell death after 6 hours (Fig. 1A) and 12 hours (data not shown). To test whether cell death was the result of apoptosis, we determined specific apoptosis (Fig. 1B). After 6 hours of exposure of PHH to anti-APO-1, specific apoptosis was induced to 51% + 6.8% (Fig. 1B). Treatment with HGF significantly decreased anti-APO-1–induced apoptosis to 22% + 4.9% (P < .005; Fig. 1B). Treatment with higher concentrations of HGF (50 and 100 ng/mL) also significantly reduced apoptosis of PHH after activation of CD95 (data not shown). Reduction of apoptosis, however, was comparable with that of 10 ng/mL of HGF.

Figure 1.

HGF inhibits CD95-mediated cell death of PHHs. PHHs were treated with the agonistic CD95 antibody anti-APO-1 (0.1 μg/mL) for (A,B) 6 hours and (C) 3 hours with or without addition of recombinant HGF ((A,B) as indicated, (C) 10 ng/mL)) on day 3 after seeding. (A) Cell viability was measured by MTT assay and is shown relative to mock treated controls. Assays were performed in triplicates with cell cultures derived from one donor and are representative for five different liver cell preparations. Values are means + SD. (B) For detection of apoptosis propidium iodide staining and FACS (fluorescence-activated cell sorter) analysis were performed according to the method of Nicoletti et al.28 Assays were performed in quadruplicates with cell cultures derived from one donor. Values are means + SD. The mean spontaneous apoptosis rate was 34%. Results are representative for experiments with 10 independent liver cell preparations. Asterisk (*) indicates significant inhibition of anti-APO-1–induced apoptosis by HGF in experiments with 10 liver cell preparations (P < .001). (C) After treatment, cellular lysates were analyzed by Western blotting for processing of PARP (116 kd uncleaved, 85 kd cleaved). α-tubulin was stained to confirm equal loading of proteins; total protein per lane, 30 μg.

Cleavage of PARP by effector caspases constitutes a hallmark of apoptosis. Thus, we analyzed the influence of HGF on CD95-induced PARP cleavage as an additional independent indicator of apoptosis. After 3 hours of anti-APO-1 treatment, the cleaved fragment of PARP was detectable by Western blot. Concomitant treatment of PHH with HGF reduced cleavage of PARP almost to background levels (Fig. 1C).

To address the question of how HGF mediates its inhibitory effects on CD95-mediated apoptosis of PHH, we first determined whether HGF influenced CD95 expression in PHH under our experimental conditions. Treatment of PHH with HGF alone enhanced the expression of CD95 after 2 hours (Fig. 2A).

Figure 2.

Effect of HGF on CD95 expression and caspase-8 cleavage. (A) PHHs were treated with HGF (10 ng/mL) for the indicated time on day 3 after seeding. After treatment, total proteins were extracted and analyzed by Western blotting for the expression of CD95 (APO-1/Fas) and α-tubulin; total protein per lane, 30 μg. The histogram shows the ratios between the densities of the CD95 band to the corresponding α-tubulin band. (B) PHHs were treated with anti-APO-1 (0.1 μg/mL) for 3 hours with or without addition of HGF (10 ng/mL) or with protein A alone (*). Cellular lysates were analyzed for processing of caspase-8 (uncleaved, 55/53 kd; cleaved, 43/41 and 18 kd) and α-tubulin; total protein per lane, 45 μg. As positive control, the CD95-expressing cell line J16 was treated with anti-APO-1 for 1 hour. Western blots are representative for three independent experiments.

Pro-caspase-8 is the apical caspase in the CD95 pathway. Thus, we asked if HGF had an effect on caspase-8 activation after the triggering of CD95. In PHH cotreated with HGF, the caspase-8 cleavage products p43/41 and p18 were not detectable by Western blot in contrast to PHH treated with anti-APO-1 alone (Fig. 2B). These results suggest that HGF is capable of inhibiting caspase-8 cleavage.

Influence of HGF on the Expression of Anti-apoptotic Bcl-2 Family Members.

Hepatocytes have been characterized as CD95-type II cells.31 Thus, mitochondria are the central integrators of the apoptotic signal, and most pro-caspase-8 cleavage occurs after mitochondrial activation and release of cytochrome c. Because Bcl-2 family members critically regulate mitochondrial activation, we investigated whether HGF influences the expression of anti-apoptotic Bcl-2 family members. HGF did not alter the expression level of the Bcl-2 family members Bcl-2 and Bcl-xL (Fig. 3A,B), both of which have been shown to play a role in hepatocyte survival.15, 32 Next, we tested expression of the Bcl-2 family member Mcl-1.33 Already after 3 hours of HGF treatment, Mcl-1 induction was detectable in lysates of PHH (Fig. 3C). In addition, we tested, whether Mcl-1 was transcriptionally regulated by HGF. Mcl-1 mRNA expression in PHH was upregulated two- to 3.5-fold after exposure to HGF (Fig. 4).

Figure 3.

HGF induces Mcl-1, but not Bcl-2 and Bcl-xL, expression in PHH. (A–C) PHH were treated with HGF (10 ng/mL) for the indicated time on day 3 after seeding. After treatment, total proteins were extracted and analyzed by Western blotting for the expression of (A) Bcl-xL, (B) Bcl-2, and (C) Mcl-1 and α-tubulin; total protein per lane, 30 μg. Blots are representative for PHH cultures of six different donors. (C) The histogram shows the ratios between the densities of the Mcl-1 band to the corresponding α-tubulin band.

Figure 4.

HGF activates Mcl-1 mRNA expression in PHH. PHH obtained from three different liver tissue samples (donors 1–3) were treated with HGF (10 ng/mL) for 1.5 hours or left untreated. mRNA expression levels of Mcl-1 and GAPDH were analyzed by real-time quantitative PCR. Expression of Mcl-1 was normalized to GAPDH in each sample. Each PCR reaction was run in triplicates. Relative Mcl-1 expression was calculated as follows: 2 −(Ct of Mcl-1 in untreated PHH − Ct of GAPDH in untreated PHH)) / (2−(Ct of Mcl-1 in HGF treated PHH − Ct of GAPDH in HGF treated PHH)).

Activation of Survival Pathways by HGF and Relevance for Mcl-1 Expression.

Mcl-1 is a known antiapoptotic member of the Bcl-2 family and possibly is involved in liver cell survival.33 Having demonstrated that HGF induces Mcl-1 expression in PHH, we next investigated how Mcl-1 expression is regulated in PHH. The mechanisms contributing to regulation of Mcl-1 expression are controversially discussed and include activation of the PI3K/Akt pathway,34 the MEK/ERK pathway,35 and the Janus kinase/STAT3 pathway,36 depending on cell type. First, we tested whether HGF also activates these pathways in PHH. HGF treatment of PHH resulted in phosphorylation of Akt at both threonine308 and serine473 already after 1 hour (Fig. 5A). In addition, after 1 hour of exposure to HGF, phosphorylation of ERK was detectable in PHH (Fig. 5B). Phosphorylation of ERK was still detectable after 6 hours of continuous exposure to HGF.

Figure 5.

HGF activates survival kinase pathways in PHH. (A–C) PHHs were treated with HGF (10 ng/mL) or interleukin 6 (10 ng/mL) for the indicated time on day 3 after seeding. Whole cell lysates were prepared, separated, and immunoblotted with phosphospecific antibodies that recognize (A) the serine473 or threonine308 phosphorylated forms of Akt, (B) Erk1/2 (p42/44) dually phosphorylated at threonine202 and tyrosine204, and (C) the serine727 or tyrosine705 phosphorylated forms of STAT3. Immunoblotting also was performed with antibodies that recognize Akt, ERK-1/2, and STAT3, regardless of its phosphorylation state (“pan,” A–C). *HGF, 50 ng/mL.

Recently, serine phosphorylation of STAT3 has been postulated to be essential for Mcl-1 expression and survival in macrophages.36 STAT3 phosphorylation could be detected in PHH after 1 hour of treatment with HGF, decreasing after 4 hours (Fig. 5C). However, HGF treatment did not lead to phosphorylation of STAT3 at tyrosine705, but did lead to phosphorylation at serine727. In contrast, treatment with IL-6, which is a well-known inducer of the STAT3-signalling in liver cells, resulted in phosphorylation of both tyrosine705 and serine727 (Fig. 5C).

To explore further the HGF-mediated induction of Mcl-1 expression in PHH, we applied kinase inhibitors to block survival pathways activated by HGF. Inhibition of the MEK/ERK pathway by using the MEK inhibitor PD98059 did not block the increase in Mcl-1 expression after 3 hours of exposure to HGF (Fig. 6A), although PD98059 clearly reduced phosphorylation of ERK. In contrast, blockage of the PI3K/Akt pathway by the PI3K inhibitor LY294002 resulted in a profound inhibition of Mcl-1 induction on HGF treatment (Fig. 6B). Pretreatment of PHH with LY294002 (25 and 50 μM) completely eliminated phosphorylation of Akt on serine473 and suppressed phosphorylation on threonine308. To substantiate further the role of the PI3K/Akt pathway, we applied adenoviruses for transduction of dominant negative Akt (AdAkt1AAA) and dominant active Akt (Ad-myrAkt1). Infection of PHH with AdAkt1AAA for 24 hours prevented HGF-triggered induction of Mcl-1 (Fig. 6C). In addition, transduction of dominant active Akt1 increased the level of Mcl-1 expression in PHH (Fig. 6D). Infection of PHH was controlled by detection of hemagglutinin-tagged protein in Western blot (Fig. 6C,D).

Figure 6.

HGF induces antiapoptotic proteins dependent on PI3K activation. (A,B) PHHs were treated with HGF (10 ng/mL) for 3 hours on day 3 after seeding. Additional cells were pretreated with (A) the MEK inhibitor PD98059, (B) the PI3K inhibitor LY294002, and DMSO, respectively, with the indicated concentrations for 1 hour followed by HGF treatment. Total cell lysates were separated by SDS-PAGE (30 μg per lane) and then probed for Mcl-1. Additionally, (A) antibodies detecting ERK1/2 (p42/44) dually phosphorylated at threonine202 and tyrosine204, and (B) phosphospecific antibodies to the serine473 or threonine308 phosphorylated forms of Akt and the respective antibodies independent of the kinase's phosphorylation state were used. (C) PHHs were infected with adenoviruses containing the dominant negative mutant of Akt1 cDNA (Akt1AAA) on day 2 after seeding with the indicated multiplicity of infection. Twenty-four hours after adenovirus infection, cells were treated with HGF (10 ng/mL) for 3 hours or left untreated before total cellular lysates were obtained. (D) PHHs were infected with adenoviruses containing the constitutively active Akt1 cDNA (MyrAkt1). Twenty-four hours after adenovirus infection, total cellular lysates were obtained. (C,D) Lysates (30 μg per lane) were probed for Mcl-1, α-tubulin, and hemagglutinin. (E–G) Total cellular lysates of PHH were analyzed for the expression of c-FLIPS and c-FLIPL by Western blotting. (E) PHHs were treated with HGF (10 ng/mL) for the indicated time on day 3 after seeding. After treatment, total proteins were extracted. (F) PHHs were treated with HGF (10 ng/mL) for 6 hours on day 3 after seeding. Additional cells were pretreated with LY294002 and total proteins were extracted. (G) Cells were treated as in (D), and total cellular lysates were obtained. (E–G) Total protein per lane, 30 μg.

In previous studies, Akt activation has been shown to induce c-FLIP proteins, which are the most upstream regulators in the signaling cascade of CD95-mediated apoptosis.37 Because caspase-8 cleavage was inhibited by HGF treatment of PHH (Fig. 2B), we next tested whether HGF-induced PI3K/Akt activation alters c-FLIP expression in PHH. HGF induced c-FLIPS expression after 3 to 6 hours dependent on PI3K activity (Fig. 6E). In addition, blockage of PI3K by LY294002 as well as blockage of Akt by adenoviral transfer of dominant negative Akt reduced the basal level of c-FLIPL (Fig. 6F,G).

Crucial Role of the PI3K/Akt Pathway for the Antiapoptotic Effects of HGF.

We identified the PI3K/Akt pathway to be central for mediating HGF-induced Mcl-1 as well as c-FLIPS expression in PHH. To assess further whether this pathway is relevant not only for the induction of antiapoptotic proteins, but also for the antiapoptotic effects of HGF on human hepatocytes, we tested whether blocking of the pathway would interfere with the effect of HGF on CD95-mediated apoptosis. Interference with the PI3K/Akt pathway by pretreating PHH with the PI3K inhibitor LY294002 abolished the inhibitory effect of HGF on cell death induction after CD95 triggering. Six hours of exposure of PHH to anti-APO-1 alone resulted in specific apoptosis of 67.4%, cotreatment with HGF significantly decreased apoptosis to 41.5% (P < .02), and pretreatment with LY294002 (50 μM) reversed the effect of HGF (Fig. 7A). Treatment with LY294002 alone resulted in approximately 10% specific apoptosis and slightly sensitized PHH to treatment with anti-APO-1 (Fig. 7A). We also used adenoviral transduction of dnAkt1 to substantiate further the role of the PI3K/Akt pathway. Treatment of PHH with anti-APO-1 (0.1 μg/mL) for 6 hours resulted in specific apoptosis of nearly 70%, HGF cotreatment reduced apoptosis by approximately 30%, and transduction of dnAkt1 reversed the antiapoptotic effect of HGF (Fig. 7B). Treatment of PHH with AdAkt1AAA alone led to a slight induction of apoptosis (Fig. 7B), but did not sensitize PHH to CD95-mediated cell death (data not shown). By contrast, blockage of MEK by PD98059 and blockage of STAT3 by a STAT3 inhibitor peptide did not abolish the antiapoptotic effects of HGF (data not shown).

Figure 7.

Inhibition of PI3K/Akt pathway reverses the antiapoptotic HGF effect. (A) PHHs were treated with the agonistic CD95 antibody anti-APO-1 (0.1 μg/mL) for 6 hours with or without addition of HGF (10 ng/mL) and LY294002 (50 μM) on day 3 after seeding. Specific apoptosis was determined according to the method of Nicoletti et al.28 Assays were performed in quadruplicates with cell cultures derived from one donor. Mean values (+SD) are presented. (B) PHHs were infected with the adenovirus AdAkt1AAA containing the dominant negative mutant of Akt1 cDNA (Akt1AAA) on day 2 after seeding with the indicated multiplicity of infection or were left uninfected. After 24 hours, cells were treated with anti-APO-1 (0.1 μg/mL) for 6 hours with or without addition of HGF (10 ng/mL). Specific apoptosis was determined according to the method of Nicoletti et al. Assays were performed in triplicate with cell cultures derived from one donor. Mean values (+SD) are presented. Mean spontaneous apoptosis rate: (A) 37%, (B) 39%. Asterisks (*) indicate significant difference in cell viability and specific apoptosis, respectively, of PHH treated with anti-APO-1 and HGF compared with those treated with anti-APO-1 alone; (A) P < .02; (B) P < .00001. Results are representative for experiments with three independent liver cell preparations.

These results strongly suggest that the inhibitory effect of HGF on CD95-mediated apoptosis of PHH is dependent on activation of the PI3K/Akt pathway.

Discussion

The aim of the present study was to investigate the effect of HGF on CD95 (APO-1/Fas)-mediated apoptosis of human hepatocytes and its molecular determinants. The results of this study show for the first time that HGF inhibits CD95-mediated apoptosis of primary human hepatocytes. Moreover, we demonstrate that the antiapoptotic effect of HGF is entirely dependent on activation of the PI3K/Akt pathway.

Apoptosis of hepatocytes is thought to be a major cause of hepatocellular injury in a variety of liver diseases such as fulminant liver failure.38 Because the death receptor CD95 is constitutively expressed in healthy liver tissue, its level of expression and its contribution to the pathogenesis of liver diseases have raised considerable interest.3, 6 The results of this study demonstrate that HGF is capable of inhibiting CD95-mediated apoptosis of PHH at a concentration approximately 40-fold higher than the serum levels of healthy subjects (10 vs. 0.25 ng/mL).39 Growth factors such as HGF and epidermal growth factor have already been identified as inhibitors of apoptosis in the past, and HGF seemed to be a physiologically important survival factor for hepatocytes.11, 12 In rodents, HGF inhibits apoptosis of hepatocytes induced by CD95 triggering,15 tumor necrosis factor-α,40 bile acids,41 and endotoxins.42 Hepatocytes from mice expressing constitutively activated HGF receptors are virtually resistant to apoptosis.43 The antiapoptotic effect of HGF on human hepatocytes as shown in this study is in contrast with studies on human hepatoma cells and human hepatocellular carcinoma cells, which often are used as a substitute for primary hepatocytes in research. In a recent study, cotreatment with high concentrations of HGF (50 ng/mL) sensitized hepatoblastoma cells to CD95-mediated apoptosis already after 6 hours.19 In our study, HGF had antiapoptotic effects even in concentrations up to 100 ng/mL.

In the presented study, HGF influenced CD95-mediated apoptosis on the level of caspase-8 cleavage as well as on the level of PARP cleavage. Because CD95 expression was not significantly decreased after 6 hours but was increased after 2 hours of HGF treatment, it is unlikely that downregulation of CD95 on the cell surface was responsible for the antiapoptotic effects of HGF. One possible mechanism of inhibition of caspase-8 cleavage is the upregulation of c-FLIPS, which occurred after 3 to 6 hours of HGF exposure and was dependent on Akt activation. Akt dependent c-FLIP expression has already been shown in tumor cells.44

A direct effect of the HGF receptor c-Met on CD95 activation in mouse hepatocytes has been proposed recently; c-Met has been suggested to directly bind and sequester CD95, thus exhibiting antiapoptotic functions in hepatic cells.19 This mechanism cannot be excluded based on the data presented in this study.

HGF activated transcription of the Mcl-1 gene in PHH. On the protein level, Mcl-1 was upregulated after 3 hours of HGF treatment. This is in line with previous studies on myeloid cells, in which peak expression of Mcl-1 has been demonstrated after 3 hours of exposure to various differentiation stimuli.33 The role of Mcl-1 for HGF-induced survival of PHH has not been addressed so far. Our results suggest Mcl-1 to be a major player in mediating the antiapoptotic effects of HGF. First, hepatocytes behave like CD95-type II cells, in which propagation of the death signal, that is, triggering of CD95, is dependent on mitochondrial amplification.8 Because Mcl-1 is a member of the antiapoptotic Bcl-2 family and is expressed in the liver,33 it likely contributes to mitochondrial integrity of hepatocytes. In this study, we show that HGF-dependent and PI3K/Akt-dependent upregulation of Mcl-1 and interference with the PI3K/Akt pathway inhibited both HGF-mediated induction of Mcl-1 and survival. Thus, induction of Mcl-1 therefore is likely to account at least in part for the antiapoptotic effect of HGF on human hepatocytes. However, an effect of HGF on apical components of the CD95 pathway, such as c-FLIPS, also may contribute to HGF-mediated protection from CD95-mediated apoptosis.

Our results agree with the model of Mcl-1 as a Bcl-2 family member that can be induced rapidly for short-term viability enhancement.33 Because hepatocytes rapidly die on CD95 triggering (Fig. 1) after destabilization of mitochondria, protection from cell death requires fast induction of antiapoptotic Bcl-2 proteins. Mcl-1 could accomplish this role in human hepatocytes. This hypothesis also is supported by the observation that HGF did not upregulate either mRNA or protein expression of two other antiapoptotic Bcl-2 family members, Bcl-xL and Bcl-2, in PHH after 6 hours of exposure to HGF (Fig. 4). Bcl-2 and Bcl-xL both have been implicated in liver cell survival in previous studies.15, 32 In contrast to our results, in primary rodent hepatocytes, as well as in mice in vivo, HGF has been shown to upregulate expression of Bcl-xL after 6 hours.15, 42 Rapid upregulation of other antiapoptotic Bcl-2 family members such as A1/Bfl-1 also may contribute to hepatocyte survival, as suggested recently.45

We identified the PI3K/Akt pathway to be critical for HGF-induced Mcl-1 expression in human hepatocytes (Fig. 6). The cell type-specific mechanisms that contribute to the regulation of Mcl-1 are controversial. Mcl-1 expression has been described to be dependent on PI3K/Akt signaling,36 MEK/ERK signaling,35 Janus kinase/STAT signaling,46 and serine727-phosphorylated STAT3.36 In our study, inhibition of MEK/ERK signaling by PD98059 did not inhibit Mcl-1 induction significantly on HGF treatment of PHH. However, inhibition of PI3K/Akt signaling either by the PI3K inhibitor LY294002 or by adenoviral transduction of dnAkt1 blocked HGF-mediated Mcl-1 induction.

It remains speculative, however, which downstream targets of the PI3K/Akt pathway are responsible for the transcriptional activation of Mcl-1 in PHH. In addition, we cannot exclude an inhibitory effect of the PI3K/Akt pathway on proteasomal degradation of Mcl-1, which also may contribute to the increase of Mcl-1 expression.

In addition to activation of the PI3K/Akt and MEK/ERK pathways by HGF, we identified activation of the STAT3 pathway, indicated by phosphorylation of STAT3 at serine727, but not at tyrosine705. In previous studies on T cells as well as on macrophages, it has been shown that phosphorylation of the serine727—and the tyrosine705—residue of STAT3 are differentially regulated and that serine727 phosphorylation is critical for STAT3 DNA binding and transcriptional activity.36, 47 In macrophages, serine727—but not tyrosine705—phosphorylation was suggested to be essential for transcriptional activation of Mcl-1 and cell survival.36 It remains to be elucidated whether serine phosphorylation of STAT3 contributes to the induction of Mcl-1 and to the antiapoptotic HGF effects in PHH.

The inhibitory effect of HGF on CD95-mediated apoptosis of PHH was reversed both by the PI3K inhibitor LY294002 and by adenoviral transduction of dnAkt1. Thus, we conclude that the PI3K/Akt signaling pathway mediates HGF-induced survival of PHH. These findings are consistent with the proposed role of PI3K and Akt in growth factor-mediated survival in other studies.11, 21 A central role of the PI3K/Akt-pathway for hepatocyte survival also has been demonstrated for bile acid-induced apoptosis of rat hepatocytes41 as well as for aflatoxin-transformed hepatocytes, in which HGF fails to activate PI3K and therefore induces cell death via activation of JNK1.24 Other studies claimed a role of the MEK/ERK pathway in growth factor-mediated cell survival after CD95 triggering.48 However, the results of our study suggest that the protective effect of HGF against CD95-mediated apoptosis is not mediated via the MEK/ERK pathway, because inhibition of MEK did not prevent the cytoprotective HGF effect.

So far, we cannot exclude that other downstream targets of PI3K/Akt signaling, apart from Mcl-1 and c-FLIPS, contribute to the antiapoptotic HGF effects as well. Several substrates of the PI3K/Akt pathway have been discussed as mediating antiapoptotic effects, including Bad,22 NF-κB,49 and glycogen synthase kinase-3β.23 Alternatively, HGF-induced activation of Akt may promote survival independent of phosphorylation events.50

In summary, the present study shows for the first time that HGF inhibits CD95-mediated apoptosis of primary human hepatocytes. The antiapoptotic HGF effects are dependent on PI3K/Akt activation. HGF induces the expression of antiapoptotic proteins such as Mcl-1 and c-FLIPS via activation of PI3K/Akt. Therefore, HGF may be of therapeutic value for the treatment of liver diseases, such as fulminant liver failure, by eliciting inhibitory functions on CD95-mediated liver cell damage.

Acknowledgements

The authors thank Kenneth Walsh, Boston, MA, for providing adenoviral Akt constructs; Helmut Friess, Andreas Untergasser, and Ulrike Protzer for collaboration in hepatocyte isolation; and Cornelius Fritsch, Rüdiger Arnold, Binje Fleischer, and Sven Baumann for their help with the computer software and for critically reading the manuscript.

Ancillary

Advertisement