Potential conflict of interest: Nothing to report.
The FasL-Fas couple is a general death mediator whose activated signals lead to caspase-8 activation and apoptosis in adult hepatocytes. Suppression of caspase-8 activation and cell death is a protective mechanism modulated by the FLICE-Like Inhibitory Protein (FLIP). Although hepatocyte growth factor (HGF) and its receptor Met are known to mediate cell survival in developing livers, the molecular mechanisms involved in this process are poorly understood. We show here that Met activation by HGF impairs Fas-triggered apoptosis of primary embryonic hepatocytes and cell survival correlates with inhibition of caspase-8 and caspase-3 activities. Furthermore, we found that HGF treatment prevents degradation of FLIPL triggered by Fas activation. In contrast to this, Met activation does not modulate FLIPL levels and its stability in untreated cells, thus showing the specificity of this regulatory mechanism for embryonic hepatocyte survival. Knocking down FLIP expression abolishes the ability of Met to inhibit Fas-triggered hepatocyte death, demonstrating the functional requirement of FLIP in HGF anti-apoptotic signals. By combining genetic and pharmacological approaches, we also demonstrate that the PI3K-Akt pathway is required in embryonic hepatocytes to prevent Fas-triggered FLIP degradation and death. Thus, Met acting on PI3K and Akt ensures high levels of FLIPL, and disruption of this pathway contributes to hepatic apoptosis and possibly to Fas-related liver diseases. (HEPATOLOGY 2007;45:1210–1217.)
The commitment of a cell to life or death involves the integration of a complex network of signaling circuitry, including stress and survival signals. Extensive studies have underlined a wide variety of signaling components that establish the proper balance between survival and death of hepatocytes. Among extracellular signals, hepatocyte growth factor (HGF), epidermal growth factor (EGF), and transforming growth factor-alpha (TGF-α) have been shown to trigger hepatocyte survival in vitro.1, 2 Remarkably, the only ligand-receptor couple identified so far that regulates hepatocyte survival in vivo is HGF and its cognate receptor tyrosine kinase Met.3, 4 Both HGF and Met are expressed at high levels in developing liver. Inactivation of either hgf or met genes interferes with liver development and results in massive apoptosis of hepatocytes.3 Production of HGF is markedly increased after liver injury, leading to rapid activation of Met and liver regeneration. Recently, the in vivo relevance of HGF-Met signaling in liver regeneration has been genetically demonstrated.5, 6
The signaling capacity of the Met receptor is mainly confined to two phospho-tyrosine residues located in the carboxy-terminal tail.4, 7In vivo mutation of these tyrosine residues (Metd) leads to embryonic lethality,8–12 with developmental defects similar to those found in met or hgf knock-out mice.3 The link between activation of distinct signaling pathways and specific biological events has been investigated using the met-specificity-switch mutant mice, in which the signaling capacity of Met is restricted to given pathways (PI3K or Src).13 Phenotypic analysis of these mutants showed that, superimposed over a generic threshold signaling levels, distinct effectors are required to achieve specific biological functions.13, 14 Knocking out the multi-adapter protein Gab1, a downstream effector of Met, has been reported to lead to massive hepatocyte apoptosis similar to met null mutants.15 Interestingly, met-specificity-switch mutants retain proper Gab1 function but show massive apoptosis of hepatocytes in developing liver.13 These results indicate that although Gab1 contributes to Met-triggered hepatocyte survival, a more complex survival network is required for proper liver development.
In contrast to HGF-Met, Fas ligand (FasL)-Fas leads to activation of death signals in hepatocytes and consequently apoptosis.16 Fas activation is considered the cause of several liver diseases, including viral hepatitis, alcoholic liver disease, and hepatocellular carcinoma.17–19 Fas is expressed in normal mouse and human hepatocytes,16 and its activation by agonistic Fas antibodies leads to massive apoptosis with fulminant liver failure in mice.20, 21 Genetic studies on fas mutant mice have revealed hepatic hyperplasia, presumably caused by reduced hepatocyte apoptosis.22In vivo silencing of the fas gene using small interfering RNA protects mice from liver failure as well as from fibrosis in a model of autoimmune hepatitis.23 A key event in Fas signaling is the formation of the death-inducing signaling complex, a multimeric protein complex formed by Fas, Fas-associated with death domain (FADD), and proCaspase-8, leading to activation of Caspase-8 and subsequently of other death effectors.22 Fas-mediated death signals can be inhibited by FLICE Inhibitory Protein (FLIP),24 which is present in cells as a long splice form (FLIPL) and a short form (FLIPS). FLIPL contains two adjacent N-terminal death effector domains with high homology to the prodomains of caspase-8 followed by an inactive enzymatic domain, whereas FLIPS retains only the two death effector domains domains.24 The death effector domains are essential for FLIP function, because they prevent death-inducing signaling complex formation and caspase-8 activation.
HGF is able to prevent Fas-triggered death of adult hepatocytes in vitro,25 and to abrogate Fas-induced fulminant hepatic failure in adult mice.26 Conditional genetic studies have recently demonstrated that met mutant mice are hypersensitive to Fas-triggered apoptosis.5 Altogether, these results unambiguously demonstrated that Met is required for hepatocyte survival and tissue remodeling in mice exposed to Fas signaling. HGF prevents Fas-triggered hepatocyte death by transcriptionally inducing the Bcl2 family member Mcl-1 and FLIPS through the PI3K-Akt pathway.27 The inhibitory effect of HGF on Fas-triggered death also occurs in a transcriptionally independent manner.25 However, the molecular mechanisms triggered by Met in this process remain elusive. Moreover, whether proper HGF-Fas signaling balance is also crucial for embryonic hepatocyte survival remains to be determined.
In the current study, we explored the relationship between HGF and Fas signaling in regulating primary embryonic hepatocyte survival. By combining genetic and pharmacological approaches, we found that HGF blocks Fas-triggered caspase-8 activation and death by preventing FLIPL degradation. This event requires PI3K-AKt activation as its inhibition interferes with FLIPL regulation by HGF. In addition, FLIP loss-of-function abrogates the survival response of embryonic hepatocytes to HGF. These results show that Met can antagonise Fas-triggered death by impairing FLIPL degradation through PI3K-Akt.
The antibodies used were anti-FLIPL (R&D System and Alexis Biomedicals), anti-FLIPS (Stressgen Biotechnologies), anti-ERKs, anti-cleaved caspase-3, anti-phospho S473-Akt and S21/9-GSK3α/β (Cell Signaling), anti-tubulin (Sigma). HGF was purchased from R&D System, agonistic anti-mouse Fas antibody (Fas-Ab; Jo2) from Alexis Biomedicals. Actinomycin D (ActD, 0.2 μg/ml) and Cycloheximide (CHX, 20 μg/ml) were obtained from Sigma.
Mice and Primary Cultures of Mouse Embryonic Hepatocytes.
Generation and genotyping of metd/d, met2P/2P and met2S/2S knock-in mutants and preparation of embryonic hepatocyte primary cultures were as previously described.8, 13 For biochemistry, cells were starved for 16 hours before HGF stimulation. For survival assays, cells were starved for 16 hours, pre-treated with HGF for 6 hours, then incubated for another 24 hours in the presence of Fas-Ab plus ActD. Cell survival was assayed by MTT and reported as the percentage of absorbance with respect to cells treated with ActD. Alternatively, cell survival was estimated by following terminal deoxynucleotidyl transferase-mediated nick-end labeling (TUNEL)-positive cells (apoptotic cells) and by counting intact DAPI-positive nuclei (surviving cells). The pharmacological inhibitors used were LY294002 (PI3K inhibitor, Calbiochem), A-443654 (Akt inhibitor, kindly provided by V. L. Giranda, ABBOTT Laboratories), epoxomicin, and lactacystin (proteasome inhibitors, Sigma). No toxic effects were observed at the concentration used. For FLIP shRNA experiments, four transductions were performed on primary cells. Western blot and survival experiments were done after 3 days of culture.
Western Blot Analysis and Caspase-8 Activity Assay.
Western blot were performed as previously described.8, 13 For caspase-8 activity assay, cells were lysed in 10 mM Tris pH 7.5, 130 mM NaCl, 1% triton 100, 10 mM NaF. Ten micrograms total protein was incubated with caspase substrate IEDT-pNA (caspase-8 Assay Kit, Calbiochem). After 3 hours at 37°C, chromophore p-NA release was measured at 405 nm with an enzyme-linked immunosorbent assay reader.
Statistical comparisons were made using analysis of variance and a Fisher's post hoc comparison was used. Data are presented as mean ± SEM and P < 0.05 as statistically significant. For survival assays reported in Figs. 4 and 5, are shown as mean ± standard deviation (SD).
HGF Prevents Fas-induced Cell Death and Caspase-8 Activation on Primary Embryonic Hepatocytes.
Fas is highly expressed in liver from postnatal stages onward.16 Moreover, FasL treatment in combination with the transcriptional inhibitor ActD induces death of adult hepatocytes.28 We first tested whether fas is expressed in embryonic livers by reverse transcription PCR. Low levels of fas expression were detected at E13 (Fig. 1A), which corresponds to the embryonic stage where apoptosis of hepatocytes is already detectable in met mutants.8 Thereafter, fas levels increased to a peak in adult livers (Fig. 1A). Consistently, activation of Fas by treatment with Fas-Ab plus ActD led to the death of E15.5 primary embryonic hepatocytes (Fig. 1B). We next investigated whether HGF was able to block Fas-induced apoptosis by pre-treating cultures with HGF. Cell death triggered by Fas-Ab was impaired by HGF in a dose-dependent manner (Fig. 1B). In contrast, HGF had no effect in preventing Fas-Ab triggered death of embryonic hepatocytes derived from metd/d mutant embryos (Fig. 3B),8 thus genetically demonstrating the requirement of intact Met signaling to block Fas and promote cell survival.
Cell death triggered by Fas is mediated by the proteolytic action of caspases. We therefore tested the ability of HGF to prevent caspase-8 activation in embryonic hepatocytes. Caspase-8 activity increased by approximately 4-fold after 2 hours of Fas-Ab plus ActD treatment, and reached a 20-fold increase after 12 hours (Fig. 1C). Pretreatment of cells with HGF for 6 hours resulted in a 60% reduction of caspase-8 activity when compared with untreated cells (Fig. 1C). Consistently, lower levels of cleaved caspase-3 were found in HGF pre-treated cells (Fig. 1D). Together, these results show that HGF prevents death and caspase activation triggered by Fas in primary embryonic hepatocytes.
HGF Selectively Impairs FLIPL Degradation in Embryonic Hepatocytes Treated With Fas, Without Acting on Its Expression.
One known negative modulator of Fas-triggered caspase-8 activation is FLIP.24 Two forms of FLIP, FLIPS and FLIPL, are expressed in cells and known to control survival. Levels and functions of FLIPS and FLIPL can vary in a cell type–dependent manner.27, 29 FLIPL is the predominant form expressed in primary embryonic hepatocytes (data not shown), as has been previously reported for adult hepatocytes.27
FLIPL protein degradation is a key event in Fas-induced cell death. Consistently, FLIPL levels were slightly reduced on 4 hours' treatment with Fas-Ab plus ActD compared with controls, and drastically reduced after 8 hours (Fig. 2A). Degradation of FLIPL by Fas was partially inhibited by proteasome inhibitors (Supplementary Fig. 1; available at: http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html), as previously described.30 Thus, FLIP levels are in part controlled by the proteasome pathway.30 HGF pre-treatment impaired FLIPL degradation triggered by Fas (Fig. 2A). We next asked whether Met activation was also modulating FLIPL expression in cells not exposed to Fas-Ab. Primary embryonic hepatocytes were stimulated with HGF, and FLIPL levels were analyzed. In contrast to the effects on its degradation, HGF did not change levels of FLIPL (Fig. 2B), as has been previously reported in human adult hepatocytes.27 We also assayed whether HGF controls FLIPL protein stability by using CHX to block nascent translation. No significant difference was seen in the half-life of FLIPL after CHX treatment in cells exposed or not to HGF (Fig. 2C). Altogether, these results demonstrate that Met specifically acts on FLIPL protein levels by preventing its degradation triggered by Fas.
PI3K-Akt Is Required to Prevent Fas-triggered FLIPL Degradation by Met.
To uncover which signals downstream Met are required to prevent Fas-triggered FLIPL degradation, we made use of genetically modified embryonic hepatocytes derived from met2P/2P and met2S/2S mutants.13 In particular, PI3K or Src are activated in met2P/2P and met2S/2S hepatocytes, respectively, upon HGF stimulation. In contrast, the Ras-MEK-ERKs pathway is defective in both mutant cells.13 We reasoned that this signaling combination allows us to genetically address the necessity of PI3K or Src to prevent Fas-triggered FLIPL degradation with reduced Ras-MEK-ERKs pathway. Fas-Ab induced FLIPL degradation in wild-type and mutant cells (Fig. 3A). HGF prevented FLIPL degradation in met2P/2P, but not met2S/2S, mutant hepatocytes (Fig. 3A). HGF pretreatment also reduced Fas-triggered caspase-8 activity by approximately 50% in wild-type and met2P/2P, but not met2S/2S, cells (Fig. 3B). Consistent with this, Fas-induced death of met2P/2P embryonic hepatocytes was significantly reduced by HGF pre-treatment as in wild-type cells (Fig. 3C), but not in met2S/2S mutants (Fig. 3C).
The requirement of PI3K downstream from Met for FLIPL regulation and cell survival was further pharmacologically investigated using the PI3K inhibitor LY294002.14 Inhibition of PI3K abolished Akt phosphorylation and FLIPL regulation by HGF in both wild-type and met2P/2P hepatocytes (Fig. 4A). Moreover, LY294002 prevented the HGF survival effect in the presence of Fas-Ab (Fig. 4B,C). LY294002 did not affect FLIPL levels or cell survival in untreated cells (Fig. 4C and data not shown).
We next investigated whether PI3K acts through Akt to prevent FLIP degradation. A-443654 is known to specifically inhibit Akt activity in cultured cells,31 as confirmed following GSK3α/β phosphorylation (Fig. 4E). Inhibition of FLIPL degradation by Fas was prevented by A-443654 in a dose-dependent manner (Fig. 4D). Altogether, these data demonstrate that Met blocks Fas-triggered FLIPL degradation and embryonic hepatocyte death in a PI3K-Akt-dependent manner.
FLIP Is Required for the Anti-apoptotic Effects of HGF in Primary Embryonic Hepatocytes.
To unambiguously demonstrate that FLIP is essential downstream of Met to antagonize Fas-triggered death, we knocked down FLIP using RNA interference. Embryonic hepatocytes transduced by lentivirus expressing FLIP shRNA showed downregulation of FLIP protein levels (Fig. 5C). Therefore, survival assays were performed in the presence or not of Fas-Ab. We found that FLIP shRNA did not affect survival of untreated cells (Fig. 5A,B). In contrast, FLIP downregulation neutralized the HGF survival response in the presence of Fas-Ab (Fig. 5A,B). No effects on FLIP levels were observed with control lentivirus (scrambled sequence, data not shown).
Thus, these studies show that FLIP is an essential mediator of HGF-dependent survival of primary embryonic hepatocytes when exposed to Fas death signal.
The ability of HGF to prevent death of adult hepatocytes exposed to Fas is supported by a growing list of reports.1, 2 However, the signaling mechanisms activated by Met and involved in this process have not been fully identified. We show here that Met signals embryonic hepatocyte survival by acting on FLIPL, a negative modulator of death-inducing signaling complex formation. Our studies place FLIP regulation as an obligatory step in Met-triggered inhibition of Fas-dependent hepatocyte death. Principal support for this comes from pharmacological and genetics strategies, and lentivirus RNA interference. Met specifically regulates FLIPL by preventing its degradation triggered by Fas, without modulating its expression (Fig. 6). Knocking down FLIP is sufficient to prevent Met-survival properties in primary cells exposed to Fas. Interestingly, PI3K and its effector Akt are required to prevent Fas-triggered FLIPL degradation in embryonic hepatocytes (Fig. 6).
A novel Met-dependent mechanism to suppress Fas-mediated cell death in hepatocellular carcinoma cell lines has been described.32 The extracellular domain of Met can directly bind to and sequestrate Fas, thus preventing both Fas self-aggregation and FasL binding. This interaction blocks Fas-mediated activation of death signals and apoptosis.32 In met signaling mutant mice,8, 13 the intact extracellular domain of Met can allow the receptor to sequester and block Fas. However, mutant embryonic hepatocytes die when Fas is activated (Figs. 1B and 3C), and massive apoptosis has been reported in developing livers.8, 13 Thus, in addition to sequestering Fas, a specific signaling capacity of Met is required to prevent Fas-triggered apoptosis. Several cytoplasmic signals are known to modulate cell death induced by Fas.22 For example, inhibition of Fas-triggered hepatocyte death occurs in a transcriptional-dependent manner by inducing Mcl-1 and FLIPS expression.27 We did not observe changes in levels of other pro- and anti-apoptotic molecules, such as FADD, DAXX, LFG, Faim, Jun, Bim, Bak, Bax, Bcl-2, and Bcl-XL, in HGF-treated primary embryonic hepatocytes (data not shown). Moreover, HGF inhibition of Fas-triggered death also occurs when transcription/translation is blocked. In this context, Met signaling can still suppress death induced by Fas through inhibition of FLIPL degradation.
In the current study, we also show that Met makes use of PI3K to prevent FLIPL degradation triggered by Fas in primary embryonic hepatocytes. The use of the Akt specific inhibitor A-443654 shows that Akt is a key signal downstream of PI3K to prevent FLIPL degradation. Studies in human bronchial epithelial cell lines have led to the proposal that nitric oxide, which can be regulated by Akt,33 also can prevent FLIPL degradation induced by Fas.34 Further work will allow identification of the signaling mechanism by which Akt prevents FLIPL degradation. Moreover, determining to what extent this mechanism is used in other cell types such as motor neurons, which are also sensitive to Fas, will be interesting.
Hepatocytes have a very high replication capacity during both development and regeneration. Under these circumstances, with different networks concomitantly activated to support cell fate decisions, including proliferation, cell cycle arrest, differentiation, migration, and survival, co-ordination is crucial. Genetic studies contributed to the discovery of several cytoplasmic modulators of hepatocyte survival, underlining the complexity of signaling networks controlling liver development. HGF-Met is the only growth factor-receptor couple that has been identified so far as being able to regulate in vivo hepatocyte survival.3, 5, 8 By studying cell survival mechanisms in hepatocytes induced by Met, we have recently provided the first direct link between Met and p53 through a novel signaling mechanism involving Mdm2 regulation via the P13K-Akt-mTOR pathway.35 Interestingly, knocking out the p53 gene did not perturb liver development, although the expression of Fas was reduced in these mutants.36 Whether during liver development, alteration of survival signals leads to upregulation of Fas by p53 to accelerate death remains to be determined. Our results together with other in vivo data on hepatocyte survival strongly support the possibility that hepatocytes possess safeguard mechanisms tightly controlling the incoming signals. p53 and Fas, which are considered as guardians of coordinated signals, are possibly ready to mediate hepatocyte death when the signaling network functions inappropriately. In the absence of genetic damage or altered survival input, both p53 and Fas guardian signals are kept inert by Met.
In conclusion, FLIP regulation by Met through PI3K-Akt emerges as a key actor in both Met signaling pathways and the control of hepatocyte survival during development. As such, and given the involvement of Fas in pathogenesis of fulminant hepatic failure,20 modulation of this mechanism may be clinically important in liver diseases. Although HGF treatment has been considered as a putative therapy in several liver diseases,2 its application must be carefully evaluated for the side effects it might have on tumor formation and progression.37 Modulation of more specific signals preventing excessive Fas activation could have relevant clinical impact and be more selective in preventing hepatocyte death, reducing side effects possibly because of excessive activation of growth and survival signals.
The authors thank F. Helmbacher and F. Lamballe for helpful discussions; L. Bardouillet with animal care and genotyping; animal care people for their work in the animal house facility. K. Dudley is particularly acknowledged for discussion and comments on the manuscript.