fax: (49) 6221-423699
Liver Biology and Pathobiology
Transforming growth factor β can mediate apoptosis via the expression of TRAIL in human hepatoma cells†
Article first published online: 16 JUN 2005
Copyright © 2005 American Association for the Study of Liver Diseases
Volume 42, Issue 1, pages 183–192, July 2005
How to Cite
Herzer, K., Ganten, T. M., Schulze-Bergkamen, H., Grosse-Wilde, A., Koschny, R., Krammer, P. H. and Walczak, H. (2005), Transforming growth factor β can mediate apoptosis via the expression of TRAIL in human hepatoma cells. Hepatology, 42: 183–192. doi: 10.1002/hep.20757
Potential conflict of interest: Nothing to report.
- Issue published online: 16 JUN 2005
- Article first published online: 16 JUN 2005
- Manuscript Accepted: 21 APR 2005
- Manuscript Received: 18 JAN 2005
- BioFuture grant from the Bundesministerium für Bildung und Forschung
- Deutsche Krebshilfe
- Tumorzentrum Heidelberg/Mannheim
- Deutsche Forschungsgemeinschaft
Transforming growth factor β (TGF-β) has been shown to induce apoptotic cell death in normal and transformed hepatocytes. However, the exact mechanism through which TGF-β induces cell death is still unknown. We examined a potential role of various death receptor/ligand systems in TGF-β–induced apoptosis and identified the tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) as a mediator of TGF-β–induced apoptosis in hepatoma cells. TGF-β–induced apoptosis is significantly impaired upon blockage of TRAIL. We show that TRAIL is upregulated in hepatoma cells upon treatment with TGF-β, whereas TRAIL receptor levels remain unchanged. In conclusion, our results provide evidence that the TRAIL system is critically involved in TGF-β–induced cell death in liver pathology. (HEPATOLOGY 2005;42:183–192.)
Transforming growth factor β (TGF-β) is a multifunctional cytokine whose numerous cell and tissue activities include cell-cycle control, regulation of early development, differentiation, extracellular matrix formation, hematopoiesis, angiogenesis, chemotaxis, immune functions, and induction of apoptosis.1 The apoptosis-inducing capacity of TGF-β has been investigated in many cell types. In hepatocytes, TGF-β1 has been shown to inhibit cell proliferation and to induce apoptosis in vitro, and to control the excessive growth and maintenance of liver size in vivo.2 In addition, TGF-β1 has been demonstrated to trigger apoptotic cell death in human and rat hepatoma cell lines. Transgenic mice overexpressing TGF-β1 suffer from continuing apoptotic cell death of hepatocytes and consequently develop hepatic fibrosis.3, 4 Exogenous administration of TGF-β in rodents also results in a significant increase in hepatic cell death.5 These data strongly suggest that apoptosis induced by TGF-β may be involved in various hepatic lesions. Given its implication in these processes, it is important to understand the biochemical mechanism of initiation of TGF-β–induced apoptosis. So far, this issue has been addressed in several studies examining the role of different and known apoptosis signaling proteins. Analyses of TGF-β–induced cell death in hepatoma cell lines have confirmed that apoptosis is accompanied by the activation of caspases and cleavage of the caspase-3–specific substrate poly(ADP-ribose)polymerase (PARP). Analysis of the apoptotic process in primary rat hepatocytes has revealed that transcriptional activation is necessary for cell death induction. The application of cycloheximide, which blocks de novo protein synthesis, results in the inhibition of TGF-β–mediated apoptosis induction.6
However, it remains unclear which target genes are crucial for TGF-β–induced apoptosis. In that respect, the role of different death receptor/ligand systems in TGF-β–induced apoptosis has not been thoroughly investigated. Essentially, these are tumor necrosis factor (TNF) and the CD95 (FAS/APO-1) and TNF-related apoptosis-inducing ligand (TRAIL) systems. TRAIL has attracted attention for its ability to preferentially kill a wide variety of tumor cell lines,7, 8 whereas most normal cells are resistant to TRAIL, both in vitro and in vivo,9, 10 TRAIL interacts with five distinct receptors. TRAIL-R1 (DR4)11, 12 and TRAIL-R2 (KILLER/DR5)8 contain an intracellular death domain necessary for apoptosis induction upon TRAIL-mediated receptor ligation. TRAIL-R313 and TRAIL-R414 cannot mediate apoptosis because of complete or partial absence, respectively, of an intracellular death domain. TRAIL-R3 and TRAIL-R4 have been suggested to act as decoy receptors, because overexpression blocks TRAIL-mediated apoptosis. Osteoprotegerin is a soluble receptor reported to bind and thereby neutralize osteoprotegerin ligand and TRAIL.15
The biochemical events leading to apoptosis induction via TNF superfamily members have been analyzed in detail.16, 17 However, for quite some time after the identification of TRAIL and its receptors, the physiological function of this novel apoptosis-inducing system remained unknown. Recently, in several studies the functional expression of TRAIL was discovered on the surface of different cells that had previously been known to induce apoptosis in target cells via an unknown mechanism. Among them are type II interferon (IFN-γ) stimulated monocytes,18 cytomegalovirus-infected fibroblasts,19 type I IFN (IFN-α and IFN-β) or TCR-stimulated T cells20 (and our unpublished observation), nonstimulated CD4+ T cells,21 IFN-α– and IFN-γ–stimulated as well as measles virus–infected dendritic cells,22 and natural killer cells.23 Interestingly, functional surface expression of TRAIL was often associated with stimulation by cytokines. Therefore, it is likely that the antitumoral effect of certain cytokines may at least be partially mediated by TRAIL-induced direct killing of TRAIL-sensitive tumor cells.
The widespread expression of death receptors and TGF-β receptors, as well as the use of a common signaling pathway via caspases for apoptosis induction, suggests that these receptor signaling mechanisms might be linked. Moreover, both receptor/ligand systems play an important role in the development of liver disease. In this study, we show that TRAIL expression is upregulated upon exposure of liver cell lines to TGF-β and that TRAIL is a major contributor to apoptosis mediated by TGF-β on hepatoma cells.
Materials and Methods
Hep3b, Huh7, HepG2, Chang, and CEM cells were obtained from the American Type Culture Collection (Rockville, MD). The liver cell lines were maintained in Dulbecco's modified Eagle medium and CEM T cells in RPMI 1640 (GibcoBRL, Karlsruhe, Germany) supplemented with 10% heat-inactivated fetal calf serum (FCS) (GibcoBRL), 10 mmol/L HEPES (GibcoBRL), 5 mmol/L L-glutamine (GibcoBRL), and 100 μg/mL gentamycin (GibcoBRL) in 5% CO2. Before stimulation, the cells were cultured for 24 hours in Dulbecco's modified Eagle medium without FCS, followed by treatment with TGF-β and as indicated.
Isolation and Culture of Primary Human Hepatocytes.
Primary human hepatocytes were isolated from healthy liver tissue obtained from patients receiving partial liver resection with a two-step perfusion technique as described.24 Before stimulation, the cells were cultured for 24 hours in Dulbecco's modified Eagle medium without FCS, followed by treatment with TGF-β and as indicated.
Antibodies and Reagents.
The monoclonal antibodies specific for the different TRAIL receptors and TRAIL were described elsewere.25 We used anti-TRAIL-R1 HS101, anti-TRAIL-R2 HS201, anti-TRAIL-R3 HS301, and anti-TRAIL-R4 HS402, all obtained from Alexis (San Diego, CA). Anti-APO-1 monoclonal antibodies were used as previously described.26 The different receptor-Fc proteins are purified fusion proteins consisting of the extracellular domain of human TRAIL-R2, CD95, or TNF-R2 coupled to the constant region of human immunoglobulin (Ig) G1 (huTRAIL-R2-Fc, CD95-Fc, TNF-R2-Fc). The soluble death receptor-Fc fusion proteins TNF-R2-Fc, CD95-Fc,26 and TRAIL-R2-Fc8 bind to TNF/LT, CD95L, and TRAIL, respectively. Leucine Zipper TRAIL (LZ-TRAIL) is a stable trimer of human TRAIL and induces apoptosis upon binding to TRAIL-sensitive cells. In brief, recombinant proteins were purified from the supernatant of Cos-7 cells transfected with LZ-TRAIL-pCDNA3.1 or the respective Fc-protein pCDNA3.1. The purification was carried out as described.8
The monoclonal antibody (mAb) against PARP (clone c-2) was derived from Biomol (Hamburg, Germany). The mAb anti-caspase-8 C15 recognizes the p18 subunit of caspase-8 and is available from Alexis. Horseradish peroxidase–conjugated goat anti-mouse IgG polyclonal antibodies were obtained from Jackson Immuno Research (Dianova, Hamburg, Germany), and biotinylated secondary goat anti-mouse antibodies were obtained from Southern Biotechnology Associates (Birmingham, AL). TGF-β was obtained from Sigma (St. Louis, MO). All other chemicals used were of analytical grade and were purchased from Merck (Darmstadt, Germany).
Reverse-Transcriptase Polymerase Chain Reaction.
RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. After 10 minutes of denaturation at 70°C, 1 μg of RNA was incubated with oligo(dT)12-18 primer (Roche, Mannheim, Germany), dNTPs (Invitrogen), and 2 U M-MLVT RT (Invitrogen) at 42°C for 1 hour to generate complementary DNA. The complementary DNA was used as a template for PCR amplification of the following gene products: TRAIL (5′-ggt cca tgt cta tca agt gct c-3′ and 5′- gac gaa gag agt agt aac agc-3′; 603 bp), TRAIL-R1 (5′-tgt tgt tgc atc ggc tca ggt tgt-3′ and 5′-gag gcg ttc cgt cca gtt ttg ttg-3′; 491 bp), TRAIL-R2 (5′-gg ccc cac aac aaa aga ggtc-3′ and 5′-cag ccc cag gtc gtt gtg agc-3′; TRAIL-R2A:602 bp, TRAIL-R2B:515 bp), TRAIL-R3 (5′-acg gcg tcg gga acc ata cc-3′ and 5′-gct aca ctt ccg gca cat ctc tg-3′; 407 bp), TRAIL-R4 (5′-ccc ccg gca gga cga agt t-3′ and 5′-ctc ctc cgc tgc tgg ggt ttt-3′; 418 bp), β-actin (5′-gtg ggg cgc ccc agg cac ca-3′ and 5′-ctc ctt aat gtc acg cac gat ttc-3′; 484 bp). Five-microliter aliquots were amplified in a DNA thermocycler (PerkinElmer Gene Amp PCRSystem 9700; PerkinElmer, Weisbaden, Germany) with 0.5 U of Taq DNA polymerase (Sigma) in a 50-μL reaction. Thirty-five reaction cycles for TRAIL and 25 cycles for β-actin were performed. Each cycle consisted of a denaturation step (94°C for 30 seconds), an annealing step (55°C for 30 seconds), and an elongation step (72°C for 30 seconds). The reaction was completed with a 72°C elongation step for 10 minutes. PCR products were analyzed on 1.5% to 2% agarose gels and visualized via ethidium bromide staining.
Quantitative Real-Time PCR.
Quantitative PCR was performed using a Lightcycler System according to the manufacturer's instructions (Roche Diagnostics, Mannheim, Germany) using the primers as indicated in the previous paragraph.
Northern Blot Analysis.
Total RNA was isolated with TRIzol reagent (Invitrogen). RNAs were separated on a formaldehyde agarose gel, transferred to a nylon filter (Millipore, Billerica, MA) and hybridized with a probe corresponding to the 3′ untranslated region of the TRAIL complementary DNA or β-actin complementary DNA. The blot was washed with 1× SSC/2% SDS before autoradiography.
Cells were harvested and lysed and lysates were processed as previously described.27
Cells were rinsed twice with phosphate-buffered saline (PBS) 36 hours after stimulation, detached from culture dishes with PBS containing 20 mmol/L EDTA, washed, and resuspended in PBS supplemented with 5% FCS. Approximately 5 × 105 cells were used per sample. Incubations with primary and secondary antibodies were performed for 30 minutes at 4°C, followed by two washing steps with PBS/5% FCS after each incubation. Analysis was performed with a Becton Dickinson FACS flow cytometer (Franklin Lakes, NJ) and Cellquest software.
Cells were incubated in 50 μL of PBS/5% FCS with TRAIL-R2-Fc for TRAIL staining or mAbs against the four surface-expressed TRAIL receptors (all of the mIgG1 isotype), or, as the respective controls, in the presence of huIgG or mIgG (Southern Biotechnology Associates) at a concentration of 10 μg/mL. Biotinylated secondary goat anti-mouse or goat anti-human antibodies (Southern Biotechnology Associates) and Streptavidin-PE (Pharmingen, Hamburg, Germany) were used in a final volume of 50 μL at a dilution of 1:200.
Determination of Cell Death.
Cell-Mediated Cytotoxicity Assay.
Cell-mediated lysis was quantitated using a standard chromium-51 release assay.30 In brief, hepatoma cells were either not stimulated or stimulated with 2 ng/mL TGF-β (Sigma) for 30 to 40 hours, subsequently washed three times with PBS, detached with 2 mmol/L EDTA, and counted. CEM cells were labeled with 100 μCi/mL 51Cr for 1 hour in culture medium. Effector cells and target cells were coincubated for 16 hours. Spontaneous release was determined by incubating target cells alone; total release was determined by directly counting labeled cells. Percent cytotoxicity was calculated as follows: % specific lysis = (experimental cpm − spontaneous cpm/total cpm − spontaneous cpm) × 100. Duplicate measurements of four-step titrations of effector cells were used for all experiments.
To block the various death ligands, effector cells were incubated with CD95-Fc, TRAIL-R2-Fc or TNF-R2-Fc fusion proteins for 30 min prior to addition of target cells at a concentration of 10 μg/mL. Human IgG1 was used as an isotype-matched control. Mean values of % specific lysis ± SD were calculated from triplicate samples.
TGF-β–Induced Apoptosis Is Mediated by TRAIL in Hepatoma Cells.
TGF-β is known as an important physiological mediator of liver cell apoptosis both in vivo and in vitro and contributes importantly to a number of liver diseases.31 Furthermore, in certain liver cell lines, TGF-β was reported to act negatively on cell growth. We could confirm apoptosis induction upon TGF-β administration in Hep3b, Huh7, and Chang liver cells, whereas the amount of apoptosis in Hep3b cells was significantly lower than in the other sensitive hepatoma cell lines. HepG2 cells, as previously known, did not die upon TGF-β exposure, because they express a mutated receptor for TGF-β32 (Fig. 1A). To test a direct implication of death receptor/ligand systems in TGF-β–induced apoptosis, we investigated whether it is possible to interfere with TGF-β–induced apoptosis in Hep3b, Huh7, and Chang cells by employing death receptor Fc proteins. TRAIL-R2-Fc blocked about 60% of TGF-β–induced apoptosis after 36 hours, whereas CD95-Fc and TNF-R2-Fc did not show any significant effect (Fig. 1B), suggesting that TRAIL might play a role in TGF-β–induced apoptosis.
Because normal cells and several tumor cell lines show resistance toward TRAIL-induced apoptosis,10 we sought to define TRAIL sensitivity of the cell lines used. To this end, hepatoma cells were incubated with different doses of TRAIL for 24 hours. We could observe that Huh7, HepG2, and Chang cells died upon exposure to LZ-TRAIL, whereas Hep3b cells did not (Fig. 2A).
To compare the kinetics of TGF-β– and TRAIL-induced apoptosis, the four hepatoma cell lines were incubated for different periods in the presence of 100 ng/mL LZ-TRAIL (Fig. 2A). We observed that HepG2 cells are highly sensitive to TRAIL exposure but insensitive to TGF-β, whereas Hep3b cells are resistant to TRAIL-induced apoptosis and at the same time slightly sensitive to TGF-β–induced apoptosis. Chang and Huh7 cells were sensitive toward both TRAIL- and TGF-β–induced apoptosis. However, TGF-β–induced apoptosis occurred after 36 hours (Fig. 1A), whereas TRAIL induced comparable amounts of apoptosis in Huh7 and Chang cells after 18 hours (Fig. 2A). Thus, in all TRAIL-sensitive hepatoma cells, TRAIL-induced apoptosis showed an earlier onset (Fig. 2B) than TGF-β–induced apoptosis (Fig. 1A), suggesting that TGF-β–induced apoptosis might require de novo synthesis of apoptosis-mediating molecules to achieve sufficient amounts of cell death.
Apoptosis induced by TGF-β as well as by death receptor ligation is mediated by the cleavage of caspases. Therefore, we investigated whether these downstream effects can be influenced by blocking TRAIL. Huh7 cells were preincubated in the presence or absence of TRAIL-R2-Fc before stimulation with TGF-β for different periods. Our data show that caspase-8 and PARP cleavage are significantly reduced 36 hours after onset of exposure to TGF-β when TRAIL is blocked (Fig. 2C). Taken together, these results indicate that TRAIL is a crucial mediator of TGF-β–induced apoptosis in hepatoma cells.
Interestingly, there also seems to be a component of TGF-β–induced apoptosis (between 5% and 15%) that is independent of TRAIL. This part corresponds to the amount of apoptosis induced in TRAIL-resistant Hep3b cells and remains elusive.
TGF-β–Induced TRAIL Expression Is Dependent on Transcriptional Processes.
We next addressed the mechanism of TGF-β–mediated recruitment of the TRAIL system. The fact that the sensitivity to TRAIL was not influenced by TGF-β treatment suggests that it is the initiation rather then a sensitizing effect downstream of TRAIL receptor crosslinking that is affected by TGF-β. However, TGF-β could influence the expression of active TRAIL at various levels, either by inducing TRAIL-regulating molecules or by directly influencing TRAIL expression on the transcriptional or translational level. To test a potential role of TGF-β in TRAIL induction on the transcriptional level, we investigated an influence of TGF-β on TRAIL messenger RNA (mRNA) levels.
Quantitative RT-PCR of RNA isolated from all TGF-β–sensitive hepatoma cell lines used after treatment with TGF-β for various times demonstrated that TRAIL mRNA was rapidly induced in Huh7 and Hep3b cells after treatment with TGF-β (Fig. 3A). However, HepG2 cells did not show any increase in TRAIL expression on the mRNA (Fig. 3A) or protein level (Fig. 4A). To investigate whether the induction of TRAIL mRNA expression by TGF-β required de novo protein synthesis, we performed Northern blot analysis and assessed the effect of the protein synthesis inhibitor cycloheximide after treatment of Huh7 cells with TGF-β. Induction of TRAIL mRNA was identical whether or not cells were pretreated with cycloheximide before stimulation with TGF-β, indicating that protein expression before mRNA induction is not necessary and thus suggesting that the TRAIL promoter is a direct and primary target of TGF-β signaling (Fig. 3B).
For other TNF family members—most prominently for TNF itself—it has been shown that mRNA stabilization is involved in upregulation.33 Therefore, we investigated whether the upregulation of TRAIL mRNA by TGF-β was a result of mRNA stabilization by examining the degradation of TRAIL mRNA in the presence or absence of TGF-β. Huh7 cells were treated with TGF-β or medium for 12 hours and were chased in the presence of the transcriptional inhibitor actinomycin D for various periods. Northern blot analysis revealed that the degradation rate of TRAIL mRNA in cells treated with TGF-β resembled that seen in untreated cells (Fig. 3C). These results indicate that the effect of TGF-β on TRAIL mRNA accumulation does not occur at the level of mRNA degradation and imply that TGF-β signaling directly induces transcription of the TRAIL gene.
TGF-β–Induced TRAIL Expression Leads to Cytotoxicity.
To determine whether the induction of TRAIL mRNA leads to an upregulation of TRAIL on the cell surface, we performed FACS analysis of TRAIL expression. We found that 36 hours after TGF-β induction, TRAIL surface expression is substantially increased (Fig. 4A). TRAIL receptor levels remained unchanged upon TGF-β treatment, both at the levels of mRNA (Fig. 3A) and surface expression (data not shown).
To examine whether TGF-β–induced TRAIL expressed on the surface of hepatoma cells is functional, we determined whether TGF-β–treated hepatoma cells were capable of killing TRAIL-sensitive CEM target cells. We observed that about 40% of CEM cells were lysed by TGF-β–stimulated Huh7 cells, but not by unstimulated cells (Fig. 4B). To test whether the increased lysis of CEM cells was in fact mediated by TRAIL or by other death ligands, we preincubated the effector cells after TGF-β stimulation with different death receptor-Fc fusion proteins. In the presence of the TRAIL-R2-Fc—which is capable of blocking TRAIL—but not with CD95-Fc, nor with TNF-R2-Fc or an appropriate control (huIgG1 antibodies), specific lysis of CEM cells induced by TGF-β–treated Huh7 cells and TGF-β–treated Chang cells (Fig. 4C) was substantially inhibited.
These data demonstrate that TGF-β–induced TRAIL on hepatoma cells is functional and thus capable of inducing apoptosis in surrounding TRAIL-sensitive cells.
TGF-β Induces TRAIL Expression in Primary Human Hepatocytes.
Because TGF-β signaling and biological effects are often different in tumor cells compared with primary tissue, we investigated whether TGF-β also influences TRAIL expression on primary human hepatocytes (PHHs). PHHs were cultivated for 48 hours after preparation and thereafter treated with TGF-β for 24 hours. In line with results in hepatoma cells, upon incubation with TGF-β we observed an increase in TRAIL mRNA levels (Fig. 5A-B) and TRAIL expression on the surface of PHHs via FACS analysis (Fig. 5C). To investigate apoptosis induction by TGF-β, PHHs were cultivated for 48 hours after preparation and thereafter treated with various doses of TGF-β for 24 hours and 48 hours. We observed only a slight apoptosis-inducing effect of TGF-β on PHHs. Even when very high doses (20 ng/mL) of TGF-β were used, after 48 hours specific apoptosis did not exceed 20% according to subdiploid DNA content analysis (Fig. 5D) and 10% in a TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling) assay (data not shown). Whereas LZ-TRAIL was not able to induce apoptosis in PHHs, anti-APO-1 mAb caused 43% of specific cell death in PHHs after 24 hours of incubation. Interestingly, it was not possible to block the TGF-β–induced apoptosis of PHHs with TRAIL-R2-Fc or CD95-Fc (Fig. 5E).
Upon coincubation of TGF-β–treated or non–TGF-β–treated PHHs with CEM cells as targets cells, the CEM cells underwent apoptosis only in contact with TGF-β–treated effectors, as determined by subdiploid DNA content analysis (Fig. 5F). To test the specificity of this effect, we preincubated the PHH effector cells in the presence or absence of TRAIL-R2-Fc or CD95-Fc before stimulation with TGF-β and subsequent coincubation with CEM target cells. Lysis of CEM target cells was almost abolished in the presence of TRAIL-R2-Fc, whereas preincubation of TGF-β–stimulated cells with CD95-Fc did not result in apoptosis inhibition (Fig. 5F). These results are in line with reports postulating PHHs to be resistant to TRAIL-induced apoptosis and suggest the involvement of TGF-β–induced TRAIL in lysis of neighboring TRAIL-sensitive tumor cells.
The capacity of TGF-β to trigger apoptosis in a variety of cells and physiological conditions is well documented.42 However, the molecular mechanisms enabling this process have not been elucidated. In particular, death receptor/ligand systems as possible contributors in TGF-β–induced apoptosis have not been thoroughly investigated so far. In the present study, we show that TRAIL plays an important role in TGF-β–induced apoptosis in liver cells.
As a model system, we used different TGF-β–sensitive liver cell lines, as well as PHHs. To examine an involvement of the death receptor/ligand systems in this context, we used death receptor-Fc fusion proteins to block the respective ligands. Addition of TRAIL-R2-Fc could significantly impair TGF-β–induced apoptosis in the TRAIL-sensitive Huh7 and Chang cells, whereas CD95-Fc and TNF-R2-Fc did not show any inhibitory effect. These data indicate that the TRAIL receptor/ligand system is intimately involved in TGF-β–induced apoptosis. However, we also observed a component of TGF-β–induced apoptosis remaining after blocking with TRAIL-R2-Fc which thus is independent of TRAIL. Interestingly, this part corresponds to the amount of apoptosis induced in TRAIL-resistant Hep3b cells and remains elusive. PHHs—as well as Hep3b cells as a result of TRAIL-resistance—show significantly reduced sensitivity toward TGF-β. Nevertheless, TRAIL expression is also significantly increased in these cells upon TGF-β exposure. These results are in line with the fact that normal cells and several tumor cell lines show resistance toward TRAIL-induced apoptosis10 despite expressing TRAIL-R1 and TRAIL-R2 on the cell surface. The molecular basis for this differential sensitivity is not clear. Ectopic expression of cFLIPL and cFLIPS inhibited TRAIL-induced apoptosis by inhibition of the DISC35 and specific downregulation of cFLIP by small interfering RNA in hepatocellular carcinoma cell lines could sensitize hepatocellular carcinoma cells for TRAIL-induced apoptosis.29 In other studies, cFLIP expression levels did not directly correlate with sensitivity for TRAIL-induced apoptosis.36 Recently, it has been suggested that proteasome inhibitors sensitize for TRAIL-induced apoptosis in some cell types by reduced levels of Bcl-xL due to decreased nuclear factor κB activity or in other cell types by an increased release of Smac/DIABLO independent of nuclear factor κB.37, 38 It has been described that nuclear factor κB activation is responsible for upregulation of TRAIL-R1 and TRAIL-R2 in HeLa cells.38 In several studies it was shown that cotreatment with chemotherapeutic agents or irradiation resulted in sensitization of TRAIL-resistant tumor cell lines.29, 39 Our own data indicate that proteasome inhibitors upregulate TRAIL-R1, TRAIL-R2, and TRAIL-R4 in primarily TRAIL-resistant cell lines and sensitize these cells for TRAIL-induced apoptosis (Ganten et al., unpublished observations, 2004). The mechanisms governing TRAIL resistance versus sensitivity are diverse, somewhat controversial, and clearly cell type–specific. We tested whether TGF-β also exhibits a TRAIL-sensitizing effect in primarily TRAIL-resistant hepatoma cell lines and PHH, but could not detect any TGF-β–mediated increase in TRAIL-induced apoptosis in any of the cell lines tested and also not in PHHs (data not shown).
Upon treatment with TFG-β, apoptotic downstream events such as caspase-8 and PARP cleavage were significantly reduced when TRAIL was blocked. To test whether induction of TGF-β–induced apoptosis by blocking of TRAIL is reflected by a change in expression of TRAIL or TRAIL receptors upon TGF-β exposure, we analyzed TRAIL and TRAIL receptor mRNA levels and cell surface expression levels in Huh7 and Chang cells with and without stimulation with TGF-β. In both cell lines, a clear increase in TRAIL expression was detectable. The induction of TRAIL expression by TGF-β and the finding that induction of TRAIL mRNA does not require new protein synthesis indicate that TRAIL is a direct target of TGF-β signaling. Taken together, these data indicate that TRAIL is a target of TGF-β signaling and is a major contributor to TGF-β–induced apoptosis in TRAIL-sensitive hepatoma cells. In contrast, further analysis revealed that upon TGF-β stimulation, neither TRAIL receptor mRNA nor protein levels on the cell surface underwent any changes. In line with this result, TGF-β did not sensitize TRAIL-resistant cells to TRAIL-induced apoptosis.
Moreover, in contrast to unstimulated hepatoma cells, those stimulated with TGF-β showed a significantly higher capability to lyse TRAIL-sensitive CEM T cells. Upon blocking of TRAIL on hepatoma cells with TRAIL-R2-Fc, lysis of CEM cells was almost completely blocked, whereas CD95-Fc and TNF-R2-Fc were ineffective, showing TRAIL specificity of this effect. Thus, functional expression of TRAIL on hepatocellular carcinomas may not only cause the autocrine suicide of hepatoma cells but may also negatively affect the immune response by inducing apoptosis in invading TRAIL-sensitive immune cells, most importantly in T cells. This effect would be especially harmful when the tumor cells have acquired a TRAIL-resistant phenotype.
Hence, TGF-β–induced TRAIL expression on tumor cells that have acquired TRAIL resistance may even support tumor progression. However, when expressed on normal tissue, TRAIL has been shown to hamper tumor development and metastasis, because the absence of TRAIL has been shown to result in an increase in metastasis in the liver.40 Thus, TRAIL upregulation by TGF-β may serve diverse purposes, depending on the environment or the TRAIL sensitivity status of the tumor, respectively. Apoptosis induction by TGF-β in cultivated primary human liver cells has not been convincingly shown so far, whereas TGF-β sensitivity of certain human hepatoma cell lines as well as rodent liver cells is well established.41, 42 Our results show that, even after exposure to high doses of TGF-β, the fraction of apoptotic human liver cells detected by subdiploid DNA content analysis or TUNEL assay (data not shown) does not exceed 20%. Furthermore, apoptosis in PHHs could be abolished neither by TRAIL-R2-Fc nor any of the other death receptor-Fc proteins used. Nevertheless, following TGF-β exposure, TRAIL expression was clearly induced in PHHs on the mRNA level as well as the protein level. These results were reproducible in six separate preparations of PHHs from six different patients and clearly underline resistance of PHHs toward TRAIL-induced apoptosis. One of the appealing features of TRAIL as a proapoptotic ligand was that it did not show liver toxicity that had precluded the systemic use of the related death-inducing ligands CD95L and TNF-α. These ligands both cause massive hemorrhagic necrosis of various tissues including the liver.43 Although this effect has not been observed with TRAIL in diverse species from rodents to primates,10, 44 it had been indicated that human hepatocytes in culture might be responsive to TRAIL, which would predict TRAIL toxicity in humans.45 Nevertheless, later studies showed that the potential of TRAIL to induce apoptosis in human hepatocytes in vitro is dependent on the recombinant form of TRAIL used46 and that human hepatocytes transferred to SCID/Alb-uPA mice were insensitive to TRAIL in vivo.47
In conclusion, many types of tumor cells develop resistance to TGF-β–induced growth inhibition and concomitantly acquire the ability to release TGF-β.48 TRAIL resistance as well as TGF-β resistance may be acquired after long-term exposure or selective outgrowth of cells with such a resistance, providing a growth advantage in the course of tumor growth and dedifferentiation. In this context, it would be interesting if TRAIL becomes overexpressed in the course of chronic liver disease. However, this issue has not been thoroughly investigated so far and should be addressed in future work. TGF-β is considered to have differential effects during tumorigenesis, acting early as a tumor suppressor, but later stimulating tumor progression through its action on tumor cells and their microenvironment.49 Our data may help to explain this switch in the function of TGF-β. It is clear that loss-of-function mutations of components in the TGF-β signaling pathway is one route toward loss of TGF-β growth control in cancer.49, 50 However, the majority of human tumors has not undergone such mutations but retains a functional TGF-β signaling system. Our data suggest that the development of a TRAIL-resistant phenotype of tumor cells may be responsible for the development of resistance toward TGF-β–induced growth inhibition.
We thank D. Koppenhöfer, J. Gersbach, and M. Pach for excellent technical assistance; R. Arnold, A. Krueger, M. Li-Weber, and M. Sprick for helpful discussions and critical reading of the manuscript; and E. Klar and T. Lehnert for making human liver tissue available.
- 20Murine TRAIL (TNF-related apoptosis inducing ligand) expression induced by T cell activation is blocked by rapamycin, cyclosporin A, and inhibitors of phosphatidylinositol 3-kinase, protein kinase C, and protein tyrosine kinases: evidence for TRAIL induction via the T cell receptor signaling pathway. Exp Cell Res 1999; 252: 96-103., , , , .