Article first published online: 8 JAN 2013
Copyright © 2012 American Association for the Study of Liver Diseases
Volume 57, Issue 2, pages 625–636, February 2013
How to Cite
Wahl, K., Siegemund, M., Lehner, F., Vondran, F., Nüssler, A., Länger, F., Krech, T., Kontermann, R., Manns, M. P., Schulze-Osthoff, K., Pfizenmaier, K. and Bantel, H. (2013), Increased apoptosis induction in hepatocellular carcinoma by a novel tumor-targeted TRAIL fusion protein combined with bortezomib. Hepatology, 57: 625–636. doi: 10.1002/hep.26082
Potential conflict of interest: Drs. Kontermann and Pfizenmaier consult for and received grants from Biontech.
This study was supported by the Deutsche Forschungsgemeinschaft (SFB Transregio 77, SFB 685, and SFB 773) and BMBF (project no.: 0315-280A; FORSYS-Partner “Predictive Cancer Therapy”).
- Issue published online: 5 FEB 2013
- Article first published online: 8 JAN 2013
- Accepted manuscript online: 18 SEP 2012 06:12AM EST
- Manuscript Accepted: 4 SEP 2012
- Manuscript Received: 19 MAY 2012
As the result of an increasing incidence and a prevalent therapy resistance of hepatocellular carcinoma (HCC), there is a strong need for novel strategies to enhance treatment responses in HCC. Tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) has been proposed as a promising anticancer drug because it can selectively induce apoptosis in cancer cells, but not in healthy cells. Nevertheless, most tumor cells show TRAIL resistance, emphasizing the requirement for apoptosis-sensitizing agents and TRAIL molecules with improved tumor specificity. In this study, we employed a recombinant TRAIL molecule, in which three TRAIL protomers were expressed as a single polypeptide chain (scTRAIL), and a novel TRAIL variant, in which scTRAIL was additionally fused to an antibody fragment recognizing epidermal growth factor receptor (EGFR) to improve its HCC-targeting properties. We analyzed the proapoptotic effects of both TRAIL versions in combination with the proteasome inhibitor bortezomib (BZB) in hepatoma cells and primary human hepatocytes as well as in intact explants from HCC and healthy liver tissue. We demonstrate that EGFR-targeted TRAIL in combination with BZB induced significantly higher caspase activation and cell death in hepatoma cells, but not in primary hepatocytes. Importantly, when incubated with fresh liver explants, the combination of EGFR-targeted TRAIL and BZB displayed selective cytotoxicity for HCC, but not for tumor-free liver tissue, which could even be verified in liver explants from the same individuals. Unlike nontargeted TRAIL, EGFR-targeted TRAIL combined with BZB exerted no toxicity in liver tissues from nonalcoholic fatty liver disease patients. Conclusion: EGFR-targeted TRAIL reveals increased antitumor activity toward HCC without inducing toxicity to tumor-free liver tissue and might therefore represent a promising novel strategy for HCC treatment. (HEPATOLOGY 2013)
Hepatocellular carcinoma (HCC) is a global health problem with increasing incidence.1 In western countries, less than 50% of patients are eligible for potential curative treatment, including resection, transplantation, or local ablation. The limited therapeutic options and its resistance to systemic chemotherapy have triggered the search for molecular-targeted therapies for liver cancer. There is evidence of aberrant activation of several signaling cascades, such as the epidermal growth factor receptor (EGFR)/mitogen-activated protein kinase (MAPK) pathway, as well as apoptosis in HCC.2
Apoptosis is triggered by two major signaling routes, namely the extrinsic death receptor and the intrinsic mitochondrial pathway.3–6 Binding of death ligands, such as tumor necrosis factor (TNF)-α, TNF-related apoptosis-inducing ligand (TRAIL) or CD95L, to their respective receptors leads to death-inducing signaling complex formation, which results in receptor oligomerization and activation of initiator caspase-8 and caspase-10. Subsequently, initiator caspases activate effector caspases, such as caspase-3 and caspase-7. In certain cell types, such as hepatocytes, the extrinsic receptor pathway is amplified by the intrinsic mitochondrial pathway through the caspase-8-mediated cleavage of Bid, which, in concert with other B-cell lymphoma 2 (Bcl-2) proteins, initiates the release of mitochondrial proapoptotic mediators, followed by activation of initiator caspase-9 and downstream effector caspases.7
In contrast to CD95L or TNF-α, TRAIL has been shown to selectively induce apoptosis in transformed, but not healthy cells, making it a promising cancer-specific agent.4,8–10 Human TRAIL can bind to four receptors. TRAIL receptor (TRAIL-R)1 and TRAIL-R2 are proapoptotic receptors that contain a cytoplasmic death domain, which is required for the recruitment of initiator caspases, whereas TRAIL-R3 and TRAIL-R4 lack a functional death domain and are incapable of triggering caspase activation. The pivotal role of TRAIL in tumor defense is underlined by the observation that TRAIL-deficient mice are more susceptible to chemically induced as well as spontaneous tumors.11–13
TRAIL-R1/2 expression in healthy cells, including hepatocytes and quiescent stellate cells, is absent or relatively low and often cytoplasmic, instead of membrane bound.14–16 In contrast, in numerous cancers, including HCC, TRAIL-R1/2 protein expression is highly up-regulated.15,17–19 Whereas proapoptotic TRAIL-R1/2 could be detected in HCC tissues, the TRAIL decoy receptors, which lack proapoptotic activity, were significantly more lowly expressed in HCC, compared to nontumor liver tissues.20 Although there is evidence that the expression of TRAIL-R1/2 is required for TRAIL-mediated apoptosis, several studies indicate that the receptor abundance or relation between functional and nonfunctional TRAIL-Rs is not directly related to apoptosis susceptibility, suggesting additional determinants of TRAIL sensitivity.4–6,17 It has become clear that many primary tumor cells are resistant to apoptosis as a result of various molecular alterations.20,21 Thus, the combination of TRAIL with TRAIL-sensitizing agents might represent a promising strategy to overcome TRAIL resistance.
Proteasome inhibitors, such as bortezomib (BZB), have recently entered the clinic for treatment of hematological malignancies.22 Ongoing clinical studies include the combination of TRAIL or agonistic TRAIL-R antibodies (Abs) with BZB or other drugs. In vitro studies recently demonstrated that proteasome inhibitors sensitize HCC cells to TRAIL-induced apoptosis.23–25 However, cotreatment of HCC liver tissue with TRAIL and proteasome inhibitors might bear the risk of hepatotoxicity to nontumor liver tissue, because TRAIL/BZB treatment caused toxicity in primary human hepatocytes at high, but clinically relevant BZB concentrations.24 Thus, there is a strong need to enhance both the apoptotic activity and tumor selectivity of recombinant TRAIL molecules.
One promising strategy to improve tumor selectivity might be the fusion of TRAIL to Abs directed against tumor-associated antigens, such as EGFR, which is up-regulated in HCC.26,27 It has been recently shown that the extracellular domain of trimeric TRAIL can be genetically fused by peptide linkers to yield highly active single-chain TRAIL (scTRAIL). This TRAIL version can be further endowed with target-dependent activity by fusion to an Ab derivate recognizing tumor-associated antigens.28,29 In the present study, we investigated the proapoptotic effects of a novel scTRAIL fusion protein targeting EGFR, in combination with BZB, in primary hepatocytes and organ-like tissues from patients with HCC, nonalcoholic fatty liver disease (NAFLD), and healthy control individuals. We demonstrate that EGFR-targeted TRAIL in combination with BZB selectively induces apoptosis in HCC, but not in healthy or nontumor liver cells and is superior to nontargeted TRAIL.
Patients and Methods
Culture of Liver Explants, HCC Cells, and Primary Human Hepatocytes.
We investigated explants from freshly isolated healthy liver (non tumor-bearing portions) of patients (n = 8; mean age, 59.8 ± 4.5; male, 62.5%) who underwent partial hepatectomy because of single metastasis of nonhepatic origin and HCC explants (n = 12; mean age, 60.5 ± 3.1; male, 58.3%; mean grading, G 2.5 ± 0.2) from patients with cryptogenic liver cirrhosis (n = 6), hepatitis C virus (HCV) infection (n = 2), hemochromatosis (n = 2), alcoholic liver disease (n = 1), and NAFLD (n = 1). We also compared tumor-free liver tissues with HCC tissues of the same patients (n = 5; mean age, 58.4 ± 4.1; male, 60%; mean grading, G 2.8 ± 0.2). Additionally, we analyzed liver tissues from nontumor NAFLD patients without cirrhosis (n = 5; mean age, 38.0 ± 6.4; male, 80%). Explants were precisely cut into 125-mm3 cubes and incubated in 24-well plates with modified Eagle's medium (Invitrogen, Carlsbad, CA), supplemented with 1% human serum, 4 U/mL of insulin, 20 mM of HEPES, 2 mM of L-glutamine, 0.2 g/L of MgCl2 × 6 H2O, 1 × vitamin solution, 20 mg/L of L-ornithine HCl, 50 mg/L of ascorbic acid, 50 µg/mL of gentamycin, and 8 µg/mL of dexamethasone (DEX). Primary human hepatocytes (PHHs) were isolated as previously described30 and cultured for 36 hours in William's medium E (Invitrogen), supplemented with 1% penicillin/streptomycin, 2 mM of L-glutamine, and, additionally, with 10% fetal calf serum (FCS) and 100 nM of DEX for the first 12 hours. Huh7 cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen), supplemented with 1 g/L of glucose, 10% FCS, and 1% penicillin/streptomycin. PHHs, hepatoma cells, and liver tissues were incubated with 100 ng/mL of scTRAIL or αEGFR/scTRAIL for 6 hours. As a positive control for apoptosis induction, 100 ng/mL of Flag-tagged CD95L were used (provided by I. Schmitz, Braunschweig, Germany). TRAIL fusion proteins were prepared as described in the Supporting Materials. BZB (500 ng/mL; Selleck Chemicals, Houston, TX) was added 2 hours before incubation with the different TRAIL versions. The caspase inhibitor Q-VD-OPh (10 µM; MP Biomedicals, Illkirch-Cedex, France) or neutralizing TRAIL Ab (2E5, 1 µg/mL; Enzo Life Sciences, Lörrach, Germany) were added 3 hours before TRAIL incubation.
Measurement of Cell Viability and Caspase Activation.
Viability of PHH and Huh7 cells was determined by methyl thiazole tetrazolium (MTT) assay and crystal violet staining. Caspase activation was measured using the luminescent substrate assay31 and immunoblotting. Details are described in the Supporting Material.
Immunohistochemistry and Flow Cytometry.
Frozen sections of healthy and HCC liver explants or liver paraffin sections were stained for active caspase-3, cytokeratin-18 (CK-18) cleavage and EGFR expression or subjected to terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining, as described previously31 and in the Supporting Materials. For flow-cytometric detection of EGFR expression, PHH and Huh7 cells were stained with phycoerythrin-labeled anti-human EGFR (BioLegend, San Diego, CA).
Statistical analyses were performed using Mann-Whitney's U test (nonequal distribution) and the unpaired Student t test (equal distribution), respectively. Data are presented as means ± standard error of the mean (SEM). A P value <0.05 was considered significant.
EGFR-Targeted TRAIL Induces Enhanced Apoptotic Caspase Activation in HCC Cells, but Not Primary Human Hepatocytes, Which Is Further Increased by BZB.
We used new TRAIL fusion proteins in which three TRAIL protomers were expressed as a single-polypeptide chain (scTRAIL) that were further fused to a humanized single-chain Fv fragment of the anti-EGFR Ab, cetuximab (αEGFR-scTRAIL). In initial experiments, we investigated EGFR expression in liver cancer (Huh7) cells and PHHs by flow cytometry. We also compared EGFR expression in HCC to healthy liver tissues using immunohistochemistry (IHC). Almost no EGFR expression was found in PHH, whereas in Huh7 cells, EGFR was strongly up-regulated (Fig. 1A, B). Similarly, in healthy liver (n = 8), we found no EGFR expression, whereas HCC patients (n = 12) revealed strong EGFR expression on the cell membrane of tumor cells (Fig. 1C, D). This observation, in line with previous reports demonstrating increased EGFR expression in the majority of HCC tissues,27 therefore suggests that EGFR is a valid tumor target in HCC.
We next compared the apoptotic activity of nontargeted scTRAIL with the construct targeting human EGFR (αEGFR-scTRAIL). Because HCC cells, as with many solid tumor cells, reveal a weak TRAIL sensitivity, sensitizing agents, such as proteasome inhibitors, have been suggested to overcome TRAIL resistance.24 Therefore, we additionally analyzed the effects of both TRAIL proteins in combination with the proteasome inhibitor BZB in Huh7 HCC cells and PHHs. Initial dose-finding experiments revealed a concentration of 100 ng/mL of the two TRAIL proteins to be the most effective for inducing apoptotic caspase-3 activation, when combined with a nontoxic concentration of BZB (500 ng/mL). Compared to BZB alone, which showed almost no effect on caspase activity, scTRAIL significantly increased caspase-3 activation (5.21- ± 1.01-fold) in Huh7 cells, which was further enhanced by BZB (17.06- ± 2.34-fold; Fig. 2A). In contrast to HCC cells, no significant caspase-3 activity was induced by treatment of PHHs with either scTRAIL alone or in combination with BZB. Compared to scTRAIL, EGFR-targeted scTRAIL even more potently increased caspase-3 activity in HCC cells (6.24- ± 1.07-fold, compared to untreated control), which was most strongly enhanced by cotreatment with BZB (50.63- ± 13.97-fold, P < 0.01; Fig. 2B). Importantly, neither αEGFR-scTRAIL alone nor its combination with BZB significantly induced caspase-3 activation in PHHs (2.19- ± 0.76- and 1.88- ± 0.77-fold; Fig. 2B). In contrast, CD95L, which served as a positive control, induced strong caspase-3 activation in PHHs (38.87- ± 10.51-fold; Fig. 2C).
We then compared apoptosis induction by nontargeted and EGFR-targeted scTRAIL in the presence or absence of BZB. In HCC cells, no significant difference in caspase-3 activation between the two TRAIL proteins was found (Fig. 2D). However, EGFR-targeted scTRAIL plus BZB induced a significantly (P < 0.01) stronger increase of caspase-3 activity in HCC cells, compared to nontargeted scTRAIL and BZB (Fig. 2D). These data implicate that EGFR targeting increases TRAIL bioactivity in HCC cells after pretreatment with TRAIL sensitizers, such as BZB. To further verify our results, we performed immunoblot analyses for death receptor–mediated activation of the initiator caspase-8. As demonstrated by equal protein loading, lower levels of full-length caspase-8 were found in PHHs, compared to Huh-7 cells (Fig. 2E). In PHHs treated with the different TRAIL versions alone or in combination with BZB, we could detect the full-length, but not the cleaved and hence activated form of caspase-8. However, treatment of PHHs with CD95L induced caspase-8 activation. Unlike PHHs, Huh7 cells revealed caspase-8 cleavage after treatment with both TRAIL proteins, which was most strongly pronounced using a combination of αEGFR-scTRAIL and BZB (Fig. 2E).
EGFR-Targeted TRAIL Combined With BZB Is Cytotoxic for HCC Cells, but Not PHHs.
We additionally analyzed the viability of PHHs and Huh7 cells after treatment with the two scTRAIL proteins. We did not find decreased viability of PHHs treated with both scTRAIL proteins either alone or in combination with BZB, as measured by MTT assay (Fig. 3A). In contrast, treatment of Huh7 cells with both TRAIL versions plus BZB significantly decreased cell viability. In agreement with the caspase activation, treatment of Huh7 cells with EGFR-targeted scTRAIL in combination with BZB resulted in an almost complete loss of cell viability, which was not observed after treatment with a single agent alone (Fig. 3B). Similar results were obtained by measurements of cell viability using crystal violet staining (Fig. 3C). These results indicate that caspase activation and cell death are most efficiently triggered by EGFR-targeted scTRAIL in combination with a TRAIL-sensitizing agent, such as BZB.
Prevention of Apoptosis by EGFR-Targeted TRAIL by TRAIL-Neutralizing Ab and Caspase Inhibitor.
To prove that the observed caspase activation was indeed mediated by TRAIL, rather than by inhibition of EGFR signaling, we performed experiments in Huh7 cells using a TRAIL-neutralizing Ab and the pan-caspase inhibitor Q-VD-OPh. The TRAIL Ab completely prevented caspase activation induced by scTRAIL either alone or in combination with BZB (Fig. 4A). Comparable results were obtained in cells cotreated with EGFR-targeted scTRAIL and BZB (Fig. 4B). Furthermore, as assessed by immunoblotting, pretreatment of HCC cells with neutralizing TRAIL Ab completely inhibited the proteolytic processing of caspase-8 induced by the combination of BZB with either scTRAIL or EGFR-targeted scTRAIL (Fig. 4C). In addition, the inhibitor Q-VD-OPh prevented caspase-3 activation as well as caspase-8 cleavage after treatment of Huh7 cells with both scTRAIL proteins and BZB (Fig. 4D-F). Therefore, these results indicate that caspase activation induced by the different TRAIL versions requires TRAIL signaling, rather than inhibition of EGFR function.
Enhanced Apoptosis by EGFR-Targeted TRAIL in Combination With BZB in Intact HCC liver Tissues, but Not Healthy Liver.
Having demonstrated the superior activity of EGFR-targeted scTRAIL, we next compared the apoptosis-inducing effects of the scTRAIL proteins in intact, unfixed tissue explants from HCC and healthy livers by measuring caspase activation in liver tissue extracts. Combined treatment of HCC tissues with scTRAIL and BZB resulted in a moderate, but not significant increase in caspase-3 activation (3.64- ± 0.92-fold of untreated control; n = 8), compared to the single treatment with both agents alone (1.86- ± 0.64- and 2.92- ± 0.72-fold, respectively; Fig. 5A). In contrast, treatment of HCC tissues (n = 11) with EGFR-targeted scTRAIL and BZB significantly (P < 0.05) increased caspase-3 activation (10.57- ± 2.80-fold), compared to BZB or EGFR-targeted scTRAIL alone (3.53- ± 0.72- and 3.46- ± 0.87-fold; Fig. 5B). Similar to our observation in HCC cells, we found a significant (P < 0.05) increase of caspase-3 activity in HCC tissues treated with EGFR-targeted scTRAIL and BZB, compared to treatment with nontargeted scTRAIL and BZB. In contrast, no significant differences in caspase-3 activation were found between targeted and nontargeted scTRAIL treatment without BZB (Fig. 5C). Importantly, neither scTRAIL nor EGFR-targeted scTRAIL alone or in combination with BZB significantly increased caspase-3 activation in intact healthy liver tissues (n = 7; Fig. 5A, B).
To further support these results, we performed IHC analyses for caspase-3 activation and caspase-mediated CK-18 cleavage in HCC (n = 5) and healthy liver tissues (n = 5) after TRAIL and BZB treatment. Almost no caspase-3 activation was found in healthy liver tissues treated with scTRAIL or EGFR-targeted scTRAIL in the presence of BZB (Fig. 6A). In contrast, HCC liver tissues treated with EGFR-targeted scTRAIL and BZB revealed a higher number of active caspase-3-positive hepatocytes, compared to scTRAIL and BZB (Fig. 6A). In line with this, HCC tissues incubated with targeted scTRAIL and BZB also showed higher levels of caspase-cleaved CK-18, compared to HCC tissues treated with nontargeted scTRAIL and BZB, whereas no CK-18 fragments were found in healthy liver tissues treated with the respective agents (Fig. 6B).
To quantify the IHC results, cells positive for caspase-3 activation or CK-18 fragments were counted at ×400 magnification in four microscopic fields of the HCC liver explants (n = 3; Fig. 6C, D). Compared to untreated HCC tissues, treatment with BZB alone resulted in no significant increase of caspase-3 activation and CK-18 cleavage, and also scTRAIL combined with BZB induced neither a significant increase of caspase-3 activation (6.33% ± 0.51%; Fig. 6C) nor of CK-18 fragments (5.35% ± 0.48%; Fig. 6D), compared to treatment with scTRAIL alone. EGFR-targeted scTRAIL significantly (P < 0.01) induced caspase-3 activation (4.04% ± 0.03%), but not CK-18 cleavage (4.79% ± 0.43%) in HCC tissues, compared to untreated control (data not shown). In contrast to scTRAIL, combined treatment with αEGFR-scTRAIL and BZB resulted in a strong and significant (P < 0.01) increase of caspase-3 activation (17.03% ± 1.20%) and CK-18 cleavage (15.09% ± 1.18%), when compared to either the single treatment with both agents alone or the combined treatment with nontargeted scTRAIL and BZB (Fig. 6C, D).
To further verify our results obtained for caspase-3 activation and CK-18 cleavage, we performed TUNEL staining to detect cell death in the HCC explants (Fig. 7A). In line with the previous results, we found significantly (P < 0.01) increased cell death in HCC tissues (n = 3) treated with EGFR-targeted scTRAIL and bortezomib (27.21% ± 0.68% TUNEL-positive cells), compared to EGFR-targeted scTRAIL alone (5.86% ± 1.57%) or to scTRAIL and bortezomib (7.81% ± 0.75%). No significant difference between scTRAIL alone and scTRAIL in combination with BZB was observed (Fig. 7B). Thus, these data indicate that caspase activation induced by the respective TRAIL versions and BZB was indeed associated with cell death.
EGFR-Targeted scTRAIL in Combination With BZB Induces Apoptosis in HCC, but Not in Tumor-Free Liver Tissues.
In a previous study, we have shown that TRAIL exerts toxicity in inflamed liver tissues from patients with chronic HCV infection or nonalcoholic steatohepatitis.32 Therefore, we asked whether EGFR-targeted scTRAIL could be toxic not only to HCC liver, but also to the adjacent tumor-free diseased liver tissue. To this end, we first analyzed tumor-free liver and HCC tissues of the same patients (n = 5) and found a strongly increased EGFR expression in HCC tissue, compared to the respective tumor-free liver tissue (Fig. 8A). Then, we compared HCC and tumor-free cirrhotic tissues of the same patients after EGFR-targeted scTRAIL and BZB treatment. Interestingly, neither EGFR-targeted scTRAIL alone nor in combination with BZB induced significant caspase-3 activation in tumor free-liver tissues of HCC patients (Fig. 8B). In contrast, combined treatment with EGFR-targeted scTRAIL and BZB exclusively induced a significant (P < 0.05) increase of caspase-3 activation in HCC tissues, but not the respective tumor-free liver tissues (11.06- ± 3.92- versus 2.51- ± 0.83-fold increase; n = 5). Only slight, but nonsignificant differences were found when HCC and tumor-free tissues were analyzed for caspase-3 activation upon single treatment with EGFR-targeted scTRAIL (4.91- ± 1.63- and 2.44- ± 0.73-fold increase, compared to untreated control) or BZB alone (Fig. 8B). Thus, our results demonstrate that the combination of EGFR-targeted scTRAIL and BZB exerts antitumor activity in HCC tissues, but shows no or only marginal cytotoxicity in tumor-free liver tissues.
To further exclude a potential toxicity of EGFR-targeted scTRAIL in the inflamed liver, we performed IHC for caspase-generated CK-18 fragments and cell death (TUNEL reactivity) in liver tissues from patients with NAFLD (n = 5; Fig. 8C, D) treated with BZB together with EGFR-targeted scTRAIL or scTRAIL. EGFR-targeted scTRAIL plus BZB induced almost no caspase-mediated CK-18 cleavage (2.59- ± 0.23-fold increase, compared to untreated control), whereas treatment with nontargeted scTRAIL and BZB resulted in significantly (P < 0.01) higher CK-18 fragment levels (8.69- ± 0.75-fold increase; Fig. 8C). Similarly, treatment with nontargeted scTRAIL and BZB induced a significantly (P < 0.01) higher cell-death rate (6.98- ± 1.00-fold increase, compared to untreated control), compared to EGFR-targeted scTRAIL combined with bortezomib (2.91- ± 0.28-fold increase; Fig. 8D). Thus, in combination with BZB, EGFR-targeted scTRAIL showed only marginal toxic effects on inflamed liver tissues, in contrast to nontargeted scTRAIL, which strongly induced toxicity in inflamed liver tissues.
Although first clinical trials using TRAIL for the treatment of various advanced cancers showed promising results, including stable disease, it becomes evident that TRAIL monotherapy most likely will not result in a sufficient response, especially in solid tumor entities.10 Under certain conditions, TRAIL treatment of solid tumors even increased tumor cell migration and metastatic spread.33,34 In such conditions, TRAIL might activate prosurvival pathways, such as the nuclear factor kappa B (NF-κB) or MAPK pathways, rather than induce apoptosis. Thus, restoring TRAIL sensitivity toward apoptosis by the combined treatment of TRAIL with TRAIL-sensitizing agents is required to increase not only the clinical benefit, but, possibly, also to prevent cancer patients from harming effects of TRAIL in apoptosis-resistant tumors.
In HCC high levels of various antiapoptotic regulators of the death receptor and mitochondrial pathways, including cellular FLICE inhibitory protein (c-FLIP), X-linked inhibitor of apoptosis protein (XIAP), and several Bcl-2 proteins, have been observed.35,36 Proteasome inhibitors, such as BZB, have been recently shown to sensitize tumor cells, including HCC cells, toward TRAIL-induced apoptosis.23,24 These agents might be superior to other TRAIL-synergizing drugs, because inhibition of the proteasome as the central regulator of protein turnover affects multiple pathways and thus increases the likelihood that various TRAIL-resistance mechanisms can be bypassed.37 BZB treatment of hepatoma cells resulted in TRAIL-R1/2 up-regulation, enhanced death-inducing signaling complex formation, and down-regulation of c-FLIP and XIAP.24 BZB not only influences the extrinsic, but also the intrinsic pathway by triggering the release of proapoptotic mitochondrial factors.24,38 In addition, proteasome inhibitors trigger cell-cycle arrest and NF-κB inhibition, thereby influencing apoptosis induction.39 In view of the various targets of proteasome inhibitors that have been identified, it must be noted that it was not our intention to further investigate the mechanisms of BZB-mediated TRAIL sensitization.
Targeting of TRAIL to the tumor site represents an additional therapeutic strategy to increase antitumoral efficacy and avoid systemic toxicity. Proof of concept for efficient tumor targeting has been recently achieved by fusing scTRAIL to Ab derivatives specific for antigens with increased expression in tumor tissues, such as the EGFR member ErbB2.28 Because our and other data26,27 show that EGFR is up-regulated in HCC, we employed a novel TRAIL protein in which scTRAIL was fused to a humanized single-chain variable fragment derived from cetuximab, a chimeric Ab directed against the extracellular domain of EGFR. This fusion protein (αEGFR-scTRAIL) selectively binds to EGFR-positive HCC cells and, through target-antigen binding, mimics membrane-bound TRAIL, which enhances stable TRAIL-R signaling complex formation.
With respect to the results of in vitro dose-finding experiments for apoptosis induction and to the pharmacokinetics of TRAIL in vivo, a concentration of 100 ng/mL of the TRAIL proteins was used in this study.40,41 Higher concentrations did not differ in their ability to induce apoptosis in the presence of BZB. For BZB treatment, we chose a concentration of 500 ng/mL, which, when combined with TRAIL, has been shown to be nontoxic in PHHs.24 We analyzed EGFR-targeted scTRAIL, compared to non-targeted scTRAIL alone and in combination with BZB, in Huh7 liver cancer cells that are known to express EGFR. Although nontargeted scTRAIL plus BZB induced significantly higher caspase activity, compared to the respective agents alone, the combination of EGFR-targeted scTRAIL with BZB was most effective.
Inhibition of EGFR signaling by itself can induce apoptosis in appropriate cell systems. However, our study shows that apoptosis induction by αEGFR-scTRAIL is entirely dependent on TRAIL signaling, rather than on inhibition of EGFR function, because TRAIL-neutralizing Abs almost completely abolished caspase activation. Furthermore, pretreatment of Huh7 cells with a caspase inhibitor strongly reduced caspase-8 and caspase-3 activation induced by targeted TRAIL, indicating that the observed proapoptotic effects were indeed the result of TRAIL-induced caspase activation. Strikingly, neither scTRAIL nor EGFR-targeted scTRAIL alone or combined with BZB induced caspase activation in PHHs, indicating that this treatment is nontoxic to healthy hepatocytes.
Because the liver is composed of multiple cell types that might modulate TRAIL susceptibility of hepatocytes, we analyzed TRAIL effects in organotypic cultures of fresh liver explants.32 We found a modest increase of caspase activity in HCC liver explants treated with scTRAIL and BZB. In contrast, EGFR-targeted scTRAIL in combination with BZB induced a significant increase of caspase activity in HCC tissues, again indicating its increased antitumor activity. We suggest that the increased antitumor activity of targeted TRAIL is largely conferred by its tumor specificity and increased bioactivity. The enhanced activity of targeted TRAIL might be explained by its action as a quasi membrane-bound form, leading to efficient formation of both TRAIL-R1- and TRAIL-R2-signaling complexes, whereas scTRAIL preferentially binds to TRAIL-R1 and TRAIL decoy receptors.42 Despite similar expression of both receptors in certain cancer cells, TRAIL-R2 appears as the primary transducer of the TRAIL death signal in many cancer cells,43 which might account for the higher proapoptotic activity of EGFR-targeted scTRAIL.
In line with the increased caspase activation induced by targeted TRAIL and BZB in HCC tissue extracts, we also demonstrated enhanced caspase-3 activity in HCC sections. Furthermore, combined with BZB, αEGFR-scTRAIL induced strong caspase-mediated CK-18 cleavage and TUNEL reactivity, indicating that the enhanced activity of αEGFR-scTRAIL in HCC tissues is the result of increased apoptosis of liver cancer cells. Consistent with our observations in PHHs, we did not find any proapoptotic effects in healthy liver tissues treated with scTRAIL or αEGFR-scTRAIL. Thus, targeted scTRAIL in combination with the selected dose of BZB induces cancer-specific apoptosis without toxicity to healthy liver tissues. We have previously reported that a nontargeted version of TRAIL shows no hepatotoxicity in the healthy liver, but exerts a significant apoptotic effect in tissues from patients with fatty liver disease.32 Interestingly, αEGFR-scTRAIL induced apoptosis neither in tumor-free cirrhotic livers of the same HCC patients nor in inflamed livers of NAFLD patients. In contrast, unlike the EGFR-targeted protein, nontargeted TRAIL induced significantly stronger apoptosis in inflamed liver tissues (e.g., from NAFLD patients). Thus, in this respect, too, targeted scTRAIL might be of superior use, compared to nontargeted versions. Given the modest survival improvement of the multikinase inhibitor sorafenib in HCC treatment, there is a strong need to unravel further molecular targets for therapeutic intervention to improve the survival of advanced HCC patients. The data of this study provide a promising concept for treatment of HCC and other EGFR-expressing solid tumors. Clinical trials are required to further evaluate tumor-selective TRAIL variants with improved specific bioactivity for anticancer treatment, which represents a future goal toward personalized medicine.
This study was supported by the Deutsche Forschungsgemeinschaft (SFB Transregio 77, SFB 685, and SFB 773) and BMBF (project no.: 0315-280A; FORSYS-Partner “Predictive Cancer Therapy”).
Potential conflict of interest: Nothing to report.
- 30The holy grail of hepatocyte culturing and therapeutic use. In: Santin M, ed. Regenerative Medicine Today. Berlin, Germany: Springer International; 2009: 283-320., , , , , .
- 37On the TRAIL to successful cancer therapy? Predicting and counteracting resistance against TRAIL-based therapeutics. Oncogene 2012 May 14. doi: 10. 1038/onc. 2012.164., , , , , .
Additional Supporting Information may be found in the online version of this article.
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.