Histone deacetylase inhibitor-mediated sensitization to TRAIL-induced apoptosis in childhood malignancies is not associated with upregulation of TRAIL receptor expression, but with potentiated caspase-8 activation

Abstract

Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) has great potential for the treatment of cancer because it targets tumor cells while sparing normal cells. Several cancers, however, fail to respond to TRAIL's antineoplastic effects. These resistant tumors require cotreatment with sensitizing agents in order for TRAIL to exert anticancer activity. Histone deacetylase inhibitors (HDACi) have been recognized as potent TRAIL sensitizers. In searching for the determinants of TRAIL responsiveness, HDACi-mediated TRAIL sensitization has been predominantly attributed to TRAIL receptor upregulation. This explanation, however, has been challenged by a few studies. The aim of the present study was to explore the relevance of TRAIL receptor expression for HDACi-mediated TRAIL sensitization in childhood tumors, i.e., in medulloblastoma, Ewing's sarcoma and osteosarcoma. In previous studies, we had shown that TRAIL and HDACi were synergistic in inducing apoptosis in medulloblastoma and Ewing's sarcoma. In the present study, we demonstrate that HDACi cooperated with TRAIL in eliciting cell death in osteosarcoma. However, HDACi treatment did not alter or even reduced cell surface expression of TRAIL receptors in the three childhood tumors. In gaining insight into the apoptotic pathway involved in TRAIL sensitization, HDACi were found to potentiate TRAIL-induced caspase-8 activation. Taken together, our findings suggest that HDACi-mediated TRAIL sensitization is not the result of TRAIL receptor upregulation, but the result of a receptor-proximal event in childhood tumor cells.

Keywords: :

• bortezomib
• caspase-8
• childhood tumor
• death receptor
• histone deacetylase inhibitors
• TRAIL

Introduction

Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL; also called Apo2L and TNFSF10) is a promising candidate for cancer treatment because it specifically kills malignant cells while leaving normal cells relatively unscathed. Consequently, clinical trials with recombinant TRAIL and agonistic anti-TRAIL receptor (TRAIL-R) monoclonal antibodies have been/are being conducted.¹ But these clinical trials have also revealed that many human tumors fail to respond to TRAIL monotherapy. TRAIL resistance, however, may be overcome by cotreatment with sensitizing agents, such as ionizing radiation or cytostatics. In particular, histone deacetylase inhibitors (HDACi), a promising new class of antineoplastic drugs that stimulate differentiation, induce apoptosis, inhibit proliferation and exert antiangiogenic and immune stimulatory activity in tumor cells,² have been documented in nearly 80 preclinical studies to enhance TRAIL's anticancer action.^3-80

TRAIL instigates the extrinsic pathway of apoptosis by binding to its receptors TRAIL-R1 (also called DR4 and TNFRSF10A) or TRAIL-R2 (also called DR5 and TNFRSF10B), resulting in receptor oligomerization and the formation of a death-inducing signaling complex (DISC) consisting of activated receptors, the adaptor protein Fas-associated death domain (FADD) and caspase-8 or -10.¹ This leads to activation of caspase-8/10 by autocatalytic cleavage, which directly activates the executioner caspases 3 and 7. In addition, TRAIL can harness the intrinsic (mitochondrial) pathway of apoptosis: activated caspase-8/10 processes the Bcl-2 family protein Bid into a truncated form that translocates to the mitochondria, promoting the release of proapoptotic factors into the cytosol. This results in the formation of the apoptosome consisting of cytochrome c, Apaf-1 and caspase-9, in turn stimulating the activation of caspases 3 and 7. The apoptotic machinery is negatively regulated at several levels by antiapoptotic factors. For instance, c-FLIP interacts with the DISC and prevents caspase-8/10 activation, antiapoptotic Bcl-2 family members, such as Bcl-2, Bcl-xL, Mcl-1 and A1, block mitochondrial events, and inhibitor of apoptosis proteins (IAPs) inhibit activated caspases.¹

A conglomeration of mechanisms have been put forward to account for HDACi-mediated TRAIL sensitization, among them downregulation of c-FLIP^{3,8,13,23-26,32,49,52,54,59,63,67,71,73,80} as well as ubiquitin/proteasome-mediated degradation of c-FLIP,^70,78 enhanced recruitment of FADD,⁶⁶ facilitated formation of the DISC,¹² downregulation of Bcl-2,^{8,21,23,25,39,52,60,73,75} Bcl-xL,^{8,21,23-25,39,40,45,52,58,60,73,75} Mcl-1^25,40,45 and A1,²³ upregulation of the proapoptotic Bcl-2 family proteins Bax and Bak,^{21,25,40,60,75} upregulation of Bid¹¹ as well as enhanced cleavage of Bid,^6,45,55,69 upregulation of the proapoptotic “BH3-only” Bcl-2 family members Bim,^{21,25,39,40,60} Noxa and PUMA,^21,25,60 downregulation of the IAPs XIAP^{8,20,21,25,40,45,73} and survivin^8,20,73,79 as well as degradation of XIAP,⁵⁷ activation of caspase-2,⁴⁴ upregulation of Apaf-1,⁷² downregulation of the antiapoptotic transcription factor NF-κB,^60,75,77 modulation of NF-κB activity^18,19,28 as well as cleavage of NF-κB,^63,68 cleavage of the prosurvival protein kinase Akt,⁶³ inhibition of the EGFR pathway,⁵⁹ downregulation of the Bcr-Abl fusion protein in chronic myeloid leukemia,⁵⁸ and, in conjunction with either IFNγ or decitabine, restoration of epigenetically silenced caspase-8 expression.^62,79

But most prominently, HDACi-mediated TRAIL sensitization has been attributed to upregulation of TRAIL-R1 and/or -R2.^{8-11,15,17-19,21,22,25,27,29,31,37,45,46,52,56,60,61,63,64,68,69,72,74} Several studies, however, have found no association of enhanced TRAIL susceptibility with increased TRAIL-R expression.^{6,12,32,34,38,47,54,55,73} Hence, in spite of extensive research into the combination of TRAIL and HDACi, the relevance of TRAIL-R expression has not been conclusively resolved. To shed more light on this issue—with an emphasis on childhood tumors—we examined HDACi effects on TRAIL-R expression in cell lines derived from medulloblastoma, Ewing's sarcoma and osteosarcoma. We found that HDACi treatment did not upregulate TRAIL-R cell surface expression, but did facilitate TRAIL-induced caspase-8 activation.

Results

Vorinostat and TRAIL synergize in inducing antineoplastic activity in osteosarcoma cells

We showed previously that HDACi and TRAIL synergized to elicit cell death in medulloblastoma and Ewing's sarcoma cells.^30,47 To confirm these observations, we reassessed the HDACi/TRAIL combination treatment in DAOY medulloblastoma cells. Cell death and Δψ_m dissipation were determined by flow cytometric analyses of propidium iodide uptake and DiOC₆(3) staining, respectively. Four hours after treatment with the HDACi vorinostat, cells were exposed to TRAIL for another 48 h. As presented in Figure 1, DAOY cells showed some responsiveness to vorinostat or TRAIL alone, but strong responsiveness to their combined application. To test a possible favorable cytotoxic interaction between HDACi and TRAIL in osteosarcoma cells, we examined cell death and Δψ_m dissipation in SaOS-2 cells. Either of the agents had little effect on the osteosarcoma cells when applied individually, but when applied together, they induced both considerable cell death and mitochondrial depolarization.

Figure 1. Vorinostat and TRAIL cooperate in inducing cell death and Δψ_m loss in DAOY and SaOS-2 cells. Four hours after administration of vorinostat, cells were exposed to TRAIL for another 48 h. Cell death and Δψ_m dissipation were determined by flow cytometric analyses of propidium iodide uptake and DiOC₆(3) staining, respectively. Means ± SEM of each three separate experiments are shown.

HDACi do not upregulate surface expression of TRAIL-R in childhood tumor cells

In former studies on the combination of HDACi and TRAIL in childhood malignancies, we could not observe HDACi to induce an increase in TRAIL-R expression, neither in Ewing's sarcoma nor in hepatocellular carcinoma or hepatoblastoma cells.^47,55 In the present study, we extended the investigation into the relevance of TRAIL-R expression for HDACi-mediated TRAIL sensitization to medulloblastoma and osteosarcoma. To begin with, we re-evaluated the effect of HDACi on TRAIL-R2 expression in WE-68 Ewing's sarcoma cells. We applied HDACi belonging to three different structural classes, the hydroxamic acid vorinostat, the cyclic tetrapeptide apicidin and the benzamide MS-275, and determined cell surface receptor expression by flow cytometric detection using monoclonal antibodies. Treatment with vorinostat or apicidin caused a significant decrease in TRAIL-R2 expression and MS-275 treatment produced a trend toward reduced expression of TRAIL-R2 (Fig. 2A). We also assessed the effects of TRAIL and vorinostat/TRAIL treatment on TRAIL-R2 expression. TRAIL alone caused a trend toward reduced TRAIL-R2 expression, while the combination of vorinostat and TRAIL had a significant effect. As a comparison, we determined the effects of the anticancer drugs etoposide and bortezomib, which have been shown to upregulate TRAIL-R2.^19,81 Etoposide treatment significantly increased TRAIL-R2 expression, whereas bortezomib treatment produced no significant effect.

Figure 2. HDACi do not upregulate TRAIL-R cell surface expression in childhood tumor cells. Cells were treated with agents for 24 h or for the indicated times and stained with TRAIL-R antibodies or isotype control. TRAIL-R cell surface expression was determined by flow cytometric analysis of viable cells. TRAIL-R mean fluorescence is depicted as percentage of untreated cells after calculation of the mean fluorescence intensities of anti-TRAIL-R-stained cells divided by isotype-matched IgG-stained cells; means ± SEM of each 3 experiments are shown (treatment vs. control: *p < 0.05, **p < 0.01; vorinostat/TRAIL vs. TRAIL: ^##p < 0.01).

Next, we determined expression of TRAIL-R1 and -R2 in DAOY medulloblastoma and SaOS-2 osteosarcoma cells. DAOY cells were treated with 2 µM vorinostat—a treatment that strongly enhances TRAIL-induced cell death in DAOY cells (see Figure 1)—and receptor levels were measured at 12 h intervals over a time course of 60 h. Figure 2B shows that TRAIL-R1 expression remained unchanged, while TRAIL-R2 expression slightly decreased. SaOS-2 cells were exposed to vorinostat, apicidin or MS-275 for 24 h: all three HDACi failed to increase TRAIL-R1 and -R2 levels. In DAOY and SaOS-2 cells, neither TRAIL alone nor the combination of vorinostat and TRAIL did have a significant effect on TRAIL-R2 expression (though a trend toward TRAIL-R2 downregulation could be observed after vorinostat/TRAIL combination treatment in DAOY cells).

In several studies, the importance of elevated TRAIL-R expression for HDACi-mediated TRAIL sensitization has been inferred from increased expression of TRAIL-R mRNA and/or protein by western blotting.^{8-11,19,60,63,64,69,74} However, HDACi have been found to upregulate TRAIL-R mRNA without concomitant change of TRAIL-R cell surface expression,³⁹ and, in Ewing's sarcoma cells, expression of TRAIL-R by western blotting did not strictly correlate with TRAIL-R surface expression.⁸² To elucidate whether HDACi upregulate TRAIL-R2 mRNA in the childhood cancer cells, we conducted real-time RT-PCR analyses. As depicted in Figure 3a 24 h incubation with 2 µM vorinostat produced inconsistent results: TRAIL-R2 mRNA was upregulated in DAOY cells, but downregulated in WE-68 cells. Hence, the HDACi effect on TRAIL-R2 mRNA expression did not regularly match that on TRAIL-R2 surface expression. By comparison, etoposide induced TRAIL-R2 mRNA upregulation, in keeping with its effect on TRAIL-R2 cell surface expression.

Figure 3. Vorinostat alters TRAIL-R2 and c-FLIP mRNA expression in childhood tumor cells. Cells were exposed to agents for 24 h. TRAIL-R2 and c-FLIP mRNA expression levels were determined by real-time RT-PCR and normalized to β-2-microglobulin mRNA levels. Means ± SEM of each two separate measurements are shown (*p < 0.05, **p < 0.01).

Next to TRAIL-R upregulation, HDACi-mediated TRAIL sensitization has been attributed most commonly to c-FLIP downregulation.^{3,8,13,23-26,32,49,52,54,59,63,67,71,73,80} We assessed c-FLIP mRNA expression by real-time RT-PCR in the childhood cancer cells. We found vorinostat treatment to affect c-FLIP gene expression only slightly in DAOY cells, but to substantially decrease it in SaOS-2 and WE-68 cells.

Vorinostat enhances TRAIL-induced caspase-8 activity

The data presented demonstrate that TRAIL-R upregulation was not essential for HDACi-mediated TRAIL sensitization in childhood tumor cells. To find out whether sensitization occurs before bifurcation into the mitochondria-dependent and -independent arm of apoptosis, we tested DAOY and SaOS-2 cells for activation of the receptor-proximal caspase-8. Caspase-8 activation was examined by measuring its catalytic activity using the fluorogenic substrate Ac-IETD-AMC and by detecting the cleaved subunit p18 on western blot. Four hours after treatment with vorinostat, cells were exposed to TRAIL for another 2 h. Both read-outs show that vorinostat augmented TRAIL-induced caspase-8 activition (Fig. 4), suggesting that HDACi potentiate TRAIL-elicited apoptosis at the level of the DISC.

Figure 4. Vorinostat potentiates TRAIL-induced caspase-8 activity in childhood tumor cells. Four hours after administration of vorinostat, cells were exposed to TRAIL for another 2 h. Caspase-8 activity was determined using the fluorogenic substrate Ac-IETD-AFC; relative caspase-8 activities are the ratio of treated cells to untreated cells. Means ± SEM of each 3 separate measurements are shown (*p < 0.05, **p < 0.01). Caspase-8 cleavage, i.e., the subunit p18, was detected by immunoblotting using anti-cleaved caspase-8 antibody.

Discussion

Almost 80 studies have demonstrated that HDACi can enhance the anticancer activity of TRAIL.^3-80 Of them, more than 20 have proposed that HDACi-mediated TRAIL sensitization primarily stems from increased TRAIL-R expression.^{8-11,15,17-19,21,22,25,27,29,31,37,45,46,52,56,60,61,63,64,68,69,72,74} Yet a few have challenged the relevance of TRAIL-R upregulation.^{6,12,32,38,47,54,55,73} In fact, HDACi have been observed to sensitize to agonistic TRAIL-R2 antibodies even in the presence of decreased TRAIL-R2 expression.³⁴ Our study lends further support to the insignificance of TRAIL-R expression. In particular, it argues against a role of TRAIL-R expression in HDACi-mediated TRAIL sensitization in childhood malignancies. We have shown that HDACi synergized with TRAIL in medulloblastoma,³⁰ in Ewing's sarcoma⁴⁷ and in osteosarcoma (this study), but did not upregulate - actually rather downregulated - TRAIL-R in these cells. This conclusion is also backed by our former study on childhood hepatocellular carcinoma and hepatoblastoma, in which HDACi synergized with TRAIL without upregulating TRAIL-R.⁵⁵

Our data presented here, however, indicate that HDACi-mediated TRAIL sensitization is a receptor-proximal event in childhood tumor cells. We found HDACi treatment to significantly decrease c-FLIP gene expression and, more importantly, to enhance TRAIL-induced caspase-8 activation. In principle, enhanced caspase-8 activity could also be the result of feedback effects of downstream executioner caspases.¹ Yet, the activity of caspase-8 was measured at 2 h post-addition of TRAIL—in fact, some increase in caspase-8 activity was observed as early as one hour after TRAIL administration (not shown)—arguing against the possibility that the potentiated activation of caspase-8 was secondary to HDACi-amplified activation of executioner caspases. This result is in line with our former study on childhood hepatoma cells showing that HDACi second TRAIL in promoting Bid cleavage⁵⁵ and it is also in line with a former study on leukemia cells showing that HDACi facilitate DISC formation and caspase-8 activation in the absence of TRAIL-R upregulation.¹²

For comparison, we also assessed the effect of the proteasome inhibitor bortezomib on TRAIL-R2 expression in Ewing's sarcoma cells. In a study on the combination of TRAIL with diverse anticancer agents, bortezomib has been reported to be the most effective one at overcoming TRAIL resistance.⁷⁶ As with HDACi, the role of TRAIL-R expression in bortezomib-mediated TRAIL sensitization has not been unambiguously unraveled: upregulation of TRAIL-R2 has both been reported to be required⁸¹ and to be not required⁸³ for bortezomib-mediated sensitization to TRAIL. Bortezomib has previously been demonstrated to synergize with TRAIL in Ewing's sarcoma.⁸⁴ Here, we have shown that bortezomib treatment only led to an insignificant increase in TRAIL-R2 expression, indicating that its upregulation is not critical for sensitization to TRAIL, at least in Ewing's sarcoma. This finding taken together with the results of the HDACi treatments suggests that TRAIL-R expression in general is not a suitable predictor for TRAIL susceptibility in childhood cancer. Support for this conclusion also comes from a study on the primary responsiveness of Ewing's sarcoma to TRAIL, in which no difference in the TRAIL-R surface expression was observed between sensitive and resistant cells.⁸⁵

In fact, key determinants, if any, of HDACi-mediated TRAIL sensitization are far from being consistently defined (see Introduction). Rather, the plethora of mechanisms reported to participate in HDACi-mediated TRAIL sensitization raise the question if a single key determinant of TRAIL responsiveness exists. In single-cell analysis of TRAIL-induced apoptosis, TRAIL susceptibility was not governed by the expression level of any individual protein, but by expression variations of multiple proteins together.⁸⁶ Fortunately, the effects of HDACi go beyond the modulation of single protein levels² and, thus, they have the potential to overcome resistance, even if the latter is not the result of unfavorable expression of any one protein in the apoptotic machinery, such as TRAIL-R.

In conclusion, our data demonstrate that TRAIL-R upregulation is not critical for HDACi to render childhood tumors susceptible to TRAIL. They also indicate that HDACi-mediated TRAIL sensitization is a receptor-proximal event, i.e., the consequence of hyperactivation of the extrinsic pathway, though we cannot rule out the possibility that HDACi boost TRAIL-induced apoptosis in addition by activating the intrinsic pathway.

Materials and Methods

Reagents

Vorinostat, apicidin, MS-275 and etoposide were purchased from Alexis. TRAIL was purchased from Peprotech. Bortezomib was purchased from LC Laboratories.

Cell lines

WE-68 cells (provided by Dr. F. van Valen, Münster, Germany) and SaOS-2 cells (a gift from Dr. T. Beckers, Konstanz, Germany) were maintained in RPMI 1640 (PAA) and DAOY cells (a gift from Dr. M. Grotzer, Zurich, Switzerland) were maintained in Improved MEM Zinc Option (Invitrogen). Media were supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 units/ml penicillin G sodium and 100 µg/ml streptomycin sulfate (PAA). WE-68 cells were cultivated in collagen-coated (5 µg/cm²; Roche) tissue culture flasks. Cells were cultivated at 37°C in a humidified 5% CO₂ incubator and routinely passaged when 90% confluent. Cell viability was determined by the trypan blue exclusion test. Cells were regularly inspected to be free of mycoplasma with the PCR mycoplasma detection kit from Applichem.

Flow cytometric analysis of cell death and mitochondrial transmembrane potential (Δψ_m)

Cell death was assessed by determining the integrity of the cell membrane by flow cytometric analysis of propidium iodide (Sigma) uptake. After harvesting, cells were incubated for 5 min in 2 µg/ml propidium iodide in PBS at 4°C in the dark. Δψ_m was assessed by determining the accumulation of 3,3′-dihexyloxacarbocyanine iodide [DiOC₆(3)] (Molecular Probes) in the mitochondrial matrix. Before harvesting, cells were incubated with 50 nM DiOC₆(3) at 37°C for 30 min. In both readouts, 10,000 cells were analyzed in each sample on a BD FACSCanto II; data were gated to exclude debris.

Flow cytometric detection of TRAIL-R1 and -R2 cell surface expression

Cells were harvested with 0.2% EDTA, washed twice with PBS, and incubated with mouse anti-TRAIL-R1/2 monoclonal antibodies (2 µg in 100 µl PBS/2.5% BSA; BioLegend) for 30 min at ambient temperature in the dark. Cells were washed twice and stained with a FITC-conjugated goat anti-mouse IgG antibody (20 µg/ml in 2.5% BSA; Dianova) for 30 min at ambient temperature in the dark. After washing, 50,000 cells were analyzed using the FACSCanto II; cells were gated on viable cells. A mouse IgG (20 µg/ml in 2.5% BSA; eBioscience) was used as isotype-matched control. Frequency histograms were generated using Flowjo software. Data are shown as fold induction after calculation of the mean fluorescence intensities of anti-TRAIL-R1/2-stained cells divided by isotype-matched IgG-stained cells.

Quantitative real-time RT-PCR

Total RNA was isolated using Peqgold Total RNA Kit including DNase digestion (Peqlab). RNA was transcribed into cDNA using Omniscript (Qiagen). Quantitative PCR for TRAIL-R2 was performed using the 7900HT Real-Time PCR system (Applied Biosystems). Expression levels were normalized to β-2-microglobulin. Reactions were done in duplicate using Applied Biosystems Gene Expression Assays (TRAIL-R2: Hs00187196_m1; c-FLIP: Hs01116280_m1; β-2-microglobulin: Hs00187842_m1) and Universal PCR Master Mix. All procedures were performed according to the manufacturers' protocols. The relative TRAIL-R2 expression was calculated by the 2^(-ΔΔCt) method.

Caspase-8 activity and cleavage

Caspase-8 activity was measured using the fluorogenic substrate Ac-IETD-AMC (Bachem). After harvesting, cells were lysed in 10 mM TRIS-HCl, 10 mM NaH₂PO₄/NaHPO₄ (pH 7.5), 130 mM NaCl, 1% Triton X-100 and 10 mM Na₄P₂O₇ and then incubated with 20 mM Hepes (pH 7.5), 10% glycerol, 2 mM DTT, 10 µM MG-132 (in order to reduce nonspecific background by inhibiting the caspase-like activity of the proteasome) and 25 µg/ml Ac-IETD-AMC at 37°C for 2 h. The release of AMC was measured on a BMG Labtech FLUOstar Omega using an excitation/emission wavelength of 355/460 nm. Relative caspase-8 activities were calculated as a ratio of emission of treated cells to untreated cells.

Caspase-8 cleavage was detected by immunoblotting. 40 µg of total cellular protein per lane were separated by standard SDS-PAGE on 12% gel and electrophoretically transferred to nitrocellulose membrane. After blocking, caspase-8 cleavage was immunodetected using rabbit anti-cleaved caspase-8 (18C8) monoclonal antibody (dilution 1:1,000; #9496, Cell Signaling). Equal loading of protein was verified by detection of GAPDH using mouse anti-GAPDH monoclonal antibody (dilution 1:25,000; Biodesign International). Peroxidase-conjugated goat anti-rabbit IgG (dilution 1:5,000; Dianova) and goat anti-mouse IgG (dilution 1:25,000; Dianova) were used as secondary antibodies followed by enhanced chemiluminescence detection (GE Healthcare) of specific signals.

Statistical analysis

Statistical significance of differences between experimental groups was determined using the paired two-tailed Student's t test.

Abbreviations:
DiOC6(3)	=	3,3′-dihexyloxacarbocyanine iodide
DISC	=	death-inducing signaling complex
HDACi	=	histone deacetylase inhibitors
TRAIL-R	=	TRAIL receptor

Acknowledgments

This work was supported by a grant from the “Wilhelm Sander-Stiftung, Neustadt/Donau.”

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.