Human Retinoblastoma Cells Are Resistant to Apoptosis Induced by Death Receptors: Role of Caspase-8 Gene Silencing

Abstract

purpose. Fas ligand (FasL) and TNF-related apoptosis-inducing ligand (TRAIL)/Apo2L are members of the TNFα family that can trigger apoptosis in susceptible cells via respective death receptors (DRs). FasL cross-links its receptor Fas, resulting in recruitment and proteolytic activation of caspase-8, which initiates the downstream apoptotic cascade. TRAIL signals through its receptors DR4 and DR5, which can activate caspase-8 as well. This study was undertaken to investigate the functional status of the FasL and TRAIL apoptotic pathways in retinoblastoma (Rb) cells.

methods. The human Rb cell lines Y79 and WERI-Rb1 were evaluated for their response to the Fas cross-linking antibody CH11 and recombinant TRAIL, as well as for cell surface presence and mutational status of Fas, DR4, and DR5 by flow cytometry and genomic DNA sequencing, respectively. The expression of caspase-8 and its inhibitor FLIP, as well as their recruitment to the DR signaling complex were studied by immunoblot analysis.

results. Rb cells express Fas, DR4, and DR5 on their surfaces, yet were resistant to DR-mediated apoptosis. This was not due to DR mutations or secretion of the soluble decoy Fas, antiapoptotic NF-κB activity, or FLIP overexpression, but to the absence of caspase-8 expression. The demethylating agent 5-aza-2′-deoxycytidine restored caspase-8 expression and sensitivity to DR-mediated apoptosis.

conclusions. Rb cells are resistant to DR-mediated apoptosis because of a deficiency in caspase-8 expression secondary to epigenetic gene silencing by overmethylation. The data help delineate the apoptotic pathways in Rb cells and suggest that the combination of demethylating agents with DR-activating modalities, such as TRAIL receptor monoclonal antibodies, may benefit patients with retinoblastoma.

Apoptosis (programmed cell death) plays a key role in normal tissue homeostasis and tumor pathophysiology. ¹ All cells contain an endogenous self-destruction program that is destined to be activated in cases of irreparable damage or impeding tumorigenesis, thus sacrificing the defective individual cell to protect the multicellular organism. Malignant cells have managed to override this internal suicide program, and, thus, inappropriate resistance to apoptosis is believed to be a necessary step for neoplastic transformation. ¹ Moreover, the antitumor surveillance of the immune system uses the apoptotic mechanisms of the target cell as part of its effector arm. Finally, apoptosis is a pivotal part of the mechanism of action of conventional chemotherapeutics and novel anticancer agents. ¹ In the past decade, the widespread recognition that cancer progression requires, in addition to increased cell proliferation, an impairment of cell death pathways ¹ has led to an exponential increase in interest in this field. Among the apoptosis inducers that have been the focus of intense study are two members of the tumor necrosis factor (TNF) family, FasL, and TNF-related apoptosis-inducing ligand (TRAIL)/Apo2L. 

FasL transmits its apoptotic signal by cross-linking its receptor, Fas (Apo-1/CD95), a transmembrane protein that contains an intracellular motif called the death domain (DD). Fas recruits another DD-containing molecule, FADD (Fas-associated death domain) ² that provides a docking surface for and facilitates the autoprocessing and activation of the proenzyme FADD-like ICE or caspase-8 (FLICE). ³ The complex of Fas, FADD, and caspase-8 is called the death-inducing signaling complex (DISC). FLICE inhibitory protein (FLIP), a protein structurally related to pro-caspase-8, which lacks enzymatic activity, competes with the latter for recruitment to the DISC and has an inhibitory effect on Fas-mediated apoptosis, thus representing a naturally occurring dominant negative form of caspase-8. ⁴ In cells capable of recruiting large amounts of pro-caspase-8 to the DISC (type I cells), that enzyme is activated rapidly and directly processes effector caspases, such as caspase-3 or -6, which are the final executioners of apoptosis. Cells with low amounts of pro-caspase-8 or insufficient recruitment of pro-caspase-8 to the DISC due to the presence of FLIP (type II cells), depend on the mitochondria to transmit and amplify the apoptotic signal, ⁵ via caspase-8–mediated cleavage of Bid, a cytoplasmic member of the bcl-2 family, which induces the release of cytochrome c from the mitochondria, ⁵ with subsequent activation of caspase-9, which in turn activates caspase-3. 

FasL is expressed in activated T lymphocytes and provides a mechanism for T-cell–mediated elimination of Fas-expressing target cells. ⁶ FasL is also present in a few tissues such as the testis, eye, brain, and placenta, where it contributes to their immune-privileged status by eliminating infiltrating lymphocytes. ^{7 8 9} Human retinal cells express FasL, which plays an immunomodulatory and antiangiogenic role. ¹⁰ Moreover, we and others have reported involvement of FasL/Fas-mediated signaling in chemotherapy-induced apoptosis ^{11 12 13} at least in some models, although this mechanism may not be operative in all types of cancers. 

TRAIL, also known as Apo2L, interacts with two apoptosis-inducing receptors, DR4 (or TRAIL-R1) and DR5 (or TRAIL-R2). Transfection experiments have shown that both DR4 and DR5 can initiate caspase-mediated apoptosis (for a review, see Ref. ¹⁴ ). Unlike FasL, whose expression is normally limited to cells of the immune system and a few immune-privileged sites, TRAIL expression has been detected in a wide range of normal fetal and adult tissues. ¹⁵ These findings suggest the existence of a protective mechanism against TRAIL-mediated cytotoxicity in normal cells, which is supported by observations that TRAIL can induce apoptosis in transformed and malignant cells, ¹⁵ but not in normal cells. ^{14 16} Many malignancies are sensitive to the apoptotic effects of TRAIL/Apo2L, including multiple myeloma, ¹⁷ certain forms of lymphoma and leukemia, ^{18 19} breast and thyroid ²⁰ carcinomas, melanoma, ²¹ Ewing’s sarcoma, ²² and malignant glioma. ²³ The effectiveness within each histologic group is also impressive and far exceeds the extent of cell killing associated with other death-inducing ligands, such as FasL and TNFα, as well as cytotoxic drugs. TRAIL-induced cross-linking of its receptors triggers recruitment and activation of caspase-8 in many models, ^{24 25 26} whereas, in other cases, caspase-10 is the apical caspase of this pathway. ^{22 27}  

Increasing evidence suggests that FasL and TRAIL are expressed on activated immune cells and constitute important natural effector molecules for host defense against transformed cells. ^{28 29 30 31 32 33} Therefore, overriding these apoptotic mechanisms could provide an evolutionary advantage to malignant cells during neoplastic transformation, cancer progression, and evolution. 

Fas expression has been reported in retinoblastoma (Rb) cells in clinical specimens, ³⁴ but the functional status of this apoptotic pathway in Rb cells is unknown. In the present study, we investigated the functional status of the DR pathways regulated by FasL and TRAIL/Apo2L in human Rb cell lines. We found that human Rb cells are resistant to these apoptotic pathways. This refractoriness could contribute to immune evasion and resistance to treatment. 

Materials and Methods

Human Tumor Specimens, Cell Lines, and Tissue Culture

Archival Bouin- or formalin-fixed and paraffin-embedded primary Rb specimens were generously provided by Winand N. M. Dinjens (Department of Pathology, Josephine Nefkens Institute, Erasmus Medical Center, Rotterdam, The Netherlands) and Miriam Marichal (Vrije Universiteit Brussel, Brussels, Belgium). Paraffin-embedded thyroid carcinoma specimens served as a positive control for caspase-8 expression and were obtained from the files of the Pathology Department, University of Athens. The human Rb cell lines Y79 and WERI-Rb1, as well as the neuroectodermal sarcoma line SK-N-MC, ³⁵ were purchased from the American Type Culture Collection (Manassas, VA) and grown in Dulbecco’s modified Eagle’s medium (DMEM; BioWhittaker, Walkersville, MD) with 100 U/mL penicillin, 100 μg/mL streptomycin, and 10% fetal calf serum (FCS; Invitrogen-Gibco, Gaithersburg, MD), unless stated otherwise. 

Materials

The Fas cross-linking activating antibody CH11 was obtained from Panvera (Madison, WI). Recombinant human TRAIL/Apo2L was from Biomol (Plymouth Meeting, PA). Additional preparations of TRAIL from Immunex Corp. (Seattle, WA) and Genentech (South San Francisco, CA) were tested and gave similar results. Mouse anti-human Fas monoclonal antibody DX2 was obtained from PharMingen (San Diego, CA); mouse MsIgG1 isotype control from Beckman Coulter-Immunotech (Miami, FL); goat polyclonal antibodies for human DR4, DR5, and DcR1 from Santa Cruz Biotechnology (Santa Cruz, CA); anti-human DcR2 rabbit polyclonal Ab from Imgenex (San Diego, CA); rabbit anti-human DcR1 polyclonal antibody from Affinity Bioreagents (Golden, CO); goat anti-human DR4, DR5, and DcR2 polyclonal antibodies from R&D Systems Inc. (Minneapolis, MN); donkey anti-goat IgG FITC-conjugated F(ab′)₂ fragment and donkey anti-rabbit IgG FITC-conjugated F(ab′)₂ fragment from Jackson ImmunoResearch Laboratories (West Grove, PA); MTT, 5-aza-2′-deoxycytidine (5-dAzaC), and cycloheximide from Sigma-Aldrich (St. Louis, MO); the inhibitory peptide SN50 from Biomol; the proteasome inhibitor MG132 from Calbiochem (La Jolla, CA); a mixture of proteinase inhibitors (Complete-TM), Ig-free normal horse serum, and SDS from Invitrogen-Life Technologies, Inc. (Gaithersburg, MD); and a chemiluminescence kit (Enhanced Chemiluminescence; ECL, which includes the peroxidase-labeled anti-mouse and anti-rabbit secondary antibodies, from Amersham (Arlington Heights, IL). 

MTT Colorimetric Survival Assay

The survival of Rb cells after triggering of the Fas or TRAIL apoptotic pathway was examined using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) colorimetric assay, as previously described. ²² Cells were plated in 24-well plates at 70% to 80% confluence and then treated as indicated. At the end of each treatment, cells were incubated with 1 mg/mL MTT for 4 hours at 37°C. A mixture of isopropanol and 1 N HCl (23:2, vol/vol) was then added under vigorous pipetting, to dissolve the formazan crystals. Dye absorbance (A) in viable cells was measured at 570 nm, with 630 nm as a reference wavelength. Cell survival was estimated as a percentage of the value of the untreated control. All experiments were repeated at least three times, and each experimental condition was repeated at least in quadruplicate wells in each experiment. 

Determination of DR Status by Flow Cytometric Analysis

Y79 cells were characterized for their surface expression of DRs by flow cytometry. Staining was performed as previously described. ³⁶ Briefly, for each cell line, 10⁶ cells were incubated with the appropriate anti-receptor Ab or a respective control (5.0 μg) for 45 minutes. Specifically, Fas cell surface expression was evaluated with the mouse anti-human Fas monoclonal antibody DX2 (using mouse IgG1 as control). DR4 and DR5 expression were assessed with goat anti-human DR4 and DR5 polyclonal antibodies, using goat IgG for control staining. DcR1 and DcR2 expression were assessed with both goat and rabbit anti-DcR1 and both goat and rabbit anti-DcR2 polyclonal antibodies, with the appropriate control. Cells were then washed with PBS and incubated for 45 minutes with 2.0 μg of goat anti-mouse IgG FITC-conjugated F(ab′)₂ fragment for anti-Fas mAb phenotypic analyses or with a donkey anti-goat (or anti-rabbit) IgG FITC-conjugated F(ab′)₂ fragment for the DcR1, DcR2, DR4, and DR5 analyses. Cells were then washed, fixed with 1% formaldehyde PBS, and analyzed on a flow cytometer (Epics-XL-MCL; Beckman-Coulter, Hialeah, FL). 

Death Receptor DD Sequencing

Genomic DNA from our Rb cell lines was extracted, and the exons encoding the DDs of Fas, DR4, and DR5 were polymerase chain reaction (PCR) amplified with the following primers: Fas (exon 9): forward (5′-GGTTTTCACTAATGGGAATTTCAT-3′) and reverse (5′-CTGAATTTGTTGTTTTTCACTCTA-3′); DR4 (exon 10): forward (5′-CTCTGATGCTGTTCTTTGAC-3′) and reverse (5′-TCACTCCAAGGACACGGCAG-3′); and DR5 (exon 9): forward (5′-CTGTCTCTGTGGCTTTCTCCAC-3′) and reverse (5′-TGACTTCCTGAAGAGAATCACAC-3′). 

Cycle sequencing of the purified PCR products was performed with one of the PCR primers, using the dye termination chemistry (Big-Dye Terminator sequencing kit; Applied Biosystems, Inc. [ABI], Foster City, CA). The Sephadex G-50-purified cycle sequencing products were analyzed on a gene analyzer (Prism 310; ABI). 

Reverse Transcription–Polymerase Chain Reaction

Total RNA from our Rb cell lines was extracted (TRIzol Reagent; Invitrogen, San Diego, CA). First-strand cDNA was synthesized from 1 μg of total RNA with oligo(dT) primer (Superscript II reverse transcriptase; Invitrogen). The cDNA products were amplified for 35 cycles using Taq DNA polymerase (Invitrogen), with the exon 3 primer Fasex3F (5′-GGATGAACCAGACTGCGTG-3′) and the exon 7 primer Fasex7R (5′-CTGCATGTTTTCTGTACTTCC-3′), both of which are common to transmembrane Fas (tmFas) and the alternatively spliced, shorter, soluble Fas (sFas) transcript. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene (GAPDHF: 5′-TGGTATCGTGGAAGGACTCATGAC-3′; GAPDHR: 5′-ATGCCAGTGAGCTTCCCGTTCAGC-3′) was used for normalization of the results. The products were run on a 3% agarose gel. 

Immunoblot Analysis

Immunoblot analysis was performed as previously described. ¹¹ The antibodies used were: mouse monoclonal antibodies for caspase-8 and FADD and polyclonal antibody for FLIP (Upstate Biotechnologies, Lake Placid, NY), and polyclonal antibodies for caspase-3, DR4, and DR5 (Santa Cruz Biotechnology, Santa Cruz, CA). 

Caspase-8 Activity Assay

Y79 and SK-N-MC cells (5 × 10⁶) were stimulated with the anti-Fas antibody CH11 (500 ng/mL) or TRAIL/Apo2L (1000 ng/mL) for 1 hour, washed in PBS and harvested by centrifugation at 800g for 10 minutes at 4°C. Caspase-8 enzymatic activity was measured with a colorimetric assay (ApoAlert Caspase-8 kit; BD-Clontech, Palo Alto, CA), normalized for protein content, and expressed in arbitrary units. 

DISC Immunoprecipitation

Immunoprecipitation of the Fas DISC was performed as described previously. ³⁷ Briefly, 10⁷ cells were either first stimulated with biotin-labeled anti-Fas APO-1 antibody (2 μg/mL; Kamiya Biomedical Co., Seattle, WA) for 10 minutes and then lysed in a lysis buffer (20 mM Tris-HCl [pH 7.4]), 1% Triton X-100, 10% glycerol, and 150 mM NaCl) supplemented with protease inhibitors (Complete; Invitrogen-Life Technologies) (stimulated cells) or first lysed and then incubated overnight with the biotin-labeled anti-Fas-APO-1 antibody (unstimulated cells). The Fas receptor and the associated proteins (DISC) were subsequently precipitated for 2 hours at 4°C with streptavidin-agarose beads (Invitrogen-Life Technologies, Inc.). After precipitation, the beads were washed five times with 1 mL lysis buffer. For Western blot analysis, 20 μg of protein were separated by 4% to 20% gradient PAGE (Invitrogen), immunoblotted with anti-caspase 8 or anti-FLIP Ab at 4°C, as described above, and visualized by ECL. 

Immunohistochemical Detection of Caspase-8 in Primary Retinoblastoma Specimens

Five-micrometer paraffin-embedded sections from three archival paraffin-embedded primary retinoblastoma specimens were deparaffinized in xylene and rehydrated in alcohol. Endogenous peroxidase activity was quenched for 30 minutes in methanol containing 0.5% H₂O₂. The sections were then subjected to antigen retrieval by incubation in an antigen retrieval solution (Dako, Carpinteria, CA) in a microwave oven. The sections were washed in PBS and immersed for 1 hour in a blocking solution consisting of 20% normal goat serum (Santa Cruz Biotechnology), 0.4% Tween 20, and 2% BSA (Sigma-Aldrich) in PBS. The primary mouse anti-caspase 8 antibody (Upstate Biotechnology, Inc.) was applied overnight at a concentration of 1:100. Subsequently, the sections were washed in PBS and incubated for 1 hour at room temperature with a biotinylated goat anti-mouse antibody (1:500; Dako). After the sections were washed with PBS, they were covered with the avidin-biotin complex reagent (Dako) for 30 minutes. The peroxidase reaction was developed with 3,3-diaminobenzidine, and the slides were counterstained with hematoxylin. Positive staining was evaluated by an experienced pathologist. 

Statistical Analysis

Statistical significance was examined by a two-way analysis of variance, followed by the Duncan post hoc test. In all analyses, P < 0.05 was considered statistically significant. 

Results

Resistance of Rb Cells to Apoptosis Induced by Fas Cross-Linking and TRAIL

We investigated the functional status of Fas and TRAIL signaling in the human Rb cell lines Y79 and WERI-Rb1. Both lines were treated with the Fas cross-linking antibody CH11 or recombinant TRAIL/Apo2L for 24 hours. No evidence of cell death was observed, either morphologically or by the MTT assay in either line (Fig. 1) . There was no cell death observed, even after prolonged exposure to CH11 or TRAIL for up to 4 days in culture (data not shown). 

Expression of DRs in Rb Cells

We next attempted to investigate the reason for the resistance of Rb cells to apoptosis through the Fas and TRAIL pathways. We first explored the expression of the respective receptors on the Rb cell surface. Flow cytometric analysis revealed that Y79 and WERI-Rb1 Rb cells express on their cell surface Fas, as well as DR4 and DR5 (Figs. 2 and 3 , respectively). Thus, lack of cell surface DR expression is not a plausible explanation for resistance to apoptosis by Rb cells. Both cell lines lack DcR1 surface expression, but express DcR2. 

Effect of Production of Soluble Fas on Fas Resistance

Synthesis of an alternatively spliced Fas mRNA, that lacks exon(s) coding for the transmembrane domain and leads to the expression of a soluble form of Fas (sFas), has been hypothesized to suppress Fas-mediated apoptosis, as sFas could function as a decoy receptor. ^{38 39} We investigated whether such a mechanism could explain the resistance of Rb cell lines to Fas-mediated apoptosis. We compared the presence of full-length and alternatively spliced Fas mRNA in Y79 and WERI-Rb1, as well as in the exquisitely Fas-sensitive SK-N-MC cell line, using semiquantitative RT-PCR. We found comparable levels of full length Fas mRNA in all three cell lines, whereas sFas mRNA was detectable essentially only in SK-N-MC cells (Fig. 4) . These data suggest that expression of an alternatively spliced soluble Fas cannot account for the resistance of Rb cells to Fas-mediated apoptosis. 

Lack of DD Mutations in Fas, DR4, and DR5

Mutations in DD of DRs have been described in various other models ^{40 41 42 43 44} and have been blamed for tumor cell resistance to respective death ligands. Sequencing of the DDs of Fas, DR4, and DR5 in Y79 and WERI-Rb1 cells did not reveal any mutations. 

Effect of Inhibition of Protein Synthesis on Resistance to DR-Induced Apoptosis

In many models of resistance to apoptosis, the protein synthesis inhibitor cycloheximide has been demonstrated to restore sensitivity to cell death. ^{26 45} This is generally regarded as evidence for the presence of a short-lived inhibitor of apoptosis. FLIP has been found to fit that description in many models. ^{26 45} We, therefore, investigated the effect of cycloheximide on the Fas and TRAIL apoptotic pathways in Rb cell lines. We found that both Y79 and WERI-Rb1 cells remained resistant to Fas cross-linking and TRAIL/Apo2L, even in the presence of cycloheximide (Fig. 5A) . 

Effect of Inhibition of NF-κB on Resistance to DR-Induced Apoptosis

The transcription factor NF-κB promotes survival in numerous models, by stimulating the transcription of several antiapoptotic genes, such as cIAP-2 and A1. ^{46 47} In several cases, NF-κB has been found to regulate negatively the apoptosis induced by FasL and/or TRAIL/Apo2L. ²⁶ We have demonstrated that the NF-κB inhibitory peptide SN50 ⁴⁸ induces apoptosis in Rb cells. ⁴⁷ We, therefore, investigated the effect of NF-κB inhibition on DR-induced apoptosis in Rb cells. We found that the peptide SN50 did not sensitize Y79 or WERI-Rb1 cells to Fas cross-linking or TRAIL/Apo2L (Fig. 5A) . On the contrary, the CH11 antibody and TRAIL/Apo2L had a minor protective effect against SN50-induced cell death (P < 0.05 for both CH11 and TRAIL/Apo2L), probably due to their ability to stimulate NF-κB activity. ⁴⁹  

The proteasome inhibitor MG132 inhibits the degradation of the NF-κB inhibitor IkB, resulting in accumulation of IkB and inhibition of NF-κB. Proteasome inhibitors induce apoptosis in several types of cancer cells, ^{50 51} including Rb cell lines. ^{52 53} We, therefore, investigated the effect of MG132 on DR-induced apoptosis in Rb cells. Again, no sensitizing effect was detected (Fig. 5B) . 

Effect of DR Cross-linking on Activation of Caspase-8

We proceeded to investigate the intracellular apoptotic machinery in our model. Caspase activity mediates apoptosis induced by DRs. In particular, caspase-8 is recruited to Fas and TRAIL receptors, and is rapidly activated on receptor cross-linking. We evaluated the activation of caspase-8 after treatment with CH11 or TRAIL in our retinoblastoma model, using the Fas-sensitive SK-N-MC cells as a positive control. We found strong activation of caspase-8 in SK-N-MC cells by both CH11 and TRAIL (P < 0.01 for both CH11 and TRAIL/Apo2L), but no effect in Y79 cells (Fig. 6) . These data suggest that the defect in DR-mediated apoptosis in Rb cells lies between the cell surface receptor and caspase-8 activation. 

Recruitment of FLIP to the DISC

We investigated further the interaction of caspase-8 with Fas in Rb cells. Y79 cells were stimulated with or without anti-Fas before cell lysis. Precipitation of Fas resulted in isolation of the DISC, which was then studied by immunoblot analysis. In the case of the Fas-sensitive SK-N-MC cells, which again served as positive control, caspase-8 was promptly recruited to the Fas complex. No recruitment of FLIP was detected. In Y79 cells, however, Fas cross-linking did not result in caspase-8 recruitment. On the contrary, FLIP was found to interact with Fas, even in the absence of Fas triggering (Fig. 7) . This suggests that the resistance of Y79 cells to Fas-mediated apoptosis is caused by an inversion of the ratio of caspase-8/FLIP molecules that are recruited to the DISC. It also confirms the structural integrity and surface localization of Fas, since Fas cross-linking resulted in FLIP recruitment. 

Lack of Caspase-8 Expression In Vitro

We then investigated the expression of caspase-8 and FLIP in Rb cells. The levels of FLIP protein were not higher in the cells than in the Fas-sensitive SK-N-MC cells, suggesting that it is not overexpression of FLIP that causes it to replace caspase-8 in the DISC. However, caspase-8 protein was absent in Rb cells (Fig. 8A) . For comparison, caspase-3 expression was comparable between Rb cells and SK-N-MC cells. The caspase adaptor molecule FADD was also equally present in Rb cells and SK-N-MC cells. The absence of caspase-8 expression in Rb cells was also confirmed at the mRNA level by RT-PCR (Fig. 8B) . 

Effect of 5-Azacytidine on Caspase-8 Expression and Sensitivity to DR-Induced Apoptosis

The demethylating agent 5-dAzaC has been reported to increase caspase-8 expression in neuroblastoma cells, thereby restoring sensitivity to DR-induced apoptosis. ^{54 55} We investigated the effect of 5-dAzaC in our retinoblastoma model. We found that treatment with 5-dAzaC resulted in restoration of caspase-8 expression in Rb cells (Fig. 9A) . Pretreatment with 5-dAzaC resulted in loss of Rb cell viability, which was potentiated, however, by treatment with CH11 or TRAIL (P < 0.003 for both CH11 and TRAIL, Fig. 9B ). This suggests that 5-dAzaC restores sensitivity of Rb cells to DR-induced apoptosis. 

Lack of Caspase-8 Expression In Vivo

To confirm that our in vitro findings are applicable to Rb cells in vivo as well, we investigated three primary retinoblastoma specimens by immunohistochemistry for caspase-8 expression. Evidence of the absence of caspase-8 expression was found in our specimens (Fig. 10A) . Thyroid carcinoma cells, known to express caspase-8, ²⁷ served as the positive control (Fig. 10B) . 

Discussion

FasL and TRAIL/Apo2L are members of the TNF family of cell surface ligands that can trigger programmed cell death (apoptosis) in susceptible cells via respective cell surface receptors. In the present study, we investigated the functional status of these apoptotic pathways in two Rb cell lines. We found resistance to DR-induced apoptosis in vitro, due to a deficiency in caspase-8 expression. Treatment with the demethylating agent 5-dAzaC restored caspase-8 expression and sensitivity to apoptosis. 

Both FasL and TRAIL are effector molecules of cell-mediated cytotoxicity and participate in the immune response against neoplastic cells. ¹⁴ Therefore, resistance of the neoplastic cells to the apoptosis-inducing activity of these ligands could contribute to their escape from immune surveillance. Fas-mediated apoptosis is one of these tumor-inhibiting pathways that must be overcome by the neoplastic cell during the carcinogenesis process. ⁶ In addition, studies in several models have implicated the Fas signaling apoptotic pathway in the mechanism of action of DNA-damaging chemotherapeutic drugs and suggested that resistance to Fas-induced cell death protects cells from anticancer chemotherapy. ^{11 12 13} Moreover, the entry of TRAIL in phase I clinical trials as an anticancer agent raises interest in the study of the regulation of its activity. ²⁶  

Several potential mechanisms for resistance to Fas-mediated apoptosis have been described in cancer cells, including downregulation of Fas protein expression, intracytoplasmic sequestration, and failure of the receptor to translocate to the cell surface, production and secretion of a soluble form of decoy receptor (either an alternatively spliced form of Fas or another soluble inhibitor of Fas activation [decoy receptor 3, DcR3]), or mutations of Fas, especially in the death domain (for a detailed review, see Ref. ¹³ ). In the case of our Rb cells, Fas was structurally intact and localized on the cell surface, and soluble Fas levels were negligible. 

A more widespread mechanism of resistance to Fas-mediated apoptosis is the overexpression of antiapoptotic proteins. FLIP is overexpressed in several types of cancers and confers resistance to FasL- and TRAIL-induced apoptosis. ¹³ In our study, FLIP was recruited to the DISC in Rb cells, instead of caspase-8. However, FLIP expression in Rb cells did not exceed that in FasL- and TRAIL-sensitive SK-N-MC sarcoma cells, suggesting that FLIP overexpression is not the primary cause of resistance to DR-induced apoptosis in Rb cells. In agreement, the protein synthesis inhibitor cycloheximide, that has been reported to restore sensitivity to DR-induced apoptosis in other models by downregulating FLIP expression, ^{26 45 56} had no sensitizing effect in our model. 

On the contrary, caspase-8 mRNA and protein were found to be absent in Rb cells in vitro. Treatment with the demethylating agent (5-dAzaC) restored expression of caspase-8 and sensitivity to DR-induced apoptosis in vitro. This suggests that resistance to apoptosis in Rb cells could be due to lack of caspase-8 expression, rather than overexpression of FLIP, and further confirms the structural integrity of the DRs themselves in our model. These in vitro findings should obviously be interpreted with caution until validated in studies involving primary tumor specimens and more cell lines. In preliminary support, immunohistochemistry for caspase-8 in a small number of primary retinoblastomas confirmed lack of expression. Moreover, prior studies have described a similar mechanism of hypermethylation leading to caspase-8 gene silencing, which may be operative in neuroblastomas and other pediatric malignancies. ^{54 55 57 58 59 60} Our data reveal that epigenetic mechanisms, such as aberrant promoter methylation, may lead to caspase-8 gene silencing and resistance to DR-induced apoptosis in Rb cells and suggest that a combination of demethylating agents with DR-activating modalities, such as TRAIL receptor monoclonal antibodies, ^{61 62 63} may benefit patients with retinoblastoma. 

In conclusion, our studies show that the DRs Fas, DR4, and DR5 are structurally intact in our retinoblastoma model. However, the respective apoptotic pathways are defective, due to deficiency in caspase-8 expression secondary to epigenetic gene silencing by hypermethylation. Our data help delineate the apoptotic pathways in Rb cells and could lead to therapeutic applications that would improve patient outcome by sensitizing Rb cells to apoptosis. 

 Supported by the Foundation Fighting Blindness.

 Submitted for publication March 21, 2004; revised August 27, 2004; accepted September 10, 2004.

 Disclosure: V. Poulaki, None; C.S. Mitsiades, None; C. McMullan, None; G. Fanourakis, None; J. Negri, None; A. Goudopoulou, None; I.X. Halikias, None; G. Voutsinas, None; S.Tseleni-Balafouta, None; J.W. Miller, None; N. Mitsiades, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Percentage of cells surviving (mean ± SD), as quantified by MTT, of Y79 (A) and WERI-Rb1 (B) cells treated with the Fas-activating antibody CH11 (500 ng/mL) or recombinant TRAIL (100–1000 ng/mL, as indicated), for 24 hours. All experiments were repeated at least three times, and each experimental condition was repeated at least in quadruplicate wells in each experiment. Results from representative experiments are shown. Both cell lines were resistant to apoptosis induced by the Fas cross-linking antibody CH11 and recombinant TRAIL/Apo2L.

Figure 2.

 

Flow cytometric analyses of surface expression of Fas (A), DR4 (B), DR5 (C), DcR1 (D), and DcR2 (E) in Y79 cells (shaded traces correspond to staining with the respective receptor antibody and unshaded traces depict the respective control staining).

Figure 3.

 

Flow cytometric analyses of surface expression of Fas (A), DR4 (B), DR5 (C), DcR1 (D), and DcR2 (E) in WERI-Rb1 cells (shaded traces correspond to staining with the respective receptor antibody and unshaded traces depict the respective control staining).

Figure 4.

 

RT-PCR for Fas (and sFas) in SK-N-MC, Y79, and WERI-Rb1 cells. GAPDH is shown for comparison.

Figure 5.

 

(A) Y79 cells were treated with recombinant TRAIL (1000 ng/mL) or Fas-activating antibody CH11 (500 ng/mL) in the presence or absence of cycloheximide (CHX) or the NF-κB inhibitor SN50. (B) Y79 cells were treated with recombinant TRAIL (1000 ng/mL) in the presence (□) or absence (▪) of the proteasome inhibitor MG132 (1, 2, or 4 μM). In all cases, the percentage of cells surviving (mean ± SD) was quantified by MTT. All experiments were repeated at least three times, and each experimental condition was repeated at least in quadruplicate wells in each experiment. Results from representative experiments are shown.

Figure 6.

 

Caspase-8 enzymatic activity in Y79 and, for comparison, in SK-N-MC cells before and after treatment with the anti-Fas antibody CH11 (500 ng/mL) or TRAIL/Apo2L (1000 ng/mL) for 1 hour. Caspase-8 enzymatic activity (mean ± SD) was measured with a colorimetric assay kit, normalized for protein content, and expressed in arbitrary units. Stimulation of DRs failed to activate caspase-8 in Rb cells.

Figure 7.

 

Recruitment of intracellular proteins to the Fas DISC in Rb cells. Y79 cells were either first stimulated with biotin-labeled anti-Fas antibody for 10 minutes and then lysed (stimulated cells) or first lysed and then incubated overnight with the biotin-labeled anti-Fas antibody (unstimulated cells). The Fas receptor and the associated proteins (DISC) were subsequently precipitated with streptavidin-agarose beads, immunoblotted, probed with anti-caspase 8 or anti-FLIP Ab, and visualized by ECL. In Y79 cells, Fas cross-linking did not result in caspase-8 recruitment. On the contrary, FLIP was found to interact with Fas, even in the absence of Fas triggering.

Figure 8.

 

(A) Immunoblot analysis for caspase-8, FLIP, FADD, caspase-3, DR4, and DR5 expression in Y79, WERI-Rb1, and SK-N-MC cells. (B) RT-PCR for caspase-8 and FLIP mRNA in Y79, WERI-Rb1, and SK-N-MC cells. Caspase-8 protein and mRNA are absent in Rb cells. SK-N-MC cells served as a positive control.

Figure 9.

 

Role of aberrant gene methylation in caspase-8 expression and sensitivity to DRs. (A) Treatment of Y79 and WERI-Rb1 cells with the methyltransferase inhibitor 5-dAzaC (4 μM for 4 days) restored mRNA expression of caspase-8, as evidenced by immunoblot analysis. (B) WERI-Rb1 cells were treated with 5-dAzaC (4 μM for 5 days) and then further treated with CH11 (500 ng/mL) or TRAIL/Apo2L (1000 ng/mL). The percentage of cells surviving (mean ± SD) was quantified by MTT. Pretreatment with 5-dAzaC sensitized Rb cells to treatment with CH11 or TRAIL (P < 0.003 for both CH11 and TRAIL).

Figure 10.

 

Immunohistochemistry for caspase-8 revealed strong expression in a papillary thyroid carcinoma (B) but not in a primary retinoblastoma (A). Magnification: (A) ×250; (B) ×400.