High Cell Surface Death Receptor Expression Determines Type I Versus Type II Signaling

Abstract

Previous studies have suggested that there are two signaling pathways leading from ligation of the Fas receptor to induction of apoptosis. Type I signaling involves Fas ligand-induced recruitment of large amounts of FADD (FAS-associated death domain protein) and procaspase 8, leading to direct activation of caspase 3, whereas type II signaling involves Bid-mediated mitochondrial perturbation to amplify a more modest death receptor-initiated signal. The biochemical basis for this dichotomy has previously been unclear. Here we show that type I cells have a longer half-life for Fas message and express higher amounts of cell surface Fas, explaining the increased recruitment of FADD and subsequent signaling. Moreover, we demonstrate that cells with type II Fas signaling (Jurkat or HCT-15) can signal through a type I pathway upon forced receptor overexpression and that shRNA-mediated Fas down-regulation converts cells with type I signaling (A498) to type II signaling. Importantly, the same cells can exhibit type I signaling for Fas and type II signaling for TRAIL (TNF-α-related apoptosis-inducing ligand), indicating that the choice of signaling pathway is related to the specific receptor, not some other cellular feature. Additional experiments revealed that up-regulation of cell surface death receptor 5 levels by treatment with 7-ethyl-10-hydroxy-camptothecin converted TRAIL signaling in HCT116 cells from type II to type I. Collectively, these results suggest that the type I/type II dichotomy reflects differences in cell surface death receptor expression.

Introduction

Death receptors play important roles in the cytotoxic effects of immune effector cells (reviewed in Refs. 1–3). Fas/CD95, which is expressed on a variety of normal and neoplastic cells, can trigger cell death when ligated by Fas ligand expressed on cytotoxic T cells (4). Likewise, death receptor 4 (DR4)⁴ and DR5 can trigger cell death when engaged by TRAIL, a cytotoxic ligand on the surface of natural killer cells and interferon-treated monocytes (1, 3, 5–7), or agonistic antibodies that recognize DR4 or DR5 (8).

According to current models (1–3, 5–7), ligation of these receptors results in a conformational change in their cytoplasmic death domains, allowing binding of multimers of the adapter protein FADD to form the so-called death-inducing signaling complex (DISC) (9–11). Oligomerized FADD in turn binds procaspase 8 and/or procaspase 10 monomers (12–14), causing their juxtaposition and activation (15–17). Once activated, these initiator caspases are released to the cytosol (15), where they cleave downstream molecules to trigger the apoptotic process.

Previous studies have shown that signaling initiated at the Fas receptor differs among various cells (18). In particular, so-called type I cells generate large amounts of DISC, activate substantial caspase 8, and directly trigger the apoptotic pathway through caspase 8-mediated cleavage of procaspase 3. In contrast, type II cells generate smaller amounts of DISC, activate less caspase 8, and rely on caspase 8-mediated cleavage of Bid followed by subsequent signal amplification through the mitochondrial apoptotic pathway for sufficient caspase 3 activation to kill the cell (18–20). As a consequence, death ligand signaling in type II cells (but not type I cells) is inhibited by overexpression of anti-apoptotic Bcl-2 family members and other changes that dampen activation of the mitochondrial pathway. In addition, death ligand-induced killing in type II cells, but not type I cells, is inhibited by treatment with phorbol esters (21), which serve as a surrogate for signaling pathways that activate protein kinase C isoforms. Thus, whether signaling proceeds through type I or type II signaling has important ramifications for potential mechanisms of resistance to death ligand-induced killing.

The biochemical basis for the dichotomy into type I versus type II cells is incompletely understood. Early studies suggested that Fas expression is similar between the two cell types (18). Subsequent publications reported that linkage of Fas to the actin cytoskeleton might be different between type I and type II cells (22). More recent studies have suggested that type II cells might have lower levels of the lipid phosphatase PTEN, which is thought to modulate Bcl-2 function (23), or differential sensitivity to the endogenous caspase inhibitor XIAP (24, 25). It is not clear, however, how these differences explain the dichotomy in DISC formation and subsequent signaling in type I versus type II cells.

In the present work we report that type II cells have less Fas receptor on their surfaces than type I cells. Building on this observation, we show that type II cells can be converted to type I cells by forced overexpression of the Fas receptor and that type I cells can be converted to type II cells by Fas down-regulation. In addition, we demonstrate that the same cells can exhibit type I signaling for Fas and type II signaling for TRAIL receptor-induced cell death, suggesting that the choice of signaling pathway is not a cell-intrinsic feature. Collectively, these observations provide new insight into an important cell type-specific difference in death ligand signaling.

EXPERIMENTAL PROCEDURES

Materials

Reagents were purchased from the following suppliers: APO-1-1 murine monoclonal anti-Fas antibody, HS-201 monoclonal anti-DR5 antibody, monoclonal caspase-8 antibody and Super Fas Ligand from Alexis (San Diego, CA); CH.11 monoclonal IgM agonistic anti-Fas antibody from Millipore (Lake Placid, NY); actinomycin D from Sigma; 7-ethyl-10-hydroxy-camptothecin (SN-38) from Tocris Bioscience (Ellisville, MO); rabbit polyclonal antibodies to XIAP, c-Flip, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as well as monoclonal anti-Bcl-x_L from Cell Signaling Technology (Danvers, MA); murine monoclonal anti-FADD and anti-caspase 3, FITC-conjugated Jo2 mouse monoclonal anti-Fas antibody, allophycocyanin (APC)-coupled annexin V, and phycoerythrin (PE) fluorescence quantification kit from BD Biosciences (San Jose, CA); murine monoclonal anti-Bcl-2 from Dako (Carpenteria, CA); goat polyclonal anti-actin antibody and rabbit polyclonal anti-Fas antibody (SC-20) from Santa Cruz Biotechnology; APC- or PE-coupled anti-mouse IgG and peroxidase-coupled anti-mouse IgG1 or IgG2a from Southern Biotechnology (Birmingham, AL); Sepharose coupled to protein G and A from GE Healthcare; enhanced chemiluminescence reagents and NHS-LC-biotin from Pierce; streptavidin-agarose from Invitrogen; recombinant human TRAIL from R&D Systems (Minneapolis, MN); broad spectrum caspase inhibitor N-(2-quinolyl)valyl-aspartyl-(2,6-difluorophenoxy)methyl ketone (Q-VD-OPh) (26) from SM Biochemicals (Pasadena, CA). Mouse monoclonal anti-heat shock protein 90 (HSP90) and rat monoclonal anti-Bid antibodies were kind gifts from D. Toft (Mayo Clinic, Rochester, MN) and A. Strasser (Walter and Elisa Hall Institute, Melbourne, Australia), respectively. Murine monoclonal anti-S peptide antibody was generated as described previously (27).

Cell Culture

Cell lines were obtained from the following sources: Jurkat and HCT-15 cells from P. Leibson and Z. Lou (Mayo Clinic, Rochester, MN), respectively; A498 from J. A. Copland (Mayo Clinic, Jacksonville, FL); Molt-3 from S. Gore (Johns Hopkins University, Baltimore, MD); SKW6.4, H9, and CCRF-CEM from American Type Culture Collection (Manassas, VA). A498 cells were maintained in Dulbecco's modified essential medium containing 10% heat-inactivated fetal bovine serum (FBS) and 1 mm sodium pyruvate. All other cell lines were maintained in RPMI 1640 medium containing 10% FBS and 2 mm glutamine.

Analysis of Cell Surface Fas, DR4, and DR5 Expression

Aliquots containing 2.5 × 10⁵ human cells were fixed with 4% (w/v) paraformaldehyde for 30 min. After 2 washes in PBS containing 2% (v/v) FBS (PBS/FBS), cells were incubated with saturating amounts (1.0 μg) of APO-1-1 murine anti-human Fas antibody, DR4, or DR5 monoclonal antibodies on ice for 45 min, washed twice with PBS/FBS, incubated with PE- or APC-conjugated anti-mouse IgG for an additional 30 min on ice, washed, and subjected to flow microfluorimetry. To quantitate cell surface antibody binding, beads conjugated with known amounts of PE (Quantibrite^TM PE beads, BD Biosciences) were run in parallel, and fluorescence histograms were analyzed as suggested by the supplier.

To further assess cell surface Fas expression, 2 × 10⁶ cells were treated with cell-impermeable biotinylating reagents as described previously (28). Biotinylated cell surface proteins were pulled down with streptavidin-agarose, separated by SDS-PAGE, transferred to nitrocellulose, and probed with anti-Fas antibody (SC-20).

Hepatocytes were also isolated from C57BL/6 mice by collagenase perfusion and purified by Percoll gradient centrifugation as previously described (29, 30). Thymocytes were isolated by passage through wire mesh as described (31). After fixation with 4% paraformaldehyde for 30 min on ice, 2.5 × 10⁵ cells were stained with 1.0 μg of FITC-conjugated Jo2 anti-mouse Fas antibody on ice for 45 min, washed, and analyzed by flow cytometry.

Plasmids

Plasmid encoding Histone H2B fused to the C terminus of enhanced green fluorescent protein (pEGFP-histone H2B) was a kind gift from J. van Deursen (Mayo Clinic, Rochester, MN). Plasmid encoding S peptide-tagged Bcl-x_L was constructed by inserting cDNA encoding codons 2–234 of the Bcl-x_L open reading frame into the AscI and EcoRI sites of pSPN (27). pCMS5A-EGFP-H2B, which encodes EGFP fused to histone H2B and also contains a multiple cloning site behind the CMV promoter, was described elsewhere (32). pCMS5A-EGFP-H2B encoding Bcl-2 was generated by ligating full-length Bcl-2 variant 1 behind the CMV promoter (33). Human full-length Fas was cloned into the EcoRI and NotI sites of pEF1, which contains the eukaryotic initiation factor 1α promoter. All open reading frames were confirmed to be free of mutations by automated sequencing.

Bid siRNA

Short oligonucleotide targeting human Bid (nucleotides 633–651 of coding sequence, GenBank^TM accession number NM_197966.1) was purchased from Ambion (Austin, TX).

Quantitative RT-PCR (qRT-PCR)

RNA was extracted using an RNeasy Mini kit (Qiagen, Valencia, CA). To produce cDNA, 2.0 μg of total RNA was processed with the Superscript III RT-PCR kit (Invitrogen) according to the supplier's instructions. qRT-PCR was performed with SYBR Green PCR Master Mix using an ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Carlsbad, CA). Experiments were performed in triplicate using at least two independent RNA preparations. The sequences of the premiers are as following: Fas forward (5′-TGAAGGACATGGCTTAGAAGTG-3′) and reverse (5′-GGTGCAAGGGTCACAGTGTT-3′); GAPDH forward (5′-ACATCGCTCAGACACCATG-3′) and reverse (5′-TGTAGTTGAGGTCAATGAAGGG-3′); RPLP0 forward (5′-AGATCCGCATGTCCCTTC-3′) and reverse (5′-CCTTGCGCATCATGGTGTT-3′); β-actin forward (5′-ACCTTCTACAATGAGCTGCG-3′) and reverse (5′-CCTGGATAGCAACGTACATGG-3′).

Human Fas Promoter Cloning and Luciferase ReporterAssay

The sequence corresponding to the FAS promoter region −1739/−19 (34) was PCR-amplified from human Jurkat genomic DNA with the following primers: 5-TTAAGGATCCCTCATCTCACTGGG-3′ (forward) and 5′-TGAGGGATCCTCCGAAGTG-3′ (reverse). The 1736-bp PCR product was digested with BamHI and inserted into BglII site of pGL3 basic vector (Promega). Reporter assays were performed using a dual luciferase strategy as recently described (35). 24 h after transfection, cell lysates were prepared, assayed for firefly luciferase activity using a Lumat LB 9501 luminometer (Berthold Technologies, Oak Ridge, TN), and normalized using Renilla luciferase activity (35).

Transfection

To assess the impact of experimental manipulations on Fas-mediated apoptosis, 1 × 10⁷ log phase cells growing in antibiotic-free medium were suspended in 400 μl of medium containing one of the following sets of additives: (i) 4 μg of pEGFP-histone H2B and 1000 nm Bid (or control) siRNA, (ii) 4 μg of pEGFP-histone H2B and 30 μg of pSPN-Bcl-x_L (or pSPN empty vector), or (iii) 30 μg of pCMS5A-EGFP-H2B-Bcl-2 (or pCMS5A-EGFP-H2B with an empty second multiple cloning site). After incubation for 5 min, cells were subjected to electroporation using a BTX 830 square wave electroporator (BTX, San Diego, CA) delivering a single 10-ms pulse at 240 V for leukemia lines or 300 V for solid tumor cell lines. Cells transfected with Bcl-x_L (or Bcl-2) were incubated for 24 h before the addition of CH.11 or etoposide. To assess the effects of Fas up-regulation, cells were incubated for 16 h in the presence of 5 μm Q-VD-OPh to minimize apoptosis induced by Fas overexpression itself, then transferred to Q-VD-OPh-free medium immediately before the addition of CH.11 or diluent.

Generation of Stable Cell Lines

To down-regulate Fas expression, A498 renal cancer cells were transduced with lentiviral particles that were packaged in HEK293T cells transfected with the MISSION® Lentiviral (Sigma) vector pLKO.1-puro containing a hairpin shRNA against human Fas mRNA (GenBank^TM NM_000043) TRCN0000038696 (R#6, CCGGGTGCAGATGTAAACCAAACTTCTCGAGAAGTTTGGTTTACATCTGCACTTTTTG), the packaging vector psPAX2 and envelope vector pMD2.G. Virus containing non-targeting shRNA (#SHC001V) was used to transduce A498 cells as a control. After viral infection on two successive days, cells were cultured in medium containing 2 μg/ml puromycin for a week, stained with APO-1-1 anti-Fas antibody, and examined for cell surface Fas expression by flow microfluorimetry.

To derive cells deficient in both Bax and Bak, lentivirus encoding Bak shRNA (25) targeting nucleotides 1713–1731 of the Bak mRNA (GenBank^TM NM_001188) was packaged and transduced into HCT116 BAX^−/− cells (36), which were selected in puromycin as described above.

Assays for Apoptosis

After the indicated treatment, cells were stained with APC-conjugated annexin V in 140 mm NaCl, 2.5 mm CaCl₂, and 10 mm HEPES (pH 7.4) as instructed by the supplier. After 20,000 events were collected on a BD Biosciences FACSCanto flow cytometer, data were analyzed by gating on EGFP-histone H2B-positive cells (typically 60–70% of transfected Jurkat cells) and assessing APC-annexin V binding. Alternatively, treated cells were stained with propidium iodide and subjected to flow microfluorimetry (37, 38) to detect cells with <2n DNA content.

SDS-PAGE

Cells were solubilized at 4 °C for 30 min in lysis buffer consisting of 1% (w/v) Triton X-100, 30 mm Tris (pH 7.4), 150 mm NaCl, 1% (v/v) glycerol, 1 mm PMSF, 10 μg/ml leupeptin, 10 μg/ml pepstatin, 100 mm NaF, 10 mm sodium pyrophosphate, 1 mm sodium vanadate, and 20 nm microcystin. After centrifugation at 14,000 × g for 15 min to remove insoluble material, aliquots of supernatant containing 50 μg of protein were separated on SDS-polyacrylamide gels, electrophoretically transferred to nitrocellulose, and probed as described previously (39).

DISC Analysis

The Fas DISC was immunoprecipitated as described previously (40). In brief, aliquots containing 5 × 10⁷ cells were treated with 500 ng/ml CH.11 anti-Fas antibody for the indicated periods of time. After cell lysates were prepared as described above, aliquots containing 4 mg of protein were precleared by sedimentation at 14,000 × g for 5 min, supplemented with 10 μg of rabbit anti-mouse IgM that was precoupled to protein A- and protein G-Sepharose using dimethyl pimelimidate, and incubated for 2 h at 4 °C. As a control, 1 ml of cell lysate from 5 × 10⁷ untreated cells was precleared, supplemented with 2.0 μg of CH.11 antibody, and treated with immobilized anti-mouse IgM in parallel with treated samples. At the end of the incubation, beads were sedimented at 14,000 × g for 3 min and washed 5 times with cell lysis buffer. Immunoprecipitated complexes were released from the beads by boiling for 5 min in SDS sample buffer, subjected to SDS-PAGE, transferred to nitrocellulose, and sequentially probed with reagents that recognize FADD and caspase-8.

RESULTS

Expression of FADD, Procaspase 8, c-FLIP, and XIAP Is Similar in Cells with Type I and Type II Signaling

Our initial experiments used a series of human lymphoid cell lines, many of which were previously employed in the original identification of type I versus type II signaling (18). To rigorously establish whether signaling was type I or type II, cells were transfected with empty vector or cDNA encoding Bcl-x_L, treated with agonistic anti-Fas antibody, and examined for apoptosis. Consistent with previous reports, Bcl-x_L diminished apoptosis induced by the agonistic anti-Fas antibody CH.11 (Fig. 1, A and B) or Fas ligand (Fig. 1C) in Jurkat cells (type II signaling) but not SKW6.4 cells (type I signaling; Fig. 1D). Bcl-2 overexpression or Bid down-regulation likewise inhibited CH.11-induced apoptosis in Jurkat cells (supplemental Fig. S1, A and B) but not SKW6.4 cells (supplemental Fig. S1C). This type of analysis demonstrated that SKW6.4 and H9 (supplemental Fig. S1D) cells exhibit type I Fas signaling, whereas Jurkat, Molt-3 (supplemental Fig. S1E) and CEM cells (supplemental Fig. S1F) exhibit type II signaling.

FIGURE 1.

Bcl-x_L rescues Fas-mediated apoptosis in type II Jurkat cells but not in type I SKW6.4 cells. A, 24 h after transfection with pEGFP-Histone H2B (to mark successfully transfected cells) and empty vector or pSPN-Bcl-x_L, Jurkat cells were treated with 25 ng/ml agonistic anti-Fas antibody CH.11 for 5 h. After staining with APC-coupled annexin V, cells were subjected to flow microfluorimetry. Values to the right of each dot plot give the ratio of upper right quadrant to the sum of lower right and upper right quadrants. B, shown is a summary of experiments performed as depicted in panel A. Error bars, mean ± S.D. of three independent experiments. Inset, shown is an immunoblot probed with antibodies to S peptide (for S peptide-tagged Bcl-x_L expression) and β-actin (loading control). C and D, 24 h after transfection, cells were treated with the indicated amounts of FasL (C, Jurkat) or CH.11 (D, SKW6.4) for 6 h. E, cell lysates (40 μg protein) from the indicated type I and type II cells were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with the indicated antibodies. F, cells incubated with 10 μm latrunculin A (Lnt A) and the indicated concentration of CH.11 for 6 h were assayed for apoptosis using APC-annexin V.

Subsequent studies examined expression of components of the extrinsic pathway as well as XIAP. Immunoblotting demonstrated that levels of FADD, procaspase 8, c-Flip, and XIAP did not distinguish type I from type II cells (Fig. 1E). Additional studies failed to demonstrate a differential effect of the actin depolymerizing agent latrunculin A on CH.11-induced apoptosis in Jurkat and SKW6.4 cells (Fig. 1F), suggesting that involvement of the actin cytoskeleton is also not a critical difference between cells with the two types of signaling. These observations prompted us to examine other potential differences between cells that exhibit type I and type II Fas signaling.

Higher Levels of Cell Surface Fas in Type I Cells

Building on previous studies of Fas trafficking to the cell surface (28), we examined cell surface Fas in this panel of lymphoid cell lines using two different approaches. First, cell surface proteins were biotinylated using cell-impermeable reagents, pulled down from cell lysates using streptavidin-agarose, and probed with anti-Fas antibody. Results of this analysis (Fig. 2A) suggested much higher levels of Fas on the surface of SKW6.4 and H9 cells (type I signaling) compared with Jurkat, CEM, and Molt-3 cells (type II signaling). To complement this approach, flow microfluorimetry was utilized to quantitate cell surface Fas levels on fixed but unpermeabilized cells. As indicated in Fig. 2B, the cells with type I signaling had 2–4-fold higher mean fluorescence intensity after staining with APO-1-1 anti-Fas antibody than cells with type II signaling. In further studies we estimated the absolute number of receptors using PE fluorescence calibration beads and approximated cell surface area based on measurements of cell diameter. Results of this analysis (Fig. 2C and Table 1) confirmed that the density of cell surface Fas was higher in type I cells even after differences in size were taken into account.

FIGURE 2.

Type I cells express more cell surface Fas than type II cells. A, 2 × 10⁶ cells were reacted with cell impermeable biotinylating agents as described previously (28). Biotinylated cell surface proteins were pulled down with streptavidin-agarose, separated by SDS-PAGE, transferred to nitrocellulose, and probed with anti-Fas antibody. Bottom panel, longer exposure of top panel. B, cell surface Fas expression was assessed by flow cytometry in type I cells (SKW6.4 and H9) and type II cells (Jurkat, CEM, and Molt-3). Numbers in each histogram indicate the value of mean fluorescence intensity (MFI) observed with anti-Fas antibody (open) minus MFI of isotype control (filled). C, the number of surface Fas molecules per cell was quantified using Quantibrite^TM PE beads after staining with APO-1-1 antibody followed with PE-anti-mouse IgG. Error bars, mean ± S.D. of three independent experiments. D, cell surface Fas expression on mouse type I thymocytes and type II hepatocytes was assessed by flow cytometry. Error bars in right panel, mean ± S.D. from 3 mice.

TABLE 1.

Cell surface Fas expression

Diameters were measured for 100 cells. Binding data are presented as the mean ± S.D. from three independent experiments performed using PE calibration beads.

Cell line	Volume	Surface Fas	Surface Fas
	μm3	molecules/cell	molecules/μm2
SKW 6.4	610 ± 220	78900 ± 14900	225 ± 87
H9	910 ± 280	85300 ± 29500	190 ± 140
Jurkat	820 ± 300	27200 ± 7300	65 ± 33
CEM	550 ± 220	8600 ± 1100	26 ± 6
Molt-3	800 ± 200	22000 ± 3300	52 ± 20

To rule out the possibility that this difference in receptor number occurred only in tissue culture cell lines, cell surface Fas expression was examined on freshly isolated mouse thymocytes and hepatocytes, which exhibit type I and type II Fas signaling, respectively (25, 41). As indicated in Fig. 2D, murine thymocytes had almost 10 times as much cell surface Fas receptor as hepatocytes despite the larger size of the latter cells.

Higher Fas mRNA in Type I Cells Reflects Increased mRNA Stability

To explore the basis for higher cell surface Fas expression in type I cells, we performed a series of experiments using the human leukemia cell lines.

First, cell lysates were examined for Fas expression. As shown in Fig. 3A, higher levels of Fas protein were detected by immunoblotting in whole cell lysates from type I cells compared with type II cells. Moreover, qRT-PCR using three different reference transcripts (Fig. 3B and supplemental Fig. S2) showed that type I cell lines (SKW6.4 and H9) contain more Fas mRNA than type II cells (Jurkat, CEM, and Molt-3). Collectively, these results suggest that type I cells express more Fas message and protein rather than trafficking a disproportionate fraction of Fas protein to the cell surface.

FIGURE 3.

Fas mRNA half-life is shorter in type II Jurkat cells than in type I SKW6.4 cells. A, cell lysates (50 μg of protein) from type I and type II cells were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with the indicated antibodies. B, qRT-PCR comparing Fas mRNA in type I and type II cells. For this graph, Fas mRNA level was normalized to GAPDH mRNA. As indicated in supplemental Fig. S2, similar results were obtained when Fas mRNA was normalized to two other long-lived transcripts. C, 24 h after transfection of Fas promoter-luciferase reporter plasmids into type I and type II cells, firefly luciferase activity was assayed and normalized to Renilla luciferase activity. Error bars, mean ± S.D. of triplicate determinations from two independent experiments. D, cells were treated with 2.4 μm actinomycin D in the presence of 5 μm Q-VD-OPh for the indicated time periods. Fas mRNA was quantified by qRT-PCR and normalized to RPLP0 mRNA. Error bars, mean ± S.D. of three independent experiments. E, half-life of Fas mRNA in type I and type II cells was determined as illustrated in panel D. Error bars, mean ± S.D. of three independent experiments.

It was reported previously that Fas expression is regulated at the transcriptional level (e.g. Refs. 42–45). To investigate the possibility that transcription factors regulating the Fas promoter might be more highly activated in type I versus type II cells, we transfected cells with a reporter construct containing firefly luciferase behind the Fas promoter. Type II Jurkat cells had 3-fold higher luciferase activity than type I SKW6.4 cells (Fig. 3C), arguing against the possibility that differences in promoter activation account for the higher Fas mRNA in type I versus type II cells.

To assess the possibility that Fas mRNA might be more stable in type I versus type II cells, we treated cells with actinomycin D to block transcription (along with Q-VD-OPh to block actinomycin-induced apoptosis), then performed qRT-PCR for Fas mRNA. Results of these experiments (Fig. 3, D and E) indicated that the mean half-life of Fas mRNA is 2.9–3.4 h in type I cells and 1.4–1.5 h in type II cells, suggesting that the higher Fas mRNA in type I cells reflects greater Fas mRNA stability.

Conversion of Type II to Type I Signaling by Fas Overexpression

If lower cell surface receptor expression were responsible for the diminished DISC formation and requirement for amplification through the mitochondrial pathway observed in type II cells, then Fas receptor overexpression should convert the cells to type I signaling. To assess this possibility, the ability of Bcl-x_L overexpression or Bid down-regulation to modulate CH.11-induced apoptosis was assessed in Jurkat cells transfected with empty vector versus Fas cDNA. This analysis demonstrated that Bcl-x_L overexpression (Fig. 4A–4C) or Bid down-regulation (supplemental Fig. S3) protected Jurkat cells transfected with empty vector but not Jurkat cells transfected with Fas cDNA. In contrast, Bcl-x_L continued to protect these cells from etoposide, a mitochondrial pathway-dependent stimulus (46–48), independent of Fas expression (Fig. 4D), ruling out indiscriminant neutralization of the Bcl-x_L in Fas-transfected cells. Consistent with type I behavior, CH.11 also induced more DISC formation and procaspase 8 cleavage in Fas-transfected Jurkat cells than in cells transfected with empty vector (Fig. 4E, DISC). As a result, cleaved caspase 3, which contributes to cellular disassembly (49), was more evident in Fas-transfected cells than in those transfected with empty vector (Fig. 4E, Extract), explaining how Fas transfection could lead to more cell death.

FIGURE 4.

Fas overexpression converts signaling from type II to type I in Jurkat cells. A, Jurkat cells were transiently transfected with pEGFP-histone H2B along with pSPN + pEF1 empty vectors (EV), pSPN-Bcl-x_L + pEF1 (Bcl-x_L), pEF1-Fas + pSPN (Fas), or pEF1-Fas + pSPN-Bcl-x_L (Fas + Bcl-x_L) and cultured in the presence of 5 μm Q-VD-OPh for 16 h. After sequential staining with APO-1-1 anti-Fas IgG and APC-conjugated anti-mouse IgG, cells were subjected to flow microfluorimetry. Numbers in each histogram indicate MFI after Fas antibody staining (open) minus MFI of isotype control (filled) in EGFP-histone H2B⁺ cells. B, cell lysates were probed with antibodies to Fas, Bcl-x_L, and as a loading control, HSP90β, in cells transfected as described in panel A. C, left panel, 16 h after transfection as described in A, Jurkat cells were washed twice with medium lacking Q-VD-OPh and treated with the indicated concentrations of CH.11 agonistic anti-Fas antibody for 5 h. Apoptosis was assayed as illustrated in Fig. 1A. Right panel, shown is a summary of experiments performed as illustrated in the left panel. Error bars, mean ± S.D. of three independent experiments. D, 16 h after transfection as described in A, Jurkat cells were washed twice with medium lacking Q-VD-OPh and treated with 50 μm etoposide for 6 h. Apoptosis was assayed as described in Fig. 1A. E, Jurkat cells were transfected with pEF1 empty vector or pEF1-Fas in the presence of 5 μm Q-VD-OPh. 16 h after transfection, cells were washed twice and treated with 500 ng/ml CH.11 for the indicated times. After cells were lysed, CH.11-bound components were precipitated with rabbit anti-mouse IgM covalently coupled to protein A/G-Sepharose beads, washed, and released in SDS sample buffer. Immunoprecipitates (DISC, left) and aliquots (50 μg) of cell lysates (Extract, right) were subjected to SDS-PAGE, transferred to nitrocellulose, and probed for the presence of the indicated polypeptides. *, splice variants of procaspase 8; **, caspase 8 or caspase-3 cleavage products. F, left panel, 16 h after transfection as described in A, Jurkat cells were washed twice with medium and treated with the indicated concentrations of TRAIL for 5 h. Apoptosis was assayed as depicted in Fig. 1A. Right panel, summary of experiments performed as illustrated in the left panel. Error bars, mean ± range of two independent experiments.

Type I Versus Type II Signaling Is Death Receptor-dependent

If type I versus type II signaling were related to cell surface DR density, cells could potentially exhibit type I signaling for one death ligand and type II signaling for another. Consistent with this possibility, Fas transfection rendered Fas signaling in Jurkat cells resistant to Bcl-x_L (Fig. 4C) but had no impact on ability of Bcl-x_L to protect these cells from TRAIL (Fig. 4F). Instead, cells transfected with either empty vector or Fas cDNA were protected from TRAIL by Bcl-x_L overexpression, suggesting that type I versus type II behavior is specific to the receptor rather than the cell line.

Threshold for Conversion of Type II to Type I Fas Signaling

To further explore the quantitative relationship between the amount of cell surface Fas and type II versus type I signaling, Jurkat cells were transfected with EGFP-H2B (to mark transfected cells) and various amounts of Fas plasmid without or with a fixed amount (30 μg) of Bcl-x_L plasmid. Increasing Fas plasmid led to increased cell surface Fas expression, which peaked at ∼110,000 Fas molecules/cell when 4–5 μg of Fas-encoding plasmid was delivered to 10⁷ cells (Fig. 5A and data not shown). Notably, transfection with 1–2 μg of Fas plasmid resulted in ∼70,000 cell surface Fas molecules/cell, a level similar to type I SKW6.4 cells. Moreover, Bcl-x_L expression did not affect Fas expression (inset, Fig. 5A), allowing us to examine the relationship between Fas expression and Bcl-x_L protection.

FIGURE 5.

Estimation of Fas threshold that converts type II to type I signaling in Jurkat cells. A, Jurkat cells were transfected with 5 μg of plasmid encoding EGFP-histone H2B along with the indicated amounts of pEF1-Fas plasmid (encodes human Fas behind the EF1α promoter) and 30 μg of empty vector or pSPN-Bcl-x_L. After incubation for 16 h in the presence of 5 μm of Q-VD-OPh, cell surface Fas levels were determined in the EGFP-histone H2B⁺ population as illustrated in Fig. 2C. Inset, a Western blot of cell lysates showing Fas expression (from non-Bcl-x_L transfectants). SKW6.4 lysate is included for comparison. B, cells transfected as in panel A were washed twice, treated with CH.11 for 5 h, and assayed for apoptosis by annexin V staining as in Fig. 1A. The percentage of apoptosis inhibition by Bcl-x_L was calculated according to the formula 100 × (1 − (Apoptosis_+Bcl-xL − Apoptosis_{+Bcl-xL,−CH.11})/(Apoptosis_−Bcl-xL − Apoptosis_{−Bcl-xL,−CH.11}), where Apoptosis_−Bcl-xL and Apoptosis_+Bcl-xL represent the percentage of annexin V⁺ cells at any particular CH.11 concentration in the absence and presence, respectively, of Bcl-x_L transfection and Apoptosis_{−Bcl-xL,−CH.11}, and Apoptosis_{+Bcl-xL,−CH.11,} represents the percentage of annexin V⁺ cells in the absence and presence of Bcl-x_L transfection but no CH.11 treatment. The graph presents results of two independent experiments. Inset, Western blots show Bcl-x_L expression in each of the indicated Fas transfectants.

Beginning 16 h after transfection, cells were treated with CH.11 and examined by annexin V staining for apoptosis in the EGFP-histone H2B⁺ population (supplemental Fig. S4). We then calculated the percent inhibition of apoptosis by Bcl-x_L (Fig. 5B). In the face of constant Bcl-x_L (inset, Fig. 5B), increasing Fas expression gradually decreased the protective effects of Bcl-x_L. At a low CH.11 concentration (12.5 ng/ml), Bcl-x_L inhibited Fas-mediated apoptosis up to 70% in Jurkat cells transfected with empty vector but only 20% in cells transfected with 4–5 μg of Fas plasmid. When cells that expressed the maximum amount of cell surface Fas were treated with a higher CH.11 concentration (100 ng/ml), Bcl-x_L completely lost its inhibitory effect.

The results shown in Fig. 5, which demonstrate that increased Fas expression results in diminished Bcl-x_L-induced protection from Fas-mediated apoptosis, are consistent with the hypothesis that high levels of Fas drive type I signaling. These results also suggest that, with the high level Bcl-x_L expression achieved in this study, the threshold for type I signaling (defined as the inability of Bcl-x_L to exert any detectable effect) was ∼95,000 cell surface Fas molecules/Jurkat cell.

Conversion of Type I to Type II Signaling by Fas Down-regulation

To rule out the possibility that the foregoing results were unique to lymphoid cells, we also examined solid tumor cell lines. Consistent with an earlier report that categorized lines based on sensitivity to soluble Fas ligand (50), we found that CH.11-induced apoptosis could be inhibited by Bcl-x_L in HCT-15 cells (type II signaling) but not A498 cells (type I signaling; see below). As was the case with the lymphoid lines, A498 cells had 3–4-fold higher levels of cell surface Fas than HCT-15 cells (Fig. 6A). Further examination indicated that forced overexpression of Fas (Fig. 6, B and the inset in C) rendered Bcl-x_L ineffective at inhibiting CH.11-induced apoptosis in HCT-15 cells (Fig. 6C), as was seen in Jurkat cells (Fig. 4C).

FIGURE 6.

Conversion from type I to type II signaling by Fas down-regulation. A, cell surface Fas expression was determined by APO-1-1 staining followed by flow cytometry. Numbers in each histogram represent the MFI after staining with anti-APO-1-1 antibody (open) minus MFI after staining with isotype control (filled) (left and center panels). Cell surface Fas molecules were also quantified using calibration beads (right panel). Error bars, mean ± S.D. of three independent experiments. B, 16 h after HCT-15 cells were transfected with the indicated plasmids, cell surface Fas was assessed. C, 16 h after transfection as indicated, HCT-15 cells were treated for 12 h with CH.11 and assayed for apoptosis as illustrated in Fig. 1A. Inset, cell lysates were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with the indicated antibodies. D, 7 days after A498 cells were transduced with lentivirus containing Fas shRNA or empty vector, Fas levels on the surface of A498 cells (top panel) were assessed by staining cells with APO-1-1 antibody followed by APC-coupled anti-mouse IgG. Filled, isotype control; solid line open, APO-1-1 staining after transduction of control non-targeting shRNA; dotted line open, APO-1-1 staining after transduction with Fas shRNA. Alternatively, cell lysates from A498 control (cont) and Fas shRNA (Fas sh)-transfected cells were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with the indicated antibodies (bottom panel). Lysates from HCT-15 cells that were transfected with the indicated plasmids for 16 h provided a basis for comparison. E, 16 h after transfection of A498 control and Fas knockdown, cells with 5 μg of pEGFP-histone H2B and 30 μg of pSPN empty vector or pSPN-Bcl-x_L were treated for 12 h with CH.11 and assayed for apoptosis as shown in Fig. 1A. Inset, a Western blot shows Bcl-x_L and Fas expression in transfected cells.

To determine whether Fas down-regulation would convert signaling from type I to type II, A498 cells were transduced with Fas shRNA. After 1 week of selection, total (Fig. 6D, lower panel) and cell surface Fas levels (Fig. 6D, upper panel) in A498 transductants were similar to those of HCT-15 cells. Fas shRNA-transduced A498 cells were less sensitive to CH.11 (Fig. 6E), consistent with lower cell surface Fas expression. Moreover, Bcl-x_L now protected these cells from CH.11-induced apoptosis (Fig. 6E), consistent with conversion to type II signaling.

Cell Surface DR5 Up-regulation Converts TRAIL Signaling in HCT116 Cells from Type II to Type I

The preceding studies suggest that as little as a 3-fold change in cell surface Fas expression is sufficient to convert signaling from type II to type I. Additional experiments were performed to determine whether this was also the case with TRAIL signaling, where it was previously suggested that DR5 up-regulation makes TRAIL-induced killing of HCT116 cells Bax-independent (51, 52).

Earlier studies suggested that TRAIL-induced killing in HCT116 colon cancer cells normally proceeds via a type II signaling pathway (51–53). Consistent with this conclusion, Bcl-x_L protected cells from TRAIL-induced apoptosis (Fig. 7A).

FIGURE 7.

Up-regulation of cell surface DR5 converts TRAIL signaling from type II to type I. A, parental HCT116 cells were transfected with 5 μg of plasmid encoding EGFP-histone H2B along with 30 μg of empty vector or pSPN-Bcl-x_L. After incubation for 16 h, cells were treated with TRAIL for 4 h and assayed for annexin V binding. B, immunoblots show Bax knockout and Bak knockdown in HCT116 cells. C, cells were treated with 50 μm etoposide (left) or the indicated concentrations of SN-38 (right) for 48 h and stained with annexin V-APC (left) or permeabilized, stained with propidium iodide (PI), and subjected to flow microfluorimetry (right). D, left panel, a histogram shows cell surface DR5 expression in HCT116 BAX^−/−/Bak shRNA cells. Filled, isotype control; dotted line open, anti-DR5 staining in cells treated with DMSO; solid line open, anti-DR5 staining in cells treated with 6.25 nm SN-38 for 48 h. Right panel, bar graph showing the MFI observed with DR5 antibody minus MFI of isotype control after HCT116 BAX^−/−/Bak shRNA cells were treated with DMSO or SN-38 for 48 h. E, after HCT116 BAX^−/−/Bak shRNA cells were treated with diluent (DMSO) or 6.25 nm SN-38 for 48 h, cells were incubated with TRAIL for an additional 4 h and assayed for annexin V binding. F, HCT116 BAX^−/−/Bak shRNA cells were transfected with 5 μg of plasmid encoding EGFP-Histone H2B along with 30 μg empty vector or pSPN-Bcl-x_L. After a 16-h incubation to allow transgene expression, cells were incubated with 6.25 nm SN-38 for 48 h, then treated with TRAIL for additional 4 h and assayed for annexin V binding.

To determine whether increased cell surface TRAIL receptor expression, like Fas expression, could convert cells from type II to type I signaling, we developed an HCT116 BAX^−/− line (36) that also had Bak stably knocked down by shRNA (HCT116 BAX^−/−/Bak shRNA, Fig. 7B). This line was resistant to induction of apoptosis by the topoisomerase II poison etoposide and the topoisomerase I poison SN-38 (Fig. 7C), two classes of agents that were previously shown to trigger apoptosis through the intrinsic pathway (46–48, 54), confirming that the mitochondrial pathway was nonfunctional in these cells.

Because enforced overexpression of DR5 was extremely toxic to HCT116 cells (data not shown), we elected to up-regulate DR5 by treating cells with SN-38, the active metabolite of the drug irinotecan that was previously shown to enhance DR5 expression (55). We treated HCT116 BAX^−/−/Bak shRNA cells with SN-38 for 48 h, then measured cell surface DR5 expression and TRAIL sensitivity. As shown in Fig. 7D, a 3-fold increase in cell surface DR5 was detected after a 48-h treatment with 6.25 nm SN-38. In the absence of SN-38 pretreatment, TRAIL did not induce apoptosis in these cells. However, after pretreatment with SN-38, TRAIL-induced apoptosis was observed (Fig. 7E). This apoptosis was not inhibited by Bcl-x_L (Fig. 7F), demonstrating that TRAIL signaling had become type I when cell surface DR5 was up-regulated 3-fold.

DISCUSSION

Since its original description, the distinction between type I and type II death receptor signaling has been widely observed in vitro and in vivo. The biochemical basis for this dichotomy, however, has remained elusive. Results of the present study provide evidence that type I versus type II signaling is determined, at least in part, by the amount of cell surface receptor present. Because type II signaling is inhibited by changes that blunt signaling through the mitochondrial pathway (18) as well as protein kinase C activation (21, 40), the present results suggest that changes in cell surface Fas expression can have implications for surfas sensitivity far beyond what would be expected by simple equilibrium binding calculations.

A number of explanations for the differences between type I and type II signaling have been previously advanced. Several studies have suggested that cells with type II signaling have elevated sensitivity to the endogenous caspase inhibitor XIAP (24, 25). In the present study, however, immunoblotting failed to demonstrate a reproducible dichotomy in XIAP expression between cells with type I and type II signaling (Fig. 1E). Moreover, it has been unclear how elevated XIAP sensitivity could cause the observed difference in DISC formation between cells with type I and type II signaling (18). Instead, it has been reported that the ability of XIAP knockdown to specifically enhance death ligand-induced apoptosis in type II cells can be explained by increased sensitivity of effector caspases to activation (56), presumably reflecting removal of the block to caspase activation that results when XIAP is bound to the apoptosome (57).

While the present work was in progress, Peacock et al. (23) reported that differences between type I and type II signaling correlate with differences in expression of the lipid phosphatase PTEN, which is thought to modulate Bcl-2 function as well as a variety of additional Akt-mediated anti-apoptotic functions (58, 59). Although treatment of Jurkat cells with the phosphoinositide 3-kinase inhibitor LY294002 was shown to diminish the unrestrained Akt signaling and increase the pace of Fas-mediated apoptosis in PTEN-deficient cells (23), the previous study did not establish whether LY294002 altered either DISC formation or the requirement for Bid during Fas-mediated killing of these cells.

In the present study we determined whether signaling was through a type I or type II pathway by directly examining the effects of Bid knockdown or Bcl-x_L/Bcl-2 overexpression. Using this original criterion for type I versus type II signaling, our results demonstrate that cells with type I signaling have a higher density of cell surface Fas than cells with type II signaling (Table 1, Figs. 2 and 6). In particular, we observed high cell surface Fas expression on type I cells using nonpermeable biotinylating agents as well as flow cytometry after staining with anti-Fas antibody. Moreover, Fas down-regulation in cells with high cell surface death receptor converted type I signaling to type II as manifested by enhanced Bcl-x_L sensitivity (Fig. 6E). Conversely, forced overexpression of Fas converted type II signaling to type I, as manifested by both increased DISC formation (Fig. 4E) and diminished Bcl-x_L sensitivity (Fig. 4C). Importantly, Fas overexpression had no impact on type II signaling by TRAIL receptors in the same cells (Fig. 4F), suggesting that the type I/type II dichotomy is not an intrinsic feature of the cell, but is instead related to the death receptor itself.

These observations provide a new conceptual framework for understanding additional differences in cells that signal through type I versus type II pathways after death ligand exposure. It has previously been reported that phorbol esters inhibit Fas-mediated apoptosis in cells with type II but not type I signaling (21). Our recent studies indicated that phorbol esters diminish ligand-induced accumulation of Fas on the cell surface (28). The present demonstration that type I cells have more cell surface Fas before ligation provides a potential explanation for successful death induction in type I cells even when phorbol esters inhibit the accumulation of additional Fas at the cell surface.

In addition to providing new insight into the action of death ligands, the present results also have potential therapeutic implications. Previous studies have shown that the mitochondrial apoptotic pathway is commonly disabled in human cancers (60–62). To the extent that death ligands such as TRAIL induce type II signaling, as appears to be the case in many cell types (52, 63–65), cancer cells would be expected to be resistant as a consequence of blocks to the mitochondrial pathway. On the other hand, additional observations have indicated that Fas and TRAIL receptors can be up-regulated by a variety of treatments, including DNA damaging agents (66–68), the spindle poison paclitaxel (69), the proteasome inhibitor bortezomib (70–72), histone deacetylase inhibitors (70, 73), farnesyltransferase inhibitors (74), and cytokines such as interferon-γ and tumor necrosis factor-α (75). These observations, coupled with our demonstration that an increase in Fas or DR expression as small as 3-fold can convert cells from type II to type I signaling, raise the possibility that appropriately chosen combinations of agents might be used to increase cell surface DR expression and overcome the mitochondrial block that prevents death ligand-induced apoptosis in cells with type II signaling.