Requirement for the CD95 Receptor-Ligand Pathway in c-Myc-Induced Apoptosis

See allHide authors and affiliations

Science  14 Nov 1997:
Vol. 278, Issue 5341, pp. 1305-1309
DOI: 10.1126/science.278.5341.1305


Induction of apoptosis by oncogenes like c-myc may be important in restraining the emergence of neoplasia. However, the mechanism by which c-myc induces apoptosis is unknown. CD95 (also termed Fas or APO-1) is a cell surface transmembrane receptor of the tumor necrosis factor receptor family that activates an intrinsic apoptotic suicide program in cells upon binding either its ligand CD95L or antibody. c-myc–induced apoptosis was shown to require interaction on the cell surface between CD95 and its ligand. c-Myc acts downstream of the CD95 receptor by sensitizing cells to the CD95 death signal. Moreover, IGF-I signaling and Bcl-2 suppress c-myc–induced apoptosis by also acting downstream of CD95. These findings link two apoptotic pathways previously thought to be independent and establish the dependency of Myc on CD95 signaling for its killing activity.

Most human tumors have genetic alterations inmyc proto-oncogene family members that result in deregulation of myc expression (1). However, growth-promoting oncogenes like c-myc are also effective inducers of apoptosis, and this lethal attribute may both restrain inappropriate cell growth and maintain normal tissue homeostasis (2). Nevertheless, the molecular mechanism by which the c-myc gene product, the transcription factor c-Myc, triggers the cell death program remains unknown. In contrast, the mechanism of induction of apoptosis by the transmembrane receptor CD95, which belongs to the tumor necrosis factor (TNF) receptor family, is relatively well characterized (3). Upon ligation by its ligand CD95L, a member of the TNF family, CD95 initiates a series of intracellular events, such as activation of a cascade of cysteine proteases called caspases (4), that lead to suicide of the affected cell (5, 6). This killing activity plays a major role in the deletion of autoreactive lymphocytes and maintenance of peripheral tolerance and immunologically privileged sites (3,5, 6). Given the ability of deregulated c-Myc to induce apoptosis, we investigated the possibility that cell killing by c-Myc and CD95 might share some common mechanisms; specifically, whether induction of apoptosis by one of these molecules might be dependent on activation of the other.

Mouse primary embryo fibroblasts (MEFs) and immortalized Swiss 3T3 (S3T3) fibroblasts expressed surface CD95 receptor (Fig. 1A) whose ligation by an agonistic antibody induced apoptotic cell death in a dose-dependent manner (Fig. 1B). Thus, an intact CD95 apoptotic signaling pathway is present in these mesenchymal cells. Both MEF and S3T3 cells also expressed low but detectable amounts of CD95 ligand (CD95L) (Fig. 1C). To examine any involvement of CD95 signaling in apoptosis induced by c-Myc, we used S3T3 cells that expressed a conditional 4-hydroxytamoxifen (OHT)–dependent c-Myc protein (S3T3 c-MycER) (7) and that die by apoptosis upon c-Myc activation in low serum concentration (8). When S3T3 c-MycER cells were incubated either with a monoclonal antibody to CD95L (CD95L mAb) that neutralizes CD95L (Fig. 2A) or with a CD95-Fc chimeric molecule that binds to CD95L and prevents its interaction with endogenous cell surface CD95 (9), c-Myc–induced apoptosis was reduced and delayed in a concentration-dependent manner (Fig. 2B). Thus, efficient induction of apoptosis by c-Myc required surface expression of CD95L.

Figure 1

Swiss 3T3 and mouse embryonic fibroblasts express CD95 and CD95 ligand. (A) Cell surface CD95 expression on fibroblast cells determined by flow cytometric analysis. Cells (1 × 106) were incubated with hamster antibody to mouse (anti-mouse) CD95 immunoglobulin G (IgG) (Jo2; Pharmingen) followed by staining with a secondary fluoroscein isothiocyanate–conjugated anti-hamster IgG (Pharmingen). The profiles obtained by CD95 antibody staining (plain line) are shown relative to secondary antibody staining alone (dotted line). (B) Sensitivity of logarithmic phase S3T3 cells growing in 10% FCS to CD95-mediated apoptosis. CD95 killing was triggered by addition of the anti-CD95 (Jo2) at various concentrations. S3T3 cells with added hamster anti-CD3 (control mAb) served as a control. Cells were observed by time-lapse videomicroscopy, and cumulative cell deaths are shown plotted against time (2, 8, 33). (C) CD95L mRNA expression analysis by reverse-transcriptase (RT)–PCR performed on total RNA isolated from thymocytes (lanes 1 and 2), S3T3 c-Myc ER growing either in the presence (lane 3 and 4) or absence (lane 5 and 6) of 10% FCS, and mouse embryonic fibroblasts (MEFs; lanes 7 and 8) (34). For thymocytes, only one-sixth of the final PCR reaction volume was loaded compared with the other samples. Lanes 2, 4, 6, and 8 represent negative control reactions without RT. A single round of PCR to amplify a 350–base pair actin fragment was used to control for equal amounts of cDNA synthesized in the RT reactions.

Figure 2

Myc killing requires the CD95 pathway. (A) c-Myc–induced apoptosis is blocked by a anti–CD95 ligand. S3T3 MycER fibroblasts were serum-deprived for 48 hours, and c-Myc was then activated by addition of 4-hydroxy-tamoxifen (4-OHT, 100 nM; Research Biochemicals International). Cells were simultaneously treated either with or without various concentrations of a hamster mAb specific for mouse CD95 ligand (CD95L mAb) (35) and analyzed by time-lapse videomicroscopy (33). Cells treated with an isotypically matched hamster anti–mouse CD3 (control mAb) and cells incubated with no antibody served as controls. (B) c-Myc–induced apoptosis is blocked by a chimeric CD95-Fc molecule. S3T3 MycER fibroblasts were treated and analyzed as in (A) after incubation with various concentrations of a chimeric mouse CD95–human Fc molecule (CD95-Fc) (9). One control consisted of cells treated with human Fc alone (Fc; Jackson Immunoresearch Labs) to determine any effect exerted by the human Fc part of the CD95-Fc molecule. The other control consisted of untreated cells. (C) Expression of dominant-negative Fadd protects S3T3 MycER cells from Myc- and CD95-induced apoptotic cell death. S3T3 MycER cells expressing the pBabe Hygro vector alone (Hygro) or DN Fadd (Hygro DN Fadd) were analyzed by time-lapse videomicroscopy for sensitivity to CD95- and c-Myc–induced death (36). Expression of DN Fadd protein in transfected cells was confirmed by protein immunoblot analysis whereas expression of the human DN Fadd was determined by immunoprobing with a mouse anti–human Fadd (Transduction Laboratories). CD95 killing was initiated by addition of anti–mouse CD95 (Jo2) at 100 ng/ml in the presence of 10% FCS. c-Myc–induced apoptosis was induced as described in (A). (D) lpr andgld MEFs are resistant to c-Myc–induced apoptosis. MEFs from wild-type, lpr, and gld mice transiently expressing c-MycER (37) were cultured in medium containing 0.5% FCS in the absence (− 4-OHT) or presence (+ 4-OHT) of hydroxytamoxifen. Cultures were inspected microscopically 72 hours after infection, and representative photographs were taken. For each cell type, a time-lapse videomicroscopy analysis of cell death was performed (right column).

To determine whether CD95 has a direct role in c-Myc–induced cell death, we blocked CD95 effector functions by expressing a dominant-negative mutant of the intracellular Fadd (also called MORT1) death-domain adapter protein (DN Fadd) in S3T3 c-MycER fibroblasts (10-12). This DN Fadd mutant can interact with CD95 but not the apoptosis effector caspase 8; DN Fadd thereby prevents activation of the apoptotic caspase cascade (13, 14). Expression of DN Fadd in S3T3 c-MycER cells blocked, with similar efficacy, apoptosis induced either by CD95 ligation or by c-Myc (Fig. 2C). Thus, CD95 action was required in effective cell killing by c-Myc. We conclude that efficient c-Myc–induced apoptosis requires cell-surface interaction between CD95 receptor and its ligand CD95L.

lpr and gld mice harbor inactivating mutations in the genes encoding CD95 and CD95L, respectively (15, 16). Fibroblasts from such animals might be expected to exhibit greater resistance to c-Myc–induced apoptosis because of their impaired CD95 signaling system. Accordingly, cells transiently infected with a retroviral c-mycER expression vector were exposed to low serum concentrations, and c-Myc was activated by addition of OHT. Activation of c-Myc increased the rate of apoptosis in wild-type (WT) MEFs; in contrast, lpr and gld MEFs showed no increased apoptosis after c-Myc activation (Fig. 2D). This result confirmed that surface expression of both CD95 and CD95L were necessary for efficient induction of apoptosis by deregulated c-Myc. These data are also consistent with, and explain, the previous observation that thelpr mice exhibit accelerated tumorigenesis in response to transgenic lymphocyte expression of the oncogene L-myc, a close relative of c-myc (17).

To investigate how c-Myc recruits the CD95-CD95L killing pathway, we made use of the Rat-1 fibroblast cell line, which harbors the c-mycER construct. These fibroblasts expressed CD95 (18) but were refractory to CD95-induced apoptosis. Nonetheless, they became sensitive to killing by a soluble CD95 ligand molecule upon expression of activated c-Myc (Fig. 3A). Thus, at least in part, c-Myc acts downstream of CD95 by sensitizing cells to the death pathway resulting from interaction between CD95 and its ligand.

Figure 3

Myc, IGF-I, and Bcl-2 regulate the CD95 pathway. (A) Rat-1 cells become sensitive to CD95–induced apoptosis only after activation of c-Myc. Rat-1 c-MycER fibroblasts were grown in 2% FCS for 36 hours. CD95 killing was induced by addition of a soluble CD95 ligand molecule (CD95Ls; Alexis Corporation) at 25 or 100 ng/ml. c-Myc was activated by addition of 4-OHT. (B) IGF-I protects S3T3 cells against CD95-mediated killing. S3T3 MycER fibroblasts were deprived of serum for 48 hours, and anti–mouse CD95 (Jo2; 4 ng/ml) was then added to the culture in the presence (100 ng/ml) or absence of IGF-I; the cells were then analyzed for cell death by time-lapse videomicroscopy. (C) Bcl-2 protects S3T3 cells against CD95-induced apoptosis. S3T3 cells stably infected with vector alone (pB Hygro) or with pBabe Hygro Bcl-2 (pB Hygro Bcl-2) were serum-deprived for 48 hours. Anti–mouse CD95 (Jo2; 100 ng/ml) was then added, and the cultures were analyzed for cell death by time-lapse videomicroscopy.

Induction of apoptosis by c-Myc in fibroblasts is suppressed by the survival factor insulin growth factor–IGF-I (8). Although some details of the anti-apoptotic signaling pathway activated by IGF-I are emerging (19), it remains unclear at what point the survival signals interfere with the apoptotic program. Having shown that CD95-CD95L participates in c-Myc–induced apoptosis, we examined the stage at which IGF-I–mediated inhibition of apoptosis acts relative to the CD95-CD95L interaction. We therefore used an agonistic antibody to CD95 to activate CD95 on serum-deprived S3T3 fibroblasts, either in the presence or absence of IGF-I. IGF-I decreased the rate and extent of apoptosis induced by CD95 ligation (Fig.3B); fetal calf serum (FCS) was even more efficient than pure IGF-I (18), consistent with the fact that serum contains multiple mesenchymal survival factors in addition to IGF-I (8). As the anti-apoptotic action of Bcl-2 functionally maps very close to that of the IGF-I signal in the kinetics of apoptosis (20) and because of the death-suppressing effect of Bcl-2 on c-Myc–induced apoptosis (21), we examined the effect of Bcl-2 in CD95-induced apoptosis. As reported previously (22, 23), expression of Bcl-2 protein in S3T3 c-MycER cells blocked apoptosis induced by antibody to CD95 (Fig. 3C). Thus, the main anti-apoptotic signaling pathway triggered by IGF-I or by the anti-apoptotic Bcl-2 protein interferes with c-Myc–induced apoptosis downstream of the CD95-CD95L interaction, a conclusion consistent with our finding that c-Myc also acts downstream of CD95-CD95L by sensitizing the cell to apoptotic CD95 signaling.

Our data argue for the requirement of an autocrine CD95-dependent signaling mechanism in the induction of apoptosis by the oncogene c-myc and suggest that the proto-oncogene c-mycserves to promote the efficacy with which the cell death signal arising from this autocrine CD95-CD95L interaction engages the cell's apoptotic machinery. The autocrine interaction between CD95 and CD95L must occur on the cell surface because it is disrupted by cell-impermeable blocking agents added to intact cells. Given the low level of CD95L present on the surface of the investigated cell lines the autocrine CD95 death signal is likely to be weak compared with that induced upon ligation by exogenous CD95L, as happens during T cell killing or upon experimental activation of CD95. This finding probably explains why both IGF-I signaling and Bcl-2 afford good protection from c-Myc–induced apoptosis yet are equivocal suppressors of classical CD95-induced apoptosis.

Precedent exists for a functional relation between c-Myc and CD95 signaling. Activation-induced T cell death, which is mediated through the CD95 signaling pathway (24), exhibits a requirement for c-Myc function (25, 26). However, the exact mechanism by which c-Myc triggers death via CD95-CD95L remains obscure. c-Myc is a basic helix-loop-helix–Zip transcription factor, and site-directed mutagenesis studies indicate that c-Myc promotes apoptosis by transcriptional modulation of target genes (2, 27). Although a component of c-Myc–induced apoptosis probably acts through sensitization of cells to CD95 signals, c-Myc may also act upstream of CD95L and CD95 by inducing expression of their cognate genes. However, changes in CD95 expression after c-Myc activation were not detected in the cell lines used (18). Thus, although we cannot exclude the possibility that Myc may also operate through enhancement of CD95L expression, our data favor the idea that Myc acts by sensitizing the cell to CD95-induced apoptosis, as it does to apoptosis induced by DNA and physical damage, inhibitors of macromolecular synthesis, and TNF (28, 29).

That both the promotion of apoptosis by c-Myc and its suppression by IGF-I and Bcl-2 signaling act downstream of CD95 ligation may reflect the fact that the “sophisticated” anti-neoplastic requirements of vertebrate cells evolved around a preexisting and evolutionarily early CD95-type death signaling pathway. It also shows that substantial tiers of regulation and modulation of cell death and survival exist distal to the CD95 receptor, whose ligation can in no way be viewed as an inevitable death sentence.

Our observations are also relevant to tumorigenesis, given the ubiquitous activation and overexpression of c-Myc in human cancers. The potent apoptotic activity of deregulated c-Myc presumably means that tumors can arise only after acquisition of compensatory anti-apoptotic mutations. This is the key to the mechanism of oncogenic synergy between c-Myc and Bcl-2. Bcl-2 suppresses c-Myc–induced apoptosis while leaving the proliferative action of c-Myc unaffected (21). Similarly, blockade of CD95 signaling by CD95L mAb or CD95-Fc in vitro blocks apoptosis but causes no detectable inhibition of cell proliferation (18). Such cells do not appear morphologically transformed, but neither do lpr andgld primary fibroblasts expressing deregulated c-Myc or fibroblasts that coexpress both c-Myc and Bcl-2 (21). It appears that cooperation between c-Myc and suppressors of apoptosis is distinct from the classical type of oncogene cooperation observed between c-Myc and activated Ras.

The possibility thus emerges that dysfunctions in the CD95 signaling pathway can cooperate with c-Myc in carcinogenesis. Indeed, oncogenic cooperation between the absence of CD95 and overexpression ofmyc has been reported in vivo (17) and is consistent with emerging evidence for a possible role of CD95 and CD95L as a tumor suppressor, not necessarily linked to the immune system (30, 31). Many tumor cells exhibit loss of sensitivity to CD95-mediated apoptosis. In addition to providing a mechanism of escape from immune surveillance (30, 31), our studies suggest that resistance to CD95 killing may also be a mechanism by which oncogene-induced apoptosis is suppressed. Dysfunctions in cell suicide pathways mediated by CD95, together with its relatives, may therefore well prove widespread in neoplasia and could offer useful targets for molecular therapy.

  • * To whom correspondence should be addressed. E-mail: hueber{at} and evan{at}


View Abstract

Navigate This Article