Linking of N-Myc to death receptor machinery in neuroblastoma cells.

The oncogene MYCN is amplified in aggressive neuroblastomas in which caspase-8, an essential component of death receptor pathways, is frequently inactivated, suggesting a critical role of death receptor-mediated apoptosis in suppression of N-Myc oncogenic activity. Elevated levels of N-Myc sensitize neuroblastoma cells to apoptosis induced by various death ligands. Using tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced apoptosis as a model, we define the mechanism underlying the sensitization effect. In neuroblastoma cells with increased expression of N-Myc, TRAIL triggers high levels of caspase-8 activation and Bid cleavage, leading to release of cytochrome c and Smac/DIABLO from mitochondria. However, the apoptotic process requires Smac/DIABLO, but not cytochrome c-mediated caspase-9 activation. N-Myc sensitizes neuroblastoma cells to TRAIL by up-regulating TRAIL receptor-2/DR5/KILLER and Bid. Moreover, DR5 mRNA is increased after N-Myc overexpression, and the human DR5 promoter contains two noncanonical E-boxes critical for the transcriptional activation by N-Myc. These findings establish a mechanistic link between N-Myc and death receptor machinery, which may serve as a checkpoint to guard the cell from N-Myc-initiated tumorigenesis.

Apoptosis is executed by caspases, which, based on their roles in apoptotic pathways, can be divided into two general groups. Initiator caspases (such as caspase-2, -8, -9, and -10) function mainly as integrators for upstream apoptotic signals. Once activated, the initiator caspases cleave and activate downstream effector caspases (such as caspase-3, -6, and -7), which are responsible for the proteolytic cleavage of many intracellular proteins, leading to the morphological and biochemical changes associated with apoptosis (1). Two major signaling pathways have been described for activation of initiator caspases in mammalian cells. The intrinsic pathway mediates apoptotic responses to various stress signals such as DNA damage, hypoxia, and growth factor deprivation. It is generally thought that these signals eventually lead to the activation of proapoptotic members of the Bcl-2 family (e.g. Bax and Bak), resulting in mitochondrial release of cytochrome c and other proapoptotic proteins such as Smac/DIABLO (2)(3)(4). Cytochrome c activates caspase-9 through Apaf-1 (5, 6), whereas Smac/DIABLO protects active caspases from inactivation by inhibitors of apoptosis (IAPs) 1 (7). The extrinsic pathway is initiated by interaction of death ligands with their corresponding death receptors including Fas (CD95/Apo1), TNF receptor-1, TRAIL receptor-1/DR4, and TRAIL receptor-2/DR5/KILLER. The interaction results in oligomerization of the receptors that in turn recruit, via adaptor proteins such as FADD/MORT1, multiple molecules of caspase-8. The high local concentration of caspase-8 molecules allows them to activate each other through their low intrinsic protease activity (8). Activated caspase-8 can process downstream effector caspases directly, leading to apoptosis. In certain cell types, however, death receptor-induced apoptosis requires amplification of death signals via a mitochondrial pathway controlled by Bcl-2 family proteins (9). Within this pathway, caspase-8 cleaves Bid, a member of the proapoptotic Bcl-2 family of proteins. The resulting truncated Bid translocates from the cytoplasm to mitochondria, where it promotes the release of cytochrome c and other apoptotic proteins (10,11).
Oncogene-induced sensitization of cells to apoptosis, as first noted in early 1990s (12,13), is an important mechanism for suppression of tumorigenesis (14,15). Tumor suppressors often play a critical role in linking oncogenes to apoptotic machinery. For example, high level expression of many oncoproteins activates the tumor suppressor p53 (16 -18). Activated p53 then targets both intrinsic and extrinsic pathways to promote apoptosis through transcription-dependent and -independent mechanisms (19,20). Alternatively, oncoproteins may directly activate the intrinsic pathway by up-regulating apoptotic proteins such as Bax and Apaf-1 (21)(22)(23). Deregulated activation of oncogenes also renders cells more sensitive to death receptormediated apoptosis (24 -27), and there is evidence that the extrinsic pathway plays an important role in suppression of tumor development (28 -30).
Neuroblastoma, a childhood tumor of the peripheral sympathetic nervous system, provides a particularly useful system for investigation of death receptor-mediated apoptosis as a tumor suppression mechanism. Neuroblastoma cells maintain a functional p53 signaling pathway (31)(32)(33)(34)(35)(36), and Apaf-1 and caspase-9, two critical components of the intrinsic pathway, are expressed and active in all neuroblastoma specimens examined (37). By contrast, caspase-8, an essential mediator of death receptor-triggered apoptosis (38 -40), is frequently inactivated in neuroblastoma cell lines and specimens, especially in those with MYCN amplification (41)(42)(43)(44). Also, overexpression of N-Myc has been shown to sensitize neuroblastoma cells to death receptor-induced apoptosis (25). Together, these findings suggest a pivotal role of death receptor-mediated apoptosis in protecting the cell from N-Myc-initiated tumorigenesis. In this study, we address the question of how elevated levels of N-Myc lead to activation of death receptor apoptotic pathways.
Cell Culture and Cell Lines-The human neuroblastoma cell line SHEP1 and various cell lines derived from SHEP1 were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum (Sigma). All cells were cultured at 37°C in a 5% CO 2 humidified incubator. Retrovirus-mediated gene transfer was performed as described previously (36,47). 24 h after the final round of retroviral infection, cells were cultured in the growth medium containing either 1.0 g/ml puromycin or 100 g/ml hygromycin B for 3 days, and drug-resistant cells were pooled. The percentage of retrovirusinfected cells ranged between 80 and 90%, as estimated in parallel infections using the retrovirus-expressing EGFP. Overexpression or down-regulation of relevant proteins was verified by immunoblotting. For transient transfection of plasmid constructs, 3 ϫ 10 5 cells were transfected with 2 g of plasmid DNA using Lipofectamine Plus reagent according to the manufacturer's protocol (Invitrogen). 48 h after transfection, the cells were selected in growth medium containing 800 g/ml G418 (Geneticin, Invitrogen) for 6 days. G418-resistant cell clones with overexpression or down-regulation of relevant proteins were pooled and used in apoptosis studies.
Apoptosis Induction and Analysis-Exponentially growing cells at 70 -80% confluence were either untreated or treated with the indicated concentrations of TRAIL (Calbiochem), 500 ng/ml of an agonistic anti-Fas antibody (clone CH-11, Upstate), 1 ng/ml human TNF-␣ (Calbiochem) plus 2.5 g/ml cycloheximide (Sigma), or 0.5 g/ml doxorubicin (Ben Venue Laboratories). 24 h after treatment, adherent and floating cells were pooled, collected by centrifugation, and washed once with ice-cold phosphate-buffered saline (PBS). Apoptotic cell death was determined by staining of the collected cells with annexin-V and 7-aminoactinomycin D (7-AAD), using a Guava cytometer according to the manufacturer's protocol (Guava).
Transient Expression and Luciferase Assays-SHEP1 and SHEP1/N-Myc cells in 6-well dishes were transfected with 1.0 g of each luciferase reporter DNA construct and 0.5 g of pSV-␤-galactosidase plasmid (Promega) using the Lipofectamine Plus kit. Cells were lysed 24 h after transfection, and luciferase and ␤-galactosidase activities were assayed using the luciferase and ␤-galactosidase assay systems according to the manufacturer's procedure (Promega). Luciferase activity was normalized to ␤-galactosidase activity to account for differences in the transfection efficiency.
Northern Blot Analysis-Total RNA was prepared from SHEP1/GFP and SHEP1/N-Myc cell lines using TRI Reagent (Molecular Research Center) according to the manufacturer's protocol. RNA samples (20 g/lane) were resolved on 1% formaldehyde agarose gels and transferred to an Immmobilon-NY membrane (Millipore). UV cross-linking was performed with Stratalinker 2400 (Strategene). Blots were hybridized to a 32 P-labeled probe specific for human DR5 (5 ϫ 10 6 cpm/ml). have functional death receptor pathways and express very low levels of endogenous N-Myc which are generally undetectable by immunoblotting (Fig. 1A), thus providing a suitable system for examining the effect of elevated levels of N-Myc on death receptor-triggered apoptosis. The human N-Myc oncogene was introduced into SHEP1 cells by retroviral-mediated gene transfer. After 3-day selection in the presence of hygromycin, infected cells were pooled, which expressed N-Myc at the level compatible to that found in IMR32 cells (Fig. 1A), a neuroblastoma cell line with MYCN amplification (49). The control, GFPexpressing SHEP1 (SHEP1/GFP) cells were relatively resistant to TRAIL-induced apoptosis (Fig. 1B). N-Myc expression markedly sensitized SHEP1 cells to TRAIL, and most of the cells exhibited morphological changes characteristic of apoptosis such as cell shrinkage within 24 h of treatment with 100 ng/ml TRAIL (Fig. 1B). The apoptotic cell death was confirmed further by an annexin-V binding assay in which annexin-V binds to externalized phosphatidylserine on the surface of apoptotic cells (Fig. 1C). In addition, SHEP1 cells expressing N-Myc (SHEP1/N-Myc) showed a marked increase in their sensitivity to an agonistic anti-Fas antibody (clone CH-11) and to TNF-␣ (Fig. 1D). These results confirmed the previous finding that N-Myc is a potent proapoptotic regulator of death receptortriggered apoptosis (25).

N-Myc Sensitizes SHEP1 Cells to Death Receptor-mediated
TRAIL Induces Apoptosis in SHEP1/N-Myc Cells through the Caspase-8-Bid-Mitochondria Pathway-We next wished to define the signaling pathway through which N-Myc sensitizes SHEP1 cells to death receptor-induced apoptosis, using the TRAIL system as a model. As expected, both FADD and caspase-8 are required for TRAIL-induced apoptosis. Expression of a FADD dominant-negative mutant (FADD-DN) (50) or CrmA, a potent inhibitor of caspase-8 (51), completely abrogated N-Myc-induced sensitization of SHEP1 cells to TRAIL (Fig. 2, A and B). By contrast, expression of CrmA had no effect on the sensitivity of SHEP1/N-Myc cells to the DNA-damaging drug doxorubicin (Fig. 2B) SHEP1/GFP cells (Fig. 2C). Concurrent with the caspase-8 activation, a decrease in the levels of full-length Bid and an increase in the levels of truncated Bid (tBid) were observed (Fig. 2C).
The observed Bid cleavage suggests a possible role of mitochondria in mediating TRAIL-induced apoptosis in SHEP1/N-Myc cells. To test this hypothesis, we examined the location of cytochrome c and Smac/DIABLO in SHEP1/N-Myc cells before and after TRAIL treatment. In untreated cells, immunofluorescent staining revealed a characteristic punctate mitochondrial pattern for both cytochrome c and Smac/DIABLO (Fig.  2D). After TRAIL stimulation, the punctate staining of cytochrome c and Smac/DIABLO became diffused in SHEP1/N-Myc cells (Fig. 2D), indicating the release of these factors from mitochondria. Overexpression of Bcl-2 in SHEP1/N-Myc cells prevented the mitochondrial release of cytochrome c and Smac/ DIABLO (Fig. 2D) and protected the cells from TRAIL-induced apoptosis (Fig. 2E). Together, these results suggest that the mitochondrial release of apoptotic factors is an essential step in TRAIL-induced apoptosis in SHEP1/N-Myc cells.
To define further the role of Bid in TRAIL-induced apoptosis in SHEP1/N-Myc cells, we generated two Bid siRNA-expressing retroviral constructs that target different regions of the Bid-coding sequence. Retroviruses produced from both con-structs were able to knock down the levels of Bid in SHEP1/N-Myc cells (Fig. 2F, inset) and to decrease their sensitivity to TRAIL-induced apoptosis (Fig. 2F), indicating that Bid plays an essential role in this apoptotic process.
Smac/DIABLO, but Not Cytochrome c-mediated Caspase-9 Activation, Is Required for TRAIL-induced Apoptosis in SHEP1/N-Myc Cells-Release of cytochrome c into the cytoplasm promotes Apaf-1-mediated caspase-9 activation (6). Indeed, caspase-9 became activated in SHEP1/N-Myc cells after TRAIL treatment, as indicated by the appearance of its processed forms, p35/37 (Fig. 3A). To investigate the role of caspase-9 in TRAIL-induced apoptosis in SHEP1/N-Myc cells, we generated a SHEP1 cell line that expresses both N-Myc and a well characterized caspase-9 dominant-negative mutant (casp9 DN) (52) (Fig.  3B, inset). The control cells expressing GFP and N-Myc were sensitive to apoptosis induced either by the DNA-damaging drug doxorubicin or by TRAIL (Fig. 3B). Expression of casp9 DN completely protected SHEP1/N-Myc cells from doxorubicin-induced apoptosis, demonstrating that casp9 DN is functional in these cells. However, casp9 DN failed to protect the cells from TRAIL-induced apoptosis (Fig. 3B). Thus, TRAIL induces apoptosis in SHEP1/N-Myc cells through a caspase-9-independent pathway downstream from mitochondria.
Because Smac/DIABLO was also released from mitochondria in SHEP1/N-Myc cells treated with TRAIL (Fig. 2D), we examined its role in TRAIL-induced apoptosis. Overexpression of Smac/DIABLO increased the sensitivity of SHEP1 cells to TRAIL (Fig. 3C). We next assessed the effect of down-regulation of endogenous Smac/DIABLO on TRAIL-induced apoptosis. SHEP1/N-Myc cells were transfected with either pcDNA3.1 or a pcDNA3.1-based plasmid that expresses Smac/DIABLO in an antisense orientation, and G418-resistant clones with significant down-regulation of Smac/DIABLO were pooled. SHEP1/N-Myc cells stably transfected with the control plasmid pcDNA3.1 remained sensitive to TRAIL-induced apoptosis (Fig. 3D). By contrast, the antisense-mediated down-regulation of endogenous Smac/DIABLO significantly decreased the sensitivity of SHEP1/N-Myc cells to TRAIL (Fig. 3D), indicating a critical role of Smac/DIABLO in mediating TRAIL-induced apoptosis in these cells.
N-Myc Sensitizes SHEP1 Cells to TRAIL by Up-regulating Apoptotic Proteins in the TRAIL Signaling Pathway-Dissection of the TRAIL apoptotic pathway in SHEP1/N-Myc cells allows us to investigate the mechanism by which N-Myc sensitizes cells to TRAIL-induced apoptosis. We first examined the levels of apoptotic proteins and regulators that may function in the TRAIL signaling pathway in GFP-and N-Myc-expressing SHEP1 cell lines. Immunoblot analysis revealed that N-Myc overexpression in SHEP1 cells resulted in a noticeable increase in the levels of DR5, Bid, and p53 (Fig. 5, A and B). We have shown earlier that Bid is essential for TRAIL-induced apoptosis in SHEP1/N-Myc cells (Fig. 2F). Also, given that these cells do not express detectable levels of DR4 (data not shown), TRAIL most likely signals through DR5 to induce apoptosis. Thus, higher levels of DR5 and Bid may account for the TRAILsensitive phenotype of SHEP1/N-Myc cells. In support of this hypothesis, overexpression of DR5 was sufficient to increase the sensitivity of SHEP1 cells to TRAIL (Fig. 5C). We examined further the requirement of DR5 induction for the TRAIL-sensitive phenotype of SHEP1/N-Myc cells. Two DR5 siRNA retroviral constructs that target different regions of the DR5coding sequence were generated. Infection of SHEP1/N-Myc cells with retroviruses produced with either construct resulted in a significant reduction in the levels of endogenous DR5 (Fig.  5D, inset), and these cells became resistant to TRAIL-induced apoptosis (Fig. 5D). By contrast, SHEP1/N-Myc cells infected with control retroviruses remained highly sensitive to TRAIL (Fig. 5D). Thus, induction of DR5 by N-Myc is essential for sensitizing SHEP1 cells to TRAIL-induced apoptosis.
DR5 Is a Transcriptional Target of N-Myc-Northern blot analysis revealed an increase in DR5 mRNA levels in SHEP1/N-Myc cells (Fig. 6A), suggesting that the induction of DR5 occurred at the transcription step. A major transcriptional activator of DR5 expression is p53 (54). SHEP1 cells express wild-type p53 (33), and consistent with a recent study (55), N-Myc expression resulted in a significant increase in the basal levels of p53 in SHEP1 cells (Fig. 5, A and B). A concomitant increase in the expression of p53 target genes such as p21 Waf1/Cip1 and HDM2 was also observed (data not shown), indicating a functional p53 signaling pathway in SHEP1/N-Myc cells. Therefore, we exam-ined the possibility that N-Myc may require p53 for induction of DR5. Down-regulation of p53 by a well defined siRNA retroviral construct (45) had no effects on the ability of N-Myc to induce DR5 (Fig. 6B) and to sensitize SHEP1 cells to TRAIL (Fig. 6C). We also obtained similar results with SHEP1/N-Myc cells expressing p53(R175H), which functions as a dominant-negative mutant (36, 56) (data not shown). Thus, N-Myc induces DR5 through a p53-independent mechanism.
Next, we examined the possibility of DR5 being a transcriptional target of N-Myc. Luciferase reporter constructs containing serial 5Ј-deletions of the human DR5 promoter have been generated previously, and the reporter construct (DR5/Ϫ1188) containing the promoter sequence spanning Ϫ1188 to ϩ1 (translation start site) was able to induce maximal levels of luciferase activity in transient transfection assays (46). We tested some of these reporter constructs in SHEP1 and SHEP1/ N-Myc cells (Fig. 6D). In both cell lines, only the DR5/Ϫ1188 reporter construct resulted in significantly higher levels of luciferase activity than the promoterless luciferase vector pGVB2 (Fig. 6E). Importantly, SHEP1/N-Myc cells transfected with the DR5/Ϫ1188 reporter construct showed a further increase in luciferase activity compared with SHEP1 cells transfected with the same construct (72-fold versus 27-fold, a 2.7-fold induction; Fig. 6E).
Inspection of the DR5 promoter sequence (Ϫ1188 to ϩ1) revealed the presence of seven noncanonical E-box elements (six CANNTG sites and one CACGCG site, see "Experimental Procedures"). All of these E-box elements have been shown previously to be able to mediate transcriptional regulation by Myc proteins (for review, see Ref. 57). Like other members of the Myc family of transcription factors, N-Myc contains a DNA binding/dimerization domain consisting of the basic, the helixloop-helix, and the leucine zipper regions (bHLHZip). The bHL-HZip domain binds to so-called E-box DNA recognition sequences with the core motif 5Ј-CANNTG (57). To determine the significance of these E-boxes in the regulation of DR5 transcription by N-Myc, SHEP1/N-Myc cells were transfected with DR5/Ϫ1188 luciferase reporter constructs containing individually mutated E-boxes (Fig. 6D). Mutation of the Ϫ221 or the Ϫ229 E-box element resulted in a significant reduction in luciferase activity compared with the wild-type DR5/Ϫ1188 reporter construct (from 91-fold to 32-fold, a 2.8-fold reduction; Fig. 6F). The levels of luciferase activity generated from the two mutant constructs in SHEP1/N-Myc cells were similar to that generated from the wild-type construct in SHEP1 cells (Fig. 6E). Together, these findings suggest that N-Myc activates DR5 transcription through the Ϫ221 and Ϫ229 E-box elements within the DR5 promoter.

DISCUSSION
Identification of signaling pathways that link oncogenes to apoptotic machinery has important implications both for our fundamental understanding of tumorigenesis and for the development of therapeutic approaches that specifically target tumor cells. In the subset of neuroblastoma with MYCN amplification, caspase-8 expression is preferentially suppressed (41). This finding, along with the observation that N-Myc sensitizes neuroblastoma cells to apoptosis induced by various death ligands (25), suggests a coupled relationship between N-Myc and death receptor-triggered apoptosis. To understand the molecular basis of this relationship, we first delineate the signaling pathway essential for TRAIL-induced apoptosis in neuroblastoma cells with increased expression of N-Myc (Fig.  7). The binding of TRAIL to its receptor DR5 results in caspase-8 activation in a FADD-dependent manner. Activated caspase-8 then cleaves Bid, and the resulting truncated Bid in turn induces mitochondrial release of Smac/DIABLO, probably by activation of Bax (58 -60). Once in the cytoplasm, Samc/ DIABLO initiates a process leading to XIAP degradation, thereby alleviating its inhibition on downstream effector caspases that are presumably activated by caspase-8. This signaling pathway is also activated by TRAIL in the human colon cancer cell line HCT116 (59), indicating a critical role of in vector (pRS) control and p53 siRNAexpressing SHEP1/N-Myc cells. ␣-Tubulin levels are shown as the loading control. C, parental SHEP1 and SHEP1/N-Myc cells infected with the vector (pRS) or p53 siRNA-expressing retroviruses were treated for 24 h with the indicated concentrations of TRAIL and analyzed for apoptosis by annexin-V and 7-AAD staining. Each point represents the average Ϯ S.D. of three independent experiments. D, schematic representation of the luciferase reporter constructs containing serial 5Јdeletions of the human DR5 promoter. pGVB2 is a promoterless luciferase vector. Also shown are mutated E-box elements and their locations within the DR5 promoter. E and F, SHEP1 (E) and SHEP1/N-Myc cells (E and F) were transfected with the indicated luciferase reporter constructs along with a ␤-galactosidase reporter plasmid. 24 h after transfection, the cells were harvested and assayed for luciferase and ␤-galactosidase activities. Luciferase values were normalized to ␤-galactosidase activity to account for differences in the transfection efficiency. Fold activation indicates the induction of luciferase activity by transfection with the DR5 promoter-luciferase reporter constructs relative to that induced by transfection with the promoterless luciferase vector pGVB2. Each point in C or bar in E and F represents the average Ϯ S.D. of three independent experiments. Smac/DIABLO in mediating TRAIL-induced apoptosis.
In this study, we define further the molecular mechanism by which N-Myc sensitizes neuroblastoma cell to TRAIL-induced apoptosis (Fig. 7). N-Myc induces DR5 and Bid, two essential apoptotic proteins in the TRAIL signaling pathway. The upregulation of DR5 and Bid is required for N-Myc-induced sensitization of neuroblastoma cells to TRAIL, because abrogation of their induction by siRNA completely reverses the TRAILsensitive phenotype of SHEP1/N-Myc cells. Mechanistically, an increase in the levels of DR5 could facilitate the formation of the DR5 death-inducing signaling complex in response to TRAIL, leading to higher levels of caspase-8 activation, as observed in this study (Fig. 2C). Moreover, elevated levels of Bid could further amplify the death signal through mitochondrial release of Smac/DIABLO. The observed induction of Bid also suggests a mechanism for N-Myc to sensitize neuroblastoma cells to other death receptor signals that require Bid for targeting mitochondria.
Several lines of evidence from this study suggest that the DR5 gene is a transcription target of N-Myc. First, DR5 mRNA and protein levels are increased after overexpression of N-Myc. Second, N-Myc is able to activate transcription from the human DR5 promoter in luciferase reporter assays. Finally, the human DR5 promoter region contains two noncanonical E-box elements critical for the transcriptional activation by N-Myc. The mechanism by which N-Myc induces Bid is currently under investigation. Nevertheless, our findings establish a direct connection between N-Myc and death receptor machinery. Conceivably, this connection makes N-Myc-amplified neuroblastoma cells a target of the surveillance network that employs death receptor-mediated apoptosis to eliminate cancer cells (28 -30). This may also be the driving force for the preferential inactivation of caspase-8 during the development of neuroblastomas with MYCN amplification.
In most cancers, p53 acts as a tumor suppressor by promoting apoptosis in response to oncogenic signals (19,20). Consistent with the notion, both this and other studies show that elevated levels of N-Myc induce p53 in neuroblastoma cells (55) (see also Fig. 5A). We also found that p53 induction is essential for N-Mycinduced sensitization to apoptosis triggered by DNA-damaging drugs (data not shown). However, overwhelming evidence suggests that p53 does not play a significant role in suppression of N-Myc oncogenic activity and of neuroblastoma development. The p53 gene is rarely mutated in neuroblastoma cell lines and primary tumor samples, with or without MYCN amplification (61)(62)(63)(64), and a number of studies demonstrate a functional p53 signaling pathway in a variety of neuroblastoma cell lines examined (31)(32)(33)(34)(35)(36). Importantly, loss of p53 has no effects on the rate and incidence of neuroblastomas in MYCN transgenic mice (65). Together, these observations suggest that the biochemical process that drives neuroblastoma development does not go through a p53-dependent checkpoint. A molecular understanding of this process may shed light on what determines the cell type dependence of p53 as a tumor suppressor and help us identify other tumor suppression mechanisms.
TRAIL is a promising agent for cancer therapy because of its selective cytotoxicity to human cancer cell lines in vitro and to tumor xenografts in immunodeficient mice (66 -68). However, the molecular basis for this differential sensitivity is poorly understood. It was initially suggested that the presence of TRAIL decoy receptors in normal, but not cancer, cells might account for the selectivity (69 -71). But subsequent studies failed to establish a correlation between the expression of the decoy receptors and the sensitivity to TRAIL (72,73). We suggest that activation of oncogenes, a common event in tumorigenesis, may be an important mechanism underlying the selective sensitivity of cancer cells to TRAIL-induced apoptosis. This model is consistent with experimental observations. In addition to N-Myc, a number of oncogenes, including E1A, c-Myc, and Ras, have been shown to sensitize cells to apoptosis induced by TRAIL (74 -76). Moreover, in a study to determine how normal human cells acquire TRAIL-sensitive phenotype during a controlled process of malignant transformation, Nesterov et al. (76) showed that both normal and immortalized human embryonic kidney cells and foreskin fibroblasts were resistant to TRAIL, whereas Ras-transformed cells were susceptible, demonstrating that the acquirement of TRAIL-sensitive phenotype depends on activation of oncogenes. Further exploring of the molecular pathways that link oncogenes to the TRAIL death machinery may offer numerous opportunities for targeting TRAIL to cancer cells.