Translation Factor eIF4E Rescues Cells from Myc-dependent Apoptosis by Inhibiting Cytochromec Release*

Eukaryotic translation initiation factor 4E (eIF4E) markedly reduces cellular susceptibility to apoptosis. However, the mechanism by which the translation apparatus operates on the cellular apoptotic machinery remains uncertain. Here we show that eIF4E-mediated rescue from Myc-dependent apoptosis is accompanied by inhibition of mitochondrial cytochrome crelease. Experiments achieving gain and loss of function demonstrate that eIF4E-mediated rescue is governed by pretranslational and translational activation of bcl-x as well as by additional intermediates acting directly on, or upstream of, the mitochondria. Thus, our data trace a pathway controlling apoptotic susceptibility that begins with the activity state of the protein synthesis machinery and leads to interdiction of the apoptotic program at the mitochondrial checkpoint.

The integration of extracellular information by a cell into a decision to live or die is of fundamental importance during development, wound healing, immune responses, and tumorigenesis (1,2). Most studies addressing the regulation of apoptosis have focused on transcriptional control, signal transduction, and other post-translational regulatory events. More recently, it has become evident that apoptosis is also subject to translational control. Cap-dependent translation involves the assembly of initiation factors at the 5Ј mRNA terminus to form the trimolecular cap binding complex, eIF4F. These factors include the cap-binding protein, eIF4E, 1 an ATP-dependent RNA helicase, eIF4A, and an eIF4G polypeptide (eIF4GI or eIF4GII), which serves as a docking site for eIF4E and eIF4A (3). The function of eIF4E is negatively regulated by members of the translational repressor family, the eIF4E-binding proteins, that sequester eIF4E in a translationally inactive complex (4). Growth factors and other pro-survival stimuli promote phosphorylation of the eIF4E-binding proteins. Hyperphosphorylated eIF4E-binding protein has a decreased affinity for eIF4E, resulting in its liberation to initiate translation.
The apoptotic program can be triggered through at least two distinct signaling pathways with the potential for cross-talk. One pathway, leading to activation of caspase-8, is triggered by ligation of specific cell surface death receptors, such as Fas/ CD95 or tumor necrosis factor ␣ receptor (20). The second pathway, initiated by various stressors such as cytotoxic drugs and radiation, is transduced through a caspase-2-mediated series of steps into mitochondrial release of cytochrome c (21). Subsequent formation of the apoptosome, a complex containing cytochrome c, adapter protein Apaf-1, and procaspase-9 leads to activation of caspase-9 (22). When activated, caspases-8 and -9 activate effectors caspases-3, -6, and/or 7, which in turn cleave critical cellular targets, resulting in death (23). Proteins of the Bcl-2 family tightly regulate mitochondrial release of cytochrome c. Proapoptotic proteins, such as Bid, Bax, Bad, and Bak, form pores in the outer mitochondrial membrane, whereas the anti-apoptotic proteins, Bcl-2 and Bcl-X L , inhibit pore formation (24,25). In most cell types, these two pathways converge, and receptor-induced activation of caspase-8 also results in mitochondrial release of death promoters with subsequent activation of the apoptosome-dependent caspase cascade (26). Thus, the mitochondria integrate a variety of cell death signals, and the ability of Bcl-2/Bcl-X L to interdict apoptosis is one hallmark of mitochondrial involvement in the apoptotic pathway (27,28).
Despite the strong connection between translational control and apoptosis, little is known about the underlying mechanisms. We have previously demonstrated that over-expressed eIF4E averts Myc-dependent apoptosis, at least in part, through a cyclin D1-dependent process (17). To provide further insight into the mechanism of eIF4E rescue, we examined which steps in the apoptotic cascade were blocked by eIF4E in rat embryo fibroblasts sensitized to apoptosis by constitutive expression of c-Myc. Here we show that cells rescued by eIF4E neither release cytochrome c from their mitochondria nor do they activate any downstream steps in the apoptotic cascade. A survey of Bcl-2 family members in rescued fibroblasts revealed most to be in the basal state. However, Bcl-X L was dramatically increased due to the selective recruitment of Bcl-X L mRNA to ribosomes by eIF4E as well as from activation of pre-translational stages of Bcl-X L production. Experiments achieving gain and loss of Bcl-X L function indicated that factors other than Bcl-X L are involved in eIF4E-dependent blockade of cytochrome c release. These data provide the first direct link between translationally mediated antiapoptotic signaling and the apoptotic machinery and underscore the pleotropic nature of survival signaling downstream of up-regulated eIF4E.

MATERIALS AND METHODS
Cell Lines-Cell lines were derived from rat embryonic fibroblasts (REF) as described (17) Analysis of Apoptosis-Cells were subjected to proapoptotic conditions as specified in the text and figure legends. Both adherent and detached cells were collected, fixed in 70% ethanol, washed with PBS, and stained with propidium iodide stain mixture (50 g/ml of propidium iodide, 0.1% Triton X-100, 32 g/ml EDTA, 2.5 g/ml RNase in PBS) for 45 min at 37°C. DNA content was determined by quantitative flow cytometry using the CellQuest program.
Caspase Activity Assay-Cells were subjected to proapoptotic conditions, and both detached and adherent cells were collected, incubated with lysis buffer, and centrifuged for 5 min at 10,000 ϫ g. Active caspase-3 was measured using a CleavaLite caspase-3 activity assay kit (Chemicon). 50 g of lysate was added to a CleavaLite Renilla luciferase bioluminescent substrate that contains the caspase-3 cleavage site, DEVD. After incubation for 1 h at 37°C, fresh luciferase substrate was added, and luminescence was read in a Lumat luminometer (EG&G Berthold). For caspase-9 activity, the cell lysate was centrifuged at 10,000 ϫ g for 1 min, supernatant was collected, and caspase-9 activity was determined by cleavage of LEHD-p-nitroanilide using a caspase-9 colorimetric assay (R&D systems).
Determination of Cytochrome c Subcellular Distribution-Adherent and floating cells were pooled and suspended in mitochondrial buffer (250 mM sucrose, 20 mM HEPES, pH 7.5, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 20 g/ml aprotinin, 1 g/ml pepstatin A). Cells were mechanically lysed using a loose-fitting Wheaton cell homogenizer, and centrifuged at 1000 ϫ g for 10 min to remove nuclei and unbroken cells. The supernatant was centrifuged at 13,000 ϫ g for 20 min to pellet the mitochondria. The resulting supernatant (cytosolic fraction) was centrifuged at 100,000 ϫ g for 1 h. Samples were diluted to obtain a total protein concentration of 0.4 g/l, and cytochrome c concentration in each sample was determined by enzyme-linked immunosorbent assay using an R&D systems Quantikine murine cytochrome c immunoassay kit.
Measurement of Mitochondrial Membrane Potential-Cells were seeded on glass cover slips precoated with fetal calf serum and incubated (24 h at 37°C) in Dulbecco's modified Eagle's medium, 10% fetal calf serum. Cultures were continued with or without 5 M lovastatin for an additional 24 h. Mitochondria were stained by exposing cells to 0.5 g/ml rhodamine 123 dye for 45 min at 37°C. Medium was changed to PBS immediately before analyzing the staining pattern of the cells.
Polyribosome Preparation-15 plates of actively proliferating cells in 100-mm dishes were treated with cycloheximide (100 g/ml) for 5 min, harvested by trypsinization, and lysed as described (50) using 60 strokes in a Dounce homogenizer. Lysate was centrifuged at 10,000 ϫ g for 10 min, and the nuclei pellet was removed. Cytoplasmic extract (1.5 mg measured at A 260 ) was layered onto a 5-ml, 0.5-1.5 M sucrose gradient. The sucrose gradients were centrifuged at 200,000 ϫ g in a Beckman SW50 rotor for 90 min at 4°C. The gradients were fractionated using an ISCO density gradient fractionator monitoring absorbance at 254 nm. Five 1-ml fractions were collected from each sample into tubes containing 100 l of 10% SDS.
Quantification of bcl-x mRNA by Real-time PCR-The RNA from each fraction of the sucrose gradient was extracted using Tri-reagent (Sigma) and quantitated. RNA (5 g) from each fraction was treated with DNase I (DNA-free TM , Ambion, Austin) according to the manufacturer's directions. cDNA was synthesized from 2 g of each RNA sample with a Taqman reverse transcriptase reagent kit (Applied Biosystems) primed with oligo-dT. Rat bcl-x DNA sequences for upper (5Ј-GGAGA-CCCCCAGTGCCATCAAT-3Ј) and lower (5Ј-AGTGCCCCGCCAAAGG-AGAAA-3Ј) primers and rat bcl-2 DNA sequences for upper (5Ј-CACC-CCTGGCATCTTCTCCTTCC-3Ј) and lower (5Ј-GCATCCCAGCCTCCG-TTATCCT-3Ј) primers were selected using the DNA STAR program (DNASTAR, Inc., Madison, WI), and the resulting sequences were synthesized in the University of Minnesota microchemical facility and purified by HPLC. Real time PCR was performed using a LightCycler FastStart DNA Master SYBR Green I Kit (Roche Molecular Biochemicals). The LightCycler PCR protocol consisted of a 10-min denaturation followed by 42 cycles of 95°C for 10 s, 68°C for 5 s, and 72°C for 16 s. Reactions were set up as recommended by the manufacturer, optimized with 4 M MgCl 2 with each primer at 1 M. A portion of the cDNA reaction (9.6 l) was used for amplification of each gradient fraction. Quantification of mRNA was carried out by comparison (linear interpolation) of the number of cycles to saturation in each sample, with the number of cycles in a concurrently run standard containing a known amount of target mRNA (5 concentrations of the standard were used which spanned the range of values for the samples).
Messenger RNA Stability-RNA stability was quantified as described previously (29). To assess stability of bcl-x mRNA, cells were treated with 5 g/ml actinomycin D (Sigma) for the time intervals specified in the text and figure legends. Total RNA was isolated and analyzed by Northern blot.
Generation of REF/Myc/Bcl-X L Clonal Cell Lines-REF/Myc cells were stably transfected with a pAPuro vector containing an 800-bp fragment encoding wild type Bcl-X L (a gift from Dr. E. Prochownik, Children's Hospital, Pittsburgh, PA) using the FuGENE 6 (Roche Diagnostics) transfection technique. Selection of transfected cells was begun after 24 h in complete medium containing 4 g/ml puromycin. Resistant clones were isolated after 12-16 days.
Bcl-X L Antisense Oligonucleotides-Phosphorothioate antisense and nonsense DNA oligodeoxynucleotides were synthesized and purified by HPLC (Operon Technologies, Inc., Alameda, CA). An 18-mer antisense oligodeoxynucleotide sequence spanning the translation start codon of bcl-x mRNA and a control scrambled sequence were (a) antisense 5Ј-CCG GTT GCT CTG AGA CAT-3Ј and (b) scramble 5Ј-CTG AAC GGA GAG ACC CTT-3Ј. Cells were seeded into chambers of 8-well glass chamber slides overnight and shifted to Dulbecco's modified Eagle's medium containing 0.1% fetal calf serum with or without 5 M lovastatin, 40 M oligodeoxynucleotides, or both. Cells were cultivated for 72 h with one media change. Cells were fixed with ice-cold 70% ethanol and stained with acridine orange, and apoptosis was quantified using morphological criteria, as described (16,17).
Cytochrome c Immunostaining-Cells were seeded at a density of 5 ϫ 10 3 /cm 2 in 24-well clusters onto glass cover slips and cultured for 24 h in complete medium. Cultures were either continued in complete medium alone or containing 5 M lovastatin for 18 or 24 h. Cells were rinsed with PBS and fixed in PBS containing 4% paraformaldehyde. The fixed cells were incubated in blocking buffer (PBS containing 5% normal goat serum, 1% bovine serum albumin, and 0.3% Triton X-100) for 30 min and for an additional 2 h in PBS containing 1% bovine serum albumin, 1% normal goat serum, and 1 g of anti-cytochrome c antibody (Promega) per ml. Cells were washed 3 times in PBS and incubated for 30 min in PBS containing 1% bovine serum albumin, 1% normal goat serum, and 1 g of fluorescein isothiocyanate-conjugated anti-mouse antibody (Sigma) per ml. Cells were rinsed 3 times in PBS, nucleistained with DAPI (0.1 mg/ml), and cover slips were mounted onto slides.

Increased Expression of eIF4E Inhibits Death Receptor-mediated and Stressor-induced Apoptotic
Pathways-We previously reported that ectopic expression of eIF4E rescues REF harboring constitutively expressed c-myc (REF/Myc) from apoptosis triggered by either serum restriction (16) or cytotoxic stress (17). To evaluate whether eIF4E also suppresses death receptor-mediated apoptosis, we examined its impact on REF/Myc cell viability after treatment with tumor necrosis factor family ligands. When eIF4E was ectopically overexpressed in REF/ Myc cells (REF/Myc/4E), apoptosis was inhibited in response to each apoptotic trigger tested (Fig. 1), indicating that eIF4E promoted suppression of the death receptor-mediated as well as the stress-induced apoptotic cascades. However, the magnitude of kill in response to tumor necrosis factor-␣ or Fas ligand was modest and not amenable to further study. In contrast, more than 64% of cells underwent apoptotic death in response to lovastatin, a value reduced to nearly the basal frequency by ectopic expression of eIF4E. Based on this result, we restricted our subsequent studies of eIF4E rescue to lovastatin-induced apoptosis.
Ectopic  (Fig. 2C). Caspase-9 followed the same pattern. Activation was noted as early as 4 h in REF/Myc cells, reaching levels more than 3-fold that of control by 16 h (Fig. 2D). In REF/ Myc/4E cells caspase-9 remained in the inactive state.
Cytochrome c normally resides between the inner and outer mitochondrial membranes and is not usually found in the cytoplasm (30). Because cytochrome c is required for activation of caspase-9 (31), we examined the subcellular distribution of cytochrome c after exposure of cells to lovastatin. At base line, cytochrome c was detectable at low levels in the cytoplasm of REF/Myc cells, a value that increased more than 3-fold after 20 h of lovastatin treatment (Fig. 3A). In contrast, cytochrome c remained near basal levels in the cytoplasm of REF/Myc/4E cells after lovastatin treatment, a result corroborated by immunoblot analysis (Fig. 3B).
Mitochondrial inner transmembrane potential collapse frequently precedes cytochrome c release and caspase activation (32). To examine whether this was the case in our system, we treated cells with lovastatin for 24 h and measured the uptake of rhodamine 123, a cationic fluorophore that enters mitochondria in direct proportion to membrane potential (33) Overexpressed eIF4E Selectively Stimulates Expression of Bcl-X L -Deregulated c-Myc mediates apoptosis in some transformed cell lines by selectively decreasing the mRNA and protein levels of the death antagonists Bcl-2 and Bcl-X L (34,35). We therefore examined whether eIF4E had any effect on this interplay between c-Myc and the Bcl-2 family proteins. Immunoblot analysis demonstrated that overexpression of eIF4E resulted in a 7-fold increase of Bcl-X L protein in REF/Myc cells without significantly affecting the expression levels of Bcl-2, Bax, or Bad (Fig. 4).
Overexpression of eIF4E Increases Recruitment of Bcl-X L mRNA to Ribosomes-To examine whether overexpression of eIF4E increased cellular levels of Bcl-X L protein by direct translational activation, total RNA from REF/Myc and REF/ Myc/4E was stratified by sucrose gradient centrifugation to separate translationally active transcripts (more bound ribosomes resulting in more rapid transit through the gradient) from less translationally active transcripts. The resulting fractions were subjected to quantitative PCR analysis for the bcl-X L transcript, comparing it to the bcl-2 transcript as a control. Examination of the absorbance pattern revealed a bias toward heavier polyribosomes, with an increased proportion of RNA in fractions 4 and 5 in cells ectopically expressing eIF4E (Fig. 5A). Real time PCR quantification indicated that the transcript for bcl-2 was distributed similarly in REF/Myc and REF/Myc/4E cells, with fraction 2 (the least translationally active) and fraction 5 (the most translationally active) containing similar amounts of transcript (Fig. 5B). In marked contrast, there was a significant increase in the quantity of bcl-X L mRNA appearing in heavy polyribosomes in REF/Myc/4E cells (Fig. 5C). The average number of bound ribosomes per bcl-X L transcript in REF/Myc/4E cells was 3.7 compared with 1.5 in REF/Myc cells, indicating a 2.5-fold increase in the rate of bcl-X L mRNA translation initiation (36,37). Thus, one mechanism for the increase in Bcl-X L protein was direct translational activation of its mRNA.
Overexpressed eIF4E Increases Cellular Levels of Bcl-X L mRNA-A growing number of studies show that eIF4E directly or indirectly participates in a variety of pre-translational stages of protein synthesis including production and processing of mRNA and its nuclear-cytoplasmic transport (38). To gain insight into the influence of eIF4E on pretranslational events in the synthesis of Bcl-X L , we measured the abundance and stability of its mRNA. Quantitative PCR analysis demonstrated a dramatic effect of eIF4E on steady state levels of bcl-X L mRNA but not on the levels of bcl-2 mRNA (Fig. 6A) Fig. 6C). These data indicate that in addition to direct translational activation, ectopic eIF4E stimulates synthesis of Bcl-X L by increasing the abundance of its transcript. Decreasing Cellular Levels of Bcl-X L Reduces eIF4E-mediated Rescue from Apoptosis-To assess whether increased expression of Bcl-X L is required for the antiapoptotic function of eIF4E, REF/Myc/4E cells were treated with an antisense oligodeoxynucleotide (ASO) directed against human bcl-X L mRNA sequences in the predicted translation initiation region. A scrambled oligonucleotide served as a control. Incubation of cells with the ASO for 72 h significantly reduced the level of Bcl-X L protein, whereas the scrambled oligonucleotide had no effect (Fig. 7A). Continuation of these cultures for an additional 24 h in the presence of lovastatin led to a significant increase in the apoptotic frequency of ASO compared with the scrambled oligonucleotide control (Fig. 7B). However, despite reduction of eIF4E completely suppressed cytochrome c release in the basal state and after lovastatin treatment (Fig. 8B). In the basal state, 15% of the cytochrome c in REF/Myc cells was in the cytoplasm, whereas after ectopic expression of Bcl-X L , virtually no cytoplasmic cytochrome c was detected. After treatment with lovastatin, levels of cytochrome c in the cytoplasm of REF/Myc cells were stable for 12 h, gradually increasing to more than 40% of total cellular cytochrome c by 24 h (Fig. 8B). For the Bcl-X L clonal lines, there was a steady translocation of cytochrome c from the mitochondria to the cytoplasm beginning at 8 h and increasing up to 24 h. At all time points examined, the magnitude of cytochrome c release in cells overexpressing Bcl-X L was significantly less than that observed in REF/Myc, with a rank order of potency reflecting the expression level of Bcl-X L (Fig. 8B). Morphological

DISCUSSION
To begin deciphering the rules governing translational control of cell death, in this report we examined which components of the apoptotic machinery were inhibited when apoptosis was suppressed by ectopic overexpression of eIF4E using Myc-dependent apoptosis in fibroblasts as a model. Here we show that eIF4E rescues cells by blocking release of cytochrome c from the mitochondria. Rescue by eIF4E was mediated in part by its ability to increase cellular levels of Bcl-X L , a key apoptotic antagonist. The eIF4E-induced increase in Bcl-X L was robust, occurring through at least two separate mechanisms, 1) direct translational activation and 2) increase in the abundance of the bcl-x L transcript. However, gain and loss of function experiments indicated that Bcl-X L did not fully account for the potent antiapoptotic activity of eIF4E. These observations indicate that the set point for cellular susceptibility to apoptosis can be governed by translational control.
The rate of protein synthesis is intimately connected with the process of programmed cell death. An ordered shut down of protein synthesis is one of the earliest events during apoptosis, and suppression of global translation can enhance apoptosis (37)(38)(39)(40). Although maintenance of global translation tends to antagonize apoptosis, cap-dependent and IRES-driven modes of translation can impact cell death differently. One mediator of IRES-regulated translation, death-associated protein 5 (DAP5/p97/NAT1), for example, promotes cell death (39,40). In contrast, eIF4E, the principal activator of cap-dependent translation, rescues cells from apoptosis (16 -18).
Our finding that eIF4E modulates the mitochondrial checkpoint for apoptosis provides a glimpse into the mechanism by which the translational apparatus links survival signaling to the apoptotic machinery. We show that Bcl-X L is regulated by direct control its translation, a finding in accord with a recent publication reporting that fibroblast growth factor-2 translationally activates both Bcl-X L and Bcl-2 through a Ras/MEK (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase)-dependent signaling pathway (41). In this regard, eIF4E is known to be activated in a MEK-dependent manner (42), thus defining a putative survival pathway from growth factor receptors through Ras activation of the translational apparatus to the Bcl-X L /Bcl-2-regulated apoptotic checkpoint.
What is the mechanism by which eIF4E selectively increases translation of the Bcl-X L transcript? It is generally agreed that activated eIF4E preferentially stimulates translation of those mRNA with a high degree of complexity in their 5Ј-untranslated region (5Ј-UTR), such as those with upstream open reading frames or a high GC-content (43,44). One documented example of translational control of Bcl-2 family proteins through a 5Ј-UTR is the bcl-2 mRNA, which contains a 35-bp upstream open reading frame (45). The 5Ј-UTR of the bcl-X L message is relatively long and highly GC-rich (46), suggesting that it too may be a good candidate for eIF4E-mediated regulation, which we found to be the case.
Pretranslational stages of Bcl-X L production were also activated by overexpressed eIF4E, suggesting that transcriptional regulation, posttranscriptional maturation, and nuclear export of its mRNA may all be targets of translational control. Transcriptional control of bcl-x is an established mechanism for regulating cellular levels of Bcl-X L (47,48). A number of transcription factors such as Stat5 (49,50), NFB (51), PAX3 (52), and Ets2 (53) may all be involved in regulating bcl-x gene expression. Although we speculate that eIF4E might increase the level of bcl-X L mRNA by activating translation of these factors or their regulators, detailed studies of the 5Ј-and 3Ј-UTR of the transcript will be needed to clarify the mechanism. Our attempts to separately quantify the impact of pre-translational and translational activation of Bcl-X L production by eIF4E using actinomycin D were unsuccessful due to the short half-life of its mRNA. 2 It should be pointed out, however, that at least two transcription factors, c-Fos and c-Myc, are known to be targets for translation control (54). Future studies will undoubtedly expand the list.
Bcl-X L is not the sole mediator of eIF4E-dependent antiapo-ptotic signaling upstream of mitochondria, since gain of Bcl-X L function could not reproduce the robust effects of eIF4E on apoptosis nor could loss of Bcl-X L function completely abrogate eIF4E mediated rescue, thus supporting the available literature that translational control of apoptosis is mediated by a set of antiapoptotic effectors (24,54,55). Because the vast majority of mRNAs encoding growth factors, their cognate receptors and signal transduction pathways have long 5Ј-UTRs or contain upstream open reading frames, it is likely that a plethora of antiapoptotic pathways emanate from activated eIF4E. In this connection, we have previously documented that one downstream effector of eIF4E, cyclin D1, is required for eIF4Emediated rescue from Myc-dependent apoptosis (17). Future investigations will be required to establish whether cyclin D1 and Bcl-X L lie on the same or different survival pathways downstream of eIF4E.
Our data help to clarify how c-Myc and eIF4E cooperate in tumorigenesis. Deregulated c-Myc triggers uncontrolled cell cycle progression and apoptosis, processes that have opposite effects on oncogenesis (56). In this regard, the oncogenic potential of c-Myc, which is masked by its own pro-apoptotic potency, requires an additional event that antagonizes apoptosis (57). Overexpressed eIF4E effectively cooperates with c-Myc in malignant conversion of rodent fibroblasts (10,58). Our previous reports (16,17) and present findings suggest that the impact of eIF4E on Myc-dependent oncogenesis is determined to a significant degree by its ability to suppress the mitochondrial events essential for Myc-activated apoptosis.
Thus, our data trace a novel pathway that controls cell susceptibility to apoptosis. The pathway originates at the protein synthesis machinery and leads to pretranslational and translational modification of the apoptotic program at the mitochondrial checkpoint. Our findings highlight the need for developing new discovery tools that accurately identify those transcripts that are translationally activated by eIF4E to mitigate a proapoptotic stress. In preliminary studies, combining polyribosome preparations to stratify mRNA by the number of bound ribosomes with gene expression microarray has begun to show promise. Independent of the identity of the entire set of translationally activated messages mediating eIF4E rescue, our data begin to explain how the translation initiation apparatus functions to suppress apoptosis and promote oncogenesis, suggesting a plausible mechanism to explain the dramatic oncogenic synergy between eIF4E and a variety of pre-neoplastic alterations that promote both uncontrolled cell cycle progression and cell death.