Induction of Cell Cycle Progression and Acceleration of Apoptosis Are Two Separable Functions of c-Myc: Transrepression Correlates with Acceleration of Apoptosis

Plasmids and vectors.

The c-MycER deletion mutants were constructed by subcloning the human c-MycER fusion gene (14) into the HindIII site of pBKS(+) vector. Deletion mutants (51) were then constructed in pBKS(+) by replacing the wild-type 5′ end of c-Myc with the various 5′ ends of mutant c-Myc alleles using the flanking EcoRI site and the ClaI site at nucleotide 784 of the c-Myc cDNA. The mutant c-Myc-ER fusion genes were then subcloned back into the HindIII site of the pMV7 retroviral vector (56). The Thr58Ala c-MycER mutant was also constructed by subcloning the 5′ end of aa-58 point mutant c-Myc gene (a gift of J. Woodgett, Ontario Cancer Center) into the EcoRI/ClaI sites of c-MycER in the pBKS(+) vector. VP16-MycER was constructed by inserting a BglII-ClaI polylinker between EcoRV (aa 47) and ClaI (aa 262) sites of c-Myc cDNA and then inserting the transactivation domain of VP16 in frame into this sites. VP16-Myc chimeric cDNA was cloned in frame with estrogen receptor (ER) ligand-binding domain into the pMV7 retroviral vector to generate pMV7-VP16MycER containing the activation domain of VP16 substituting for aa 47 to 262 of c-Myc. The pBabePuro retroviral vector was used to deliver T58AMycER and wild-type MycER (35) into mouse embryo fibroblasts (MEF).

Cell culture.

Rat1a fibroblasts were grown in Dulbecco modified Eagle medium (DMEM) with 10% fetal calf serum (FCS). Cells were infected with either pMV7, pMV7–c-MycER, or pMV7-mutant c-MycER retroviruses as described before (56). Cells with integrated viruses were selected in 500 μg of G418 per ml and maintained in phenol red-free DMEM with 10% certified low-estrogen content FCS (Atlanta Biologicals). For apoptosis assays, cells were plated at density of 100,000 cell per 3-cm dish. Quantitation of apoptosis by DAPI (4′,6′-diamidino-2-phenylindole) staining was performed as previously described (29). For cell cycle progression cells were plated at high density of 10⁶ cells per 10-cm dish to prevent spontaneous apoptosis during prolonged serum deprivation period. MEF were derived from 14.5-day embryos of C57/B6 mice. Passage three MEF were infected with pBabePuro retroviruses followed by selection with 1 μg of puromycin per ml.

Flow cytometric analysis.

For flow cytometric analysis, cells (approximately 10⁴ per time point) were quiesced by serum deprivation (0.5% FCS) for 60 h in phenol red-free DMEM. Subsequently, 4-hydroxytamoxifen (4-OHT) was added to the media for 18 h prior to harvesting cells by trypsinization. Cells were then pelleted in a clinical centrifuge at 1,000 rpm, resuspended in 300 μl of phosphate-buffered saline (PBS), and then fixed by adding 700 μl of cold ethanol while vortexing the mixture. Fixed cells were then repelleted, resuspended in 1 ml of PBS, and pelleted again. Cells were resuspended in 1 ml of PBS containing 10 μg of propidium iodide and 100 μg of RNase-free DNase per ml. Cell cycle profiles were determined using the Lysis II program on the FACScan flow cytometer (Beckton-Dickinson).

BrdU analysis.

Cells were plated on sterile glass coverslips in 30-mm-diameter wells and starved for 60 h in phenol red-free DMEM containing 0.5% FBS. Cells were then treated for 18 h with 10⁻⁷ M 4-OHT. During the last hour of treatment 10 μM 5-bromo-2′-deoxyuridine (BrdU) was added to the media. Immunofluorescent detection of BrdU incorporation as performed according to the manufacturer's protocol (Boehringer Mannheim). Nuclei were counterstained with DAPI, and slides were visualized by fluorescent microscopy and photographed with a digital camera. Cells stained with DAPI and with BrdU incorporated were then counted by printing the digitized images of random fields and scoring at least 300 cells per experimental group.

RNA analysis.

Rat1a cells were grown to 80% confluence and serum starved in 0.5% FCS with phenol red-free DMEM for 60 h. Cells were then treated with 10⁻⁷ M 4-OHT, and RNA was harvested 8 h later. RNA was isolated by using the Qiagen method according to the manufacturer's description and then fractionated on 1% agarose–6% formaldehyde gels and transferred to Duralon-UV membranes (Stratagene) by capillary blotting. After UV cross-linking, membranes were hybridized sequentially to cDNA probes for ODC, gadd45, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) that had been labeled with [³²P]dCTP by random priming (Stratagene). Northern blot analysis for ODC mRNA was done by using a 2-kb EcoRI/BamHI fragment from pBS-ODC (25) and for GAPDH was done by using a 1.4-kb PstI fragment of pBS-GAPDH. Quantitation of autoradiograph signals was performed by using the NIH Image software.

For RNase protection, cells were treated exactly as for Northern analysis except that 15 μg of RNA was used for the RNase protection assays (55). Briefly, RNA was prepared from quiescent cells treated with 4-OHT for 6 h, and equal amounts of RNA per experimental group were then hybridized with 450-bp rat gadd45 (a gift of Linda Penn, Ontario Cancer Center) and 300-bp rat GAPDH probes (55).

Western blot analysis.

Equal numbers of exponentially growing Rat1a cells expressing the pMV7 control, wild type, or c-Myc mutants were lysed in 2× Laemmli buffer (28). After sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transfer to nitrocellulose, equal loading of protein was confirmed by Ponceau S staining. Nitrocellulose was then blotted with the rat monoclonal anti-ER antibody H222 (a kind gift of Geoffrey Greene), detected by Anti-Rat-HRP (Zymed), and developed using chemiluminescence according to the manufacturer's instructions (Amersham).

Cytochrome c immunostaining.

Cells were plated at a density of 125,000 cells per 3-cm dish on glass coverslips. Cells were fixed in PBS containing 4% formaldehyde–0.2% saponin and stained with 1 μg of Hoechst 33258 per ml for 20 min. The fixed cells were incubated in blocking buffer (PBS containing 10% FCS and 0.2% Triton X-100) for 30 min and then for an additional 30 min in PBS containing 0.2% saponin, 2% bovine serum albumin (BSA), and 1 μg of anti-cytochrome c antibody (clone 6H2.B4; PharMingen, La Jolla, Calif.) per ml. Cells were then washed three times with blocking buffer and incubated for 30 min in PBS containing 2% BSA, 0.2% saponin, and 1 μg of tetramethyl rhodamine isocyanate (TRITC)-conjugated anti-mouse antibody (Jackson ImmunoResearch, West Grove, Pa.) per ml. Cells were then rinsed three times with blocking buffer, and coverslips were mounted onto slides.

Soft agarose assays.

Cells (10⁵) were plated subconfluently in 60-mm plates in 0.7% agarose on a 1.4% agarose bed in the presence of 4-OHT with various serum concentrations. Plates were fed once per week with 2 ml of 2% or 10% FCS-DMEM plus 4-OHT. Colonies were scored by projecting plates onto a dry erase board using an overhead projector. Only colonies with a diameter of >1.0 mm were scored.

The ability of c-Myc to accelerate apoptosis does not correlate with the integrity of Myc's known transcriptional activating domains.

Ectopic expression of c-Myc accelerates apoptosis upon growth factor withdrawal (6, 8, 15, 56). The ability of c-Myc to accelerate apoptosis is thought to be associated with its transcriptional activation capability and with its ability to induce cell cycle progression (40). However, treatment of cells with cycloheximide does not diminish c-Myc-accelerated apoptosis (15, 54), suggesting that transcriptional activation and the subsequent induction of de novo protein synthesis by c-Myc are not required to accelerate apoptosis. To directly address this apparent paradox, we initiated studies to define the transcriptional regulatory domains required for c-Myc-accelerated apoptosis.

As a first step in these analyses, we generated polyclonal Rat1a cell lines stably expressing equivalent amounts of a panel of c-Myc mutants that are schematically illustrated in Fig. 1A. The c-Myc deletion mutant dl41-53 (MB I deletion) has been shown previously to have diminished transactivating properties, whereas the mutant dl106-143 (MB II deletion) has been shown previously to be defective in transrepression. The deletion mutant dl41-178 eliminates both MB I and MB II and is unable to transactivate target genes. The chimeric VP16-Myc was constructed by replacing the c-Myc aa 47 to 262 with the acidic VP16 transcriptional domain and fusing this domain in frame to the remaining DNA binding domain of c-Myc. The c-Myc T58A mutant is an alanine substitution for the threonine at aa 58, a highly conserved phosphorylation site in MB I. This site and its immediately flanking amino acids are mutational hotspots for c-Myc in Burkitt's lymphoma primary tumor samples and in many Burkitt's lymphoma cell lines (1, 7, 58). In addition to being a mutational hotspot in Burkitt's lymphomas, the equivalent amino acid is replaced with a nonphosphorylatable residue in the highly transforming v-myc MC29, OK10, and MH2 strains of avian leukemia retroviruses, suggesting that this residue may play a critical role in c-Myc-mediated transformation (52, 53). In experimental tissue culture systems, mutation of c-Myc Thr58 to a nonphosphorylatable amino acid and other mutations clustered near this site have been associated with an increased transforming potential of c-Myc (21, 23, 52). This phosphorylatable residue is conserved in the vertebrate c-Myc and in N-Myc and L-Myc (20).

To define the domains of c-Myc that are necessary for acceleration of apoptosis, we used a well-characterized inducible system in which c-Myc is activated by the addition of the ER ligand, tamoxifen (4-OHT). Retroviral vectors encoding a panel of amino-terminal deletion and point mutations of c-Myc fused in frame with the hormone-binding domain of the ER were constructed. Retrovirus stocks were generated and used to infect Rat1a cells as described in Materials and Methods. Expression of similar levels of the mutant and wild-type c-Myc proteins, in the presence of OHT, was verified by Western blot analysis using antibodies directed at the hormone-binding domain of the ER (Fig. 1B) and an anti-Myc antibody 9E10 (data not shown).

To evaluate c-Myc-mediated acceleration of apoptosis, Rat1a cell pools expressing either wild-type or mutant Myc proteins were plated subconfluently and allowed to adhere. The next day, cells were subjected to serum deprivation, and 4-OHT was added to activate c-Myc via the ER fusion protein. Quantitation of apoptosis was done using DAPI and UV microscopy to score apoptotic nuclei with condensed chromatin as previously described (29). As shown in Fig. 2 and consistent with previous results (29, 54), 36 h after serum deprivation c-Myc accelerates cell death of Rat1a cells to 90% in comparison with the 30% of cell death observed in Rat1a cells that do not ectopically express c-Myc. Interestingly, the transactivation-defective deletion mutant dl41-53 was still able to accelerate apoptosis to approximately the same extent as had wild-type c-Myc. Thus, aa 41 to 53 are not required for c-Myc to accelerate apoptosis. In contrast, expression of dl106-143 (deletion of MB II, Fig. 2) did not accelerate apoptosis, suggesting that MB II, and possibly transcriptional repression, plays a critical role in c-Myc-accelerated apoptosis. In fact, expression of dl106-143 consistently resulted in diminished apoptosis compared with the pMV7 Rat1a control cell line. This was also previously observed by others (57). This may account, in part, for the ability of this mutant to act as dominant negative and to attenuate proliferation. Indeed, this mutant modestly inhibits entry into S phase of cell cycle (Table 1). The attenuation of growth is probably dependent on having an intact MB I and deletion of MB II because the deletion of aa 41 to 178 does not exhibit this phenotype (data not shown) and substitution with VP16 with both MB I and MB II deleted also does not exhibit this phenotype. VP16-Myc expression, on the other hand, showed a level of apoptosis similar to that of the parent cell line. Taken together, these results suggest that there is no correlation between the regions required for transactivation by c-Myc (MB I) and its ability to accelerate apoptosis. The integrity of the conserved region implicated in transcriptional repression (MB II), however, does correlate with the ability of c-Myc to accelerate apoptosis. Surprisingly, the T58A mutation that leaves the transactivation domains intact diminishes apoptosis to approximately 50% of either the wild type or dl41-53 (Fig. 2). This is even more significant considering that the T58A mutation may augment the c-Myc protein half-life, as was recently reported using transient transfection of mutant and wild-type c-Myc proteins (47). The results could be explained by the reduced ability of the Thr58Ala mutant to repress transcription and to release cytochrome c from mitochondria (see below).

c-Myc-accelerated apoptosis correlates with transrepression and not transactivation of gene expression.

The results described above led us to examine directly the correlation between transactivation, transrepression, and apoptosis functions of c-Myc. The most well-characterized target of c-Myc transcriptional activation is the ODC gene. ODC has been shown to be induced by c-Myc in transient and stably expressing cell populations (reviewed in reference 13). Both the rat and human ODC genes contain two consensus CACGTG Myc-binding sites located in the same position in the first intron. We have previously shown 4-OHT-dependent activation of ODC expression in BALB/c 3T3 and Rat1a fibroblasts following expression of the conditionally active c-MycER chimeric protein (55). Therefore, we measured the relative induction of endogenous ODC gene expression and the acceleration of apoptosis in cells expressing individual mutant c-MycER proteins. To determine the extent of ODC transactivation by c-Myc, Rat1a cell pools were cell cycle arrested by serum deprivation, and c-Myc was activated by the addition of 10⁻⁷ M 4-OHT. Analysis of the induction of ODC mRNA transcript normalized for GAPDH expression (Fig. 3) revealed that transactivation of ODC in cells expressing the MB I mutant dl41-53 was inhibited approximately 30% compared with cells expressing wild-type c-Myc. Thus, the MB I deletion mutant dl41-53 was unable to transactivate ODC as robustly as wild-type c-Myc, even though it retained the ability to accelerate apoptosis. In contrast, expression of the phosphorylation site mutant T58A resulted in potent transactivation of ODC and a significant loss of apoptotic function. Similarly, the VP16-Myc chimeric protein also induced significant ODC expression (Fig. 3A) and yet expression of this protein had no effect on apoptosis (Fig. 2). In summary, these results indicate that c-Myc-mediated transcriptional activation has no direct correlation with acceleration of apoptosis and suggest that induction of apoptosis may be mediated by mechanisms independent of transactivation using ODC as a readout for transactivation.

The results presented in Fig. 2 and 3 strongly suggest that MB II, a region associated with transcriptional repression, may play an essential role in the ability of c-Myc to augment apoptosis. In order to determine whether transcriptional repression correlates with apoptosis, we studied repression of the well-established c-Myc target gene gadd45 (36) under conditions similar to those for the ODC measurements. RNA was harvested from cells undergoing serum deprivation for 60 h, followed by 4-OHT stimulation for 6 h. Figure 4 shows an RNase protection assay in which endogenous gadd45 mRNA levels are clearly repressed 6 h after activation of wild-type c-Myc expression (lane 2). Repression of gadd45 mRNA is not seen following activation of dl106-143 (lane 4), a finding consistent with the essential role that this region (MB II) plays in mediating transcriptional repression. Interestingly, the MB I T58A mutant of c-Myc consistently showed reduced transcriptional repression of gadd45 (lanes 5 and 6), while the dl41-53 (lanes 7 and 8) deletion mutant showed wild-type levels of gadd45 repression. The RNase protection assay was repeated twice with consistent results. These results imply that dephosphorylation of Thr58 diminishes transrepression by c-Myc while preserving transactivation and suggest that phosphorylation may play an important role in c-Myc function as a transcriptional repressor. Furthermore, while c-Myc-mediated transactivation does not correlate with the ability to accelerate apoptosis, the transrepression function of c-Myc correlates directly with Myc's apoptotic function.

Induction of cell cycle progression and acceleration of apoptosis are separable functions of c-Myc.

The c-Myc mutants enabled us to determine whether there is a strict correlation between the ability of c-Myc to induce cell cycle progression and to accelerate apoptosis. In order to measure cell cycle progression and apoptosis simultaneously, cells expressing different c-Myc alleles were serum deprived for 60 h so that approximately 80% of cells were in G₀/G₁ prior to c-Myc activation (Table 1). At 18 h after c-Myc activation with 4-OHT, we employed two techniques to measure induction of S phase: (i) evaluation of individual cell BrdU incorporation, as measured by indirect immunofluorescent microscopy, and (ii) fluorescence-activated cell-sorting (FACS) examination of cell populations (10⁴ cells) for DNA content (Fig. 5). Table 1 illustrates that the two techniques correlated well; with either method cells evaluated 18 h after activating ectopically expressed mutant c-Myc proteins either showed robust induction of S phase similar to wild-type c-Myc or no S-phase induction (Table 1). Wild-type c-Myc and dl41-53 and T58A mutants were able to induce S phase equally well, while expression of VP16-Myc, dl106-143, and dl41-178 did not stimulate cell cycle progression (Table 1 and Fig. 5) and the cells remained arrested indefinitely (data not shown). Since VP16-Myc led to a robust transactivation of ODC but could not induce cell cycle progression, ODC induction is clearly not sufficient to drive fibroblasts into the cell cycle. Also, despite diminished ODC induction in dl41-53-expressing cells, these cells were able to enter the cell cycle as efficiently as cells expressing wild-type c-Myc. Most interestingly, S-phase entry and apoptosis induced by c-Myc could be genetically dissociated in cells expressing the T58A c-Myc protein. This lymphoma-associated mutation in MB I showed robust induction of S-phase, diminished acceleration of apoptosis, and diminished ability to repress transcription. Taken together, these results suggest that the induction of cell cycle progression and acceleration of apoptosis are separable functions of c-Myc.

The results showing that T58A mutant has diminished ability to accelerate apoptosis were obtained in immortalized Rat1a cells that have wild-type p53 (54) and therefore are likely to have a lost p19-ARF during the process of immortalization (59). Because it was shown that in primary MEF the ability of c-Myc to elicit apoptosis is dependent, at least in part, on its ability to transactivate p19-ARF and the subsequent stabilization of p53 (59), we have analyzed the T58A mutant in nonimmortalized, early-passage MEF. MEF at passage 3 were infected with pBabePuro(WTMycER), pBabePuro(T58AMycER), and pBabePuro retroviruses. Stably infected pools of cells expressing T58AMyc-ER, WTMyc-ER, and vector alone were selected. As shown in Fig. 6, both WTMyc and the T58A mutant accelerate apoptosis in MEF grown in either 0.5 or 2% FCS. However, the ability the T58A mutant to accelerate apoptosis is also diminished in MEF, although not to the same extent as was observed in Rat1a cells (see Discussion).

The diminished ability of Thr58Ala mutant to accelerate apoptosis correlates with its reduced ability to induce the release of cytochrome c from mitochondria.

In most cases, apoptosis is initiated by the loss of integrity of mitochondria and the release of cytochrome c (47). The released cytochrome c acts as a cofactor to initiate the apoptotic cascade and the activation of caspases that execute apoptosis. We and others have recently shown that c-Myc is able to induce the release of cytochrome c from mitochondria in a caspase-independent manner (27, 30). It was shown that the mechanism by which c-Myc is sensitizing cells to apoptotic stimuli is dependent on its ability to release cytochrome c from mitochondria (27). We therefore compared the abilities of wild-type c-Myc and T58A mutant to release cytochrome c. Quantitation of cytochrome c release was performed as previously described (30). After activation of c-MycER with OHT the cells were maintained in 0.5% FCS in the presence the caspase inhibitor zVAD for 3 h and then fixed and immunostained with cytochrome c antibodies. As a negative control in this experiment we used the cells expressing the deletion mutant dl106-143 that showed even less apoptosis than did the control Rat1a cells (Fig. 2). Rat1a cells coexpressing Bcl-2 and wild-type c-Myc served as a positive control. As shown in Fig. 7, the release of cytochrome c is directly correlated with the level of apoptosis (Fig. 2). The T58A mutant is about 70% less effective in releasing cytochrome c in comparison with wild-type c-Myc (Fig. 7). These results suggest that the primary defect in the ability of T58A mutant to elicit apoptosis is its inability to effectively mediate the release cytochrome c from mitochondria.

The transformation potential of wild-type and mutant c-Myc proteins reflects a balance between c-Myc-induced cell cycle progression and apoptosis.

The balance between cell cycle progression and apoptosis is thought to contribute to the ability of cells to establish tumors and to anchorage-independent growth (38, 46, 50). The transformation potential of cells expressing the various mutant c-Myc proteins was assessed by measuring anchorage-independent growth in low serum, in which apoptosis is accelerated in wild-type c-Myc-expressing cells, versus high serum, in which growth and survival factors protect against c-Myc-accelerated apoptosis. Rat1a cells can be transformed to anchorage-independent growth by wild-type c-Myc alone (49, 51). Previously published reports of transformation assays comparing cells expressing different mutant c-Myc alleles were done in high (10% FCS) under conditions in which the differences in apoptotic potential were suppressed. We therefore assessed the effect of variable serum concentrations (i.e., high [10%] or limited [2%] FCS) and, in turn, variable apoptotic rates on anchorage-independent growth of colonies by performing the agarose assays in high and low serum concentrations. As a control a polyclonal cell line that coexpresses MycER and Bcl-2 was also included in these studies. The level of MycER expression in the MycER/Bcl-2 cell line was comparable to the level in the MycER cell line (data not shown). We found that, in 10% serum, the cell pools expressing wild-type Myc, T58A mutant, and Myc/Bcl-2 formed similar numbers of colonies on soft agarose and the dl41-53 mutant formed colonies 76% of the wild type (Fig. 8). The deletion of aa 41 to 53 does not affect apoptosis and cell cycle progression, but it inhibited transactivation of ODC and cellular transformation, implying that activation of ODC or other growth-related genes is required for full transformation by c-Myc.

Under low-serum (2%) conditions (predicted to accelerate apoptosis because of limiting growth and survival factors), the cell pools expressing c-Myc proteins capable of initiating S phase but protected from apoptosis (cells expressing c-Myc/Bcl-2 and the Thr58Ala mutant) exhibited significantly more anchorage-independent colonies than did wild-type c-Myc (Fig. 8). Thus, the ability of these Rat1a pools to grow on soft agarose is directly correlated to their relative sensitivity to apoptosis (Fig. 2). The deletion mutants dl106-143 and dl41-178 were unable to transform cells under any conditions, presumably because of an inability to initiate DNA synthesis and proliferate in soft agarose. The VP16-Myc chimera, although a very efficient transactivator of endogenous ODC (Fig. 3), was unable transform cells or initiate DNA synthesis (Fig. 8). Taken together, these data imply that cellular transformation by c-Myc depends on the balance between c-Myc-induced cell cycle progression and apoptosis and that these two functions of c-Myc can be independently executed. At low serum concentrations, T58A exhibited increased cellular transformation relative to wild-type c-Myc, a result consistent with this mutant's intact ability to accelerate DNA synthesis and its reduced ability to accelerate apoptosis in low serum concentrations.

How does c-Myc induce cell cycle progression?

Many studies have shown that ectopic expression of c-Myc can drive cells into all phases of the cell cycle in the absence of growth factors (20). However, the mechanism by which Myc drives cell cycle progression remains elusive. The emphasis in the past was to identify genes that can be transcriptionally activated by c-Myc. Indeed, there are many target genes that can be transcriptionally activated by c-Myc and are required for cell cycle progression (reviewed in reference 13). In agreement with previous reports (12, 57), the data presented here suggest that transcriptional repression by c-Myc is also important for c-Myc-mediated cell cycle progression. Several genes that are negative regulators of cell cycle progression have been identified as c-Myc targets, and these include the cyclin kinase inhibitors p27 (11) and p21 (37; A. Gartel and N. Hay, unpublished results), gadd45 (36), and gas1 (33). Identification of other genes that are transcriptionally repressed by c-Myc may uncover key downstream effectors of c-Myc-induced cell cycle progression and acceleration of apoptosis.

How does c-Myc accelerate apoptosis?

Our results suggest that the ability of c-Myc to activate the transcription of cellular genes is unlikely to be the main mechanism by which c-Myc sensitizes immortalized cells to apoptosis. The acceleration of apoptosis induced by c-Myc could be through repression of transcription of certain cellular genes and/or through interaction with cellular proteins that are associated with apoptosis. We have previously shown that the execution of c-Myc-accelerated apoptosis is dependent on wild-type p53 (54). More recently, it was shown that the acceleration of apoptosis by c-Myc is mediated, at least in part, through the stabilization of p53 as a consequence of elevating p19-ARF expression by c-Myc (59). While this mechanism clearly contributes to the acceleration of apoptosis by c-Myc in primary cells, it might not be the major mechanism by which c-Myc accelerates apoptosis in immortalized cells. This is mainly because c-Myc was shown to accelerate apoptosis in a variety of established cell lines that lack functional ARF. Furthermore, we have previously shown that overexpression of p53 in primary cells in which endogenous p53 was spontaneously deleted is not sufficient to elicit apoptosis. Only the overexpression of both p53 and c-Myc could elicit apoptosis in these cells (54). Another mechanism by which c-Myc accelerates apoptosis could be through the activation of FADD, the downstream effector of Fas-mediated apoptosis (24). Although the precise molecular link between c-Myc and FADD has not been established yet, it is possible that c-Myc can activate this pathway by repressing the expression of negative regulators of this pathway such as FLIP (26). Finally, it has been recently shown that the primary function of c-Myc in acceleration of apoptosis is the enhancement of the release of cytochrome c from mitochondria (27, 30). The induction of cytochrome c release is not sufficient by itself to accelerate apoptosis, and p53 and FADD are required after cytochrome c release (27). We showed here that the T58A mutant of c-Myc, which is defective in transrepression, has also significantly reduced ability to induce cytochrome c release, accounting for its diminished ability to accelerate apoptosis in Rat1a fibroblasts. Because the decreased ability of T58A mutant to accelerate apoptosis is more profound in immortalized cells, it is possible that in immortalized cells cytochrome c release is more critical for Myc-mediated apoptosis, whereas in primary MEF activation of p19-ARF/p53 pathway is more critical.

c-Myc may have an impact on mitochondrial integrity through both changes in cellular and mitochondrial metabolism. Indeed, it was shown that c-Myc increases metabolism and glycolysis, at least in part, through the direct activation of LDH-A expression (48). It remains to be determined whether other cellular genes that are repressed by c-Myc have an effect on mitochondrial metabolism and apoptosis.

The results presented here and in a recent report (39) suggest that phosphorylation of certain residues in the amino-terminus MB I of c-Myc could modulate its apoptotic function. Interestingly, Thr58 was reported to be phosphorylated by glycogen synthase kinase 3 (GSK3) (45), which is inactivated by Akt-PKB that has been shown to effectively protect from Myc-accelerated apoptosis (28, 29). Indeed, both GSK3 activity and the apoptotic function of c-Myc are augmented upon growth factor withdrawal. Further experiments are required to assess the possibility that GSK3 activity mediates Myc-accelerated apoptosis through phosphorylation of Thr58.

Induction of cell cycle progression and apoptosis are two separable functions of c-Myc.

Previous results showed that the ability of c-Myc to accelerate apoptosis is not dependent on the cell cycle phase of the cells (15, 19, 41, 54). However, the ability of c-Myc to induce cell cycle progression has always coincided with its ability to accelerate apoptosis. The analysis presented here allowed the separation of these two functions of c-Myc. The results obtained with the various deletion mutants in the amino-terminus region of c-Myc and with the T58A mutant clearly show that the ability of c-Myc to induce entry into cell cycle and to accelerate apoptosis are separable functions of c-Myc. These properties of c-Myc have a remarkable resemblance to the ability of E2F-1 to induce entry into cell cycle and apoptosis. Analysis of E2F-1 mutants showed that, although DNA binding is required, transcriptional activation is not necessary for the induction of apoptosis by E2F-1 (42). It was concluded that E2F-1 can show independent cell cycle progression and apoptotic functions. Interestingly, E2F-1 also activates the ARF-p53 pathway (25) and as with c-Myc this may account, at least in part, for the ability of E2F-1 to induce apoptosis in primary cells but not in established cell lines.

The analysis of E2F-1 mutants suggests that induction of apoptosis by E2F-1 may be mediated through alleviation of E2F-1-dependent transcriptional repression rather than activation of E2F-responsive genes (42). In the case of E2F-1 this transcriptional repression is mediated through the interaction with pRb. It remains to be determined whether a similar scenario is true for c-Myc. What are the cellular protein(s) that interact with c-Myc to mediate transrepression, and how might the disruption of these interactions affect apoptosis? Proteins that interact with c-Myc and may have these properties include p107, TRAP, and Bin1 (43). Further identification of targets of c-Myc-mediated transcriptional repression is likely to contribute to our understanding of c-Myc-mediated acceleration of apoptosis.