Oncogenes Induce and Activate Endogenous p73 Protein*

The identification of upstream pathways that signal to TP73 is crucial for understanding the biological role of this gene. Since some evidence suggests that TP73 might play a role in tumorigenesis, we asked whether oncogenes can induce and activate endogenous TP73 . Here, we show that endogenous p73 a and b proteins are up-regulated in p53-deficient tumor cells in response to overexpressed E2F1, c-Myc, and E1A. E2F1, c-Myc, and E1A-mediated p73 up-regulation leads to activation of the p73 transcription function, as shown by p73-respon-sive reporter activity and by induction of known endogenous p73 target gene products such as p21 and HDM2. Importantly, E2F1-, c-Myc-, and E1A-mediated activation of endogenous p73 induces apoptosis in SaOs-2 cells. Conversely, inactivation of p73 by a dominant negative p73 inhibitor (p73DD), but not by a mutant p73DD, inhibits oncogene-induced apoptosis. These data show that oncogenes can signal to TP73 in vivo . Moreover, in the absence of p53, oncogenes may enlist p73 to induce apoptosis in tumor cells.

The identification of upstream pathways that signal to TP73 is crucial for understanding the biological role of this gene. Since some evidence suggests that TP73 might play a role in tumorigenesis, we asked whether oncogenes can induce and activate endogenous TP73. Here, we show that endogenous p73 ␣ and ␤ proteins are upregulated in p53-deficient tumor cells in response to overexpressed E2F1, c-Myc, and E1A. E2F1, c-Myc, and E1A-mediated p73 up-regulation leads to activation of the p73 transcription function, as shown by p73-responsive reporter activity and by induction of known endogenous p73 target gene products such as p21 and HDM2. Importantly, E2F1-, c-Myc-, and E1A-mediated activation of endogenous p73 induces apoptosis in SaOs-2 cells. Conversely, inactivation of p73 by a dominant negative p73 inhibitor (p73DD), but not by a mutant p73DD, inhibits oncogene-induced apoptosis. These data show that oncogenes can signal to TP73 in vivo. Moreover, in the absence of p53, oncogenes may enlist p73 to induce apoptosis in tumor cells.
TP53 is a crucial tumor suppressor for preventing the malignant transformation of cells. Surprisingly, despite TP53's central role in carcinogenesis, no related genes were known for 20 years. In 1997, two novel family members were identified and termed TP73 and TP63 (1)(2)(3)(4). p73 shares 63% identity with the DNA-binding region of p53 including the conservation of all DNA contact residues, 38% identity with the tetramerization domain, and 29% identity with the transactivation domain. In contrast to TP53, human TP73 produces six C-terminal splice variants (p73 ␣-⌽) (1,5,6). For example, TP73 ␣ encodes all 14 exons, while TP73 ␤ lacks exon 13. In addition, mouse TP73 has an alternative promoter in intron 3, which encodes a p73 protein that lacks the transactivation domain (⌬N p73) and acts as a dominant negative suppressor of p73 ␣ (7). When ectopically overexpressed in cell culture, p73 ␣ and ␤ closely mimic p53 activities. Ectopic p73 ␤, and to a lesser extent p73 ␣, transactivate many p53-responsive promoters, although relative efficiency differences on a given promoter are observed (8 -10). Like p53, p73 forms a complex with p300/ CBP, which mediates transcription by p73 (11). Ectopic p73 also promotes apoptosis irrespective of the p53 status (1), and overexpression of p73 ␣, ␤, and ␦ suppresses focus formation, while p73 ␥ does not (5,8). The suppressor activities of isoforms ⑀ and ⌽ have not been determined.
Despite this experimental evidence, the role of TP73 in tumorigenesis is as yet unclear. Current genetic data have ruled out that TP73 is a Knudson-type tumor suppressor. Although TP73 maps to chromosome 1p36.3, which undergoes frequent loss of heterozygosity in breast cancer, neuroblastoma, and several other cancers (12), mutations in the TP73 gene are extremely rare in human tumors. Initially, imprinting of the TP73 locus was thought to be an explanation to satisfy the two-hit hypothesis in tumors with loss of heterozygosity but no mutations. However, imprinting is highly variable from patient to patient and tissue to tissue (6,(13)(14)(15). In fact, in lung, esophageal, gastric, and renal carcinoma, the second TP73 allele is specifically activated in the tumor compared with the normal tissue of origin (loss of imprinting) (16 -19). Furthermore, p73-deficient mice lack a spontaneous tumor phenotype but have neurological and immunological defects (7).
Both differences and similarities to p53 are found with respect to p73 inactivation by viral oncoproteins. SV40 T antigen, adenovirus E1B 55-kDa protein, and HPV E6, which all target and inactivate p53 during host cell transformation, do not target the p73 protein physically or functionally (20 -22). Indeed, ectopic p73, but not p53, induces apoptosis in E6-transformed cells, highlighting TP73's potential for gene therapy in HPV-mediated cancers (23). However, the adenovirus E4orf6 oncoprotein specifically represses p73 but not p53 transactivation in some experimental systems (22,24), indicating that adenoviral transformation targets both p53 and p73 via different viral products. Stable transfectants of a human rhabdoid tumor cell line expressing E1A or adenovirus 5 large E1B showed increased p73 levels, although functional activation of the protein was not tested (24). The Tax protein of the human T-cell leukemia virus type 1 represses the transactivation function of p73 ␣ and ␤ via a p300-dependent mechanism (25). Moreover, analogous to p53, Mdm2 suppresses p73 transactivation function via a negative feedback loop (26 -28). However, in contrast to p53, cellular Mdm2 and the HPV E6 protein do not mediate degradation of exogenous p73 (26,27,29), suggesting that the regulation of p73 degradation might be distinct from the one regulating p53.
Unlike p53 protein, which becomes stabilized and activated in response to a very broad spectrum of cellular stresses, little is known about the upstream signals that induce a p73 response. p73 is not activated by UV, actinomycin D, doxorubicin, and mitomycin C (1,37), all of which stabilize and activate p53. However, recently it has been shown that endogenous p73 is activated for apoptosis in response to cisplatin and ␥-ionizing irradiation in a pathway that depends on the nonreceptor tyrosine kinase c-Abl (38 -40). This DNA-damage-dependent upregulation of p73 by c-Abl may be partly responsible for p53independent apoptosis. What is already evident is that TP73, at least qualitatively, utilizes the same or very similar effector pathways as TP53. Complete identification of all upstream signals of TP73 that operate physiologically will be very important in elucidating its normal function and role, if any, in tumorigenesis. We therefore asked whether deregulated oncogenes, which are a preeminent signal for triggering p53-dependent transactivation and apoptosis, also induce and activate p73 function. We report that overexpression of cellular and viral oncogenes do up-regulate endogenous p73 proteins and activate their transactivation function. Moreover, E2F1-, c-Myc-, and E1A-mediated activation of endogenous p73 induces apoptosis in p53-deficient tumor cells. Disruption of p73 function by a dominant negative p73 inhibitor (p73DD), but not by a mutant version thereof, inhibited oncogene-induced apoptosis. These data show that oncogenes can signal to TP73 in vivo.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-The human lung carcinoma line H1299 and the human osteosarcoma line SaOs-2 each carry a homozygous deletion for TP53 and were used for transfection. Other cell lines were SK-N-AS (human neuroblastoma), MRC5 (human diploid fibroblasts), and COS (monkey kidney cells). All cells were cultured in Dulbecco's modified Eagle's medium plus 10% fetal calf serum. Cisplatin was purchased from Sigma.
pcDNA3-p73DD and pcDNA3-mtp73DD were gifts of Dr. William Kaelin and are described in detail in Ref. 43. Briefly, pcDNA3-p73DD expresses T7-tagged amino acids 327-636 of human p73 ␣ and acts as a dominant negative p73 in vivo. The corresponding loss-of-function mutant named mtp73DD, which contains a L371P point mutation, is inactive as inhibitor. Green fluorescent protein expression plasmid (CLONTECH) was cotransfected in transient transfections to verify relative transfection efficiency. The p53/p73-responsive reporter construct PG13-Luc and its mutant counterpart MG15-Luc (gift of B. Vogelstein) were used for luciferase assays. p73 ␤ exhibited 80% of the activity of p53 (on a molar basis) in transactivating the PG13-Luc reporter and no activity with the MG15-Luc reporter (data not shown).
Transfections-H1299 cells were plated in 60-mm dishes and grown overnight to 80% confluence. For transient transfections, 2 g of expression plasmid or empty vector was cotransfected with 200 ng of green fluorescent protein-encoding plasmid using the LipofectAMINE Plus reagent (Life Technologies, Inc.). Cells were collected after 24 h. Stable transformants were seeded into P100 plates (1 ϫ 10 7 cells) and selected for 21 days in medium containing 0.5 mg/ml G418 (Life Technologies), ring-cloned, and expanded into single cell clones. For the SaOs-2 experiments (see Fig. 5D), transformants were pooled after 3 weeks of G418 selection and transiently transfected with c-Myc prior to parallel TUNEL assays and immunofluorescence staining 24 h later. For luciferase assays, cells were seeded into 24-well plates and transfected with an expression vector or empty vector (400 ng) together with the p53-responsive PG13-Luc from firefly (80 ng) and pRL-TK Renilla luciferase cDNA (8 ng). To test the effect of the dominant negative inhibitor and its mutant, each well was transfected with 100 ng of expression vector plus 300 ng of empty vector, with 100 ng of expression vector plus 300 ng of p73DD, or with 100 ng of expression vector plus 300 ng of mtp73DD, together with the reporter constructs as above. Luciferase activity was measured after 24 h by the dual luciferase reporter assay (Promega), and transfection efficiency was standardized against Renilla luciferase.
For apoptosis, SaOs-2 cells were seeded in duplicates into eight-well chambers 48 h prior to transfection. At about 70% confluence, cells in duplicate wells were transfected with expression plasmid (150 ng) plus empty vector (350 ng) or expression plasmid (150 ng) plus p73DD (350 ng) or expression plasmid (150 ng) plus mtp73DD (350 ng). After 24 h, cells were fixed and processed in parallel for the TUNEL assay (Roche Molecular Biochemicals) and for immunofluorescence using the appropriate antibodies against p73, E2F1, c-Myc, or E1A. Expression was reproducibly about 30%, similar among all constructs and evenly distributed throughout the wells. For each construct, TUNEL-positive cells (494 fields at 40ϫ) and transfected cells (30 fields at 40ϫ; around 200 cells) from duplicate chambers were counted, and the percentage of apoptosis of transfected cells was determined after correction for background with vector alone (500 ng per well). Experiments were performed 3-6 times, depending on the construct.

RESULTS
Endogenous p73 ␣ and ␤ Proteins Are Induced in Response to E2F1, c-Myc, and E1A-The great majority of functional and regulatory p73 studies to date have used ectopically expressed p73 proteins. To reliably detect endogenous p73 proteins, we used three different p73-specific antibodies. They comprised a p73 ␤-specific monoclonal (GC15), a p73 ␣-specific polyclonal raised against a C-terminal peptide (poly-p73␣) and a pan-p73 polyclonal raised against an N-terminal peptide (poly-p73N). The antibodies detected endogenous full-length p73 ␣ and ␤ in several tumor cell lines including the p53-deficient human H1299 line (Fig. 1) Next we tested whether viral and cellular oncogenes, which are major upstream signals for TP53 activation, are also physiologically relevant for triggering the induction of endogenous TP73. To this end, H1299 cells were transiently transfected with various oncogene-encoding plasmids. Their expression was verified by immunoblotting with the respective antibodies ( Fig. 2A). Both p73 ␣ and ␤ proteins were markedly induced after expression of E2F1, c-Myc, and adenoviral E1A when compared with empty vector. A representative experiment is shown in Fig. 2, B and C. By molecular weight standards and reactivity with the polyclonal p73N antibody, only full-length proteins were observed, with p73 ␣ migrating at about 83 kDa and p73 ␤ at about 75 kDa. The equal loading of immunoblots was confirmed by reblotting the membranes with an antibody specific for vimentin. Since transfection efficiency in these transient assays ranged between 30 and 40% (judged by coexpressed green fluorescent protein), the actual degree of induction is higher than the one detected here by immunoblots, since this assay is subject to dilution by untransfected cells. Significant induction of endogenous TP73 was also seen after oncogene transfection in SaOs-2 cells (see Fig. 5B).
Stable H1299 Clones Overexpressing c-Myc Recapitulate the Up-regulation of p73 Proteins-Since transient overexpression of oncogenes induces the accumulation of p73 ␣ and ␤ (Fig. 2), we next asked if stable oncogene overexpression similarly would lead to long term up-regulation of p73. To this end, vector control and c-Myc-transfected H1299 cells were selected in G418 for 3 weeks. Of the surviving c-Myc foci, seven were randomly picked, ring-cloned, and successfully expanded into stable sublines. As shown in Fig. 3, all seven clones overexpressed c-Myc, albeit to various degrees compared with vector control. Clones 1, 2, and 4 showed the highest c-Myc expression. Cell extracts were then probed for p73 protein levels. As already seen with transient c-Myc transfections, p73 ␣ and ␤ were found to be induced above base line in all seven subclones (Fig. 3). p73 ␤ induction appeared proportional to the level of c-Myc expression in the individual clones, with clones 1, 2, and 4 showing the highest p73 ␤ accumulation. Interestingly, for reasons that remain to be elucidated, p73 ␣ reproducibly behaved inversely to ␤ (i.e. whenever ␤ was high, ␣ was low, and vice versa). Taken together, this result indicates that stable deregulation of the c-Myc oncogene recapitulates the p73 overexpression already seen with transient deregulation of c-Myc.
Oncogene-mediated Up-regulation of Endogenous p73 Protein Leads to p73 Transcriptional Activation-We then tested whether the oncogene-mediated up-regulation of endogenous p73 translates into activation of p73 transcriptional function. To this end, we carried out luciferase reporter assays in transiently transfected H1299 cells using the p53/p73-responsive PG13-Luc reporter. As shown in Fig. 4A, all three oncogenes were able to activate p73 reporter activity. E2F1 exhibited a 16.5-fold, c-Myc a 10.3-fold, and E1A a 13.9-fold induction of the p73-responsive reporter compared with vector controls. These data indicate that oncogenes induce the transcriptional activation of endogenous p73.
Oncogene-mediated Up-regulation of Endogenous p73 Leads to Activation of TP73 Response Genes-TP73 shares many response genes with TP53 in vivo. This has been shown in several cell systems using transient or inducible expression of ectopic p73 (1,8,9,38). To further support our previous results, we tested whether oncogene-mediated accumulation of endogenous p73 leads to the induction of TP73 target gene products. When p53-deficient H1299 cells were transiently transfected with expression plasmids for E2F1, c-Myc, and E1A, endogenous p73 ␤ protein was again up-regulated (Fig. 4B, top panel,   lanes 2-4, compare with empty vector in lane 1). Oncogene expression was confirmed by immunoblots (data not shown). The p73 up-regulation was accompanied by the induction of the TP73 response gene products Waf1p21 and HDM2 (Fig. 4B, middle panels, lanes 2-4; compare with empty vector in lane 1). Lanes 5 and 6 are positive controls after direct transfection of p73 ␣ and ␤ expression plasmids. Although the induction is modest, it could be due to the fact that we are relying on endogenous rather than ectopic p73. Furthermore, since transfection efficiency was only between 30 and 40%, the actual induction is probably higher than the one detected here by immunoblots, due to the dilutional effect by the untransfected majority of cells. E2F1 reproducibly caused a stronger transactivation of the p21 and HDM2 genes than c-Myc and E1A. Previous studies have shown that Waf1p21 and HDM2 are direct in vivo targets of ectopic p73, as demonstrated by detecting their products in response to inducibly expressed p73 ␣ and ␤ in EJ (mutant p53) and p53-deficient H1299 cells (9,38). In p53-expressing cells, transactivation of HDM2 in response to a broad spectrum of overexpressed oncogenes including the panel used here has been shown to be indirect and strictly dependent on p53 (for a review, see Ref. 47). Transactivation of p21 by c-Myc and E1A is also p53-dependent, although E2F1, in addition to activating the p21 promoter through p53, can transactivate p21 directly (47). Overall, these data strongly suggest that with the partial exception of E2F1, the induction of p21 and HDM2 in response to oncogenes is mediated through p73 in H1299 cells. Together with the reporter assays, these results are consistent with the idea that in the absence of p53, oncogene-induced endogenous p73 is capable of activating its target genes.
Oncogene-mediated Activation of Endogenous p73 Induces Apoptosis in p53-deficient Tumor Cells; Conversely, Inactivation of p73 Inhibits Oncogene-induced Apoptosis-The activation of the p73 transcription function by oncogenes suggested that these upstream signals might also induce the activation of the apoptotic function of p73. To test this prediction, we performed apoptosis assays on transiently transfected SaOs-2 cells using the in situ TUNEL assay.
To confirm that the oncogene-induced apoptotic activity is mediated through TP73, we tested the effect of a coexpressed dominant negative inhibitor of p73 (p73DD) (43). p73DD is modeled after the dominant negative p53 inhibitor (p53DD) (48) and encodes amino acids 327-636 of human p73 ␣. p73 DD acts as a specific inhibitor of p73 ␣ and ␤-dependent transactivation (Fig. 5A) but not of p53-dependent transactivation (data not shown). Moreover, p73DD binds to p73 ␣ and ␤ proteins in vitro and in vivo but not to p53 protein (43). When coexpressed with p73 ␤, p73DD suppressed PG13-Luc reporter activity by 98%, making it an efficient specific inhibitor (Fig.  5A). In contrast, an inactive point mutant called mtp73DD, carrying a L371P exchange, does not bind p73 ␣ and ␤ and does not block p73 ␣and ␤-dependent transactivation (43). mtp73DD was completely incapable of suppressing the reporter activity of p73 ␤ (only 4% inhibition compared with p73 ␤ alone) (Fig. 5A). Together, these results indicate the specificity of these two reagents.
Prior to doing apoptosis assays, we needed to demonstrate that oncogenes induce endogenous p73 in p53-deficient SaOs-2. As seen in Fig. 5B, expression of E2F1, c-Myc, and E1A in these cells markedly induced p73 levels, analogous to what we already saw in H1299 cells (see Fig. 2). Importantly, the induced p73 protein was functionally active (Fig. 5C). After transient transfection, all three oncogenes induced apoptosis in SaOs-2 cells, which resembled the one seen after transfecting p73 ␤ directly (light grey columns). Moreover, the apoptotic activity of  5 and 6). Four independent experiments gave similar results. each oncogene was greatly suppressed or abrogated when oncogenes were coexpressed with the dominant negative inhibitor p73DD, with 84% suppression for E2F1, 96% for c-Myc, and 72% for E1A (black columns) (Fig. 5C). The suppression of oncogene-mediated apoptosis by a p73-specific inhibitor strongly suggests that E2F1, c-Myc, and E1A mediate their apoptotic effects through p73. This conclusion is further confirmed by the lack of significant suppression when p73DD is exchanged for the functionally inactive inhibitor mtp73DD (dark grey columns) (Fig. 5C). Furthermore, the dominant negative p73DD inhibitor also prevented apoptosis when it was stably expressed in SaOs-2 cells, while its mutant counterpart failed to prevent apoptosis (Fig. 5D). Also, E2F1-and c-Mycinduced apoptosis was found to be partially mediated through p73 in transiently transfected HeLa cells (data not shown; E1A not tested). Taken together, these data show that E2F1-, c-Myc-, and E1A-mediated apoptosis in p53-deficient tumor cells largely depends on p73. DISCUSSION Our results show that during short term exposure, overexpression of three different oncogenes induce the endogenous TP73 gene and activate it for transcriptional and apoptotic function in p53-deficient cells. We demonstrate that disruption of p73 function inhibits oncogene-induced apoptosis in p53deficient human tumor cells. Collectively, our findings identify an important novel afferent signaling pathway to p73 that appears to operate in vivo. This conclusion is in agreement with evidence from several other experimental systems. For example, Myc-induced apoptosis in p53 Ϫ/Ϫ mouse embryo fibroblasts is only attenuated but not abrogated compared with wild type p53 cells, supporting the idea that part of the apoptotic response is p53-independent (49). Based on our results, this p53-independent apoptosis may be due, at least in part, to p73. The demonstration that endogenous p73 is induced and activated by oncogenes is in line with the known p73 activation by cisplatin and ␥-ionizing irradiation (38 -40). Together, these data suggest that TP73 can be a component of a tumor surveillance pathway, which in the absence of p53 might respond to different types of incoming signals in a p53-compensatory fashion.
While this work was ongoing, W. G. Kaelin's group obtained similar data on the relationship of E2F1 and p73 (43). They also found that p73 mediates E2F1-induced apoptosis in SaOs-2 cells and p53 Ϫ/Ϫ mouse embryo fibroblasts. Moreover, pendent transactivation by the dominant negative inhibitor p73DD but not by the inactive mtp73DD (L371P) mutant. H1299 cells were transiently transfected with either p73 ␤ (100 ng) plus empty vector (300 ng), p73 ␤ (100 ng) plus p73DD inhibitor (300 ng), or p73 ␤ (100 ng) plus mtp73DD mutant (300 ng), together with 80 ng of PG13-Luc and 8 ng of Renilla luciferase. Luciferase activity was measured 24 h later and standardized for Renilla activity. Results are the average Ϯ S.D. of three independent experiments. B, immunoblot of SaOs-2 cells after transient transfection with empty vector or expression vectors for E2F1, c-Myc, and E1A. The blot was developed with a mixture of GC15 and vimentin antibodies. C, SaOs-2 cells in duplicate wells were transfected with either expression plasmid (150 ng) plus empty vector (350 ng), expression plasmid (150 ng) plus p73DD inhibitor (350 ng), or expression plasmid (150 ng) plus mtp73DD mutant (350 ng). After 24 h, cells were processed in parallel for TUNEL and for immunofluorescence to determine expression. The percentage of apoptosis of transfected cells is shown after correction for background with vector alone (500 ng/well). The results represent the average Ϯ S.D. of three independent experiments. D, SaOs-2 cells were transfected with empty vector or expression vectors for p73DD or mtp73DD. Cells were selected for 3 weeks in G418 (550 g/ml) prior to transfection with c-Myc or empty vector. Cells were seeded into duplicate eight-well chamber slides. After 24 h, cells were processed in parallel for TUNEL and for immunofluorescence to determine expression. The percentage of apoptosis of transfected cells is shown after correction for background with vector alone (500 ng/well).
FIG. 5. Oncogene-mediated activation of endogenous p73 induces apoptosis in p53-deficient tumor cells. Inactivation of p73 inhibits oncogene-induced apoptosis. A, inhibition of p73 ␤-de-they identified the mechanism of this interaction to be directly transcriptional; i.e. E2F1 induces transcription of full-length p73 ␣ and ␤ via E2F1 binding sites in the P1 promoter of TP73 (43). A functional interaction between E2F1 and p73 was also recently shown in T-cell receptor activation-induced cell death (44). A direct transcriptional mechanism of TP73 induction by E2F1 is in complete agreement with two predicted E2F1 binding sites at positions Ϫ284 and Ϫ1862 of the TP73 promoter (50). Moreover, the c-Myc-mediated up-regulation of TP73 that we observed most likely also operates via E2F (51)(52)(53). E2F1 (as well as E2F2 and E2F3a) is induced by c-Myc via a group of E-box elements in its promoter conferring positive Myc responsiveness. The TP73 promoter itself has not been reported to contain Myc binding sites. In cells with functional Rb such as in H1299 cells, E1A-mediated up-regulation of TP73 might again be mediated through E2F1, albeit indirectly via inactivation of Rb by E1A (54). SaOs-2 cells, which were also used in our study, are deficient for Rb, indicating an additional Rb-independent mechanism of TP73 induction by E1A. Taken together, our paper independently confirms the E2F1-TP73 signaling pathway and extends it to additional oncogenes, thus broadening the concept of TP73 activation by oncogenes.
In contrast to transcriptional activation of TP73 by oncogenes, cisplatin activates endogenous p73 by a posttranscriptional mechanism (38). Cisplatin increases protein stability from 45 min to 2 h without altering TP73 transcript levels. Protein stability of endogenous and exogenous p73 ␣ has been shown to be regulated in part through proteasome-dependent degradation, since cells accumulate p73 ␣ after treatment with proteasome inhibitors (27,29,55). Furthermore, ␥-ionizing irradiation activates p73 through c-Abl-mediated tyrosine phosphorylation without protein stabilization (39,40). Taken together, it appears that the regulation of p73 activity occurs both on a transcriptional and posttranscriptional level and might depend on the specific activating stimulus.
We also find that stable deregulation of c-Myc in H1299 subclones recapitulates the p73 overexpression seen with transient deregulation of c-Myc. One might ask how stable clones can be generated, given that in transiently transfected tumor cells deregulated c-Myc was able to activate the transcriptional and apoptotic activity of p73. One possibility is that clonal outgrowth of cells with stable overexpression of oncogenes selects for loss of p73 transactivation function. In keeping with this idea, a preliminary analysis on three clones with the highest p73 ␤ overexpression suggested that this might be the case. In contrast to transient Myc transfection, Myc clones 1, 2, and 4 ( Fig. 3) failed to show evidence of p21 and HDM2 induction compared with vector. The same clones also failed to show transactivation activity in a p73-responsive reporter assay. 2 Moreover, in a previous study on five human breast cancer cell lines with p73 overexpression (four of the five lines were mutant for p53), we found no correlation with p21 mRNA and protein levels (6). While it is tempting to speculate, it is important to point out that more extensive analyses need to be done before definitive statements about the functional consequence of constitutive p73 overexpression in tumor cells can be made. Nevertheless, the sustained p73 overexpression in stable Myc clones is highly reminiscent of the fact that multiple primary human tumor types and tumor cell lines overexpress p73. This includes tumors of breast, neuroblastoma, esophagus, stomach, lung, colon, bladder, ovary, ependymoma, hepatocellular carcinoma, and myeloid leukemia (6,13,14,(17)(18)(19)(31)(32)(33)(34)(35)(36).
Most human tumors harbor deregulated oncogenes, includ-ing the Myc gene (56). In particular, many human tumors have suffered a deregulation of the E2F1 activity through mutations that inactivate the Rb pathway (57)(58)(59), thus derepressing E2F1-responsive genes. Our finding could provide a framework for the fact that p73 is frequently overexpressed in human tumors. Our stable clones will be a helpful tool in future studies aimed at determining the functional relevance of constitutive p73 overexpression in tumors.