Specific isoforms of p73 control the induction of cell death induced by the viral proteins, E1A or apoptin

p73 isoforms have differential effects on the induction of cell death by E1A or Apoptin

To determine whether p73 isoforms differ in their effects on the induction of cell death by E1A or Apoptin, we used the doxycycline-inducible Saos-2 cells expressing either the dominant negative ΔNp73α or the TAp73 isoforms. The inducible Saos-2 cells expressing TAp73α, β, γ or ΔNp73α were infected with either a recombinant adenovirus vector expressing wild-type E1A (Ad-E1A) or the control virus Ad-Del³⁷ in the presence or absence of p73 inducer, doxycycline. Cell survival was measured by the MTT assay at 24 and 48 h post-infection. 24 h infection with Ad-E1A resulted in a significant reduction of cell survival in Saos-2 cells overexpressing TAp73β (50.4% cell survival) and TAp73γ (59% cell survival), but not in those overexpressing TAp73α. ΔNp73α-expressing cells infected with Ad-E1A had a slightly increased growth rate compared to the control (Fig. 1A, top). 48 h after expression of E1A in Saos-2 cells overexpressing TAp73α, TAp73β and TAp73γ resulted in dramatic cell death, particularly in Saos-2 cells expressing TAp73β and TAp73γ (Fig. 1A, bottom). Survival rates of Saos-2 cells overexpressing TAp73α, TAp73β and TAp73γ infected with Ad-E1A were reduced to 34.1%, 12.2% and 19.3%, respectively. In contrast, almost 60% of Saos-2 cells overexpressing ΔNp73α and infected with Ad-E1A remained viable, indicating that the cells overexpressing TAp73 isoforms, but not ΔNp73α, significantly enhanced E1A-induced cell death (p < 0.001). The control virus Ad-Del had no significant effect on the survival of cells with induced expression of various p73 isoforms.

We then investigated the effect of different p73 isoforms on the sensitivity of inducible Saos-2 cells to Apoptin. Cells were infected with either adenovirus expressing Apoptin fused to GFP (Ad-GFP-Apoptin) or the control adenovirus expressing GFP (Ad-GFP) (Fig. 1B). Similar to the effect observed with E1A, after 24 h, expression of Apoptin in cells induced for TAp73β and TAp73γ showed a significant reduction of cell survival as compared with Ad-GFP-Apoptin-infected cells induced for ΔNp73α (Fig. 1B, top; p < 0.001). At 48 h post-infection, a significant increase in cell sensitivity to Apoptin-induced cell death was observed in cells expressing TAp73 isoforms, in particular TAp73β (46.3%, 19% and 31.9% survival with the expression of TAp73α, TAp73β and TAp73γ, respectively). Furthermore, cell survival in those cells was significantly less than Ad-GFP-Apoptin-infected cells expressing ΔNp73α (62.8% cell survival, p < 0.001) (Fig. 1B, bottom). The control virus Ad-GFP had no significant effect on the survival of cells expressing various p73 isoforms. Western blot analysis at 24 h post-infection with Ad-GFP-Apoptin in the doxycycline-induced Saos-2 cells expressing TAp73α, TAp73β and TAp73γ, but not ΔNp73α, showed increased levels of the apoptotic marker PARP p85, as compared with non-induced, Ad-GFP-Apoptin-infected cells (Fig. 1C, Lanes 3 and 6). The level of cleaved PARP fragment increased with combined expression of Apoptin with TAp73α (Fig. 1C, upper left, Lanes 4 and 6) but not with TAp73β or TAp73γ (Fig. 1C, upper right and lower left, Lanes 4 and 6). This may be due to the TAp73β or TAp73γ isoforms being more potent inducers of apoptosis than TAp73α, therefore, the fragmentation of PARP reaching the maximum level at this time point. Indirect immunofluorescent study also showed increased nuclear fragmentation by DAPI staining in Saos-2 cells expressing the TAp73 isoforms, but not ΔNp73α (Fig. 1D).

Anti-tumor effects of Ad-TAp73β in head and neck cancer xenografts

The above experiments suggested the TAp73β to be the most potent TAp73 isoform in enhancing sensitivity to E1A or Apoptin-induced cell death; we therefore evaluated the potential therapeutic benefit of TAp73β in p53-mutated HSC-3 head and neck cancer xenografts. We injected HSC-3 cells (5 × 10⁶) into the right flank of nude mice as described in Materials and Methods. Two weeks later, the mice were grouped (n = 7/group) and injected into tumors of 5–6 mm in diameter with either control (PBS), Ad-GFP or adenovirus carrying GFP tagged TAp73β (Ad-GFP-TAp73β) as described in Materials and Methods. As depicted in Figure 1E, tumor growth was significantly reduced in mice treated with Ad-GFP-TAp73β compared to PBS or the control Ad-GFP (p < 0.001).

E1A transactivates p73 and PUMA independently of the p53-binding sites in p53-deficient cancer cells

We have previously shown that E1A highly transactivates TAp73 and pro-apoptotic targets such as Noxa, a ‘BH3-only’ member of the Bcl-2 family.²⁴ Here, we have investigated the effect of E1A on the transactivation of another ‘BH3-only’ family member, PUMA, an important p53 target activated by DNA damage.³⁸ Two PUMA luciferase reporter constructs were used; one in which both putative p53-binding sites are intact, and one in which the p53-binding sites were deleted (Fig. 2A). SH-SY5Y neuroblastoma cells with aberrant p53 were co-transfected with PUMA reporters together with E1A13S, E1A12S or empty vector DNA and the luciferase activity was measured 24 h post-transfection. Both E1A13S and E1A12S expression were shown to activate luciferase expression from the PUMA promoter with deleted p53-binding sites by 21 and 13.5 folds, respectively. Unexpectedly, there was only a small increase in the luciferase activity from the PUMA promoter with intact p53-binding sites (2.7 and 1.8 folds by E1A13S and E1A12S, respectively) (Fig. 2B). To verify that the p53-binding independent induction of the PUMA promoter by E1A was not cell-type specific, the two PUMA reporters were expressed in the p53-null lung carcinoma H1299 cells together with E1A13S or empty vector DNA. At 24 h post-transfection, luciferase reporter assay showed a 3.3- to 3.7-fold activation of the PUMA promoter by E1A13S, irrespectively of the presence of the p53-binding sites (Fig. 2C). Taken together, these data indicate that the transactivation of the PUMA promoter by E1A is independent of the p53-binding sites.

E1A induces the endogenous expression of PUMA

The transcriptional effect of E1A on the endogenous PUMA gene was further investigated in SH-SY5Y and H1299 cells by Western blotting. Cells were infected with Ad-E1A or control Ad-GFP at a MOI of 10 and were lysed at 0, 6, 12, 24, 36 and 48 h post-infection. PUMA protein expression was examined using a rabbit anti-PUMA antibody. In Ad-E1A-infected SH-SY5Y cells, a 4.7-fold increase in levels of PUMA protein was observed as early as 6 h post-infection compared with the baseline level at 0 h. The highest level of PUMA expression was observed at 12 h post-infection (8.3-fold), after which PUMA protein levels were reduced to 7.7, 5.1 and 4.8 folds at 24, 36 and 48 h, respectively (Fig. 2D, left). In agreement with the reporter gene assay, which showed an approximately three-fold transactivation of the PUMA promoter in H1299 cells (Fig. 2C), Western blot analysis showed a small increase in endogenous PUMA protein levels with the highest level of 1.4 fold at 12 h post-infection. At 36 h post-infection, when extensive apoptosis was clearly detectable, a downregulation of endogenous PUMA protein was detected (Fig. 2E, left). Co-transfection with Ad-GFP control did not result in increased levels of PUMA protein (Fig. 2D and E, right).

Apoptin induces expression of p73 and its downstream target PUMA

We next investigated whether, similar to E1A, Apoptin was able to induce p73 and its downstream pro-apoptotic targets including PUMA at the transcriptional and protein levels. SH-SY5Y cells were transfected with the luciferase reporter plasmids under the control of the TAp73, ΔNp73 or PUMA promoters in combination with pCMV-Apoptin or the control empty vector plasmid. Apoptin was found to have no effect on the transactivation of all the promoters tested (Fig. 3A). We then examined the effect of Apoptin on the induction of protein levels of p73 and PUMA in p53-mutated HNSCC cells, H357. This cell line was chosen because of its high sensitivity to Apoptin cytotoxicity. H357 cells were left either untreated or infected with Ad-GFP-Apoptin or a control Ad-GFP and total cell lysates were collected at 6, 12 and 48 h post-infection. Western blot analysis using anti-p73α antibody (p73SAM), which has been previously shown to react with only TAp73α and ΔNp73α but not with other p73 isoforms or with TAp63α,³⁹ demonstrated increased levels of endogenous TAp73α as early as 6 h post-infection. The increase in the level of TAp73α was particularly high at 48 h post-infection. Interestingly, the level of anti-apoptotic protein ΔNp73α isoform was significantly reduced at 48 h post-infection (Fig. 3B). The sizes of TAp73α isoform and ΔNp73α isoform detected in H357 cells by anti-p73α antibody were comparable to those observed in Saos-2 cells inducible for TAp73α and ΔNp73α (Fig. 5C).

The effect of Apoptin on p73 pro-apoptotic target, PUMA, was also examined. An almost 2-fold induction of endogenous PUMA was observed at 6 h after the expression of Apoptin and the level of PUMA was then reduced; however, remained higher than those in the control non-infected cells or Ad-GFP-infected cells, to 1.4 fold at 12 h post-infection. At 48 h after infection, the level of PUMA was reduced approximately 40% as compared with the control non-infected cells, this coincided with the clear presence of apoptosis at 48 h, detected by cleaved PARP p85 antibody, in cells expressing Ad-GFP-Apoptin (Fig. 3B). Expression of GFP in these cells did not have any significant effect on the protein level of both p73 and PUMA (Fig. 3B).

TAp73-compromised c-Abl knockout cells are resistant to E1A-induced transactivation of TAp73 and p53/p73 apoptotic targets

The direct involvement of p73 in E1A-induced apoptosis was investigated using c-Abl knockout MEFs. p73 protein has been shown to be stabilized and activated after DNA damage through a pathway involving the c-Abl tyrosine kinase.^36,40,41 c-Abl knockout MEFs are therefore p73 function deficient and do not respond to treatment by chemotherapeutic drugs such as cisplatin.³⁶ c-Abl wild-type and c-Abl knockout MEFs were transfected with the luciferase reporter plasmids for TAp73, Bax or Noxa in combination with a plasmid expressing E1A13S or empty vector DNA. In c-Abl wild-type MEFs, expression of E1A13S resulted in an 11-fold induction of the TAp73 promoter, while in c-Abl knockout cells, expression of E1A13S resulted in only a four-fold induction (a ~60% reduction compared to wild-type cells). Similar results were obtained with the Bax and Noxa reporters, with the ability of E1A to transactivate Bax and Noxa in c-Abl knockout MEFs being reduced by 33% and 49.9%, respectively, as compared with those in c-Abl wild-type cells (Fig. 4A).

TAp73 function is required for E1A-induced cell death

We further examined the induction of cell death by E1A in c-Abl wild-type and c-Abl knockout MEFs. Cells were infected with either Ad-E1A or control Ad-Del and survival was measured by the MTT assay at 24, 48 and 72 h post-infection. Substantial reduction in cell survival was observed in c-Abl wild-type MEFs infected with Ad-E1A, with 60% and 90% reduction observed at 48 and 72 h post-infection, respectively. In contrast, c-Abl knockout MEFs appeared to be resistant to E1A-induced apoptosis as almost 100% of cells infected with Ad-E1A survived suggesting that the c-Abl/p73 pathway is crucial for E1A-induced apoptosis (Fig. 4B and Suppl. Fig. 1). To confirm that this was not due to cell-type differences in the efficiency of transfection and expression of the transgene, the expression of E1A in c-Abl wild-type and c-Abl knockout MEFs were detected by immunofluorescent staining and Western blotting (Fig. 4C and D). Both cell lines appeared to have similar levels of E1A expression (detected in green in Fig. 4C) and E1A was expressed mainly in the nucleus in both cell lines. Figure 4C demonstrated apoptotic nuclei detected by DAPI staining in E1A-expressing c-Abl wild-type MEFs whereas normal nuclear morphology was observed in E1A-expressing c-Abl knockout MEFs.

Specific p73 isoforms show important differences in efficiency to induce apoptosis, transactivation of target genes and protein stability

The above results demonstrated that p73 isoforms were different in response to cell death induced by both E1A and Apoptin. We therefore examined the different characteristics of specific p73 isoforms in relation to the ability to induce apoptosis, transactivation of target genes and protein stability among p73 isoforms.

The ability of TAp73 expression in inducing efficient apoptosis in p53-deficient cells was investigated using doxycycline-inducible Saos-2 cells for specific p73 isoforms. The MTT assay (Fig. 5A) showed that 72 h after doxycycline treatment cell viability in the TAp73-positive cells was significantly reduced compared with non-induced or cells expressing ΔNp73α (p < 0.001). TAp73β showed the highest induction of apoptosis (41% cell survival) followed by TAp73γ (56.4% cell survival) and TAp73α (63.5% cell survival). In contrast, ΔNp73α had an insignificant effect on the induction of cell death. Furthermore, we performed the transient transfection study to further confirm the efficient ability of TAp73 in inducing apoptosis in a p53-independent manner. Saos-2 cells were transiently transfected with plasmids encoding TAp73α, β, γ, δ or ΔNp73α and cells were analyzed at 24 and 48 h post-transfection. Indirect immunofluorescence studies showed that each p73 isoform was expressed mainly in the nucleus although cells expressing TAp73α showed cytoplasmic expression at 24 h (Suppl. Fig. 2A and B). All TAp73 isoforms, but not the dominant negative ΔNp73α, significantly induced cell death in Saos-2 cells (p < 0.05). The percentages of apoptotic cells induced by TAp73 isoforms varied from 10.9 to 26.6 (Suppl. Fig. 2A–C).

Specific p53/p73 targets have been shown to be differentially activated by various p73 isoforms.⁸ In this study we used the inducible Saos-2 cells to investigate the effect of each p73 isoform on the activation of several p53/p73 targets. Saos-2 cells expressing the doxycycline-inducible p73 isoforms were transfected with the luciferase reporter plasmids for Bax, PIG3, Pidd, MDM2, p21^WAF1/CIP1 and PUMA and the luciferase activity was measured after 24 h. Expression of each p73 isoform was confirmed by Western blot analysis after 24 h-doxycycline treatment (Fig. 5C).

As shown in Figure 5B, induction of TAp73γ isoform substantially increased Bax and PIG3 promoter activity by 82 and 30 folds, respectively, and resulted in an almost eight-fold increase in the activity of the p21^WAF1/CIP1 promoter. TAp73β expression resulted in about 20-fold increase in the MDM2 promoter activity, while the Pidd promoter was highly induced by both TAp73α and TAp73γ by 85 and 60 folds, respectively. Expression of either TAp73α or TAp73γ resulted in a 2.5- and a 2.3-fold induction of the PUMA promoter. In contrast, overexpression of ΔNp73α had no effect on transcription from any of the promoters tested. In agreement with the luciferase reporter assays, Western blot analysis demonstrated the induction of p53 downstream targets including p21^WAF1/CIP1, MDM2 and PUMA by all three TAp73 isoforms tested, whereas ΔNp73α expression failed to induce those targets. However, there were some differences in the levels of induction of various targets between the luciferase and Western blotting methods. This could be due to the effect of post-translational modification and/or protein stability of tested targets. All TAp73 isoforms-expressing Saos-2 cells underwent apoptosis but not in cells expressing ΔNp73α (Fig. 5C). In agreement with the MTT assay, the TAp73β and TAp73γ expression were more apoptotic than TAp73α as evident from PARP p85 levels.

Several studies have investigated the mechanisms involved in the regulation of p73 activities and protein stabilities. The stability of p73 proteins, which in most cases have been studied under conditions of transient overexpression, varies among different studies.^22,42 We chose to investigate the half-life of TAp73α, TAp73β, TAp73γ and ΔNp73α when expressed from stably transfected inducible constructs in Saos-2 cells.

As shown in Figure 5D, TAp73α had the longest half-life, followed by TAp73β, TAp73γ and ΔNp73α. Over 50% of TAp73α protein remained undegraded at approximately 3.5 h after cycloheximide treatment, while TAp73β had a half-life of almost 1.5 h and TAp73γ and ΔNp73α had turnovers of less than 50 min. At 4 h after cycloheximide treatment, 40% of TAp73α protein remained intact, whereas the other three isoforms were barely detectable.

Cell lines and plasmids

Osteosarcoma cell line Saos-2, wild-type or knockout c-Abl MEFs,⁵⁷ neuroblastoma SH-SY5Y, non-small cell lung cancer cell line H1299, head and neck cancer cell line HSC-3 and human primary embryonal kidney 293A adenoviral packaging cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS, 50 μg/ml streptomycin, 100 μg/ml penicillin and 1 mM sodium pyruvate, all purchased from Sigma (Gillingham, UK). HNSCC cell line H357 was cultured in Nut Mix (Invitrogen, Paisley, UK) supplemented with 5% FCS, 4 mM L-glutamine, 25 μg/l hydrocortisone, 50 μg/ml streptomycin, 100 μg/ml penicillin and 1 mM sodium pyruvate. TAp73α, β, γ and ΔNp73α-inducible Saos-2 cells⁵⁸ were grown in RPMI medium with 10% tetracycline free FCS, G418 0.5 mg/ml, hygromycin 0.25 mg/ ml and induced with 2.5 μg/ml doxycycline for 24 h. All of the p73 isoforms were haemagglutinin (HA) tagged. Cultures were incubated at 37°C and 5% CO₂.

The pcDNA3-TAp73α, β, γ, δ and ΔNp73α, all are HA-tagged, were obtained from Prof. Gerry Melino. pcDNA4-HA-E1A13S and pcDNA4-HA-E1A12S have been described elsewhere.²⁴ The TAp73 and ΔNp73 luciferase reporter plasmids and luciferase reporter constructs of p53/p73 target genes were provided by Prof. Gerry Melino and Prof. Thierry Soussi (Institut Curie and Université P. and M. Curie, Laboratoire de Génotoxicologie des Tumeurs, Paris, France) and have been described elsewhere.²⁴ The PUMA-luciferase constructs were from Prof. Bert Vogelstein (The Howard Hughes Medical Institute and Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, Baltimore, MD).⁵⁹ pCMV-Apoptin expression plasmid has been described elsewhere.³² The following replication-incompetent adenoviruses were used: AdE1A (dl324) containing functional E1A (both E1A12S and E1A13S) but with complete deletion of E1B and E3, and Ad-Del (dl312) as a control containing a complete deletion of E1A and E3.³⁷ Ad-GFP-TAp73β expressing TAp73β fused to GFP was obtained from Prof. Karen Vousden (Beatson Institute for Cancer Research, Glasgow, UK). Ad-GFP was obtained from Prof. Bert Vogelstein. Ad-GFP-Apoptin has been described previously.⁶⁰

Adenovirus amplification and purification

Approximately 1 × 10⁹ 293A cells were infected with different adenovirus constructs. Cells were harvested when a complete cytopathic effect was observed, i.e., 95–100% of the cells were rounded and 5–10% were floating. Cells were lifted from the plate by pipetting gently and collected by centrifuging 10 min at 200×g. Purification and titration of the adenovirus was essentially done as described previously.⁶¹

MTT cell proliferation assay

Cell survival was measured by the MTT assay as described previously.³¹

Transient transfection and fluorescence microscopy

Cells were seeded in Falcon 8-well culture slides (Becton Dickinson, Oxford, UK) and transfected at 50–80% confluency with 400 ng plasmid DNA preincubated with 1.4 μl LipofectAMINE 2000 (Invitrogen). Cells were washed in PBS and directly fixed in 2% paraformaldehyde for 30 min and permeabilized in 0.2% Triton X-100 for 15 min at room temperature. p73 expression was detected by anti-HA monoclonal antibody (Sigma) followed by goat anti-mouse IgG (whole antibody) fluorescein isothiocyanate conjugate (FITC) (Sigma). For Saos-2 inducible cells, p73 expression was detected by anti-HA monoclonal antibody (Sigma) followed by horse anti-mouse Texas Red (Vector Laboratories, Peterborough, UK). Subsequently, cells were mounted in Vectashield mounting medium (Vector Laboratories) containing 4′,6-diamidino-2-phenylindole (DAPI), for nuclei staining. Cells were visualized using an Olympus AX70 fluorescence microscope at high power (x400).

Luciferase reporter gene assays

Cells were seeded at 4 × 10⁴ cells per well in 96-well plates. Cells were transfected in the next day with 40 ng/well of the various luciferase reporter plasmid and 200 ng of inducer expression plasmid preincubated with 0.6 μl LipofectAMINE 2000 in 80 μl DMEM without FCS and antibiotics. In doxycycline-inducible Saos-2 cells harbouring the various p73 isoforms, cells were transfected with 40 ng/well of reporter construct followed by treatment with 2.5 μg/ml doxycycline. For normalizing transfection efficiency, 13.2 ng of the pRL-null Renilla luciferase reporter (Promega, Southampton, UK) was transfected into each well. At 24 h post-transfection, the medium was removed and the firefly and Renilla luciferase activities were sequentially assayed using Dual-Glo reagents (Promega) in a Wallac Trilux 1450 luminometer (PerkinElmer Life Sciences, Massachusetts, USA).

Western blot analysis

Cells were washed in cold PBS and lysed in Laemmli Sample buffer (62.5 mM Tris-Cl pH 6.7, 100 mM β-mercaptoethanol, 2% SDS, 1 μg/ml aprotinin, 1 μg/ml leupeptin, 100 μg/ml phenylmethylsulfonylfluoride). The lysate was passed through a 25G needle and boiled for 5 min. Lysates were resolved by gel electrophoresis for 1.5–2 h at 80 volts on 10–12% SDS-polyacrylamide gels. Western blotting was performed as previously described.⁶² The antibodies used for Western blot analysis were: mouse anti-E1A clone M58 (Pharmingen, BD Biosciences, Oxford, UK), mouse anti-HA (Sigma), rabbit anti-p73α (p73SAM)³⁹ (obtained from Prof. Gerry Melino), mouse anti-Abl (Pharmingen, BD Biosciences), rabbit anti-PUMA (abcam^®, Cambridge, UK), mouse anti-tubulin (Sigma), mouse anti-MDM2 (Oncogene Research Products, Calbiochem, Nottingham, UK), mouse anti-p21^WAF1/CIP1 clone EA10 (Oncogene Research Products, Calbiochem), rabbit anti-phospho-Apoptin (T108) (anti-108-P) (Eurogentec, Seraing, Belgium), rabbit anti-PARP p85 fragment clone G734A (Promega), secondary anti-mouse (Sigma) and anti-rabbit antibodies linked to horseradish peroxidase (Amersham, UK).

Determination of p73 half-life

p73 protein turnover was determined by adding 20 μg/ml cycloheximide to TAp73α, β, γ or ΔNp73α-inducible Saos-2 cell lines 24 h after induction with 2.5 μg/ml doxycycline. Protein levels were determined by collecting cells at the different time points and performing Western blotting as described above. The relative amount of p73 protein was evaluated by using Version 2.0 of Aida 2D Densitometry software and normalized to tubulin.

Animal study

Six-week-old female BALB/c nude mice were injected with 5 million HSC-3 head and neck cancer cells in 0.1 ml PBS. Cells were injected subcutaneously into the right flank of each mouse. Animals were monitored twice weekly for tumor growth. Two weeks later when tumor size reached 5–6 mm in diameter, mice were placed into groups of seven and were injected orthotopically with either PBS, 2 × 10⁹ pfu of Ad-GFP or Ad-GFP-TAp73β into several areas within the tumor mass. Tumor growth was monitored twice weekly. The tumors were measured with calipers for two perpendicular diameters and tumor size was calculated based on average dimensions.

Statistical analysis

For statistical analyses, the Student's t-test or one-way analysis of variance was carried out. Statistically significant difference was defined as p < 0.05.