Enhanced tumor cell kill by combined treatment with a small-molecule antagonist of mouse double minute 2 and adenoviruses encoding p53

The tumor suppressor p53 is the central coordinator of cellular responses to danger signals, such as oncogene activation, aberrant growth signals, and DNA damage. In response to these danger signals, p53 triggers cell cycle arrest, DNA repair, and apoptosis (1). Abrogation of the p53 pathway is the most common change in human cancers. A dysfunctional p53 pathway seems critical in the pathogenesis and progression of tumors, as it fails to protect against sustaining mutations culminating into malignant transformation. In the majority of cancer cells, p53 function is lost due to mutation of the p53 gene (2). In most other cases, p53 function is compromised by inhibition of the p53 protein. Inhibitors of p53 include several cellular ubiquitin ligases as well as oncogenic viral proteins. In most cell types, the endogenous ubiquitin ligase mouse double minute 2 (MDM2) is regarded as the major negative regulator of p53 function (3, 4). MDM2 binds to p53, which inhibits p53 transcriptional activity, induces p53 nuclear export, and targets p53 for proteosomal degradation. High expression of MDM2, inactivating p53, is associated with tumor development (5, 6). MDM2 is amplified in ∼7% of cancers, with the highest frequencies found in soft-tissue sarcomas, osteosarcomas, and esophageal carcinomas (7). Tumors may also express MDM2 at high levels without gene amplification (8, 9).

Several anticancer therapeutic strategies aim to restore normal p53 functions. In cells expressing mutant p53, this includes introduction of the normal p53 gene or reactivating mutant p53 with molecules capable of stabilizing p53 protein folding. In p53 wild-type (WT) cells, this is done by inhibiting expression of p53 inhibitors or disrupting the interaction between p53 and its inhibitors (10). The central role of the p53-MDM2 interaction in regulating p53 activity has directed much attention to developing compounds that interfere with this interaction. Such molecules were derived from the p53 peptide interacting with the p53-binding site in MDM2 and were synthesized by computer-aided design or identified by screening libraries of microbial extracts or synthetic chemicals (11, 12). Recently, a class of small molecules, the Nutlins, was identified that inhibits the interaction between MDM2 and p53 with particular potency and selectivity (13). Nutlins showed antitumor activity in vitro and in vivo against cancer cells with WT p53 gene. Obviously, in tumors with mutated p53 where the p53-MDM2 interaction is redundant, treatment with a Nutlin has no effect. In these cells, p53 function can best be restored by introduction of exogenous WT p53 (e.g., by adenovirus vector–based gene transfer). This treatment is already in advanced clinical development phase for a variety of cancer types. However, high MDM2 expression was shown to hamper the anticancer effect of p53 gene therapy (14, 15), suggesting that this treatment is less suitable for MDM2-overexpressing cancers.

Oncolytic adenoviruses, which selectively replicate in cancer cells, are also effective gene delivery agents. Moreover, their capacity to lyse cancer cells (referred to as oncolysis) with subsequent spread of viral progeny to neighboring tumor cells makes them particularly attractive anticancer agents (16). Importantly, we found previously that exogenous p53 expression enhanced the cancer cell killing potency of an oncolytic adenovirus (17). Recently, however, we found that high expression of MDM2 limited the anticancer effect of exogenous p53 expressed by oncolytic adenoviruses (18).

Hence, high MDM2 expression is a determinant of resistance against p53 gene therapy and p53-enhanced oncolytic virus therapy. Here, we test the hypothesis that these therapies could be made more effective against cancer cells with p53 WT or mutant genotype by concomitant disruption of the p53-MDM2 interaction using the MDM2 antagonist Nutlin-3.

SJSA-1 and HCT116 cells were obtained from the American Type Culture Collection. The cell lines SaOs-2, CAL-72, U2OS, and MKN45 were kindly provided by Dr. F. van Valen (Westfalische Wilhelms-Universität Münster, Münster, Germany), Dr. J. Gioanni (Laboratoire de Cancerologie, Faculté de Medecine, Nice, France), Dr. S. Lens (Dutch Cancer Institute, Amsterdam, the Netherlands), and Dr. M. Tsujii (Osaka University School of Medicine, Osaka, Japan), respectively. Adenovirus E1-complementing 911 cells were obtained from IntroGene. SJSA-1 cells were maintained in RPMI 1640 supplemented with 10% FCS, 50 infectious units (IU)/mL penicillin plus 50 μg/mL streptomycin (PS), 1 mmol/L sodium pyruvate, 10 mmol/L HEPES, and glucose to a concentration of 4.5 g/L. HCT116 cells were cultured in McCoy's 5A medium supplemented with 10% FCS and PS. SaOs-2, CAL-72, and U2OS cells were cultured in F-12–supplemented DMEM with 10% FCS and PS. MKN45 cells were grown in RPMI 1640 with 10% FCS and PS. All cells were cultured at 37°C in a humidified, 5% carbon dioxide atmosphere.

The small-molecule antagonist of MDM2 Nutlin-3a and its inactive enantiomer Nutlin-3b, which serves as a control for nonspecific effects, were obtained from Hoffmann-La Roche. Nutlins were dissolved in DMSO at a concentration of 10 mmol/L and further diluted for experiments in complete medium.

The replication-deficient adenovirus vector Adwtp53 that harbors a human p53 gene driven by a SV40 virus early promoter in place of the E1 region has been described previously (19). The conditionally replicative adenovirus AdΔ24 that lacks 24 bp corresponding to amino acids 122 to 129 in the CR2 domain of E1A necessary for binding to Rb protein (20) and has a deleted E3 region and its derivative AdΔ24-p53 that carries a SV40 virus early-p53 expression cassette inserted in place of the deleted E3 region were described before (17).

Cells were harvested and lysed in 200 μL radioimmunoprecipitation buffer (50 mmol/L Tris-HCl, 1% Igepal, 0.5% sodium deoxycholate, 150 mmol/L NaCl, 0.05% SDS) supplemented with 1 μL of 200 mmol/L Pefabloc (Roche Diagnostics) and incubated for 30 min on ice. Lysates were cleared by centrifugation and protein concentrations were determined using bicinchoninic acid protein assay kit (Pierce). For detection of p21 or p53, 15 μg protein was fractionated on a NuPAGE 10% Bis-Tris gel (Invitrogen Life Technologies) and transferred to a polyvinylidene difluoride membrane (Bio-Rad). For detection of MDM2 protein, 35 μg protein was loaded on a 7% Tris-acetate gel (Invitrogen) and transferred to a polyvinylidene difluoride membrane. Proteins were detected by enhanced chemiluminescence Western blotting detection reagent (Amersham Biosciences) after incubation with primary antibodies for p53 (DO7, DAKO), p21 (SX118, DAKO), MDM2 (SMP14, DAKO), or β-actin (AC-15, Sigma) followed by secondary antibody anti-immunoglobulin G-horseradish peroxidase conjugate (DAKO). Protein levels were quantified by densitometric scanning on a VersaDoc 4000 imaging system using Quantity One software (Bio-Rad Laboratories).

Cells were plated at a density of 50% to 70% confluency in a 96-well plate and transfected either with the p53-dependent reporter plasmid PG13-Luc or with negative control construct MG15-Luc as described previously (17). After 3 h, cells were treated with or without adenovirus (Adwtp53, AdΔ24, or AdΔ24-p53) at a multiplicity of infection of 100 IU/cell, Nutlin-3a or Nutlin-3b (8 μmol/L), or a combination of virus and Nutlin. Forty-eight hours after transfection, cells were harvested and luciferase expression in cells was measured using the Luciferase Chemiluminescent Assay System (Promega). p53 transcriptional activity is expressed as the relative luciferase activity of PG13-Luc–treated cells over MG15-Luc–treated cells.

Cells were seeded in a 96-well plate at a density of 5 × 10³ cells per well. The next day, cells were inoculated with Adwtp53 at 100 IU/cell (SaOs-2 and U2OS) or 300 IU/cell (CaL-72, HCT116, SJSA-1, and MKN45), Nutlin-3a or Nutlin-3b (1 μmol/L for SJSA-1 and MKN45; 3 μmol/L for HCT116; and 8 μmol/L for SaOs-2, CAL-72, and U2OS), or a combination of both. Five days after treatment, culture medium was replaced by 100 μL of 10% WST-1 (Roche Diagnostics) in culture medium and cell viability was determined by colorimetric WST-1 conversion assay. Formazan dye formation was measured at 450 nm using a Bio-Rad model microplate reader. Relative WST-1 conversion in treated cells compared with untreated control cells was calculated after subtraction of WST-1 conversion in the absence of cells. Dose response graphs of Nutlin treatment were analyzed using the GraphPad Prism 4 program (GraphPad Software, Inc.). The IC₅₀ values of Nutlin treatment on different cell lines are given as mean values of three independent experiments with SD.

Cells (5 × 10³) were seeded per well in a 96-well plate and cultured overnight. The next day, cells were treated with Adwtp53 or the control vector AdCMV-Luc (SaOs-2 and U2OS at 50 IU/cell and CAL-72, HCT116, SJSA-1, and MKN45 cells at 300 IU/cell) for 1.5 h. Subsequently, cells were treated with Nutlin-3a or Nutlin-3b (SaOs-2, CAL-72, and U2OS at 8 μmol/L; HCT116 at 3 μmol/L; and SJSA-1 and MKN45 at 1 μmol/L). On days 1 to 5 after infection, cells were harvested. Cell suspensions are mixed 1:1 with 0.4% trypan blue solution (Sigma) and viable cells were counted using a cell counting chamber (Marienfeld). The detection limit in this assay was 289 viable cells per culture.

SaOs-2 and U2OS cells were seeded at 1 × 10⁵ cells per well in a six-well plate and cultured overnight. The next day, cells were treated with Adwtp53 (100 IU/cell), Nutlin-3a (8 μmol/L), Nutlin-3b (8 μmol/L), or a combination of virus and Nutlin. Forty-eight hours after treatment, adherent and nonadherent cells were collected. Cells were washed in PBS and fixed in 70% ethanol and stored for a minimum of 1 h at −20°C. Before analysis, cells were washed in PBS again. The pellet was resuspended in propidium iodide solution [50 μg/mL propidium iodide and 1 mg/mL RNase A (Sigma) in PBS] for 1 h and analyzed by flow cytometry on a FACScan (Becton Dickinson). Cell cycle distribution was analyzed by ModFitLT cell cycle analysis software (Verity Software House). This analysis excludes the sub-G₁ cell population before calculating the percentage cells in G₀-G₁, S, and G₂-M phases of the cell cycle.

U2OS cells (5 × 10³) were seeded per well in a 96-well plate. The next day, cells were infected with AdΔ24-p53 at 100 IU/cell for 1.5 h at 37°C after which infection medium was replaced with fresh medium with or without Nutlin-3a or Nutlin-3b (8 μmol/L). Cells and supernatant were harvested separately 1 to 8 days after infection. Virus was released via three freeze-thaw cycles. The number of infectious particles produced in the cells and in the medium was determined by 3-fold limiting dilution infection on 911 cells in triplicate. After 48 h, cells were fixed in methanol and stained for hexon expression using the mouse anti-hexon antibody and rat anti-mouse immunoglobulin (horseradish peroxidase conjugate) from the AdenoX rapid titer kit (BD Biosciences).

Cells were seeded at a density of 5 × 10³ cells per well for most cell lines and 5 × 10² cells per well for SJSA-1 cells in a 96-well plate. The next day, cells were infected with AdΔ24 or AdΔ24-p53 at the indicated multiplicity of infection in medium containing 2.5% FCS for 1.5 h at 37°C. Subsequently, complete medium with or without Nutlin-3a or Nutlin-3b (8 μmol/L for SaOs-2, CAL-72, and U2OS; 3 μmol/L for HCT116; and 1 μmol/L for SJSA-1 and MKN45 cells) was added. Cells were harvested 8 to 15 days after infection with 50% medium (with or without Nutlin) refreshment once or twice weekly. After removal of the culture medium, cells were fixed in 4% formaldehyde in PBS and stained using 1% crystal violet dye in 70% ethanol. After several washes, culture plates were air dried and scanned on a Bio-Rad GS-690 imaging densitometer.

The Student's t test was used for statistical evaluation of differences between treatment groups. Differences were considered to be significant when P < 0.05.

For our studies, we selected six human cancer cell lines representing different p53/MDM2 phenotypes. Figure 1A shows Western blot analysis for expression of p53, p21, and MDM2 in these cells. Expression levels were normalized by that of β-actin. As expected, in p53-null SaOs-2 and CAL-72 osteosarcoma cells, p53 was absent and expression levels of its transcriptional targets p21 and MDM2 were low or undetectable. The other four cell lines (i.e., SJSA-1 and U2OS osteosarcoma cells, MKN45 gastric carcinoma cells, and HCT116 colorectal carcinoma cells) all expressed p53, consistent with their documented p53 WT genotypes. The p53 target p21 was expressed in all four cell lines, albeit at low level in MKN45 cells. MDM2 expression was lower in U2OS cells than in HCT116, SJSA-1, and MKN45 cells.

We then subjected these cancer cells to treatment with the specific small-molecule antagonist of the p53-MDM2 interaction Nutlin-3. This cis-imidazoline derivative was synthesized as racemic mixture from which the enantiomers 3a and 3b were separated with the use of a chiral column. Nutlin-3b is 150 times less potent in binding MDM2 than Nutlin-3a (13). Therefore, Nutlin-3b was included in all experiments to serve as control for non-MDM2–related activities. Cells were incubated with Nutlin-3a or Nutlin-3b for 5 days and IC₅₀ values were then determined (Fig. 1B). Consistent with its mode of action, Nutlin-3a exerted selective toxicity against p53 WT cells only. Nutlin-3a–specific IC₅₀ values were not reached for p53-deficient SaOs-2 and CAL-72 cells, where antagonizing the p53-MDM2 interaction should be irrelevant.

Next, we investigated the effects of p53 gene therapy using replication-defective adenovirus vector Adwtp53 or MDM2 inhibition using Nutlin-3 on p53 activity in cancer cells. We subjected the six cancer cell lines to Adwtp53, Nutlin-3a, or Nutlin-3b and measured p53 transcriptional activity using a p53-dependent reporter construct. As can be seen in Fig. 1C, all cell lines displayed p53 activities that were consistent with their genetic status and with the results from the Western blot analysis. WT p53 cells had p53-specific transactivation values ranging from 10 to 80, whereas in p53-deficient cells, these values were <2. Transduction of p53-deficient cells with Adwtp53 resulted in very high p53 activities. Adwtp53 also increased p53 transcriptional activity in p53 WT cells, but on average to a lesser extent than in p53-deficient cells. This suggested that p53 inhibitors including MDM2 that are expressed in p53 WT cancer cells also inhibited exogenous Adwtp53-encoded p53 to some extent. Cells were also infected with the control virus AdGFP, a replication-deficient adenovirus vector expressing green fluorescence protein. This did not affect p53 transactivation (data not shown). As expected, Nutlin-3a induced p53 activity in p53 WT cells (by at least 2-fold) but not in p53-deficient cells. The absolute increase in p53 activity was substantially lower in HCT116 cells than in the other three cell lines. In addition to MDM2, several other endogenous p53 inhibitors have been identified (21). Perhaps, multiple p53 inhibitors are expressed in HCT116 cells, in which case disrupting the p53-MDM2 interaction would only partially restore p53 activity. This would be in line with the previous observation that silencing MDM2 in HCT116 cells by RNA interference did not stabilize p53 (22). Taken together, Nutlin-3a and Adwtp53 each contributed to restoring p53 functions in cancer cells in their own way (i.e., by reactivating endogenous p53 and by providing exogenous p53 activity, respectively).

Because MDM2 seemed to inhibit both endogenous and exogenous p53 and because exogenous p53 may induce MDM2 expression in p53-deficient cells with an intact MDM2 gene, we hypothesized that combined treatment of cancer cells with Adwtp53 and Nutlin could be more effective than either treatment alone. Cancer cells were treated with Adwtp53, Nutlin-3a, or a combination of both agents. As controls, AdCMV-Luc, a replication-deficient adenovirus vector expressing firefly luciferase, and Nutlin-3b were used. Five days later, cell viabilities were measured by WST-1 conversion assay (Fig. 1D). As single agents, Adwtp53 decreased viability only of p53-null SaOs-2 and CAL-72 cells and Nutlin-3a decreased viability only of p53 WT cell lines. Most importantly, combination treatment with Adwtp53 plus Nutlin-3a was more effective than the best of either single treatment on SaOs-2, CAL-72, U2OS, HCT116, and SJSA-1 cells (P < 0.05). The increased cell killing by combination treatment was dependent on introduction of exogenous p53 and disruption of the p53-MDM2 interaction by Nutlin-3a because neither AdCMV-Luc with Nutlin-3a nor Adwtp53 with Nutlin-3b caused similar combination effects.

These findings were corroborated by counting viable cells 1 to 5 days after start of treatment (Fig. 2). Nutlin-3a alone did not affect expansion of SaOs-2 cells and had only minor effect on CAL-72 and MKN45 cells (2%, 21%, and 29% reduction in viable cell numbers on day 5, respectively). In contrast, it strongly reduced U2OS, HCT116, and SJSA-1 viable cell output compared with controls (86%, 91%, and 76% reduction, respectively; all P < 0.005). However, also in these cultures, absolute cell numbers increased over time, suggesting that Nutlin-3a delayed cell proliferation rather than killed cells. Adwtp53 treatment as monotherapy killed SaOs-2 cells (absolute loss of viable cells) and delayed proliferation of CAL-72 (P < 0.001) and U2OS (P < 0.005) cells. In line with the observations in the WST-1 assay experiment, Adwtp53 with Nutlin-3a was the most effective treatment on all six tested cell lines (P values compared with the best single agent treatment: P < 0.05 for SaOs-2, U2OS, HCT116, SJSA-1, and MKN45; P < 0.01 for CAL-72). On most cell lines, combination treatment reduced the absolute viable cell count, indicating that cells were killed. Again, we did not observe similar effects when treated with AdCMV-Luc+Nutlin-3a or Adwtp53+Nutlin-3b.

To explain the combined effect of activating the p53 pathway in cancer cells with Adwtp53 and Nutlin-3a, we investigated the effect of this treatment on cell cycle arrest and apoptosis induction. SaOs-2 and U2OS cells were subjected for 2 days to 100 IU/cell Adwtp53 or 8 μmol/L of the active or inactive Nutlin-3 enantiomer or to a combination of Adwtp53 plus Nutlin-3. Cell cycle distribution analysis (Fig. 3) showed that Adwtp53 treatment decreased the S phase population (P < 0.005) in both cell lines, indicating activation of cell cycle checkpoints. As single agent, Nutlin-3a, but not 3b, caused a substantial shift in cell cycle populations only in p53 WT U2OS cells. Accumulation of cells in G₁ phase (P < 0.01) and depletion of cells in S phase (P < 0.005) indicated arrested cell cycle progression mainly at the G₁-S cell cycle checkpoint. Combined treatment with Adwtp53 and Nutlin-3a caused the highest accumulation in G₂-M phase in both cell lines, although this did not differ significantly from single agent treatments. As measure for cells in apoptosis, we analyzed sub-G₁ events (Fig. 3A). Due to variation between experiments, differences were not significant. However, whereas Nutlin-3a–treated cells did not show more apoptosis than controls, Adwtp53 consistently increased the number of sub-G₁ events in both cell lines (average, 7–38% in SaOs-2; 4–10% in U2OS). Combination treatment further increased the fraction of cells in apoptosis (average, 59% in SaOs-2; 25% in U2OS). Thus, protecting Adwtp53-encoded p53 from MDM2-mediated inactivation seemed to augment apoptosis induction.

An effective way to kill cancer cells is by means of oncolytic adenoviruses. Previously, we observed profound anticancer effects by the p53-expressing oncolytic adenovirus AdΔ24-p53 (17, 23). AdΔ24-type oncolytic adenoviruses encode E1A proteins that are incapable of sequestering pRb from E2F-1, due to a deletion of eight amino acids in their pRb-binding CR2 domain (20). This deprives the virus of its capacity to induce cell cycle progression in normal cells with intact pRb control. Consequently, AdΔ24-type adenoviruses exhibit selective replication in cells with dysfunctional pRb control, where E2F-1 mediates cell cycle progression through the G₁-S phase checkpoint. Because Nutlin-3 activated this checkpoint (see Fig. 3) and could thus potentially inhibit replication of AdΔ24-type oncolytic adenoviruses, we investigated if Nutlin treatment affected replication of AdΔ24-p53. U2OS cells were cultured in the presence of 8 μmol/L Nutlin-3a or Nutlin-3b and infected with 100 IU/cell AdΔ24-p53. At various time points after infection, the amount of infectious virus particles produced was determined. Figure 4A shows that over a period of 8 days, the titer of AdΔ24-p53 rose to a final yield of ∼2 × 10⁴ IU/cell. The virus production profiles in the presence of the active or the inactive Nutlin enantiomer were almost identical. Thus, Nutlin-3 did not inhibit oncolytic adenovirus replication. To investigate if Nutlin-3 affected oncolysis and release of progeny virus, we measured AdΔ24-p53 titers in cells and culture medium separately. Figure 4B shows that the release of AdΔ24-p53 progeny into the medium of infected U2OS cells was accelerated in the presence of Nutlin-3a compared with Nutlin-3b. Taken together, these data showed that in the presence of Nutlin-3a, the production of infectious AdΔ24-p53 in U2OS cells was not changed, but AdΔ24-p53 progeny release was induced sooner after infection.

We next sought to determine if Nutlin-3 protects oncolytic adenovirus-encoded p53 from MDM2-mediated inactivation. In adenovirus-infected cells, adenovirus proteins including E1B-55K and E4orf6 interact with p53. This is associated with p53 redistribution over cellular compartments and degradation (24, 25). In the face of adenovirus proteins interacting with p53, the binding of p53 to MDM2 might therefore become bypassed. To investigate if Nutlin-3 would still affect p53 activity in this context, cells were infected with AdΔ24 or AdΔ24-p53 and treated with Nutlin-3a or Nutlin-3b. Two days later, p53 transcriptional activity was measured. As can be seen in Fig. 5, AdΔ24 infection severely decreased p53 activity in p53 WT cells (compare with Fig. 1C). In three of these cell lines, Nutlin-3a protected endogenous p53 at least partially from adenovirus protein-mediated inactivation. Nutlin-3a also increased p53 transcriptional activity in AdΔ24-p53–infected p53 WT cell lines, to levels exceeding those in AdΔ24/Nutlin-3a–treated cells. This suggested that also in the face of oncolytic adenovirus replication, Nutlin-3a protected exogenous p53 from inactivation. However, AdΔ24-p53–encoded p53 transcriptional activity remained low (compare with Adwtp53-encoded p53; see Fig. 1C), suggesting that most exogenous p53 interacted with AdΔ24-encoded or induced proteins.

Finally, we explored if combined treatment with oncolytic adenovirus and Nutlin-3a would accomplish a more effective eradication of cancer cells. The six cancer cell lines were inoculated with AdΔ24 or AdΔ24-p53 in a dilution titration. The viruses were allowed to replicate for up to 15 days before remaining cells were stained with crystal violet (Fig. 6). AdΔ24 and AdΔ24-p53 exhibited dose-dependent cytotoxicity on all cancer cell lines. Oncolytic efficacies against SJSA-1 cells, which express low levels of the primary adenovirus receptor CAR (data not shown), were less than against the other cell lines. Nutlin-3a did not clearly augment oncolytic adenovirus efficacy against the two p53-deficient cell lines. In contrast, treatment with Nutlin-3a, but not with Nutlin-3b, augmented oncolytic propagation of AdΔ24 and AdΔ24-p53 on all four WT p53 cell lines. Under the experimental conditions used, enhancement of AdΔ24 oncolytic efficacy by Nutlin-3 treatment ranged from an estimated 10-fold on HCT116, SJSA-1, and MKN45 cells to 100-fold on U2OS cells. Combined AdΔ24-p53 with Nutlin-3a treatment was somewhat more effective, with efficacy increases up to 100-fold on MKN45 cells and 1,000-fold on U2OS cells (see Fig. 6).

We investigated if restoring p53 functions by concomitant p53 gene delivery and disruption of the p53-MDM2 interaction would provide an effective means to kill cancer cells irrespective of their p53 status. Separately, adenovirus-encoded p53 and MDM2 antagonist Nutlin-3 increased functional p53 expression and induced cancer cell death in line with their known modes of action. Together, exogenous p53 and Nutlin-3 elicited the most effective death of p53-positive and p53-negative cancer cells. Nutlin-3 thus enhanced endogenous as well as exogenous p53-mediated cell kill. Interestingly, the effect of Nutlin-3 on exogenous p53 seemed somewhat different from its documented effect on endogenous p53. Endogenous WT p53 reactivated by Nutlin-3 treatment arrested U2OS cells only in G₁, whereas addition of exogenous p53 also activated the G₂-M cell cycle checkpoint. Moreover, although U2OS did not exhibit noticeable apoptosis in response to Nutlin-3 Adwtp53/Nutlin-3 combination treatment induced extensive apoptosis, killing all cells within 2 days. Thus, p53-dependent apoptotic signaling was not impaired in U2OS cells and could be exploited to kill these cells by Adwtp53/Nutlin-3 combination treatment. It remains to be elucidated which factors govern differential cell cycle and apoptosis responses to endogenous versus exogenous p53.

Nutlin-3 also markedly enhanced the potency of oncolytic adenoviruses against WT p53 cell lines. It accelerated cell death and viral burst, culminating in a more effective eradication of these cells. This is in line with previous observations that p53 promotes oncolytic adenovirus propagation in cancer cells (17, 23, 26, 27). Nutlin-3 thus presumably augments oncolytic adenovirus propagation by relieving p53 from MDM2 inhibition. Interestingly, in this context, reactivation of endogenous p53 seemed more important than protecting exogenous p53 from degradation. Recently, it was found that adenovirus E1A protein stabilizes p53 in the nucleus through forming a tricomplex with the structural MDM2 homologue MDMX (28). Notably, Nutlin-3 does not inhibit the MDMX-p53 interaction (29). This suggests that combined treatment with Nutlin-3 will not interfere with this potentially important p53 regulation by oncolytic adenovirus.

The precise mechanism by which p53 augments adenovirus-induced killing of cancer cells remains poorly understood. We found that Nutlin-3 reproducibly augmented AdΔ24 and AdΔ24-p53 propagation on HCT116 cells, without increasing p53 transcriptional activity. This could suggest that the oncolysis-enhancing effect of p53 is independent of its transcriptional activity. Interestingly, it was reported recently that during early stages of adenovirus infection, a fraction of p53 accumulates in mitochondria (25), where p53 is known to exert direct transcription-independent proapoptotic activity (30). It could thus be speculated that oncolysis enhancement by p53 relies primarily on its apoptogenic effects at mitochondria, rather than on nuclear p53 driving target gene expression. However, Nutlin-3 also increased AdΔ24 oncolytic efficacy on U2OS cells, where endogenous p53 reactivation did not promote apoptosis. This would point at a p53-dependent effect on adenovirus propagation entirely distinct from apoptosis induction. Clearly, these hypotheses await experimental verification.

In the aggregate, we have shown herein that restoration of p53 functions in cancer cells by concomitant exogenous p53 gene delivery and MDM2 inhibition provides efficacious cancer cell kill by increased induction of cell death and augmented adenovirus-induced oncolysis.