Conditionally replicative adenovirus expressing degradation-resistant p53 for enhanced oncolysis of human cancer cells overexpressing murine double minute 2

Conditionally replicative adenoviruses (CRAd) represent a rapidly expanding novel class of anticancer agents. They are designed to selectively replicate in tumor cells and to destroy these cells by inducing lysis (1–3). The oncolytic effect of CRAds is amplified by in situ virus multiplication and spread to adjacent cells following lysis of infected tumor cells. The first clinical trials with CRAds have shown the safety of CRAds but have also shown that their oncolytic potency needs to be improved. Preclinical studies have shown several ways to accomplish this, including tropism modification, therapeutic transgene expression, and combination with other therapies (3).

Previously, we and others found that adenoviruses can be made into more effective oncolytic agents by expressing exogenous p53 tumor suppressor protein in infected cancer cells (4–6). The human p53-expressing CRAd AdΔ24-p53 more effectively killed most human cancer cell lines tested in vitro than did its parent AdΔ24 (4). Recently, we expanded this finding on primary human glioma and neuroblastoma cells in vitro and on xenografts in vivo (7, 8). Only few cancer cell lines were refractory to the oncolysis-enhancing effect of exogenous p53. Among these were cell lines with p53 wild-type, mutant, and null genotypes (4).

In search for a molecular determinant of resistance against the oncolysis-enhancing effect of exogenous p53, we focused our attention on the major negative p53 regulator oncoprotein murine double minute 2 (MDM2). MDM2 overexpression, due to gene amplification, is most frequently found in soft tissue sarcomas, osteosarcomas, and esophageal carcinomas (9). MDM2 directly binds to p53 and inhibits its transcriptional activity (10–12). In addition, MDM2 binding induces p53 nuclear export (13) and targets p53 for proteolytic degradation (14, 15). Hence, high MDM2 expression might inhibit p53-mediated oncolysis enhancement by efficient inactivation of exogenous p53.

Here we show that several cell lines that poorly responded to oncolysis enhancement by p53 constitutively expressed high levels of MDM2. To obviate MDM2-mediated inactivation of CRAd-encoded p53, we constructed the new CRAd AdΔ24-p53(14/19) that expresses a p53 variant incapable of binding to MDM2 (16). This new CRAd was more effective in killing cancer cells with high levels of human MDM2. This finding suggests that knowledge on molecular defects in cancer cells can be used to design tailored CRAds with augmented oncolytic potency.

A549 lung carcinoma cells and 293 E1-complementing cells were obtained from the American Type Culture Collection (Manassas, VA). 911 E1-complementing cells were provided by IntroGene (Leiden, the Netherlands). SaOs-2 and MNNG-HOS, U2OS, MG-63, and CAL-72 osteosarcoma cells were gifts from Dr. F. van Valen (Westfalische Wilhelms-Universität, Munster, Germany), Dr. S. Lens (Dutch Cancer Institute, Amsterdam, the Netherlands), Dr. C.W. Lowik (Leids Universitair Medisch Centrum, Leiden, the Netherlands), and Dr. J. Gioanni (Laboratoire de Cancerologie, Faculté de Medecine, Nice, France), respectively; MKN45 gastric carcinoma cells, A2780 ovary carcinoma cells, and SF763 astrocytoma cells were gifts from Dr. M. Tsujii (Osaka University School of Medicine, Osaka, Japan), Dr. E. Boven (VU University Medical Center, Amsterdam, the Netherlands), and Dr. D. Dougherty (Neurosurgery Tissue Bank, University of California San Francisco Brain Tumor Research Center, San Francisco, CA), respectively. MKN45 cells were grown in RPMI 1640 with 10% FCS and antibiotics; all other cell lines were maintained in F12-supplemented DMEM with 10% FCS and antibiotics (all from Gibco BRL Life Technologies, Paisley, United Kingdom).

CRAds AdΔ24 and AdΔ24-p53 were previously described (4). To construct AdΔ24-p53(14/19), two single nucleotide substitutions (T>A and T>G) that change amino acids 14 and 19 from L to Q and from F to C, respectively, were introduced into the p53 gene on plasmid pABS.4-p53 (4) by PCR-mediated site-directed mutagenesis using Pfu polymerase (Stratagene, La Jolla, CA). First, two PCR amplification products were made from pABS.4-p53 using upstream primer 5′-CGTTTCCCGTTGAATATGGC-3′ and mutation primer 5′-CTGAACATGTTTCCTGACTCTGAGGGGGCTC-3′, and downstream primer 5′-GAAGTCTCATGGAAGCCAGC-3′ and mutation primer 5′-CCTCAGAGTCAGGAAACATGTTCAGACC-3′, respectively. Next, the two amplification products were mixed and amplified using the upstream and downstream primers to generate a full-length amplification product containing the nucleotide substitutions. The wild-type Simian virus 40 early promoter (SVE)-p53 expression cassette from pABS.4-p53 was subcloned into the KpnI and SalI sites in the multiple cloning site of a pBluescriptSK− derivative with deleted SmaI restriction site to create pBSKΔSma-p53. The 568 bp KpnI/SmaI fragment from the PCR product encompassing the two mutations was used to replace the corresponding wild-type p53 fragment in pBSKΔSma-p53, creating pBSKΔSma-p53(14/19). Correct introduction of the two nucleotide substitutions without any other changes in the p53 gene sequence was confirmed by DNA sequencing. The SVE-p53(14/19) expression cassette from pBSKΔSma-p53(14/19) was cloned into KpnI/SalI-digested pABS.4 (Microbix Biosystems, Toronto, Canada) to create pABS.4-p53(14/19). The 4 kb PacI fragment of pABS.4-p53(14/19) was inserted into PacI-digested pBHG11 (Microbix Biosystems). A clone with the SVE-p53(14/19) cassette on the adenovirus L-strand was isolated and the kanamycin resistance gene was removed by SwaI digestion and self-ligation, yielding pBHG11-p53(14/19)-L. AdΔ24-p53(14/19) was made by homologous recombination in 911 cells between pXC1-Δ24 (a kind gift from Dr. R. Alemany, Gene Therapy Center, University of Alabama, Birmingham, AL) and pBHG11-p53(14/19)-L. Viruses were plaque purified, propagated on A549 cells, and CsCl gradient purified according to standard techniques. The E1Δ24 mutation and SVE-p53 or SVE-p53(14/19) insertion were confirmed by PCR and restriction analysis on the final products. Functional plaque forming unit (pfu) titers were determined by limiting dilution plaque titration on 293 cells according to standard techniques.

Cells (where indicated, 24 hours after infection with AdΔ24-p53 or AdΔ24-p53(14/19) at 100 pfu/cell) were harvested from plate cultures and lysed in 200 μL of 140 mmol/L NaCl, 0.2 mol/L triethanolamine, 2 g/L deoxychelate, 1 mmol/L phenylmethanesulfonylfluoride, and 50 μg/μL antipain by three freeze/thaw cycles. Lysates were cleared by centrifugation and protein concentrations were determined using the BCA Protein Assay Kit (Pierce, Rockford, IL). Equal amounts (15 μg for p53 and β-actin blots or 50 μg for MDM2 blots) of protein were separated on 7.5% or 10% SDS-PAGE gel and transferred to polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA). Immunoblots were processed according to standard procedures, using mouse monoclonal antibody SMP14 (DAKO, Glostrup, Denmark) for MDM2, PAb 1801 (NeoMarkers, Fremont, CA) directed against p53 amino acids 32 to 79, or AC-15 (Sigma, St. Louis, MO) for β-actin, followed by anti-immunoglobulin G-horseradish peroxidase conjugate (DAKO) and ECL or ECL^PLUS chemiluminescence detection reagent (Amersham Biosciences, Roosendaal, the Netherlands).

MG-63 cells were grown in six-well cultures until ∼70% confluence. The cells were subjected to 250 pfu/cell AdΔ24-p53 or AdΔ24-p53(14/19) for 1 hour and incubated in fresh medium at 37°C and 5% CO₂. Twenty-four hours postinfection, the medium was replaced with medium containing 30 or 60 μg/mL cycloheximide (Sigma) in two experiments and cells were cultured for 0, 1, or 4 hours before harvesting for Western blot analysis as above, using 35 μg protein and primary monoclonal antibodies 1801 and AC-15. Band intensities on X-ray film were measured on a Bio-Rad GS-700 imaging densitometer and analyzed using Quantity One software (Bio-Rad Laboratories, Veenendaal, the Netherlands). Approximate p53 half-lives were calculated from the average β-actin normalized signals from two independent experiments.

The reporter cell line SaOs-2/PG-13 was made by cotransfecting SaOs-2 cells with the p53-dependent reporter plasmid PG13-Luc (17), carrying the firefly luciferase gene driven by a p53-dependent promoter and pTK-hyg (Clontech, Palo Alto, CA); carrying the hygromicin resistance gene driven by a minimal herpes simplex virus thymidine kinase promoter in a 10:1 molecular ratio using Lipofectamine PLUS (Life Technologies) according to the method described by the manufacturer; and selecting stable transfectants with 200 μg/mL hygromicin B (Roche Diagnostics) for 2 weeks. To assess p53-specific transactivation as a result of adenovirus infection and exogenous p53 introduction, SaOs-2/PG-13 cells were seeded at 5 × 10⁴ per well in 24-well plates and infected the next day with AdΔ24-p53 or AdΔ24-p53(14/19) for 1 hour at 100 pfu/cell. After 24 hours of culture at 37°C, the luciferase expression in the cells was measured using the Luciferase Chemiluminescent Assay System (Promega, Madison, WI) and a Lumat LB 9507 luminometer (EG&G Berthold, Bad Wildbad, Germany). Transactivation is expressed as relative luciferase activity compared with mock-treated SaOs-2/PG-13 cells.

Cells were seeded 5 × 10⁴ cells per well in 24-well plates and cultured overnight. The next day, they were infected with AdΔ24, AdΔ24-p53, or AdΔ24-p53(14/19) at a dose range of 1 to 10⁻³ pfu/cell for 1 hour at 37°C. The cells were subsequently cultured for up to 3 weeks at 37°C with 50% medium changes twice weekly for most cell lines, or every other day for MKN45 cells. Cultures were maintained until cytopathic effects were evident in AdΔ24-infected cultures (i.e., for 12–20 days for different cell lines). The culture medium was then removed and adherent cells were stained with crystal violet dye as described (4).

Previously, we tested a large panel of human cancer cell lines for susceptibility to treatment with AdΔ24 or AdΔ24-p53 (4, 7, 8). Among these, we identified only five cell lines (i.e., A2780, MKN45, SF763, MNNG-HOS, and MG-63) that were not more rapidly killed by AdΔ24-p53 than by AdΔ24. The p53 genotypes of these cell lines did not hint at an explanation for their unresponsiveness to exogenous p53 expression. We envisioned that effective degradation of exogenous p53 in these cancer cell lines could perhaps have limited the efficacy of AdΔ24-p53 treatment. This prompted us to analyze expression of the major negative p53 regulator MDM2 by Western blot analysis. In addition, with two of the refractory cell lines being osteosarcoma cell lines, of three osteosarcoma cell lines tested, this cancer type seemed overrepresented. Because MDM2 is often amplified in osteosarcoma (9, 11), we also analyzed the osteosarcoma cell line SaOs-2, which is moderately sensitive to p53-mediated oncolysis enhancement (4), and osteosarcoma cell lines U2OS and CAL-72, which we had not included in our studies before. Figure 1 shows that the cell lines expressed various MDM2 isoforms, consistent with the existence of different MDM2 splice variants in cancer cells (18). All five refractory cell lines and CAL-72 cells expressed higher levels of 90 kDa full-length human MDM2 than did U2OS and SaOs-2. In addition, all cell lines except U2OS expressed an MDM2 species with slightly reduced molecular weight, possibly representing the ∼85 kDa isoform expressed from the full-length transcript (19, 20). MKN45, A2780, and SF763 cells displayed particularly high MDM2 levels with, in addition to the high molecular weight species, several smaller molecules. Together, these findings showed that resistant cell lines constitutively overexpressed MDM2, suggesting that this might be a determinant of resistance against the oncolysis-enhancing effect of exogenous p53.

To obviate a possible inhibitory effect of MDM2, we decided to construct a new AdΔ24-derived CRAd expressing a p53 variant resistant to MDM2-mediated degradation. Several such p53 variants have been described, including p53(14/19) and p53(22/23) with amino acid substitutions in the MDM2 binding domain (16), p53(d 13–19) with a deletion in the MDM2 binding site (15), and the chimeric p53 analogue CTS1 in which the entire NH₂-terminal domain comprising the MDM2 binding site was replaced by the herpes simplex virus VP16 transactivation domain (21). The variants p53(14/19), p53(d 13–19), and CTS1 have already been introduced into cancer cells with high MDM2 expression by means of replication-deficient adenovirus vectors and shown to confer growth inhibition and apoptosis induction in these cells (22–25). In this respect, p53(22/23) that lacks transactivation function was less effective (25, 26). The MDM2 binding domain in p53 partially overlaps with the binding site for adenovirus E1B-55K protein and comprises amino acids required for the transactivation function of p53 (16). Because the mechanism of oncolysis enhancement by p53 has not yet been resolved, we decided to use a p53 variant that leaves its transactivation and complex formation functions intact as much as possible. The p53(14/19) variant was reported to retain ∼50% transactivation activity and 60% E1B-55K binding (16). We chose this variant and constructed the new CRAd AdΔ24-p53(14/19), which is identical to AdΔ24-p53 except for amino acid substitutions L14Q and F19S. AdΔ24-p53(14/19) was characterized in comparison with its parent AdΔ24-p53 by infecting p53-null, MDM2-low SaOs-2 cells and measuring p53 expression and transactivation activity 24 hours later (Fig. 2). Western blot analysis using monoclonal antibody 1801, which is directed against a p53 epitope distinct from the mutations in p53(14/19), showed comparable p53 signals in AdΔ24-p53– and AdΔ24-p53(14/19)–infected cells (Fig. 2A). Assuming that the mutation did not affect antibody binding affinity, this suggested that the two variants accumulated to similar levels. Transactivation function was analyzed by measuring the activity of luciferase driven by a p53-dependent promoter (Fig. 2B). For this, we used SaOs-2/PG13-Luc, a cell line derived from SaOs-2 by stable transfection with the p53 reporter construct PG13-Luc (17). As can be seen, infection with AdΔ24-p53 caused p53 promoter transactivation, increasing luciferase expression by ∼7-fold. AdΔ24-p53(14/19) infection elevated luciferase activity ∼3-fold, indicating that the amino acid substitutions in p53(14/19) partially affected its specific transactivation function. This was consistent with the moderate loss of transactivation activity previously reported for this variant (16).

In adenovirus-infected, MDM2-overexpressing cells, wild-type p53 is degraded via E1B-55K– and E4orf6-mediated ubiquitination (27) as well as through MDM2-mediated ubiquitination, whereas p53(14/19) is resistant to MDM2-mediated degradation (16, 25, 26), but is still subject to adenovirus protein–mediated ubiquitination. We investigated relative stabilities of CRAd-encoded p53 and p53(14/19) in the face of MDM2 overexpression. MDM2-high, p53-null MG-63 cells were infected with the CRAds at a high multiplicity of infection and, 24 hours later, cycloheximide was added to inhibit de novo protein translation. Immediately thereafter, and 1 and 4 hours later, cell lysates were prepared and p53 was detected by Western blot analysis (Fig. 3). As can be seen, the amounts of CRAd-encoded wild-type p53 and p53(14/19) both declined. In two independent experiments, p53 was lost faster than p53(14/19). Approximate half-lives estimated from these experiments were 38 and 80 minutes for p53 and p53(14/19), respectively. The longer presence of p53(14/19) might suggest that adenovirus protein–mediated p53 degradation is less efficient than MDM2-mediated degradation. Alternatively, the moderate decrease in binding affinity of p53(14/19) to E1B-55K (16) may have affected adenovirus protein–mediated degradation.

Finally, we determined the oncolytic potencies of AdΔ24, AdΔ24-p53, and AdΔ24-p53(14/19) against A2780, MKN45, SF763, MG-63, and MNNG-HOS cells, known to be similarly susceptible to AdΔ24 and AdΔ24-p53 (i.e., cells not responsive to p53-mediated oncolysis enhancement), and against SaOs-2 cells that are more effectively killed by AdΔ24-p53 than by AdΔ24 (4). In addition, we included MDM2-low U2OS and MDM2-high CAL-72 cells that we had not previously evaluated for susceptibility to CRAd treatment. Cells were inoculated with each virus in a range of 1 to 0.001 pfu/cell and cultured for up to 3 weeks until cytopathic effects became evident in AdΔ24-infected cultures. Remaining attached cells were then stained with crystal violet to assess relative CRAd cytotoxicity. At the low multiplicities of infection used, AdΔ24 did not exert detectable toxicity on A2780 and MG-63 cells within 3 weeks. These cells required 100 pfu/cell or more to be eradicated (not shown). As can be seen in Fig. 4, AdΔ24-p53 was clearly more effective than AdΔ24 in killing MDM2-low U2OS cells, eradicating cell monolayers after infection with a 100-fold lower CRAd inoculum. MDM2-low SaOs-2 cells were also more effectively killed by AdΔ24-p53 than by AdΔ24, but not in all experiments. On average, oncolysis enhancement in SaOs-2 cells was less than 10-fold. In contrast, wild-type p53 expression did not at all enhance oncolysis of any of the MDM2-high cell lines, consistent with earlier observations. Hence, only MDM2-low cells were more susceptible to the wild-type p53-expressing CRAd, substantiating the notion that MDM2 overexpression correlates with resistance to p53-dependent oncolysis enhancement. Most importantly, the new CRAd AdΔ24-p53(14/19) expressing MDM2-binding deficient p53 was ∼10 times more effective than either AdΔ24 or AdΔ24-p53 in killing MDM2-high cells. AdΔ24-p53(14/19) was, however, less effective than AdΔ24-p53 against MDM2-low cells. This was confirmed on a larger panel of cancer cell lines previously shown to be highly susceptible to p53-mediated oncolysis enhancement (not shown). These findings showed that the efficacy enhancement through abrogation of MDM2 binding was specific and could not be explained by minor differences in viral titers. On MDM2-low cells, where p53 inhibition through MDM2 binding should be insignificant, mutation of the MDM2 binding site in p53 reduced CRAd efficacy, possibly due to reduced p53 transcriptional activity (see above). The mutations introduced into p53 to enhance CRAd efficacy against MDM2-high cancer cells were thus at the expense of a general decline in its oncolysis-enhancing activity. Therefore, the contribution of MDM2-binding inhibition to p53-mediated oncolysis enhancement in MDM2-overexpressing cancers was probably even underestimated. In the aggregate, AdΔ24-p53(14/19) seems a more effective agent for virotherapy treatment of the subset of MDM2-high cancers, rather than a generally more efficacious anticancer agent. For most cancers, previously constructed AdΔ24-p53 remains the preferred agent.

In conclusion, our results support the general notion that expression of functional p53 tumor suppressor protein augments the anticancer efficacy of CRAds. We designed tailored CRAd AdΔ24-p53(14/19) for a more effective oncolysis of cancer cells with constitutive MDM2 overexpression. Similar strategies might perhaps be followed to increase CRAd efficacy against cancers expressing other p53 inhibitors, such as human papillomavirus E6 protein, hepatitis B virus X protein, p53-associated Parkin-like cytoplasmic protein, or COP1 (28–31). This way, a library of CRAds targeted for different genetic aberrations in cancer could be compiled. Linking this repertoire of targeted therapeutics to the promise of functional genomics and proteomics technologies might contribute to the advance of individualized molecular medicine, wherein each patient is treated with a tailored CRAd designed to meet the genetic changes in that patient's particular tumor, resulting in a more effective virotherapy.