A Single Recombinant Adenovirus Expressing p53 and p21-targeting Artificial microRNAs Efficiently Induces Apoptosis in Human Cancer Cells

Purpose: Gene transfer involving p53 is viewed as a potentially effective cancer therapy, but does not result in a good therapeutic response in all human cancers. The activation of p53 induces either cell cycle arrest or apoptosis. Cell cycle arrest in response to p53 activation is mediated primarily through the induction of the cyclin-dependent kinase inhibitor p21. Because p21 also has an inhibitory effect on p53-mediated apoptosis, the suppression of p53-induced p21 expression would be expected to result in the preferential induction of apoptosis. However, p21 also has tumor-suppressive properties. In this study, we developed an adenovirus vector that expresses p53 and suppresses p21 simultaneously to enhance p53-mediated apoptosis.

Experimental Design: We constructed a replication-deficient recombinant adenovirus (Ad-p53/miR-p21) that enabled cocistronic expression of the p53 protein and artificial microRNAs that targeted p21, and examined the therapeutic effectiveness of this vector in vitro and in vivo.

Results: The levels of p21 were significantly attenuated following infection with Ad-p53/miR-p21. In colorectal and hepatocellular carcinoma cells, infection with Ad-p53/miR-p21 augmented apoptosis as compared with an adenovirus that expressed p53 alone (Ad-p53/miR-control). Ad-p53/miR-p21 also significantly increased the chemosensitivity of cancer cells to adriamycin (doxorubicin). In a xenograft tumor model in nude mice, tumor volume was significantly decreased following the direct injection of Ad-p53/miR-p21 into the tumor, as compared with the injection of Ad-p53/miR-control.

Conclusion: These results suggest that adenovirus-mediated transduction of p53 and p21-specific microRNAs may be useful for gene therapy of human cancers.

p53 is one of the most important tumor suppressor genes. In approximately half of all human cancers, p53 is inactivated as a direct result of mutations in the p53 gene (1, 2). In other cancers, p53 is inactivated through its association with viral oncoproteins, or as a result of alterations in genes that are involved in the p53 signaling network (3). Furthermore, mutation or deletion of p53 is related to poor prognosis and resistance to chemotherapy and radiation (4–7). Vector-mediated gene transfer of p53 is viewed as a potentially effective cancer therapy. In fact, clinical trials of adenovirus-mediated p53 gene therapy are ongoing in patients with head and neck (phase III), non–small cell lung (phase II), breast (phase II), and esophageal (phase II) cancer (8). However, gene transfer of p53 does not always have a good therapeutic outcome in all cancers (9–11).

The activation of p53 is induced by a variety of cell stresses, such as DNA damage, oncogene activation, spindle damage, and hypoxia. Activated p53 transactivates a number of genes, many of which are involved in DNA repair, cell cycle arrest, and apoptosis (12). Depending on the cell type and intensity of stress, p53 activation induces either cell cycle arrest or apoptosis (13). The precise mechanism that regulates whether a cell undergoes cell cycle arrest or apoptosis is unclear. Cell cycle arrest induced by p53 is mediated primarily through the induction of expression of p21 (14–17), which is involved in restoring genomic integrity by functioning in an antiapoptotic manner (18–20). Thus, the suppression of p53-induced expression of p21 might lead to the preferential induction of apoptosis rather than cell cycle arrest following p53 activation. Because resistance to apoptosis is a major factor in carcinogenesis, the attenuation of p21 may also enhance the effectiveness of chemotherapeutic agents.

There are several lines of evidence that p21 also has tumor-suppressive properties. Tumor susceptibility is increased in p21-null mice (21–24), and mice that lack p21 are more prone to developing malignant skin tumors following exposure to carcinogens (25, 26). Following a single dose of γ-irradiation, p21-deficient mice develop more tumors, and the tumors have increased metastatic potential (27). In addition, it has been shown that the suppression of p21 induces cell cycle progression, resulting in increased cell proliferation (28, 29). Thus, the suppression of p21 independently of p53 expression might increase the risk of tumor progression in some cancers. To avoid this risk, p21 suppression must be linked to p53 expression.

In the current study, we constructed a recombinant adenoviral vector that enabled the cocistronic expression of the p53 protein and microRNAs (miRNA) that targeted p21. We examined the therapeutic effectiveness of this vector in vitro and in vivo.

Cell culture. The human embryonic kidney cell line HEK293 was obtained from the Riken Cell Bank. The colorectal cancer cell lines DLD-1 and SW480, and the hepatocellular carcinoma cell line Hep3B were purchased from the American Type Culture Collection. The hepatocellular carcinoma cell line HLF was from the Health Science Research Resource Bank. HCT116 (p53−/−) cells were kindly provided by Dr. Bert Vogelstein (Johns Hopkins University). HEK293 cells were cultured in DMEM supplemented with 10% FCS. SW480 cells were cultured in Leibovitz L-15 medium with 10% FCS. All other cell lines were cultured in RPMI-1640 medium with 10% FCS.

Plasmids. Three pre-miRNA sequences were designed that targeted the 3′ untranslated region (UTR) of the human p21 mRNA using an online tool, Invitrogen's RNAi Designer.³

The engineered pre-miRNA sequences were designed as a mimic of the endogenous murine miR-155. The double-stranded DNA oligonucleotides corresponding to the three different p21-specific pre-miRNAs and a control sequence were individually cloned into the parental vector pcDNA6.2-GW/miR (Invitrogen) to generate pcDNA6.2-miR-p21A, pcDNA6.2-miR-p21B, pcDNA6.2-miR-p21C, and pcDNA6.2-miR-control, respectively. The three p21 pre-miRNAs were also cloned in tandem into one plasmid by multiple rounds of chaining to generate pcDNA6.2-miR-p21 which enabled cocistronic expression of multiple miRNAs. An expression vector for a FLAG epitope fusion protein of p53 (pCMV-Tag2-FLAG-p53) was generated from pCMV-Tag2-FLAG (Stratagene). The coding region of the human p53 gene was cloned into pcDNA6.2-miR-p21 and pcDNA6.2-miR-control, to generate pcDNA6.2-p53/miR-p21 and pcDNA6.2-p53/miR-control, respectively. The oligonucleotide sequences of the engineered pre-miRNA and adjacent flanking regions used for plasmid construction were as follows: miR-p21A: 5′-TGCTGTAGGGTGCCCTTCTTCTTGTGGTTTTGGCCACTGACTGACCACAAGAAAGGGCACCCTA-3′ (forward), 5′-CCTGTAGGGTGCCCTTTCTTGTGGTCAGTCAGTGGCCAAAACCACAAGAAGAAGGGCACCCTAC-3′ (reverse); miR-p21B: 5′-TGCTGAGCTGCCTGAGGTAGAACTAGGTTTTGGCCACTGACTGACCTAGTTCTCTCAGGCAGCT-3′ (forward), 5′-CCTGAGCTGCCTGAGAGAACTAGGTCAGTCAGTGGCCAAAACCTAGTTCTACCTCAGGCAGCTC-3′ (reverse); miR-p21C: 5′-TGCTGAATACTCCAAGTACACTAAGCGTTTTGGCCACTGACTGACGCTTAGTGCTTGGAGTATT-3′ (forward), 5′-CCTGAATACTCCAAGCACTAAGCGTCAGTCAGTGGCCAAAACGCTTAGTGTACTTGGAGTATTC-3′ (reverse); miR-control: 5′-TGCTGAAATCGCTGATTTGTGTAGTCGTTTTGGCCACTGACTGACGACTACACATCAGCGATTT-3′ (forward), 5′-CCTGAAATCGCTGATGTGTAGTCGTCAGTCAGTGGCCAAAACGACTACACAAATCAGCGATTTC-3′ (reverse).

Recombinant adenovirus. Recombinant adenovirus was produced using the ViraPower Adenoviral Expression System (Invitrogen), according to the manufacturer's instructions. Briefly, the recombination region of each pcDNA6.2-GW/miR–based expression vector was transferred to the Gateway Vector pAd/CMV/V5-DEST using the transfer vector pDONR221 in an in vitro recombination reaction. The recombined adenoviral plasmids generated from pAd/CMV/V5-DEST in this manner were transformed into competent DH5α (Toyobo). After selection, a single clone of DH5α was isolated and expanded. The recombinant adenoviral plasmid was purified, and then transfected into 293A cells. After a sufficient cytopathic effect was observed in 293A cells, adenovirus was purified using the Adeno-X Virus Purification Kit (Clontech). The recombinant adenoviruses Ad-p53/miR-21, Ad-p53/miR-control, Ad-mock/miR-p21, and Ad-mock/miR-control were generated from pcDNA6.2-p53/miR-21, pcDNA6.2-p53/miR-control, pcDNA6.2-miR-p21, and pcDNA6.2-miR-control, respectively. All insertion sequences were confirmed by nucleotide sequencing. Detailed information about the construction of recombinant adenoviruses is available from the authors upon request.

Adenovirus titer in plaque-forming units was determined by plaque formation assay following infection of HEK293 cells. The multiplicity of infection was defined as the ratio of the total number of plaque-forming units to the total number of cells that were infected. We titrated adenovirus from duplicate samples in order to confirm the reproducibility of the experiments.

Western blot analysis. The anti-p21 (Ab-1) mouse monoclonal antibody was purchased from Calbiochem, antiactin mouse monoclonal antibody was from Chemicon, anti-FLAG M2 mouse monoclonal antibody was from Sigma-Aldrich, and anti-p53 (DO-1) mouse monoclonal antibody was from Santa Cruz Biotechnology. Total cell lysate was extracted at 4°C with radioimmunoprecipitation assay buffer (150 mmol/L NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mmol/L Tris HCl, pH 8.0). Samples were fractionated by SDS-PAGE and transferred onto Immobilon-P membranes (Millipore). Immunoreactive proteins were detected using enhanced chemiluminescence (Amersham).

Immunofluorescence microscopy. p53(−/−) HCT116 cells were grown on poly-L-lysine-coated coverslips (Asahi Technoglass). After being fixed with 4% paraformaldehyde, the cells were incubated with anti-FLAG rabbit polyclonal antibody (Sigma-Aldrich) and anti-p21 mouse monoclonal antibody (Calbiochem) overnight at 4°C. Following incubation with Alexa Fluor 488-labeled goat antimouse IgG and Alexa Fluor 594-labeled goat antirabbit IgG (Invitrogen), the coverslips were inspected with a fluorescence microscope (Keyence).

Flow Cytometry. Cells (1 × 10⁶) were plated in 6-well plates. Twenty-four hours after plating, the cells were incubated with purified virus in 1 mL of medium supplemented with 1% FCS with brief agitation every 10 min. For flow cytometry, at various times after infection, cells were harvested by trypsinization and pelleted by centrifugation. Pelleted cells were fixed in 90% cold ethanol, treated with RNase A (500 units/mL), and then stained with propidium iodide (50 mg/mL). Samples were analyzed on a FACSCalibur flow cytometer (BD Bioscience). Experiments were repeated at least three times, and 50,000 events were analyzed for each sample. Data were analyzed using FlowJo software (Tree Star). For combination therapy, 24 h after infection, cells were treated with 0.5 μg/mL of doxorubicin, and then analyzed by flow cytometry after 24 h, as described above.

Determination of caspase-3 activity. Caspase-3 activity was determined by colorimetric assay using a caspase-3 assay kit (Biovision), according to the manufacturer's instructions. The kit utilizes synthetic tetrapeptides labeled with p-nitroanilide. Briefly, cells were lysed in the lysis buffer that was supplied with the kit. The supernatants were collected and incubated at 37°C with reaction buffer containing dithiothreitol and substrates. Caspase-3 activity was determined by measuring changes in absorbance at 405 nm using a microplate reader.

Animal models. All animals were maintained under specific pathogen-free conditions and treated in accordance with guidelines set by the Animal Care and Use Committee of Sapporo Medical University. To evaluate the effect of treating established tumors, 24 female BALB/c nude mice were injected s.c. into both flanks with 2 × 10⁶ SW480 or DLD1 cells. When the tumor size reached 100 mm³, the mice received direct an intratumoral injection of 1 × 10⁹ plaque-forming units (in 100 μL of PBS) of the indicated adenovirus a total of three times, on days 0, 1, and 2. Three mice were used for each treatment group. Tumor formation in mice was monitored for up to 4 wk. The tumor volume was calculated using the equation V (mm³) = a × b²/2, where “a” represents the largest dimension and “b” is the perpendicular diameter.

Expression of p53 and suppression of p21 induction by a single plasmid vector. We designed three different artificial pre-miRNA sequences (miR-p21A, B, and C) that targeted the 3′ UTR of the p21 mRNA (Fig. 1A). The pre-miRNA sequences were designed as a mimic of the endogenous murine miR-155. The structure consisted of the following three parts: (a) 21-nucleotide core sequence which was completely complementary to target sites in the p21 3′UTR, (b) 19-nucleotide sequence derived from murine miR-155 to form a terminal loop, and (c) the sense target sequence that removed two nucleotides (positions 9 and 10) to form a short internal loop that resulted in more efficient knockdown. The pre-miRNAs were cloned individually into a plasmid vector pcDNA6.2-GW/miR, in which expression of the engineered pre-miRNA was driven by the human cytomegalovirus (CMV) immediate early promoter. In HEK293 cells, the basal level of p21 expression was suppressed by transfection using each miRNA vector individually (Fig. 1B). To determine whether the induction of p21 expression following the activation of endogenous p53 was also suppressed by p21- specific miRNAs, we transfected a mixture of the three miRNA vectors into HCT116 colon cancer cells, in which wild-type p53 is activated by treatment with adriamycin (doxorubicin). The induction of p21 gene expression in response to adriamycin was evident in cells transfected with the control miRNA vector, whereas it was suppressed in cells transfected with the mixture of p21-specific miRNA vectors (Fig. 1C). The parental miRNA plasmid used in these studies is unique in that the Pol II promoter enables cocistronic expression of multiple miRNAs in one primary transcript, which enables the knockdown of multiple target sequences using a single vector. To determine whether the three p21-specific miRNAs functioned in a synergistic manner when expressed from a single vector, we cloned these three miRNAs in a tandem array into pcDNA6.2-GW/miR to generate pcDNA6.2-miR-p21. We then examined whether the combined expression of all three miRNAs was able to suppress the induction of p21 induced by the overexpression of exogenous p53. In HEK293 cells that overexpressed p53, the induction of p21 was suppressed by cotransfection with pcDNA6.2-miR-p21 (Fig. 1D).

In cotransfection experiments involving several vectors, all vectors are not transfected with equal efficiency into each cell. Therefore, in some cells p53 is overexpressed in the absence of p21 suppression, whereas in other cells p21 is suppressed in the absence of exogenous p53 expression. Several reports indicate that p21 suppression enhances cell growth through the derepression of cell cycle arrest (28, 29) and induces tumorigenesis (21–27). To avoid enhancing cancer cell proliferation, p21 suppression and p53 overexpression must be induced simultaneously in each cell.

A unique feature of the parental miRNA vector system using Pol II promoter is that a protein coding sequence is incorporated into the vector such that the miRNA insertion site is in the 3′ UTR of the protein coding sequence. This feature enables cocistronic expression of a protein of interest and an artificial miRNA that suppresses a specific target gene. We inserted the coding region of the p53 gene upstream of cluster of multiple p21-specific miRNAs, or a control miRNA sequence, to generate pcDNA6.2-p53/miR-p21 or pcDNA6.2-p53/miR-control, respectively (Fig. 2A). In HEK293 cells, the transfection of pcDNA6.2-p53/miR-p21 was sufficient to express p53 and fully inhibit the induction of p21 (Fig. 2B). In colon cancer cells SW480 and p53(−/−) HCT116, transfection with pcDNA6.2-p53/miR-p21 resulted in the expression of p53 and the suppression of p21 induction even in the presence of adriamycin (Fig. 2C).

Enhanced induction of apoptosis by a single adenovirus expressing p53 and p21-specific miRNAs in vitro. We constructed several adenoviral vectors based on the p53 and/or miR-p21 expression plasmids. To test whether the adenoviral vectors functioned in a similar manner as the plasmid vectors, we infected p53(−/−) HCT116 cells with an adenovirus that expressed p53 alone (Ad-p53/miR-control), or an adenovirus that expressed both p53 and a cluster of multiple p21-specific miRNAs (Ad-p53/miR-p21). The p53 protein level was increased following infection with Ad-p53/miR-control or Ad-p53/miR-p21 in a dose-dependent manner. Similar to the results of the transfection experiments, however, infection of cells with Ad-p53/miR-p21 resulted in the suppression of p21 induction efficiently (Fig. 3A, Supplementary Fig. S1). The expression of p53 and suppression of p21 induction was also confirmed by immunofluorescence staining (Fig. 3B).

To determine the effect of adenoviral infection on apoptosis, we infected the hepatocellular carcinoma cell lines HLF (carrying mutated p53) and Hep3B (p53-null), and the colorectal carcinoma cell line DLD1 (carrying mutated p53) with adenoviruses. Western blot analysis confirmed that p53 was expressed and the induction of p21 was suppressed in these cells (Fig. 3C). When we examined the cells by flow cytometry, cells that were infected with Ad-p53/miR-p21 had a significantly greater sub-G₁ fraction, which is indicative of apoptotic cell death, as compared with cells infected with Ad-p53/miR-control (Fig. 3D).

Not all cancer cells in which p53 is mutated are sensitive to exogenous p53-mediated apoptosis. In a previous study, we showed that SW480 colorectal cancer cells are relatively resistant to the apoptotic effect of adenovirus-mediated p53 gene transfer (30). To determine whether the combined expression of p53 and p21-specific miRNAs increased the susceptibility of SW480 cells to exogenous p53-induced apoptosis, we measured the sub-G₁ fraction and caspase-3 activity of cells infected with adenovirus in the presence or absence of adriamycin (Fig. 4). The expression of p53 and suppression of p21 induction were confirmed by Western blot (Fig. 4A). Infection of cells with Ad-p53/miR-p21 increased caspase-3 activity (Fig. 4B) and the number of cells in sub-G₁ (Fig. 4C) as compared with cells infected with Ad-p53/miR-control.

Therapeutic effect of adenovirus-mediated expression of p53 and p21-specific miRNAs in vivo. To determine whether the effect of p53 expression and p21 suppression on apoptosis in vitro correlated with a therapeutic effect in vivo, we examined the activity of our novel combination adenoviral vector in a xenograft model of tumorigenesis. SW480 and DLD1 cells were injected s.c. into nude mice. When tumor volume reached a consistent size, adenovirus was injected directly into the tumor at days 0, 1, and 2 (Fig. 5, arrows). Over time, the volume of both SW480- and DLD1-derived tumors that were injected with Ad-p53/miR-p21 was less than tumors injected with Ad-p53/miR-control (Fig. 5). Note also that the injection of Ad-mock/miR-p21 resulted in an increase in tumor volume of SW480-derived tumors compared with the injection of Ad-mock/miR-control. These results indicated that p21 suppression in the absence of p53 overexpression increases the risk of tumor progression in some types of cancer, and suggested that p21 suppression should be simultaneously induced along with p53 expression in tumor cells for effective and safe cancer therapy.

There have been several reports that p21 has an inhibitory effect on p53-induced apoptosis, particularly in combination with chemotherapeutic agents (18–20), which suggests that the simultaneous expression of p53 and suppression of p21 might have a synergistic effect on apoptosis in cancer cells. However, p21 plays an important role in cell cycle arrest and has tumor-suppressive properties (21–29). Thus, in the absence of p53 expression, the suppression of p21 may actually increase the risk of cancer progression.

RNA interference is a widely used technique for the suppression of a specific target gene (31, 32). There are several RNA interference systems, including double-stranded RNA oligonucleotides, i.e. small interfering RNAs that can be transfected directly into cells, and short hairpin RNAs that are expressed using a vector-based expression system (33–35). Short hairpin RNA has a short length structure like G-N18-Loop-N'18-C without 5′-cap and polyA tail. Therefore, the length of transcript must be strictly regulated in short hairpin RNA expression by vector. For this type of transcription, a pol III promoter that can regulate the length of transcript is required. Therefore, Pol II promoters with strong transcriptional activity, such as the CMV promoter, cannot be utilized for the expression of short hairpin RNA. Recently, a vector system for the expression of miRNAs (36, 37), which is a third type of RNA interference system, was developed. miRNA is transcribed as a long mRNA (pri-miRNA) with 5′-cap and polyA tail like a transcript of coding genes. Therefore, pol II promoter is available for the expression of miRNA by vector. In this study, the pre-miRNA sequences were designed as a mimic of the endogenous murine miR-155.

The miRNA expression vector has several advantageous features. First, open reading frame can be incorporated into the vector such that the pre-miRNA insertion site is in the 3′ UTR of the coding sequence. Pol II promoter enables cocistronic expression of a protein of interest and an artificial miRNA engineered to suppress a specific target gene in mammalian cells. Using this system, we expressed p53 and p21-specific miRNAs simultaneously from a single vector. In this manner, we successfully avoided the risk of cancer cell proliferation by p21 suppression in the absence of p53 expression. The second advantage of parental miRNA vector is that multiple miRNA sequences can be inserted in tandem. This feature enables cocistronic expression of multiple miRNAs from a single construct. Actually, some endogenous miRNAs are expressed in clusters in long primary transcripts driven by Pol II promoter. We inserted three different miRNA sequences that targeted the p21 mRNA into a single vector to achieve a synergistic effect. The third advantage of this system is that a miRNA plasmid vector can readily be converted into a recombinant adenoviral vector, thus providing a versatile system for therapeutic applications.

In the current study, we designed an adenoviral vector that expressed p21-specific miRNAs. However, miRNAs against negative regulators of p53 could also be utilized. The MDM2 oncogene is amplified in several cancers. MDM2 is a p53 target gene, and has been shown to function as a negative-feedback regulator of p53 (38–41), mediating the ubiquitination and degradation of the p53 protein (42, 43). Missense mutations of p53 are found in many cancers, and exert a dominant negative effect on exogenous wild-type p53 (44–46). Thus, one could design artificial miRNA sequences that target the 3′ UTR of the p53 mRNA, which would specifically knock down endogenous mutant p53, but have no effect on exogenous wild-type p53 expressed from an expression vector, which would contain only the coding region of p53. One would expect that if additional negative regulators of p53 are knocked down by artificial miRNAs through the use of this vector system, the apoptotic effect of p53 might be enhanced.

We previously reported that the p53 family members p63 and p73 have a significant apoptotic effect in cancer cells (47, 48). The incorporation of p53 family members together with multiple, tandem artificial miRNAs targeting negative regulators of these family members in a single recombinant adenovirus could also be effective.

Vector-mediated gene transfer of p53 is viewed as a potentially effective cancer therapy. Therefore, many clinical trials of adenovirus-mediated p53 gene therapy are ongoing (8). However, gene transfer of p53 does not always have a good therapeutic outcome in all cancers (9–11). Our results suggest that simultaneous expression of p53 and suppression of p21 by single vector should be explored as a way to overcome cellular resistance to p53 gene therapy.

No potential conflicts of interest were disclosed.

We thank Ms. Naomi Yasuda for excellent technical assistance.