Array-based genome-wide RNAi screening to identify shRNAs that enhance p53-related apoptosis in human cancer cells.

p53 transduction is a potentially effective cancer therapy but does not result in a good therapeutic response in all human cancers due to resistance to apoptosis. To discover factors that overcome resistance to p53-induced apoptosis, we attempted to identify RNAi sequences that enhance p53-induced apoptosis. We screened a genome-wide lentiviral shRNA library in liver cancer Huh-7 and pancreatic cancer Panc-1 cells, both of which resist p53-induced apoptosis. After the infection of adenovirus expressing p53 or LacZ as a control, shRNA-treated populations were analyzed by microarray. We identified shRNAs that were significantly decreased in p53-infected cells compared with control cells. Among these shRNAs, shRNA-58335 was markedly decreased in both cancer cell lines tested. shRNA-58335 enhanced p53-related apoptosis in vitro and augmented the inhibitory effect of adenoviral p53 transduction on tumor growth in vivo. Furthermore, the enhanced apoptotic response by shRNA-58335 was also confirmed by treatment with PRIMA-1, which reactivates mutant p53, instead of adenoviral p53 transduction. We found that shRNA-58335 evokes the apoptotic response following p53 transduction or functional restoration of p53 with a small molecule drug in cancer cells resistant to p53-induced apoptosis. The combination of p53 restoration and RNAi-based drugs is expected to be a promising novel cancer therapy.


INTRODUCTION
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]. Furthermore, mutation or deletion of p53 is related to poor prognosis and resistance to chemotherapy and radiation [2]. Therefore, vector-mediated gene transfer of p53 has been viewed as a potentially effective cancer therapy. However, gene transfer of p53 does not always result in good therapeutic outcomes in all cancers [3,4]. The inactivation of apoptotic machinery is considered as one of the reasons that p53 transduction fails to induce apoptosis in cancers. In such cancers, p53 status does not necessarily correlate with the responsiveness to chemotherapy [5].
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 not only coding genes but also non-coding RNAs, including microRNAs (miRNAs) and large intergenic noncoding RNAs (lincRNAs), many of which are involved in DNA repair, cell cycle arrest and apoptosis [6,7]. Although the precise mechanisms that regulate whether a cell undergoes cell cycle arrest or apoptosis are unclear, cell cycle arrest induced by p53 is mediated primarily through the induction of the expression of p21 [8], which is involved in restoring genomic integrity by functioning in an anti-apoptotic manner [9]. Thus, the suppression of p53-induced expression of p21 may lead to the preferential induction of apoptosis rather than cell cycle arrest following p53 activation. In our previous study, we constructed a replication-deficient recombinant adenovirus that encoded co-cistronic p53 and artificial microRNAs targeting endogenous p21, resulting in simultaneous p53 expression and the suppression of p21 induction. In colorectal and hepatocellular carcinoma cells, infection with the adenovirus vector augmented apoptosis and suppressed tumor growth compared with an adenovirus that expressed p53 alone in vitro and in vivo [10]. Thus, the combination of p53 and RNAi was shown to be a potential novel cancer therapy.
To discover factors that modulate p53 pathways, we attempted to identify RNAi sequences that enhance p53induced apoptosis. In this study, we screened a genomewide lentiviral shRNA library in cancer cells resistant to p53-related apoptosis. We identified specific shRNAs that were significantly decreased in p53-infected cells but not in control cells. These results indicated that specific shRNAs evoked the apoptotic response following p53 transduction or functional restoration of p53 in cancer cells resistant to p53-induced apoptosis, suggesting that the combination of p53 restoration and treatment with specific shRNAs may be an effective cancer therapy.

Genome-wide shRNA screening in p53-resistant cancer cells
To identify shRNAs that can relieve resistance to p53-induced apoptosis in cancer cells, we screened a genome-wide lentiviral shRNA library in liver cancer Huh-7 cells and pancreatic cancer Panc-1 cells, as summarized in Figure 1A and briefly described below. The Huh-7 and Panc-1 cell lines have mutated p53 and a weak apoptotic response following wild-type p53 transduction. The human shRNA library, comprising approximately 200,000 shRNA sequences targeting 47,400 human mRNA transcripts, was packaged into lentivirus particles and pooled. First, we infected the cell lines with the pooled lentiviral shRNA library. Seventy-two hours after lentiviral infection, the cells were infected with adenovirus expressing p53. Forty-eight hours after adenoviral infection, total mRNA was extracted, and mRNA, including shRNA sequences derived from lentiviral vectors, was amplified by PCR. In this shRNA library, each lentiviral vector shRNA sequence corresponds to a probe sequence found on an Affymetrix microarray (GeneChip). Therefore, the shRNA populations can be quantified by cDNA microarray to determine the expression levels of each shRNA; as a reference, the shRNA population in each cell line infected with adenovirus expressing LacZ was also analyzed.

Identification of shRNA-decreased populations in p53-transduced cells
We identified the shRNAs that were significantly decreased in p53-infected cells compared with control cells (Fig. 2A). Several shRNAs evoked the apoptotic response following p53 transduction in cancer cells resistant to p53-induced apoptosis. In Huh-7 cells, 547 shRNAs were decreased more than 4-fold in cells with p53 transduction compared with control cells. In Panc-1 cells, 1418 shRNA were decreased more than 16-fold in cells with p53 transduction compared with control cells ( Fig. 2A, areas encircled by dashed line). Furthermore, 28 shRNAs were commonly decreased in both cell lines (Fig. 2B, Supplementary Table S1). Among these shRNAs, shRNA-58335 in the presence of p53 was markedly decreased (66-fold in Huh-7 cells and 436-fold in Panc-1 cells). Therefore, we further analyzed shRNA-58335.

Enhancement of p53-induced apoptosis by shRNA-58335
Next, we stably infected lentivirus expressing shRNA-58335 or a control sequence into Huh-7 cells and quantified p53-induced apoptosis by evaluating the sub-G 1 population. In these cells, p53 transduction induced a strong apoptotic response in shRNA-58335infected cells compared with control cells (Fig. 3A, B). Additionally, treatment of adriamycin significantly   enhanced the apoptotic response (Fig. 3A, B). Increased caspase-3 cleavage, which serves as another indicator of apoptosis, was also observed by western blotting (Fig.  3C). In colorectal cancer SW480 cells, which also have mutated p53 and a weak apoptotic response following p53 transduction, shRNA-58335 sensitized the p53-induced apoptotic response (Fig. 3D). These results indicated that shRNA-58335 improved the apoptotic response after p53 transduction in other cancer cell lines.

Inhibition of tumor growth by p53 transduction with shRNA-58335 in vivo
To confirm whether the improved p53-induced apoptotic response in vitro was correlated with a therapeutic effect in vivo, we examined the tumor suppressive effect of p53 transduction in a xenograft model of tumorigenesis. shRNA-infected Huh-7 cells were injected s.c. into nude mice. When tumor volume reached a consistent size, adenovirus expressing p53 or LacZ was injected directly into the tumor at days 0, 1 and 2 ( Fig.  4, arrow). In control shRNA-infected Huh-7 cells, the injection of Ad-p53 did not affect tumor volume (Fig. 4, left). Conversely, in shRNA-58335-infected Huh-7 cells, tumor volume was significantly decreased by injection of Ad-p53 compared with Ad-LacZ (Fig. 4, right). These results suggest that combined treatment with p53 and a specific shRNA may be an effective cancer therapy in vivo.

Altered gene expression pattern by shRNA-58335 in p53-transduced cells
According to the NCBI BLAST search (http://blast. ncbi.nlm.nih.gov/), the seed sequence of shRNA-58335 is perfectly matched only to AP2γ (TFAP2C) among all human transcripts in the RefSeq database. AP2γ belongs to the AP-2 family of transcription factors and overexpressed in human breast cancers and the overexpression related with reduced patient survival [15]. A recent report demonstrated that AP2γ is a key transcriptional regulator for maintaining the luminal phenotype in human breast carcinoma [16]. AP2γ is also a direct transcriptional target of p53 [17]. We confirmed a statistically significant but slight decrease in AP2γ mRNA levels by shRNA-58335 (Supplementary Figure S1). However, we were unable to detect AP2γ protein using Western blotting. Therefore, it remains controversial whether shRNA-58335 works through the suppression of AP2γ.
To explore the effect of shRNA-58335 on p53induced apoptosis, we evaluated gene expression by cDNA microarray analysis in p53-transduced Huh-7 cells. We found that the expression of 350 genes was decreased more than 2-fold in shRNA-58335-infected Huh-7 cells compared with control-shRNA infected cells (Supplementary Table S2). Interestingly, gene ontology analysis revealed that these genes were significantly associated with cell cycle progression and cell division ( Table 1). These results indicate that shRNA-58335 evoked an apoptotic response following p53 transduction by modulating the gene expression associated with the process of cell division.

DISCUSSION
We demonstrated shRNAs that enhanced p53related apoptosis by screening a genome-wide shRNA library. Conventional shRNA library screens have shown shRNAs that resist p53-induced apoptosis through the presence of p53-resistant cellular colonies [18]. However, this method is not able to identify shRNAs that enhance p53-related apoptosis because cells that include such shRNAs are lost. Recently, array-based shRNA screening was developed, allowing the direct quantification of the amount of each shRNA. We applied this new method here and successfully identified shRNAs that enhanced p53related apoptosis.
Although the shRNA library used in this study included approximately 200,000 shRNA sequences targeting 47,400 human mRNA transcripts, the knockdown efficacy of each shRNA has not been validated. Therefore, the screened shRNAs, the sequences of which match specific transcripts, are not necessarily able to effectively knockdown their target genes. In fact, only weak suppression of the TFAP2C gene by shRNA-58335 was confirmed, although the seed sequence of shRNA-58335 completely matched the AP2γ mRNA sequence (Supplementary Figure 1). Furthermore, populations of other shRNAs that target AP2γ mRNA (shRNA-58334, 58336 and 58337) were not significantly decreased in p53-infected Huh-7 or Panc-1 cells compared with LacZinfected control cells. However, shRNA-58335 effectively enhanced the p53-induced apoptotic response not only by p53 transduction (Fig. 3, 4) but also by the functional restoration of p53 with small molecule PRIMA-1 (Fig. 5). These results may suggest that shRNA-58335 operates as a molecular modulator in p53-related apoptosis through other unknown mechanisms rather than through the knockdown of the TFAP2C gene. Our microarray analysis revealed that shRNA-58335 with p53 transduction modulates gene expression associated with the process of cell division (Table 1), whereas no genes were significantly decreased more than twofold in shRNA-58335-infected cells without p53 transduction compared with control-shRNA-infected cells. These results also support our conclusion that shRNA-58335 does not knockdown any specific genes. However, the molecular mechanism of shRNA-58335-mediated enhancement of p53-related apoptosis is unclear. Further functional analysis should be performed in future studies.
We demonstrated that the apoptotic response induced by adenoviral p53 transduction was enhanced by shRNA-58335 in cancer cells resistant to p53-induced apoptosis (Fig. 3, 4). However, the clinical application of adenoviral vectors is presently limited because adenoviral vectors must be administered locally [4]. Therefore, we replaced adenoviral p53 transduction with PRIMA-1 treatment to reactivate mutant p53 [13] and successfully confirmed the enhancement of an apoptotic response by PRIMA-1 with shRNA-58335 treatment (Fig. 5). PRIMA-1 can be intravenously administered in human clinical studies [14]. Furthermore, non-viral delivery systems, such as nanoparticles, for RNAi-based drug treatments are also being developed and are in various stages of clinical investigation [19,20].
Most of clinically used chemotherapeutic agents cause DNA damage and target dividing cells in S or M phase. Therefore, chemotherapeutic agents not only kill cancer cells but also injury normal dividing cells, resulting in adverse effects of cancer treatment. In order to protect normal cells from chemotherapeutic agents, pretreatment of drugs which induce cell cycle arrest in normal cells, socalled "cyclotherapy" is considered as an effective therapy [21,22]. Recent reports suggest that the pretreatment of small-molecule p53 activators such as nutlin-3 protect normal cells from chemotherapeutic agents [23,24]. The combination of p53 restoration, RNAi-based drugs and p53-based cyclotherapy is expected to be a promising novel cancer therapy in the near future. www.impactjournals.com/oncotarget

Cell culture
The hepatic cancer Huh-7 cell line, the pancreatic cancer Panc-1 cell line and the colorectal cancer SW480 cell line were purchased from the American Type Culture Collection. Huh-7, Panc-1 and SW480 cells were cultured in DMEM, RPMI-1640 and Leibovitz L-15 medium with 10% FCS, respectively. The construction, purification and infection of replication-deficient recombinant adenoviruses containing p53 (Ad-p53) or the bacterial lacZ gene (Ad-LacZ) have been described previously [11].
shRNA library screening shRNA library screening was performed according to the manufacturer's protocol. Briefly, 3×10 6 Huh-7 and Panc-1 cells were plated on a 10-cm plate. Twenty-four hours after plating, the cells were infected with lentivirus particles from the GeneNet siRNA Library (System Biosciences). Seventy-two hours after lentiviral infection, Ad-p53 or Ad-LacZ was infected into these cells at an MOI of 50. Forty-eight hours after adenoviral infection, total RNA was extracted with TRIZol (Invitrogen). cDNA was synthesized from extracted RNAs. siRNA inserts were amplified from cDNA using two rounds of PCR with biotinylated primers. PCR products were treated with lambda exonuclease to remove the sense non-biotinylated strand. Ten micrograms of biotinylated samples was hybridized to an Affymetrix GenChip microarray (HG-U133+ 2.0). The obtained data were analyzed with GeneNet siRNA Library Data Analysis Software (System Biosciences).

Antibodies and reagents
Adriamycin and VP-16 were purchased from Sigma. PRIMA-1 was purchased from Cayman Chemical. The anti-p53 (DO-1), anti-MDM2 (SMP14) and anti-p21 (F-5) mouse antibodies were purchased from Santa Cruz Biotechnology. The anti-actin mouse antibody was purchased from Millipore, and the anti-caspase-3 (8G10) rabbit antibody was purchased from Cell Signaling Technology.

Flow cytometry
Cells (1×10 6 ) were plated on 6-well plates. After infection or treatment, the 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). The samples were analyzed on a FACSCalibur flow cytometer (BD Bioscience). The experiments were repeated at least three times, and 50,000 events were analyzed for each sample. Data were analyzed using FlowJo software (Tree Star).

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 effects of treating established tumors, 24 female BALB/c nude mice were injected subcutaneously (s.c.) with 2×10 6 Huh-7 cells infected with lentiviral shRNA-58335 or control vector into both flanks. When the tumor size reached 100 mm 3 , the mice received a direct intratumoral injection of 1×10 9 p.f.u. (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 the mice was monitored for up to 2 weeks. The tumor volume was calculated using the equation V (mm 3 ) = a×b 2 /2, where "a" represents the largest dimension and "b" is the perpendicular diameter. www.impactjournals.com/oncotarget

Microarray analysis
Huh-7 cells stably expressing shRNA-58335 or control shRNA were infected with adenovirus expressing p53 (Ad-p53) or LacZ (Ad-LacZ) as control at 200 moi. Seventy-two hours after infection, total RNA was extracted by RNeasy mini kit (QIAGEN) according to manufacturer's protocol. Total RNA was labeled with Cy3, hybridized to a microarray (Agilent SurePrint G3 Human GE) and scanned by Agilent SureScan according to the manufacturer's protocols. The obtained data were normalized using Limma (R package). The microarray data were deposited in the NCBI Gene Expression Omnibus (GEO, accession number: GSE56156).