DU145 and LNCaP cell line was obtained from American Type Culture Collection (Manassas, VA) and cultured as per the supplier’s protocol. C4-2 a generous gift from Dr. John Issacs Laboratory (Johns Hopkins University) were cultured in RPMI (Cellgro Mediatech, Herndon, VA) supplemented with 10% FBS, and maintained at passages 50 to 60. Virus packaging cell line DPL-S11 described previously [9] was maintained in DMEM (Cellgro Mediatech, Herndon, VA) with 5% FBS and 200 μg/ml G418. The 293HEK cell line (Quantum Biotech) was used for viral titration and was maintained in DMEM with 10% FBS. All media were supplemented with Ciprofloxacin Hydrochloride 5 μg/ml (US Biological, Swampscott, MA) and Gentamicin 50 μg/ml (Quality Biological Inc., Gaithersburg, MD). All cells were maintained at 37˚C in an atmosphere containing 5% CO₂. All restriction enzymes used are from NEB (New England Biolabs, MA). Primary mouse monoclonal antibodies for androgen receptor (AR 441) and E1A (M73) were bought from Santa Cruz Biotech (USA). The majority of the chemicals and reagents used in this study were from Sigma-Aldrich unless otherwise specified.

Viruses (Ad5-PSE/PBN-E1A, Ad5-PSE/PBN-E1A-miR- 34c, Ad5-Delta-E1B, Ad5-CMV-miR-34c and Ad5-CMVhTR) were generated using Ad-Easy system (Stratagene, CA) through recombination of linearized shuttle plasmids in DPL-S11 cells. Briefly shuttle plasmids RpsToadPSE/PBN-E1A, pAd-TrackCMV-hTR, or pAdTrack-CMVmiR34c that carries prostate-specific enhancer and rat probasin promoter driving E1A, and CMV driving miR- 34c or hTR (control miRNA) were linearized with pme1 restriction endonucleases. After gel purification the linearized vectors were separately transformed into the electro-competent AdEasier-1 (BJ5183-AD-1) cells for homologous recombination. The desired clones (pAd5- PSE/PBN-E1A or pAd5-TrackCMV-miRNA34c or pAd5- TrackCMV-hTR) after screening were transformed into DH10B cells for large-scale DNA amplification. For viral propagation the recombinant plasmids were linearized with EcoR1 and transfected into adenovirus packaging cell line DPL-S11 to generate recombinant adenoviruses.

Similarly for generating a prostate-specific conditionally replicating adenovirus that carries CMV driving miR-34c after the fiber region, an adenoviral pFEX system [17] was used to recombine a shuttle plasmid RpsToad-PSE/PBN-E1A using homologous recombination between the two LTRs of the adenovirus in BJ5183 bacterial competent cells. After getting the desired recombinant, the fiber-less viral DNA that carried prostate-specific enhancer and rat probasin promoter (pAd5-FexPSE/PBN-E1A) was linearized with Pac1 and transfected into the fiber expressing FBJ cell line (a generous gift of Dr David Johns, Johns Hopkins University) to generate prostate-specific pseudo-typed Ad5-FexPSE/PBN-E1A virus. The Ad5-PSE/PBN-E1A miR-34c that carries miR- 34c under the control of CMV promoter was produced by the transfection and infection experiment using a transient vector RpUC-WT-FIB CMV-miR-34c and a pseudo-typed fiber-less adenovirus Ad5-FexPSE/PBN-E1A in 293Cre57 cells. Once Ad5-PSE/PBN-E1A-miR-34c was made and purified, RCA (replication competent adenovirus) was ruled out using PCR amplification with a set of primers E1A 5’-CGTTCCGGGTCAAAGTTGG 3’ E1A 5’-CCTCCGTGGCAGATAATATGTC-3’ spanning the wild type E1A promoter and E1A gene.

Cells were washed with 1× PBS and re-suspended with five volumes of cold lysis buffer (50 mM Tris-HCl, pH 7.5, 250 mM NaCl, 5 mM EDTA, 50 mM NaF, 0.5% NP-40) supplemented with protease inhibitor cocktail (Roche, Indianapolis, IN). The cell lysate was incubated on ice for 30 min and then centrifuged for 10 min at 4˚C. Equal amounts of proteins were separated by SDS-PAGE, and the resolved proteins were then transferred to a nitrocellulose membrane. After blocking with 5% nonfat milk in TBST overnight at 4˚C, the blot was incubated with primary antibody at 1 h at room temperature. The membrane was then probed with HRP-conjugated secondary antibody for 1 h and developed (ECL-Plus system, Amersham Pharmacia, Piscataway, NJ) using the manufacturer’s protocol.

We have applied a reporter system to quantify adenoviral replication in real time by linking green fluorescent protein (GFP) expression to the viral major late gene fiber through an Internal Ribosome Entry Site (IRES). In brief, this replication-deficient reporter virus, FFIG expresses GFP in a replication-dependent manner when co-infected with a replicating adenovirus. For reporter experiments, cells were plated into 96-well plates at 1 × 10⁴ per well overnight, and co-infected the next day with replication competent adenoviruses at an MOI of 2 PFU per cell and the reporter virus FFIG at an MOI of 10 - 20 PFU per cell. The GFP fluorescence signals were measured at the indicated time points using a multi-plate fluorometer (FLUOstar Optima Microplate Reader), using an excitation wavelength of 485 ± 20 nm and an emission wavelength of 530 ± 25 nm. Background fluorescence was measured in cells, which were infected with FFIG virus alone. GFP data are plotted as fold GFP induction relative to the mock-infected cells. All samples were performed in either triplicate or quadruplicate with error bars representing the s.e.m.

The metabolic viability of the cells was monitored using an MTS assay kit (CellTiter 96) from Promega (Madison, WI). A mixture of MTS and phenazine methosulfate (an electron-coupling reagent; final concentrations, 333 μg·ml⁻¹ and 25 μM) was added, and cells were incubated for 1 hour at 37˚C. Formazan formed from the reduction of MTS was quantified by measurement of absorbance of the medium at 490 nm using a microplate reader. Data obtained from the experiments were normalized to the background signals and plotted as percent cell kill.

All experiments were done in triplicate or quadruplicate and plotted with s.e.m. All statistical analysis was performed using Excel running on IBM-PC compatible computer on the Windows XP-operating system. Statistical comparisons for paired data were analyzed by the Student’s t-test for the in vitro assays. Statistical significance was defined as P < 0.05.

In order to identify candidate miRNAs that have the potential to suppress the proliferation of prostate cancer cells and an additive effect in combination oncolytic therapy, we performed a miRNA library screen using miRIDIAN microRNA 810 Mimics (Thermo Fisher Scientific, USA). Prostate cancer C4-2 cells plated at the density of 4 × 10⁴ cells were transfected in duplicate with individual miRNAs. Twenty four hours post transfection cells were infected with CN702 (a wild type adenovirus) using 0.2 MOI. MTS assay was performed at 72 h in duplicate to evaluate percent cell kill by each miRNA in the presence of virus. Plates were normalized to mock miRNA tranfected and wild type CN702 infected C4-2 cells. All data were plotted as a waterfall plot to show any cell survival or death when treated with individual miRNA in combination with CN702 virus (Figure 1(a)). Using a cutoff value of at least 20% cell kill above the wild type (CN702) virus and mock transfected miRNA cells, a total of twelve individual miRNAs were identified (Figure 1(b)). Among these twelve therapeutic miRNAs, we specifically choose miR-34c for consequent studies as it has been shown to function downstream in the P53 apoptotic pathway and has also been shown to suppress the proliferative growth of prostate cancer cells. In order to validate that the cell kill was due to miR-34c, we performed the colony formation assay independent of CN702 virus. Briefly equal number of C4-2 cells transfected with miR-34c or control hTR plasmid were mixed with DMEM containing 15% fetal bovine serum and 0.5% agar and plated on six-well plates for 14 days. As shown in Figure 1(c), cells that were transfected with miR-34c have approximately 2.5 fold lower number of colonies than that of controls (Figure 1(c)). These data confirmed the previously known anti-proliferative role of miR-34c in prostate cancer [13]. To further investigate whether overexpression of miR-34c in combination with a non-replicative adenovirus Ad5-Delta E1B causes programmed cell death, C4-2 cells transfected with miR-34c (3 μg/well), Ad5-Delta E1B alone, or in combination with Ad5-Delta E1B virus were infected with 1 MOI. Cells were harvested 48 h post infection (PI) and were subjected to western blot analysis for caspase-3 levels. As shown in the Figure 1(d), cells that were both transfected and infected had lower levels of caspase-3 protein levels compare to the miR-34c alone or virus alone. These data support the anti-proliferative role of miR-34c

in C4-2 cells.

To demonstrate that the therapeutic role of miR-34c was not limited to androgen dependent cells but also affects androgen resistant cells, we constructed a replication incompetent adenovirus that expresses miR-34c and a control hTR virus using CMV promoter. DU145 cells that do not express AR were plated in 96 well plates at the density of 2 × 10⁴ cells. Equal MOI of each non-replicating adenovirus (2 MOI of Ad5CMV-miR-34c or Ad5CMV-hTR) were used to infect cells for five days followed by MTS cell cytotoxicity assay. As shown in Figure 2(a) there was no difference among the two treatments by the end of 24 h, however, by 120 h a significant cell kill was only observed in the Ad5CMVmiR34c infected DU145 cells compared to control Ad5- CMV-hTR infected cells. To further validate these observations, we performed a crystal violet assay using these viruses in DU145 cells. Briefly, 7 × 10⁵ cells plated into 6 well plates were infected with different MOIs (0.5, 5 and 20 MOI) of either Ad5CMV-miR-34c or control Ad5-CMV hTR virus for 10 days. By the end of the experiment wells were stained with crystal violet to evaluate cell cytotoxicity. As shown in Figure 2(b), wells that were infected with miR-34c virus have fewer cells compared to hTR virus (Figure 2(b)). These data further confirm the anti-proliferative role of miR-34c in androgen resistant prostate cancer cell line.

Although tissue specificity that is achieved through AR dependent promoters reduces toxicity by restricting replication to prostate tissue this poses a problem for those

patients on androgen ablation therapy or in tumors that have loss AR expression. Thus an alternative method is needed to rescue the use of prostate specific CRAds that are dependent on androgen for their replication. Furthermore, using tissue specific promoters and enhancer (PSA or PBN and PSE) in the viral backbone utilize much of the genomic space that can be accommodated within the viral capsid. To overcome these problems, we investigated if a single therapeutic miRNA (miR-34c) that has shown therapeutic effect in our in-vitro cell line models will rescue the viral oncolysis in AR negative cells without compromising replication potential in androgen dependent cells. A prostate specific CRAd (Ad5- PSE/PBN-E1A-miR34c) that expresses miR-34c from CMV promoter was engineered using the pFEX system. To demonstrate that miR-34c expressing CRAd retains its replication potency we performed the FFIG reporter assay system, which places a GFP gene under the control of the major late promoter (MLP), allowing us the ability to measure viral replication in real time in a non-invasive high throughput assay. C4-2 cells plated at a density of 5 × 10⁴ cells per well were infected with MOI of 2 with Ad5-PSE/PBN-E1A-miR34c in the presence of a reporter virus FFIG (MOI 5) in complete medium supplemented with 5 nM of R1881. As shown in Figure 3(a), the replication kinetics, measured by GFP signal, was observed as early as 36 h PI, suggesting that Ad5-PSE/ PBN-E1A-miR34C retains it replicative potency. Similarly, to assess the oncolytic potential of Ad5-PSE/PBNE1A-miR34c, MTS assay was performed by 96 h PI in C4-2 cells, as shown in Figure 3(b) cells that were infected with Ad5-PSE/PBN-E1A-miR-34c had more than 80 percent cell death compared to the FFIG reporter virus.

To further evaluate whether the addition of miRNA (miR-34c) in androgen dependent prostate specific CRAd can rescue the oncolysis of virus in AR negative DU145 cells, we took LNCaP (AR positive) and DU145 cells (AR negative) and stably expressed GFP or RFP with the help of lentiviral vectors. Stably selected LNCaP-GFP or DU145-RFP cells mixed in 3:1 ratio (LNCaP and DU145) were infected with 0.5 MOI of Ad5-PSE/PBN-E1A or Ad5-PSE/PBN-E1A-miR-34c or a non replicating control Ad5-PSE/PBN-Luc virus. At the end of 96 h viral oncolysis was measured using florescent microscopy. As shown in the Figure 3(c), the prostate specific CRAd that does not express miR-34c was only able to kill the green AR positive LNCaP-GFP cells but was unable to cause any significant toxicity in the DU145-RFP cells when compared to the control Ad5-PSE/PBN-Luc virus. However, Ad5-PSE/PBN-E1A-miR34c that was engineered to express the miR-34c was able to cause significant oncolysis of both cell types providing evidence of its therapeutic potential in a heterogeneous androgen dependent and independent in-vitro prostate cancer cell model.