Adenovirus-mediated Delivery of p16 to p16-deficient Human Bladder Cancer Cells Confers Chemoresistance to Cisplatin and Paclitaxel1

We have previously established the efficacy of adenoviral gene delivery vectors for the treatment of bladder carcinoma in vivo. In the present work, we developed a gene therapy strategy for bladder cancer based on the replacement of the tumor suppressor pi6, which is known to be mutated or deleted in a variety of human tumors, including those derived from the bladder. Previous reports have demonstrated that reconstitution of p16 has marked effects on the proliferative capacity of tumor cell lines both in vitro and in vivo, and that p16 expression causes resistance to some chemotherapeutic agents. In the present study, we describe the construction of the recombinant adenovirus Adpl6, expressing thepl6 gene, to evaluate the effects of transient p16 replacement in the context of bladder carcinoma. To identify candidate target cell lines, we screened a panel of bladder cancer cell lines for p16 and retinoblastoma (RB) protein expression. We demonstrate that Adpl6 can mediate highefficiency p16 replacement to the p16-negative cell lines EJ and UMUC-3. In addition, the reconstitution of p16 to the highly transducible p16-negative, RB-positive bladder cancer cell line EJ caused a profound inhibition of cell proliferation mediated by arrest in the G1 phase of the cell cycle. In contrast, the p16-positive, RB-negative cell line J82 was unaffected by this treatment. However, when adenovirally mediated p!6 replacement was combined with the chemotherapeutics cisplatin and paclitaxel, a marked chemoresisReceived 1/28/97: revised 8/5/97: accepted 8/I 1/97. The costs of publication of this article were defrayed in part by the payment of page charges. This article must thereftre be hereby marked adtertise,nt’,it in accordance with I S U.S.C. Section 1 734 solely to indicate this fact. This work was supported in part by a Veterans Administration Research Advisory Group grant (to R. A. K.) and by NIH Grant 1 ROI CA 68245-OIAI (to D. T. C.). 2 To whom requests for reprints should be addressed, at Gene Therapy Program. University of Alabama at Birmingham. 1824 6th Avenue South, Room WTI 620. Birmingham. AL 35294. Phone: (205) 9348627: Fax: (205) 975-7476. tance was observed in genetically corrected cells. This work has implications for future gene therapy strategies based on p16 replacement. INTRODUCTION Cancer is increasingly recognized as a genetic disease that results from a defined set of lesions that activate or inactivate normal cellular proteins. As such. proteins that are involved in the control of the cell cycle and the integrity of the genome. such as p53. cyclins. and cdks.3 often show altered expression patterns in human tumors and tumor-derived cell lines (reviewed in Ref. 1 ). In this regard. the recently identified tutnor suppressor p16 (also known as MTSI. CDK4I, or CDKN2) is central to the control of cell cycle progression through its inhibitory interaction with cdks. and. indirectly. with RB (2). Evidence for the involvement of p 1 6 in the progression of human cancer includes the fact that this protein has been found to be mutated or deleted in tumors derived from a variety of tissues, including pancreas. skin, lung, brain, and bladder (3-5). In the latter case. deletions on chromosome 9 represent common events in bladder cancer (6-8), and it has been shown that one target of these deletions is the tumor suppressor p16 (3-S. 9). Initial reports demonstrated loss of this protein in human bladder cancer cell lines but failed to determine the significance of this deletion in primary tumors. We have shown recently that 13 of 20 primary bladder tumors contained no detectable p16 by immunohistochemistry ( 10). In addition, Cairns et a!. ( I 1 ) reported total homozygous deletion of this chromosotnal region in 7 1 % of primary bladder cancers. a higher percentage than any other tumor evaluated. These results affirm the role that p16 plays in the malignant phenotype and the progression of human bladder cancer and provide a rational basis for anticancer gene therapy strategies that involve the replacement of this protein. Because the mutation or deletion of tumor suppressor genes provides a conceptual framework in which to attempt anticancer gene therapy strategies, a variety of tumor suppressor replacement protocols have been developed for the treatment of human malignancies. To date. most of these have involved the replacement of the tumor suppressor p53 ( 1 2). which represents one of the genes most commonly lost as a consequence of tumor progression (13). In this regard. Roth’s group demonstrated that intratumoral retrovirally-mediated p53 gene delivery could greatly inhibit or abrogate established lung tumors in animal I The abbreviations used are: cdk, cyclin-dependent kinase: RB. retinoblastoma: EMEM. Eagle’s MEM: MOl. multiplicity of infection: MTS. 3-(4,5dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl )-2-(4-sulfaphenyl)-2Htetrazoliuni: PMS, phenazine methosulfate: CMV. cytomegalovirus. on April 7, 2017. © 1997 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from 2416 Adpl6-induced Chemoresistance in Bladder Cancer Cells models of human disease (14). In fact, p53-based antitumor therapy has been approved for clinical trial by the Food and Drug Administration and is currently being evaluated in human subjects (reviewed in Ref. 15). Thus, genetic correction of tumor cells represents an established treatment paradigm to approach neoplastic disease by gene therapy. In the context of these mutation-compensation gene therapy strategies, it has been noted that, in certain disease models, genetic correction of cellular neoplastic lesions restores or enhances tumor cell chemosensitivity (16). In this regard, p.53 loss is clearly correlated with resistance to standard therapies (17, 18), and replacement of this gene has been shown to enhance lung cancer sensitivity to the chemotherapeutic agent cisplatin (19). Delivery of the adenovirus E1A protein can induce a similar chemosensitivity in cell lines derived from a variety of tissues (20). In addition, our group has shown that genetic abrogation of the erbB-2 oncoprotein can sensitize tumor cells to cisplatin in both in vitro and in vivo therapy models (21). Thus, it appears that in many instances, genetic correction of the tumor cell by restoration of tumor cell chemosensitivity may allow the implementation of combined modality therapies with gene therapy as their basis. Previous experiments have demonstrated that the replacement of the p16 protein by DNA transfection methodologies can inhibit the progression of the cell cycle and thus inhibit the growth of human cancer cell lines, including those derived from the bladder (22-24). To further investigate the involvement of p16 in carcinoma of the bladder and to evaluate the utility of p16-based antitumor gene therapy for this disease, we constructed a recombinant adenovirus expressing the p16 gene. This vector efficiently delivers this gene to target cell lines, resulting in a marked decrease in the proliferative capacity of p16negative bladder cancers. However, when used in conjunction with the chemotherapeutic agents cisplatin and paclitaxel, the cell cycle arrest induced by p16 expression confers a marked chemoresistance. These studies indicate that gene therapy strategies based on the replacement of the pitS gene will likely reduce the efficacy of standard treatment modalities, greatly limiting their utility in the clinical setting. MATERIALS AND METHODS Cell Lines. The human bladder cancer cell lines J82, TCC, and UMUC-3 were obtained from the American Type Culture Collection (Rockville, MD). The human bladder cancer cell line EJ was a kind gift from Kevin Scanlon (Berlex, Richmond, CA). The non-small cell lung cancer cell line H2009, which is p16 positive and RB negative, has been described previously (25, 26). The mesothelioma cell lines H246l and H2373, which are p16 negative and RB positive, have been described previously (27). The small cell lung cancer cell line H209, which contains a missense RB mutation resulting in a nonphosphorylatable form of this protein, has been described previously (27). The low-passage adenoviral transcomplementing cell line 293 was a kind gift from Dr. Frank Graham (McMaster University, Hamilton, Ontario, Canada). Cell lines were maintained in either complete EMEM (El, J82, TCC, and UMUC-3), complete RPMI (H2009, H246l, H2373, and H209), or complete DMEMIF12 (293). These consisted of appropriate medium formulations of EMEM, RPMI, or DMEM/F12 (Mediatech, Washington, DC) supplemented with 10% FCS (FCS; HyClone Laboratories, Inc., Logan, UT), L-glutamine (300 p.g/ ml; Mediatech), penicillin (100 lU/mI; Mediatech), and streptomycin (25 p.g/ml; Mediatech). In addition, complete EMEM contained nonessential amino acids ( 1 X ; Sigma Chemical Co., St. Louis, MO) and sodium pyruvate (1 mM; Mediatech). Adenoviral Vectors. The p 16-expressing adenovirus Adpl6 was constructed and validated according to standard techniques (28). Briefly, the p16 expression plasmid pT7-p16 was a kind gift from Greg Otterson (National Cancer Institute, Bethesda, MD) and has been described previously (27). This plasmid encodes a truncated form of p 16. This truncated form, which lacks the first five amino acids of the wild-type construct, has been shown to retain the functional properties of the fulllength wild-type protein (24). A 451-bp EcoRI fragment contaming the p16 open reading frame was cloned into the adenoviral shuttle vector pACCMVPARLS(+). The resulting construct was cotransfected along with the adenoviral packaging plasmid pJM 17 into 293 cells. Plaques were then isolated and tested for the presence of the p16 cDNA through the use of the PCR with p16-specific primers. Three rounds of plaque purification were performed to eliminate contaminating wild-type adenovirus. The recombinant adenovirus AdCMVLacZ, cxpressing the Escherichia coli 3-ga1actosidase gene, was a kind gift from Dr. De-Chu Tang (University of Alabama at Birmingham) and has been described previously (29). Evaluation of p16 and RB Expression in Human Bladder Carcinoma Cell Lines. Cell lines were grown to 80% confluence in complete medium in 100-mm dishes and then lysed by adding 1 ml of radioimmunoprecipitation assay buffer [150 m i NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 20 msi EDTA, and 50 msi Tris (pH 7.4)] to the dishes and incubating them for 30 mm on ice. The cells and lysate were then harvested using cell scrapers and cleared by centrifugation at 14,000 rpm at 4#{176}C for 15 mm. The supernatants were retained, and protein concentration was analyzed by a modified Bradford assay (Bio-Rad, Hercules, CA). Total protein (100 .Lg) was separated on either a 15% SDS-PAGE gel (p16) or a 10% SDS-PAGE gel (RB) and then transferred to a polyvinylidene difluoride membrane (Bio-Rad). The blot was probed with a primary antibody against p16 (PharMingen, San Diego, CA) or RB (PharMingen) at a dilution of 1 : 1000 and a horseradish peroxidase-conjugated secondary antibody (Jackson Laboratories, Bar Harbor, ME) at a dilution of 1 :5000 and visualized using the Renaissance reagent system (DuPont, Boston, MA). For studies of RB phosphorylation, Western blots were performed as described previously (30). Briefly, whole-cell lysates were prepared from 5 X 106 cells in 1 ml of lysis buffer [50 m i Tris-HC1 (pH 7.5), 250 mM NaC1, S msi EDTA, 0.1% NP4O, 50 mM NaF, and 1 m t phenylmethylsulfonyl fluoride], and 75-150 p g of total protein were electroblotted onto nitrocellulose following separation on a 7.5% SDS-PAGE gel. Nitrocellulose filters were incubated overnight at a 1 :200 dilution of the anti-RB monoclonal antibody (PharMingen). The filter was then exposed to a rabbit antimouse secondary antibody (PharMingen), followed by 2.5 X 10 cpm ‘251-labeled protein A and autoradiography. on April 7, 2017. © 1997 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from Clinical Cancer Research 2417 Adenoviral Vector-mediated Delivery of p16. To verify that the recombinant adenovirus could deliver the p16 gene, resulting in immunologically detectable p16 protein, cells were seeded at 10 ’ cells/lOO-mm dish in appropriate medium contaming 10% FCS. After 24 h, cells were infected with either AdCMVLacZ or Adpl6 at a MOI of 100 in 2 ml of medium containing 2% FCS. One h later. the infection was terminated with 8 ml of 10% medium. One dish was left untreated as a control. The cells were lysed at 24 h postinfection as outlined above. Protein quantitation and immunoblotting were performed as outlined above. Analysis of Cell Growth Kinetics. Growth curves were generated to analyze the growth kinetics of Adpl6 treated and untreated cells. For this analysis, cells were seeded at 25,000 cells/well in six-well plates in serum-free medium to allow cell cycle synchronization. Cells were then infected 24 h after seeding with either AdCMVLacZ or Adpl6 in 1 ml of 2% medium at a MOl of 100. After 1 h, the infection was terminated with 3 ml of 10% medium. Medium was changed 24 h postinfection to prevent nonspecific viral toxicity. At time points of 1. 3. 5, 7, 9, and 1 1 days postinfection. cells were rinsed twice with PBS (137 msi NaCI, 2.7 mM KCI, 4.3 m i Na,HPO4, and 1.4 msi KH2PO4) and then harvested with 1 ml of trypsin (Mediatech) for 5 mm at 37#{176}C. After the addition of 1 ml of 10% medium. cells were counted with a Coulter counter. Cell Cycle Analysis. Cells ( 1 X l0 ) were seeded into 100-mm dishes and infected with either AdCMVLacZ or Adpl6 in 1 ml of 2% medium at a MOI of 100. After 1 h. the infection was terminated with 10 ml of 10% medium. Cells were harvested 72 h later by trypsinization. The cells were washed twice in 1 X PBS. The cell pellet was resuspended in 1 ml of solution A (0.1% sodium citrate and 0.1% Triton-X). Propidium iodide (Sigma; 50 ig/ml) and DNase-free RNase (Boehringer Mannheim: I ig/ml) were added to the suspension and incubated at 37#{176}C for 30 mm. Cell suspension vials were wrapped in aluminum foil and kept at 4#{176}C for 1-4 h. Flow cytometric analysis was performed on a Becton Dickson FACScan. Determination of Tumor Cell Chemosensitivity. To evaluate the effects of Adpl6 in combination with the chemotherapeutics cisplatin and paclitaxel, cells were seeded into 96-well plates at 5000 cells/well in serum-free medium. At 24 h after seeding. the cells were infected with either AdCMVLacZ or Adpl6 at a MOI of 100 in 2% medium. Uninfected cells were used as controls. Medium was removed 24 h postinfection. and fresh medium containing various concentrations of cis-diamminedichloroplatinum (cisplatin: Bristol-Myers Squibb. Princeton, NJ) or paclitaxel (Taxol: Bristol-Myers Squibb) were added. After 5 days, medium was replaced with 100 p.1 of 10% medium and direct analysis of cell viability was carried out using the Cell Titer 96 AQ Non-Radioactive Cell Proliferation assay (Promega. Madison, WI) according to the manufacturer’s instructions. This assay is based on the ability of viable cells to reduce MTS to a formazan compound that is soluble in tissue culture medium and can be measured spectrophotometrically at an absorbance of 490 nm. Briefly. MTS solution (2 ml) was mixed with 100 p.1 of PMS immediately before addition to cells in the 96-well plate. This MTS-PMS solution was then added into each well maintaining a ratio of 20 i.l MTS-PMS: 100 p.1 of medium. After 30 mm, the reduced product was measured at an p16 C 0 L) , ‘ e -I