Adenoviral-mediated Delivery of the Herpes Simplex Virus Thymidine Kinase Gene Selectively Sensitizes Human Ovarian Carcinoma Cells to Ganciclovir1

One strategy used for gene therapy of cancer is molecular chemotherapy. This approach is based on selective expression of an encoded toxin in cancer cells to achieve their eradication. One potential advantage of this strategy derives from a phenomenon, termed the bystander effect, whereby only a fraction ofcells needs to be transduced to eradicate a tumor population. Despite the theoretical advantages of this phenomenon, it has only been described in a few cellular targets. Therefore, we undertook strategies to develop a molecular chemotherapy approach for ovarian carcinoma utilizing the herpes simplex virus thymidine kinase (HSV-TK) gene. Initially, we established that human ovarian carcinoma cell lines could be transduced at high efficiency with adenoviral vectors encoding reporter genes. We next determined that the human ovarian cell line SKOV3 could exhibit bystander killing by stably transducing it to express HSV-TK and performing cell mixing experiments with varying percentages of HSV-TK-expressing and HSVTK-nonexpressing cells. Based on these findings, we constructed a recombinant adenovirus encoding HSV-TK and utilized it to induce human ovarian carcinoma cell lines to the sensitizing effects of ganciclovir. In addition, primary cultures of ovarian carcinoma cells were found to be highly transducible with recombinant adenoviral vectors and could be induced to the sensitizing effects of ganciclovir after induction of HSV-TK expression by the adenoviral vector. These studies indicate that molecular chemotherapy using a recombinant adenoviral vector expressing HSV-TK may provide a rational strategy for human ovarian carcinoma. Received 3/28/95: revised 6/13/95: accepted 8/17/95. I This research was supported by Grant DAMD I 7-94-J-4398 froni the Department of the Army. U. S. Army Medical Research and Development Command (Fort Detnick. Frederick. MD). 2 To whom requests for reprints should he addressed. at University of Alabama at Birmingham. Gene Therapy Program. 1824 6th Avenue South. Room 620, Birmingham. AL 35294. INTRODUCTION Ovarian carcinoma is the leading cause of death from gynecological malignancies in women. It is anticipated that in 1995 there will be 26,600 new cases of ovarian cancer, and it will account for approximately 14,500 deaths ( 1 ). This disease has a high fatality rate due to the lack of effective screening strategies and to the lack of symptoms associated with early stage disease. Consequently. 70% of women with ovarian cancinoma present with advanced stage disease at the time of diagnosis. Current standard treatment consists of surgical debulking in collibination with chemotherapy. However, extensive intnaabdominal disease is difficult to completely eradicate by surgery. Most patients will only transiently respond to adjunctive therapies. and long-term survival for patients with advanced stage disease rarely exceeds 15-30% ( I ). Based on these sunvival statistics. a variety of novel therapeutic approaches have been proposed for this disease. These experimental approaches include the use of monocbonal antibodies, immunoconjugates. cytokines. and radioimmunotherapy (2. 3). To date. however, these strategies have not offered evidence of long-term efficacy. Gene therapy approaches may offer unique methods to achieve anticancer effects. To this end, a variety of distinct strategies have been developed and used. The approach of molecular chemotherapy is designed to achieve selective eradication of carcinoma cells via an expressed toxin gene. In this regard, it has been previously demonstrated that such a gene product can selectively sensitize tumor cells to an agent not ordinarily toxic (4-6). The toxin gene therapy system most widely used to accomplish cell killing utilizes the HSV-TK3 gene. Expression of the HSV-TK gene allows the tumor to have an enhanced sensitivity to nucleoside analogues such as GCV and acycbovir. The mechanism that allows utilization of viral thymidine kinase for cytotoxic effect is preferential monophosphorylation of GCV by the HSV-TK gene (7). GCV will then be further phosphorybated by cellular kinases to diphosphates and then to triphosphate forms that become incorporated into the cellular DNA. It is the incorporation of the triphosphate form of GCV into cellular DNA that causes inhibition of DNA synthesis and RNA polymerase, which eventually leads to cell death (8). The specificity of this system is the differential metabolism of GCV preferentially by the viral thyrnidine kinase gene. Because the mammalian enzyme has a much lower affinity for the drug. normal cells tend to be resistant to the toxic effects of GCV, The abbreviations used are: HSV-TK, herpes simplex virus thymidine kinase: GCV. ganciclovir: AdCMVLuc, recombinant adenovirus cxpressing luciferase: AdCMVLacZ, recombinant adenovirus expressing 3-gaIactosidase: CMV. cytomegalovirus: AdCMVHSV-TK, recombinant adenovinus expressing HSV-TK. Research. on April 13, 2017. © 1995 American Association for Cancer clincancerres.aacrjournals.org Downloaded from 1572 HSV-TK Gene Sensitizes Ovarian Carcinoma Cells to GCV when administered at low bevels. Thus, the tumor cells transduced to express the viral gene have enhanced sensitivity for cell killing. In addition to selective sensitization, another advantage of this system was the finding of a bystander effect, in which every tumor cell does not need to express HSV-TK to eradicate a tumor population (9, 10). However, despite the apparent advantage of the bystander effect, this phenomenon has been described in relatively few neoplastic targets such as glioma (9), mesothelioma (11-13), fibrosarcoma (10, 14), and melanoma (15, 16). From a mechanistic standpoint, previous work has demonstrated the necessity of cell to cell contact as a prerequisite for bystander-mediated cell killing (10, 14). Therefore, the tumor models utilized to date have been confined to body compartrnents such as the cranial vault and the thoracic cavity, where this phenomenon could be exploited. Because >80% of ovarian carcinomas are confined to the pentoneal cavity, this disease context would theoretically be similar to other compartmentalized tumor models demonstrated to be responsive to molecular chemotherapy strategies. We, thus, investigated the efficiency of gene transfer in human ovarian carcinoma cells in vitro, and examined whether bystander-mediated cell killing occurred. The strategy described here demonstrates that an adenoviral vector is suitable for efficient gene transfer into both ovarian carcinoma cell lines and primary neoplastic cells obtained from women with ovarian cancer. Furthermore, these studies indicate that bystander killing, in the context of ovarian cancer, is feasible and, therefore, provides the rationale for the use of molecular chemotherapy as a gene therapy strategy for human ovarian carcinoma. MATERIALS AND METHODS Cell Lines. The human ovarian carcinoma cell lines SKOV3, PA-l, SW626, CaOv3, and the human cervical carcinoma cell line HeLa were obtained from the American Type Culture Collection (Rockville, MD). The human ovarian carcinoma cell line OV-4 was kindly provided by Timothy J. Eberlein (Brigham and Women’s Hospital, Harvard Medical School, Boston, MA) and the SKOV3 derivative line SKOV3.ipl was kindly provided by Janet Price (Baylor University, Houston, TX). These cell lines were maintained in DMEM (Mediatech, Herndon, VA) supplemented with L-glutamine (200 jig/ml), penicillin (100 jig/mb), streptomycin (25 jig/mb), and 10% heat-inactivated FCS (PAA Laboratories, New Port Beach, CA) (complete media) at 37#{176}C in a humidified, 5% CO2 atmosphere. The human ovarian carcinoma cell line OVCAR-3 was kindly provided by Donald Buchsbaum (University of Alabama at Birmingham, Birmingham, AL) and maintained in RPM! 1640 media (Mediatech) supplemented with 20% heat-inactivated FCS, L-glutamine, antibiotics, and 10 pig/mb bovine insulin (GIBCO Life Technologies, Gaithersburg, MD). Primary Cells. Human ascites fluid was obtained from patients with ovarian carcinoma at the time of laparotomy or paracentesis. Ovarian carcinoma cells obtained from malignant ascites samples were prepared for short-term maintenance in the following manner. Ascites fluid was aliquoted into 10-mb sampIes and contaminating RBCs were lysed with 5 ml ACK buffer (0.15 M NH4CI, 1.0 msi KHCO3. 0.1 msi EDTA) for 2 mm at room temperature. Lysis was terminated by the addition of 5 ml complete media. Cells were pelleted by a 5-mm centrifugation at 190 X g, resuspended, and plated in complete media. Media was changed daily for 7 days during expansion of the carcinoma cells. Ovarian carcinoma cells were harvested by trypsinization and centrifuged through a Ficoll-Hypaque gradient (1077: Sigma Chemical, St. Louis, MO). Viable cells were removed from the interface and replated for subsequent studies. Recombinant Adenoviruses Encoding Reporter Genes. To analyze gene transfer efficiency, recombinant adenoviral vectors encoding reporter genes were used. AdCMVLuc and AdCMVLacZ are E1AIB-deleted, replication-incompetent, recombinant adenoviruses previously described ( 17). AdCMVLuc encodes the firefly luciferase reporter gene and is under the control of the human CMV promoter/enhancer (provided courtesy of R. Gerard, University of Texas-Southwestern Medical Center, Houston, TX). AdCMVLacZ encodes the reporter gene Escherichia co/i 3-galactosidase and is under the control of the CMV promoter/enhancer (provided courtesy of R. Gerard). Construction of Recombinant Adenoviral Vector Encoding HSV-TK. The adenovirus expressing HSV-TK was prepared using standard homologous recombination techniques (18). In brief, a DNA fragment containing the HSV-TK gene (provided by R. Garver, University of Alabama at Birmingham. Birmingham, AL) was subcloned into the polylinker of the adenoviral shuttle vector pACCMVpLpARS (+ ) (provided courtesy of R. Gerard). This plasmid provides prornoter/initiation signals derived from the CMV early promoter/enhancer and polyadenylation signals from SV4O. The resulting recombinant adenovirus shuttle plasmid pAC-HSV-TK was used to derive an El-deleted, replicationincompetent, recombinant adenovirus using standard niethodologies ( 18). In brief, the shuttle plasmid and the adenoviral packaging plasmid pJMl7 (provided by F. Graham, McMaster University, Hamilton, Ontario, Canada) were cotransfected into the E1A trans-complementing cell line 293 using the comrnercial cationic biposome vector DOTAP (GIBCO Life Technobogies). Transfected cells were maintained until the onset of eellular cytopathic effects. The newly generated recombinant adenovirus was plaque purified three times. Validation of single plaques was accomplished by direct PCR. The recombinant adenovirus encoding the HSV-TK gene AdCMVHSV-TK was expanded within 293 cells and purified by CsCI gradient centrifugation (18). Genomic DNA derived from the recombinant adenovirus was subjected to digestion with various restriction endonucleases and analyzed by agarose gel electrophoresis. Wild-type adenovirus WT300 (provided by T. Shenk, Princeton University, Princeton, NJ) was used as a control for analysis of genornic DNA derived from AdCMVHSV-TK. Adenoviral vectors were titered within the cell line 293 using plaque assay techniques for direct determination of viral plaque-forming units. Establishment of Stable Clones Expressing the HSV-TK Gene. A stable human ovarian carcinoma cell line expressing the HSV-TK gene from a retroviral vector was produced by standard methods (19). The GP envAm-12 cell line, kindly provided by Arthur Bank (Columbia University. New York, NY), was utilized with the STK netroviral vector construct Research. on April 13, 2017. © 1995 American Association for Cancer clincancerres.aacrjournals.org Downloaded from Clinical Cancer Research 1573 kindly provided by Frederick Moolten (E. N. Rogers Memorial Veterans Hospital, Bedford, MA). STK is a Moboney murine retrovinal construct that encodes the HSV-TK gene driven by an SV4O promoter and a neoinvcin-resistance gene driven by a long terniinal repeat (20). The netrovirus was propagated by transiently transfecting the packaging cell line with 6 jig of the STK construct. The following day retrovirus was obtained from the culture media and syringe filtered through a 0.2-jim filter. An aliquot of SKOV3 cells ( I X 106/60-rnrn dish) that had been plated 24 h previously was infected with 100 jib of virus and polybrene (Sigma Chemical) added to a final concentration of 8 jig/nil. The following day complete media were changed without the addition of polybrene. At 2 days postinfection. cells were split I :5 into selective media containing 0.5 mg/mI geneticin (G418: GIBCO Life Technologies). After 2 weeks of continual selection, G4l8-resistant clones were expanded and tested for GCV sensitivity (Cytovene: Syntex Laboratories, Inc., Palo Alto, CA). Sensitivity of HSV-TK-expressing Clones to GCV. To validate the expression of HSV-TK, G4 I 8-resistant SKOV3 clonal cells. along with untransduced SKOV3 cells, were analyzed k)r GCV sensitivity. Cells were plated in triplicate at 5000 cells/well in 96-well plates, and 24 h later complete media containing varying concentrations of GCV (0-100 jiM) were added. Cells were cultured at 37#{176}Cand in 5% CO2 for 5 days. Sensitivity to GCV was evaluated using a colonimetric cell proliferation assay that measured conversion of tetrazolium salt to formazan by viable cells as described by the manufacturer (Cell Titer 96 Aqueous Nonradioactive Cell Proliferation Assay: Prornega, Madison, WI). In brief, 20 jib assay mixture were added to each well of cells, and the plates were incubated for I to 4 h at 37#{176}C before absorbance was measured at 490 nm in a 96-well plate reader (Molecular Devices, Menlo Park, CA). The cbonal population of cells expressing HSV-TK utilized for all experiments is designated SKRV4 (SKOV3 TK retroviral clone 4). Cell viability from GCV sensitivity assays were also performed by crystal violet staining. Cells seeded in 6-well plates were infected with the HSV-TK virus as described below. Media were removed 5-7 days postinfection. and cells were fixed with I ml 10% formalin (Fisher Scientific, Pittsburgh, PA) for 10 mm at room temperature. Cells were stained for 20 mm at room temperature with 1 ml crystal violet (70% ethanol containing 1 % crystal violet; Fisher Scientific) and then washed five to six times with water to remove dye. Recombinant Adenoviral Infections. Cells were infected with recombinant adenoviral vectors as previously descrihcd (21). In brief, cells were plated in 6-well plates and infected 24-48 h later at a confluency of 75-95%. Viral infections were carried out in DMEM media containing 2% FCS and were allowed to proceed at 37#{176}Cfor 1 h. after which time complete media were added to each culture. For the AdCMVLuc vector, cells were infected with 1.5 X l0 partides/well ( I 500 viral particles/cell ). For AdCMVLacZ and AdCMVHSV-TK vectors, cells were infected with 1 X l0 particles/well (100 viral particles/cell). Analysis of Reporter Gene Expression. Following adenovirus infection, cells were analyzed for expression of the encoded reporter genes to assess the relative efficiency of gene transfer. Harvesting of cells, cell lysis. and analysis of luciferase gene expression were performed as described by the manufactuner (Prornega Luciferase Assay System: Prornega. Madison, WI). Lysates from transfected cells were obtained by aspirating the media and adding 150 jil 1 X cell culture lysis reagent to each well of cells. Cells were lysed at room temperature for 10 to 15 mm, and cellular debris was removed by centrifugation at 13.000 x g for 3 mm. Cell extract (20 jil) was added to 100 jib luciferase assay reagent and analyzed for emitted light in a Lumat LB95O1 lurninorneter (Berthold Systems Inc., Aliquippa, PA). Primary cell extracts from ascites fluid were prepared as above, except extracts were stored at -20#{176}Cuntil several sampIes could be analyzed concurrently. The primary cell extracts were thawed and quantitated for protein concentration at the time of analysis. For each sample. 15 jig extract were added to 100 jil of the luciferase assay reagent and measured for relative light units as described above. L .icZ reporter gene expression was also utilized to examine transduction frequency. Cells infected with AdCMVLacZ virus were harvested 72 h posttransduction and stained for 3-galactosidase expression. Cell monolayers were washed with PBS and fixed with 0.5% glutaraldehyde for 10 mm at 37#{176}C.ells were washed twice for 15 mm each in PBS containing I msi MgCI.,, then stained for approximately 6 h in a PBS solution containing 5 mM K3Fe(CN)6, 5 m K4Fe(CN)6, I rn i MgC12. and I mg/mI X-gal (5-bnomo-4-chloro-3-indolyl-h-o-galactopyranoside; Sigma Chemical). After removal of the X-gal solution. cells were overlaid with 70% glycerol and stored at 4#{176}C. Induction of GCV Sensitivity by an Adenovirus Expressing HSV-TK. The human ovarian carcinoma cell lines and primary cells to be infected were initially seeded into 6-well tissue culture plates. When cells reached the appropriate density, they were infected, as described above, with either the HSVTK-expressing virus or the luciferase-expressing virus. Twentyfour h postinfection three wells from each sample were treated with 20 jiM GCV. Cell viability was determined with crystal violet staining 5-7 days postinfection. Toxin Gene-killing Experiments. To analyze the bystander effect, cell mixing experiments were performed. These were accomplished by mixing HSV-TK-expressing (SKRV4) and HSV-TK-nonexpnessing (SKOV3) cells at varying percentages. Cells were plated in 96-well plates in triplicate wells at 20,000 cells/well to ensure cell to cell contact. One-half of the samples were treated 24 h later with 20 jiM GCV: the remaining cells had complete media changed. Cell viability was determined 5 days later using the cell proliferation assay as described above. In these experiments SKOV3 cells were plated on the day of the assay to generate a standard curve. Cells were removed by trypsinization and plated in triplicate wells at the following densities: 100.000. 50,000. 20,000, 10,000. 5,000, 2,000, and 0 cells/well. respectively. From the standard curve, viable cell numhers could be calculated for experimental groups using the SOFTmax computer software (Molecular Devices. Menlo Park, CA). RESULTS Human Ovarian Carcinoma Cell Lines Were Transduced at High Efficiencies with Recombinant Adenoviral Vectors. Recombinant adenoviral vectors have been shown to be effective vehicles for the in viva transduction of a variety of Research. on April 13, 2017. © 1995 American Association for Cancer clincancerres.aacrjournals.org Downloaded from