Targeted Gene Therapy of Ovarian Cancer Using an Ovarian-Specific Promoter

Gynecologic Oncology 84, 228 –234 (2002) doi:10.1006/gyno.2001.6490, available online at http://www.idealibrary.com on

Targeted Gene Therapy of Ovarian Cancer Using an Ovarian-Specific Promoter Rudi Bao, M.D., Muthu Selvakumaran, Ph.D., and Thomas C. Hamilton, Ph.D. 1 Ovarian Cancer Program, Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111 Received June 18, 2001; published online December 13, 2001

Objectives. The “suicide” gene therapy of cancer using promoters such as cytomegalovirus could cause severe toxicity to normal tissues due to a lack of specificity of prodrug activation. Therefore, we investigated gene therapy of ovarian cancer using ovarianspecific promoter (OSP1) to limit the synthesis of the prodrug activating enzyme HSVtk to ovarian cancer cells. Methods. The HSVtk expressing plasmid pOSP1–HSVtk was created and transfected into an ovarian cancer cell line OVCAR3. The ganciclovir (GCV) sensitivity of the stable transfectants was evaluated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide method. Tissue specificity of this promoter was evaluated by comparing the sensitivity to GCV between ovarian and nonovarian cancer cell lines after they were transfected with pOSP1–HSVtk. One transfectant sensitive to GCV was implanted intraperitoneally to immunocompromised mice which were treated subsequently with GCV. Furthermore, this ovarian cancer survival model was used to evaluate the in vivo efficacy of cationic lipid mediated pOSP1–HSVtk gene delivery followed by GCV treatment. Results. Stable transfectants of OVCAR3 cells bearing OSP1– HSVtk became more sensitive to GCV treatment compared to the parental cell line and vector transfected OVCAR3 cell line. OSP1– HSVtk could specifically sensitize the OVCAR3 ovarian cancer cell line to GCV. SCID mice transplanted with the OVCAR3 transfectant and treated with GCV survived longer than the mice without GCV treatment (P ⴝ 0.032). In vivo gene delivery mediated by a cationic lipid (GL67) followed by GCV treatment yielded a longer survival in the OVCAR3 survival model (P ⴝ 0.016). Conclusions. The OSP1 promoter can selectively direct suicide gene therapy of ovarian cancer and the in vivo efficacy is improved by using a cationic lipid GL67 as delivery vehicle as opposed to the direct injection of plasmid. © 2001 Elsevier Science Key Words: ovarian cancer; promoter; Herpes simplex virus thymidine kinase; experimental therapy.

INTRODUCTION Ovarian cancer remains the leading cause of death from gynecological malignancies. Despite improved surgical tech1 To whom correspondence and reprint requests should be addressed at Ovarian Cancer Program, Fox Chase Cancer Center, 7701 Burholme Avenue, Philadelphia, PA 19111. Fax: 215-728-2741. E-mail: [email protected].

0090-8258/01 $35.00 © 2001 Elsevier Science All rights reserved.

niques and chemotherapy, the long-term prognosis for patients with advanced disease has not markedly improved over the past 10 years [1]. Hence, better therapies must be investigated. Molecular chemotherapy, also called gene-directed enzyme prodrug therapy, is the delivery of “suicide” genes into target cells with the suicide genes activating subsequently administered prodrug which after activation kills the target cells. One of the combinations of suicide gene/prodrug that is frequently studied is the HSVtk/GCV system. Herpes simplex virus thymidine kinase (HSVtk) is expressed in the target cells and preferentially phosphorylates an antiviral agent ganciclovir (GCV) into monophosphates. Mammalian cellular kinase then converts the monophosphates into active triphosphates which inhibit DNA elongation during cell replication and cause cell death. In this way, tumor cells that have been transfected with this suicide gene will be killed [2]. The confinement of ovarian cancer within the peritoneal cavity in the majority of cases suggests the possibility of effective local gene therapy. Recently, three protocols of HSVtk-based molecular chemotherapy were approved for clinical trials in ovarian cancer patients by the Recombinant DNA Advisory Committee. Two of these protocols made use of the retrovirus vector as a delivery system based on the possibility that a higher percentage of proliferative cells in the cancer would facilitate gene transduction [3, 4]. Another protocol employed an adenovirus which was able to infect both proliferating and nonproliferating cells [5]. In either case, the potential for the transfer of foreign genes to normal tissues of the peritoneal cavity still exists. Unwanted toxicity to normal tissue may then develop as a result of a lack of selectivity of the universal promoter. The use of tissue-specific or tumor-specific promoters to restrict suicide gene expression to the cancer cells is one possible way to avoid this problem [6]. By linking the suicide gene to transcription control elements selective for a particular tumor or tissue type, significant transcription of the suicide gene can often be activated only in the tumor cells even though both normal and malignant cells are transduced. Examples of tissue-specific promoters include tyrosinase promoter in targeting melanocytes [7], MUC-1 (polymorphic epithelial mucin) promoter in targeting breast tissue [8], and glial fibrillary

228

OVARIAN-CANCER-SPECIFIC GENE THERAPY

229

FIG. 1. Construction of the HSVtk (herpes simplex virus thymidine kinase) expression vector under control of the OSP1 promoter (pOSP1–HSVtk). PA, polyadenylation signal. NeoR, neomycin resistance gene.

acidic protein promoter (gfa 2) in targeting astrocytes [9]. In addition to the use of tissue-specific promoters, some tumorspecific promoters including the ␣-fetoprotein promoter in hepatocarcinoma [10], upstream sequences of the erbB-2 oncogene in breast cancer [11], carcinoembryonic antigen promoter in gastrointestinal carcinomas [12], and the survivin promoter in most cancers [13, and unpublished results] are being studied. The promoter used in this study was isolated from retrovirallike genomic elements of rats. These elements are termed ovarian-specific transcription (OST) units and are repetitively distributed in the rat genome and transcriptionally active only in the rat ovary [14]. The U3 portion of the 5⬘ long terminal repeat, which contains the binding motifs that regulate transcription, was reconstructed from the U3 sequence present in the 3⬘ portion of one transcript of the OST family and named ovarian-specific promoter-1 (OSP1), based on its ability to drive reporter gene expression in cells of ovarian lineage [15]. In this study we used a HSVtk expression construct driven by the OSP1 promoter to evaluate the selectivity and efficacy of GCV in the OVCAR3 human ovarian cancer cell line in vitro and in vivo. MATERIALS AND METHODS Cell lines and cultures. The human ovarian cancer cell line, NIH: OVCAR3, was established in our laboratory from the malignant ascites of a patient with progressive ovarian adenocarcinoma after combination chemotherapy with cyclophosphamide and cisplatin [16]. Two human colon adenocarcinoma cell lines, HT29 and HCT116, were also used in this study for comparison of tissue selectivity of OSP1-HSVtk gene therapy. All of these cell lines were maintained at 37°C in a humidified incubator containing 5% CO 2. The culture medium was RPMI 1640 supplemented with 10% (V/V) fetal bovine serum, 100 ␮g/ml streptomycin, 100 units/ml penicillin, 0.3 mg/ml glutamine, and 0.25 units/ml pork insulin. Plasmid construction. The promoter OSP1 was isolated from the retroviral-like genomic elements of the Fisher 344 rat [14]. The OST30-1 cDNA was amplified by PCR with the forward primer 5⬘-CAGAGGTACCTAAAACAAGTTG-3⬘ and the reverse primer 5⬘-AGACTGGCGCGCCTAGCAGAGCC-3⬘. The 462-bp PCR fragment was cloned into the multiple cloning sites of the pc3 vector which had been modified from pcDNA3 plasmid (Invitrogen, Carlsbad, CA) by removal of the cytomegalovirus (CMV) promoter [15]. The HSVtk gene was excised from pAB109 plasmid with BglII/

PvuII and then subcloned into pc3 which already had the OSP1 promoter. The 5.6-kb construct, designated pOSP1–HSVtk, contained a neomycin resistance gene (Neo) for selection of stable transfectants (Fig. 1). In vitro transfection of cancer cell lines. The pOSP1– HSVtk plasmid and the control plasmid, pc3, were linearized with ScaI which has a cleavage site in the ampicillin resistance gene (Amp). The linearized DNA was purified using the phenol– chloroform method. For stable transfection, subconfluent OVCAR3 ovarian cancer cells were trypsinized, washed with phosphate-buffered solution (PBS), and transferred into an electroporation cuvette (10 ⫻ 10 6 cells in 0.7 ml PBS for each cuvette). A total of 10 ␮g of the plasmid was added to the cell suspensions. After 10 min on ice, the cells were electroporated by using a Gene Pulser II system (Bio-Rad Laboratories, Hercules, CA). The conditions used were voltage 250 V/cm and compacitance 975 ␮F. Next, the cells were plated in 100-mm petri dishes with complete medium. G418 was added to the culture 24 h later to reach a final concentration of 500 ␮g/ml. One week later, the culture medium was changed. Fresh medium contained G418. Clones surviving G418 selection were isolated using cloning cylinders. HSVtk-positive clones were confirmed by Northern blot analysis and in vitro GCV cytotoxicity using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) method as described below. In order to evaluate the ovarian specificity of the OSP1 promoter, two colon cancer cell lines, HT29 and HCT116, were also transfected with pOSP1–HSVtk. After G418 selection, the surviving clones were pooled and compared with pooled clones of OVCAR3 as to their sensitivity to GCV using the MTT method. Northern blot analysis. After electroporation and G418 selection, 20 clones of OVCAR3 were selected randomly and cultured separately. The remaining clones were pooled and named OVCAR3 MIX. Total RNA was extracted from these individual transfectants and analyzed by Northern blot for HSVtk expression. Briefly, 10 ␮g total RNA from each sample was electrophoresed on formaldehyde-containing gel and transfered to a nylon membrane as described previously [14]. The HSVtk fragment was used as a probe. The parental cell line OVCAR3 and vector transfected OVCAR3 cell line were used as controls. In vitro sensitivity of stable transfectant to GCV (MTT assay). Two thousand cells were plated into each well of 96-well plates. Following overnight incubation, cells were

230

BAO, SELVAKUMARAN, AND HAMILTON

exposed to various concentrations of GCV (Roche, Nutley, NJ). Seventy-two hours after GCV exposure, 40 ␮l MTT solution (5 mg/ml) was added into each well. After 2 h of incubation, the cells were lysed by 100 ␮l lysis buffer (20% SDS and 50% N,N-dimethylformamide) and incubated overnight. The absorbance at 570 nm was measured using a Model 3550 microplate reader (Bio-Rad Laboratories). The IC 50 was the concentration of GCV that resulted in 50% reduction in absorbance relative to untreated cells. Triplicate experiments were used for the final analysis. In vivo sensitivity of OVCAR3 transfectant to GCV. Immunocompromised C.B.17/ICR severe combined immune deficiency (SCID) mice were used in the animal study. They were bred in the laboratory animal facility at Fox Chase Cancer Center and kept under specific pathogen-free conditions in plastic cages equipped with air filters. The animals received commercial food and water ad libitum. All SCID mice used in this study were female, 10 –12 weeks of age, with body weights between 18 and 22 g. The animal care and use were in accord with the institutional guidelines. One of the stable transfectants, OVCAR3 OSP1–HSVtk, was used for the in vivo sensitivity study. OVCAR3 OSP1–HSVtk cells growing as a monolayer were harvested with trypsin, washed with complete medium, and resuspended with PBS. A total of 10 ⫻ 10 6 OVCAR3 OSP1–HSVtk cells were diluted in 1.0 ml PBS and injected intraperitoneally (ip) into each SCID mouse (n ⫽ 10). From Day 4 through Day 11 after tumor implantation, 5 mice received an intraperitoneal injection of GCV (50 mg/kg) and another 5 mice received an intraperitoneal injection of physiological saline as controls. The side effects including body weight changes were observed closely. Survival was used as an endpoint for the efficacy of this therapy. Cationic lipid mediated gene transfer in the OVCAR3 survival model. The in vivo transfection efficiency of the cationic lipid mediated gene transfer was evaluated using a secreted reporter system in the OVCAR3 xenografts. Then the efficacy of gene therapy using HSVtk driven by the OSP1 promoter was evaluated in the OVCAR3 survival model. Delivery of the reporter vector (pCMV–SEAP) and pOSP1–HSVtk was with a cationic lipid GL67 ((GL67:DOPE:DMPEPEG(5000), 1:2:0.05) provided by Genzyme (Framingham, MA). Just before use, dried lipid films were hydrated for 10 min with sterile water and vortexed for 2 min to generate liposomes. The cationic lipid/pOSP1–HSVtk complex was prepared by mixing equal volumes of liposomes and the plasmid (molar ratio 1/4). The complex was kept at room temperaure for 15 min before in vivo transfection. The OVCAR3 survival model was established from a variant of the OVCAR3 cell line which underwent a series of in vitro and in vivo selections so that it became more potent in producing ascites and solid tumors after it was injected into the peritoneal cavity of immunocompromised mice. This pattern of growth closely mimics clinical ovarian cancer [16].

The ascites cells were aseptically collected from OVCAR3 bearing mice, suspended in PBS, and centrifuged at 300 rpm for 5 min with a tabletop centrifuge. A total of 0.2 ml packed cells diluted in 1 ml PBS was injected intraperitoneally into each SCID mouse. Thirteen days after tumor implantation, formation of intraperitoneal tumors and ascites were confirmed by performing autopsy on two euthanized mice. The rest of the mice were randomized into three groups with five mice in each group. The control group received only an intraperitoneal injection of water (224 ␮l water per mouse). The plasmid therapy group received an intraperitoneal injection of pOSP1– HSVtk plasmid (50 ␮g in 224 ␮l water per mouse). The lipid/plasmid therapy group received an intraperitoneal injection of GL67/pOSP1–HSVtk complex (50 ␮g plasmid in 224 ␮l lipid solution per mouse). The gene delivery started on Day 13 at intervals of 12 h for a total of three dosages. On Day 15, an intraperitoneal injection of GCV (50 mg/kg body wt, every 12 h) was administered to all animals and continued for 8 days. Survival was used as an endpoint for the efficacy evaluation of this gene therapy. Statistical analysis. The Wilcoxon test for paired samples was used to analyze survival differences between animal groups. RESULTS Generation of stable transfectant of OVCAR3 and its in vitro sensitivity to GCV. Twenty clones were analyzed by Northern blot for HSVtk expression. A representative HSVtk-positive clone is shown in Fig. 2A. Both the parental cell line and the vector transfected cell line were negative for HSVtk transcript (Fig. 2A). The in vitro sensitivity of this stable transfectant to GCV was next evaluated by the MTT method. As shown in Fig. 2B, compared with its parental cell line (OVCAR3) and vector control cell line (OVCAR3 vector), the OVCAR3 OSP1–HSVtk transfectant cell line was more sensitive to GCV treatment in a dose-dependent manner. Its IC 50 value was only 0.26 ␮M. On the contrary, the parental cell line OVCAR3 and the vector control cell line OVCAR3 VECTOR were very resistant to GCV treatment with only 8 and 20% cell death, respectively, at 10 ␮M of GCV concentration. Specific killing of ovarian cancer cell line using HSVtk/GCV system with OSP1 promoter. To determine tissue specificity of the OSP1 promoter, the sensitivity of OVCAR3 was compared with nonovarian cancer cell lines (HT29, HCT116) as to their sensitivity to GCV after they were transfected with pOSP1–HSVtk. The G418-resistant clones were pooled for each cell line. The MTT method was used to determine their sensitivity to GCV. As shown in Fig. 3, the pooled clones of OVCAR3 were sensitive to GCV treatment in a dose-dependent manner. GCV (10 ␮M) caused nearly 50% cell death in the pooled clones of OVCAR3 (OVCAR3 MIX), but less than 10% cell death in the pooled clones of HT29 (HT29 MIX) and

OVARIAN-CANCER-SPECIFIC GENE THERAPY

231

FIG. 3. GCV cytotoxicity of pooled clones of OVCAR3 (OVCAR3 MIX), HT29 (HT29 MIX), and HCT116 (HCT116 MIX) after they were transfected with pOSP1–HSVtk and selected with G418. One day before GCV exposure, 2 ⫻ 10 6 cells were plated into each well of a 96-well plate. The sensitivity to GCV was determined by the MTT method. The nontransfected OVCAR3 cell line was also included as a control.

FIG. 2. Stable transfectant of the OVCAR3 cell line. A. Northern blot analysis of HSVtk transcript in the stable transfectant (OVCAR3 OSP1–HSVtk) compared with its parental cell line (OVCAR3) and vector transfected control (OVCAR3 VECTOR). B. In vivo GCV sensitivity of the stable transfectant (OVCAR3 OSP1–HSVtk) compared with its parental cell line (OVCAR3) and vector transfected control (OVCAR3 VECTOR). One day before GCV treatment, 2 ⫻ 10 6 cells were plated into each well of a 96-well plate. Three days following GCV treatment, the living cells were measured using a microplate reader after the cell was exposed to MTT and lysis buffer.

HCT116 (HCT116 MIX) cell lines. The OVCAR3 parental cell line was also resistant to GCV treatment, with only 8% cell death at 10 ␮M GCV concentration. In vivo sensitivity of a stable HSVtk transfectant of OVCAR3 to GCV. To test the in vivo sensitivity of the OVCAR3 transfectant to GCV, OVCAR3 OSP1–HSVtk was injected intraperitoneally into the SCID mice to establish intraperitoneal disease. Four days after tumor implantation, daily GCV (50 mg/kg) was delivered intraperitoneally for a total of 7 days. At this dose level, the mice tolerated the treatment without loss of body weight or other noticeable side effects. This genetically modified cancer cell line retained the biological properties of its parental cell line, such as formation of ascites and peritoneal nodules, unchecked growth, and metastasis, which finally led to host death. Without GCV treatment, animals bearing OVCAR3 OSP1–HSVtk developed progressive ascites and peritoneal tumors and finally died from the complications 2–3 months after tumor implantation. On the contrary, if the OVCAR3 OSP1–HSVtk bearing ani-

mals were treated subsequently with GCV, they lived significantly longer than the control mice (P ⫽ 0.032). The median survivals of the control and GCV treatment groups were 60 and 137 days, respectively. One mouse in the GCV treatment group was even free from cancer at 400 days after gene therapy based on necropsy of the euthanized mouse (Fig. 4). In vivo efficacy of cationic lipid mediated pOSP1–HSVtk gene therapy in the OVCAR3 survival model. On the basis of the in vitro results demonstrating selective cytotoxicity of pOSP1–HSVtk therapy and in vivo results exhibiting significant efficacy of pOSP1–HSVtk therapy in prolongation of survival of animals bearing OVCAR3 OSP1–HSVtk, we next evaluated whether the direct in vivo transfection of pOSP1–HSVtk

FIG. 4. In vivo sensitivity of the OVCAR3 OSP1–HSVtk to GCV. Ten SCID mice were inoculated intraperitoneally with 5 ⫻ 10 6 OVCAR3 OSP1–HSVtk cells. GCV at 50 mg/kg was injected intraperitoneally to 5 mice from day 4 until day 11 following tumor implantation. Survival was compared with the control group receiving an intraperitoneal injection of physiological saline (n ⫽ 5).

232

BAO, SELVAKUMARAN, AND HAMILTON

followed by GCV treatment could inhibit the development of intraperitoneal cancer by OVCAR3 cells and prolong animal survival. The in vivo transfection efficiency of the GL67-mediated gene delivery system was first evaluated by transfecting the intraperitoneal ovarian tumor with a reporter plasmid pCMV– SEAP [17]. This construct was a SEAP (secreted human alkaline phosphatase) expression vector with a strong, constitutive CMV promoter. The secretability of the SEAP product made it possible to measure activity in the extracellular compartment. The OVCAR3 cell line (2 ⫻ 10 6 cells per mouse) was implanted into two CB17/ICR SCID mice to establish the intraperitoneal disseminated ovarian tumors. Reporter gene transfection was carried out 21 days after tumor implantation. A total of 5 ␮g pCMV–SEAP plasmid was mixed with an equal volume of GL67 lipids (molar ratio 4/1) and the mixture was injected into the mouse abdominal cavity. At time points of 0, 24, and 48 h after gene transfection, the abdominal cavity was washed with 0.5 ml normal saline. The aspirated ascitic fluid was centrifuged and 2 ␮l supernatant was used for SEAP assay [17]. As shown in Fig. 5A, the SEAP activity in the ascites increased at 48 h after gene transfection even though no SEAP activity was detected at 24 h. This result indicated that the intraperitoneal ovarian tumor could be transfected with the GL67-mediated, intraperitoneal gene delivery system. Based on this result, the in vivo efficacy of the GL67mediated pOSP1–HSVtk gene therapy was evaluated in the ovarian cancer model. The pOSP1–HSVtk plasmid was mixed with cationic lipid and injected intraperitoneally into OVCAR3 ip xenografted mice after the existence of intraperitoneal tumors was confirmed. As shown in Fig. 5B, the control group, which was treated only with water followed by GCV, had a median survival of 22 days, a typical life span of mouse for this tumor model. The plasmid therapy group which was treated with plasmid and GCV had a median survival of 40 days, which was an 82% increase in survival. However, the difference did not reach statistical significance (P ⫽ 0.075). The lipid/plasmid therapy group, which was treated with the lipid/ pOSP1–HSVtk complex followed by GCV, had a median survival of 51 days, which is a 132% increase in median life span over the control group. Statistically, this survival was significantly longer than that of controls (P ⫽ 0.016). Additionally, no apparent toxicity including body weight loss was observed in any group. DISCUSSION Ovarian cancer spreads mainly by intraperitoneal dissemination which makes it an appropriate disease for local gene therapy [3]. Even though many kinds of gene therapies are exploited in preclinical and clinical studies of ovarian cancer and some have shown promising results, significant limitations are still apparent. One of the problems is the nonspecific transfection of normal tissues, which is the basis of severe

FIG. 5. Cationic lipid mediated gene delivery in the OVCAR3 xenografts. A. SEAP activity in the ascites of SCID mice xenografted with OVCAR3. The reporter plasmid was injected intraperitoneally together with the GL67 lipid on day 21 following tumor implantation. The ascites was taken at time points of 0, 24, and 48 h. SEAP activity was measured with a chemoluminescence method. B. Survival of the OVCAR3 carrier mice after they were treated with the cationic lipid mediated OSP1–HSVtk gene therapy. The control (n ⫽ 5) received an intraperitoneal injection of water followed by GCV treatment. The plasmid therapy group (n ⫽ 5) received an intraperitoneal injection of pOSP1– HSVtk plasmid and GCV. The lipid/plasmid therapy group (n ⫽ 5) received an intraperitoneal injection of GL67/pOSP1–HSVtk complex and GCV. The gene delivery started on day 13 at an interval of 12 h for a total of 3 dosages. The intraperitoneal injection of GCV (50 mg/kg body wt, every 12 h) was administered from day 15 until day 23.

toxicity. One approach to overcoming this problem is the use of tissue/tumor-specific promoters to direct expression of the therapeutic gene. In this way, normal cells are spared from killing because only the targeted cells can express the toxic genes. In the gene therapy of ovarian cancer, several promoters have been reported that could be used to control gene expression in ovarian cancer cells. Among these promoters, a few of them are currently under clinical evaluation. hCG (human chorionic gonadotropin) is normally only produced in the pituitary axis, placenta, and various fetal tissues but is overexpressed in many gynecologic tumors. Lidor et al. cloned the hCG promoter into a retrovirus encoding the diphtheria toxin a chain gene and transfected the construct into ovarian cancer cell lines. Selective killing of a number of ovarian cancer cell lines was achieved with minimal toxicity in normal ovarian cells and fibroblasts [18]. The secretary leukoprotease inhibitor

233

OVARIAN-CANCER-SPECIFIC GENE THERAPY

(SLPI) was shown to be overexpressed in many cancers including ovarian cancer. The SLPI promoter was used to control HSVtk expression in ovarian cancer cell line SKOV3 and achieved specific killing [19]. Other ovarian tissue/cancerspecific promoters under active study are MUC1 (mucin) promoter, L-plastin promoter, human ␣-folate receptor promoter [20], and the survivin promoter [unpublished results]. In the course of efforts to identify genes involved in the etiology of ovarian cancer, we discovered a family of retroviral-related transcripts in rat ovarian RNA whose genomic templates are widely distributed in the rat genome. Northern blot analysis showed that these elements are transcriptionally active in the rat ovary, but not in a wide range of other normal rat tissues (such as brain, heart, kidney, lung, spleen, uterus, fallopian tubes, and thymus). In situ hybridization to ovarian tissue sections revealed that the elements were expressed in granulosa and theca internal cells and in the surface epithelial cells adjacent to preovulatory follicles. The genomic template was reconstructed from a full-length retrovirus-like transcript and the upstream regulatory sequence was cloned and named ovarian-specific promoter-1. Analysis of this sequence revealed 50 potential regulatory elements. Based on comparison to the upstream regulatory sequences of some other genes of importance to ovarian physiology, such as estrogen receptor, FSH receptor, and LH receptors, 13 overlapping elements were found [14]. Based on these findings, we speculate that binding by one or a combination of ovary-specific transcription factors to these regulatory units within the OSP1 sequence contributes to its tissue specificity. To test OSP1’s utility in gene therapy of ovarian cancer, we created an HSVtk expression construct driven by the OSP1 promoter and transfected the construct into the OVCAR3 human ovarian cancer cell line. The GCV sensitivity of pooled G418-resistant clones of OVCAR3 was much greater than that of similarly transduced colon cancer cell lines. This supports our belief that OSP1 is an ovarian-specific promoter. We had previously tested reporter gene expression driven by the OSP1 promoter in human cancer cell lines using the CAT reporter. Compared with nonovarian cell lines, such HeLa, HT29, and NIH3T3, 30 to 80 times higher expression was found in A2780, OVCAR2, and OVCAR10 ovarian cancer cell lines [15]. The current data further indicate that OSP1 is an ovarianspecific promoter and that the correct combination of transcription factors specific to the ovarian tissue can bind the transcription factor binding motifs in OSP1 and drive downstream gene expression. Furthermore, our in vitro and in vivo studies suggest that OSP1 has sufficient specificity to be used to drive suicide gene therapy of ovarian cancer without toxicity to the host. Several approaches are being evaluated to deliver genes for therapeutic purposes. In this study, we used a cationic lipid. This is one of the in vivo gene delivery vehicles under active evaluation based on its safety and lack of immune response after in vivo administration. Cationic lipids, by virtue of their

positive charges, can complex with the negatively charged DNA and facilitate its entry into cells in vitro and in vivo [21–23]. GL67, the lipid used in our study, is one of the cationic lipids recently developed by Genzyme Corp. and has been demonstrated to be 100-fold more active than other cationic lipids such as DMRIE or DC-Chol in mediating gene delivery. Furthermore, it is comparable to recombinant adenovirus vectors at low multiplicities of infection in delivery capability [24, 25]. In this study, we demonstrated that GL67mediated, intraperitoneal gene delivery was effective in transfecting intraperitoneal cancer cells in the OVCAR3 xenograft model. We believe the results from the cationic lipid mediated OSP1–HSVtk gene therapy in the OVCAR3 survival model are encouraging. The efficacy of this gene therapy in prolonging animal survival and the low toxicity of the therapy indicate that OSP1 is functional in targeting ovarian cancer in vivo and that the GL67 cationic lipid is effective in mediating in vivo gene transfection. These data support the further evaluation of OSP1–HSVtk as a suicide gene therapy for clinical ovarian cancer. ACKNOWLEDGMENTS Dr. Hamilton is supported by grants from the National Institutes of Health, CA06927, CA51228, CA56916, and CA84242, a Specialized Program of Research Excellence (SPORE) CA83638, an appropriation from the Commonwealth of Pennsylvania, the Adler Foundation, and the Evy Lessin Fund. We thank Dr. Ronald Scheule for kindly providing us with the GL67 cationic lipids used in this study. We also thank Dr. Samuel Litwin for statistical analysis of the data related to this study.

REFERENCES 1. Ozols R. Treatment of gynecologic cancer: the US experience. Tumori 1999;85:S5–11. 2. Robertson M, III, Barns M, Rancourt C, Wang M, Grim J, Alvarez R, Siegal G, Curiel D. Gene therapy of ovarian cancer. Semin Oncol 1998; 25:397– 406. 3. Freeman SM, Abboud CN, Whartenby KA, Packman CH, Koeplin DS, Moolten FL, Abraham GN. The “bystander effect”: tumor regression when a fraction of the tumor mass is genetically modified. Cancer Res 1993; 53:5274 – 83. 4. Link C, Moorman D. Clinical protocols: a phase I trial of in vivo gene therapy with the herpes simplex thymidine kinase/ganciclovir system for the treatment of refractory of recurrent ovarian cancer. Cancer Gene Ther 1995;2:230 –1. 5. Rosenfeld ME, Feng M, Michael SI, Siegal GP, Alvarez RD, Curiel DT. Adenoviral-mediated delivery of the herpes simplex virus thymidine kinase gene selectively sensitizes human ovarian carcinoma cells to ganciclovir. Clin Cancer Res 1995;1:1571– 80. 6. Hart IR. Tissue specific promoters in targeting systemically delivered gene therapy. Semin Oncol 1996;23:154 – 8. 7. Siders WM, Halloran PJ, Fenton RG. Melanoma-specific cytotoxicity induced by a tyrosinase promoter-enhancer/herpes simplex virus thymidine kinase adenovirus. Cancer Gene Ther 1998;5:281–91. 8. Patterson A, Harris AL. Molecular chemotherapy for breast cancer. Drugs Aging 1999;14:75–90.

234

BAO, SELVAKUMARAN, AND HAMILTON

9. Vandier D, Rixe O, Brenner M, Gouyette A, Besnard F. Selective killing of glioma cell lines using an astrocyte-specific expression of the herpes simplex virus-thymidine kinase gene. Cancer Res 1998;58:4577– 80. 10. Mawatari F, Tsuruta S, Ido A, Ueki T, Nakao K, Kato Y, Tamaoki T, Ishii N, Nakata K. Retrovirus-mediated gene therapy for hepatocellular carcinoma: selective and enhanced suicide gene expression regulated by human alpha-fetoprotein enhancer directly linked to its promoter. Cancer Gene Ther 1998;5:301– 6. 11. Harris JD, Gutierrez AA, Hurst HC, Sikora K, Lemoine NR. Gene therapy for cancer using tumour-specific prodrug activation. Gene Ther 1994;1: 170 –5. 12. Cao G, Kuriyama S, Cui L, Nagao S, Pan X, Toyokawa Y, Zhang X, Nishiwaki I, Qi Z. Analysis of the human carcinoembryonic antigen promoter core region in colorectal carcinoma-selective cytosine deaminase gene therapy. Cancer Gene Ther 1999;6:572– 80. 13. Li F, Altieri DC. Transcriptional analysis of human survivin gene expression. Biochem J 1999;344:305–11. 14. Godwin A, Miller P, Getts L, Jackson K, Sonoda G, Schray K, Testa J, Hamilton T. Retroviral-like sequences specifically expressed in the rat ovary detect genetic differences between normal and transformed rat ovarian surface epithelial cells. Endocrinology 1995;136:4640 –9. 15. Selvakumaran M, Bao R, Crijns AP, Connolly DC, Weinstein JK, Hamilton TC. Ovarian epithelial cell lineage-specific gene expression using the promoter of a retrovirus-like element. Cancer Res 2001;61:1291–5. 16. Hamilton TC, Young RC, McKoy WM, Grotzinger KR, Green JA, Chu EW, Whang-Peng J, Rogan AM, Green WR, Ozols RF. Characterization of a human ovarian carcinoma cell line (NIH:OVCAR-3) with androgen and estrogen receptors. Cancer Res 1983;43:5379 – 89.

17. Bao R, Selvakumaran M, Hamilton TC. Use of a surrogate marker (human secreted alkaline phosphatase) to monitor in vivo tumor growth and anticancer drug efficacy in ovarian cancer xenografts. Gynecol Oncol 2000;78:373–9. 18. Lidor Y, Lee W, Nilson J, Maxwell I, Su LJ, Brand E, Globe L. In vitro expression of the diphtheria toxin A-chain gene under control of human chorionic gonadotropin gene promoters as a means of directing toxicity to ovarian cancer cell lines. Am J Obstet Gynecol 1997;177:579 – 85. 19. Garver RI, Jr., Goldsmith KT, Rodu B, Hu PC, Sorscher EJ, Curiel DT. Strategy for achieving selective killing of carcinomas. Gene Ther 1994; 1:46 –50. 20. Casado E, Nettelbeck DM, Gomez-Navarro J, Hemminki A, Gonzalez Baron M, Siegal GP, Barnes MN, Alvarez RD, Curiel DT. Transcriptional targeting for ovarian cancer gene therapy. Gynecol Oncol 2001;82:229 – 37. 21. Aoki K, Yoshida T, Matsumoto N, Ide H, Hosokawa K, Sugimura T, Terada M. Gene therapy for peritoneal dissemination of pancreatic cancer by liposome-mediated transfer of herpes simplex virus thymidine kinase gene. Hum Gene Ther 1997;8:1105–13. 22. Cooper MJ. Noninfectious gene transfer and expression systems for cancer gene therapy. Semin Oncol 1996;23:172– 87. 23. Ledley FD. Nonviral gene therapy: the promise of genes as pharmaceutical products. Hum Gene Ther 1995;6:1129 – 44. 24. Scheule RK. Gene therapy for lung cancer—an application for cationic lipid- mediated gene delivery? J Natl Cancer Inst 1998;90:1118 –9. 25. Scheule RK, St George JA, Bagley RG, Marshall J, Kaplan JM, Akita GY, Wang KX, Lee ER, Harris DJ, Jiang C, Yew NS, Smith AE, Cheng SH. Basis of pulmonary toxicity associated with cationic lipid-mediated gene transfer to the mammalian lung. Hum Gene Ther 1997;8:689 –707.