Introduction

Rho proteins are low molecular weight GTPases that belong to the Ras superfamily and are intimately involved in diverse cellular processes and diseases.1, 2 Rho proteins are pivotal in the regulation of actin cytoskeleton organization, such as lamellipodia, membrane ruffles and stress fibers. They also regulate signal transduction proteins involved in the mitogen- and stress-signaling pathways, such as extracellular signal-regulated kinase, p38 and stress-activated protein kinase.3 More importantly here is their involvement as mediators of proliferation and malignant transformation.4

Although most Rho proteins are involved in oncogenesis, invasion and/or metastasis, recent evidence suggests a tumor suppressive role for RhoB (Ras homologous gene B). Ectopic expression of RhoB inhibits tumor growth in vitro and in vivo and metastasis, and also inhibits oncogenic signaling.5, 6, 7, 8 Moreover, rhoB−/− transformed cells are very aggressive and RhoB knockout mice show increased sensitivity to chemically induced tumors.9 Ectopic expression of RhoB induces apoptosis in some cell models,8, 10, 11 whereas rhoB−/− cells show resistance to apoptosis induced by radiation and cytotoxic agents.12 Notably many oncogenes such as the epidermal growth factor receptor Ras, and Akt suppress the expression of RhoB in tumor cells.9, 13 Moreover, RhoB expression is dramatically decreased in numerous cell lines, including ovarian, lung or breast cancer,14, 15 as well as in brain, head and neck cancer where the tumors become more aggressive and highly invasive.7, 16, 17

These studies suggest that RhoB could play a critical role in suppressing malignant transformation by blocking oncogenic and tumor survival pathways. These observations prompted us to investigate the potential use of gene modification of tumor cells by RhoB, as has been suggested with other tumor suppressor genes, such as p53, Rb or WWOX.18, 19

Ovarian cancer is the most lethal of the gynecological malignancies with 75–80% of patients presenting with advanced-stage disease wherein the survival rate is only 15–20%. Patients undergoing an optimal tumor reductive surgery at presentation increase the disease-free interval, but most finally die of the disease. Approximately 70–80% of patients with advanced-stage disease respond to initial treatment with platinum and/or paclitaxel-based chemotherapy. However, the majority of patients show tumor relapse and progression within 2 years of treatment. Chemotherapy drugs given either intravenously or intraperitoneally have offered only minimal increase in survival rates over the past 15–20 years. Novel ways of treating residual or recurrent disease in the peritoneal cavity may improve long-term survival.18, 20, 21, 22, 23

We are currently investigating molecular mechanisms that may influence the growth and progression of epithelial ovarian cancers. Our aim is to develop therapeutic agents designed to correct defects at the molecular level and ultimately provide innovative treatment options for patients not responding to standard therapies. With the disease generally remaining confined to the abdominal cavity throughout its course, ovarian cancer represents a particularly good candidate for gene transfer therapy using viral vectors via intraperitoneal (i.p.) delivery. This approach has the advantage of producing high viral loads close to the tumor while minimizing adverse events due to a delay before the vectors reach the systemic circulation. Promising preclinical data on the in vitro transfection of wild-type p53 into ovarian cancer cells bearing p53 mutations and into xenograft models, coupled with evidence from phase I trials, led to the introduction of this innovative and encouraging technique for front-line treatment of patients with advanced ovarian cancer.18, 24, 25 However, dominant-negative crosstalk between ectopic wild-type p53 and recently identified dominant p53 mutants as well as splice variants of p63 and p73 (frequently overexpressed in ovarian cancers) could seriously compromise the effectiveness of p53 gene therapy.

In the present study, we constructed replication-defective adenovirus (Ad) vectors expressing RhoB. Ad vectors encoding RhoB were used in vitro and in vivo. We aimed to assess the efficiency of RhoB gene transfer in inducing apoptosis of tumor cells and abolishing tumor cell proliferation in vivo using a murine model that mimics the natural progression of the human disease.

Materials and methods

Reagents

Fetal calf serum (FCS) and culture media were purchased from Pan (Invitrogen, Cergy Pontoise, France) and Cambrex Biosciences (Lonza, Verviers, Belgium), respectively. Mouse monoclonal anti-Apo2.7 antibody (2.7A6A3 clone) was purchased from Immunotech (TEBU, Le Perray en Yvelines, France), mouse anti-caspase-3 (CPP32) monoclonal antibody from Chemicon International (Upstate Direct, Hampshire, UK) and rabbit polyclonal anti-human RhoB from Santa Cruz (TEBU, Le Perray en Yvelines, France). Fluorescein isothiocyanate peroxidase-conjugated anti-rabbit and anti-mouse secondary antibodies, the anti-β-actin and fluorescein isothiocyanate-labeled anti-mouse monoclonal antibodies and a liquid alkaline phosphatase detection kit were all purchased from Sigma (St Quentin Fallavier, France).

Cell culture

The 293 cymR cells (a transformed human embryonic kidney cell line) were obtained from Qbiogen (Evry, France). Human ovary adenocarcinoma cell line NIH-OVCAR3 (OVCAR-3) and HeLa cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The human ovary adenocarcinoma cell line IGROV-1 was a generous gift from the Institut Gustave Roussy, Villejuif. The 293 cymR cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FCS and 2 mM L-glutamine (Cambrex Biosciences, Lonza, Verviers, Belgium). OVCAR-3 and IGROV-1 cells were routinely cultured in RPMI-1640 medium containing 10% heat-decomplemented serum, supplemented with 20 ng ml−1 epidermal growth factor (Roche, Meylan, France), 10 μg ml−1 insulin (Roche) and 2 mM L-glutamine (Cambrex Biosciences). Cells lines were maintained as monolayers at 37 °C in a humidified 5% CO2 atmosphere.

Adenoviral constructs and transduction protocol

Replication-defective (ΔE1,E3) Ad vectors expressing RhoB under the transcriptional control of the cytomegalovirus (CMV) promoter were constructed with the AdEasy System (Qbiogen). Initially, a BglII–BamHI fragment containing RhoB, obtained from a digestion of the pGR5-2-RhoB plasmid already described, was subcloned into BamHI-digested pAdenoVATOR-CMV5(CuO)-ires green fluorescent protein (GFP) transfer vector (Qbiogen) to obtain the shuttle vector pAdRhoB. This vector allowed transgene expression driven by the cumate-inducible CMV5(CuO) promoter. An internal ribosome entry site sequence ensured coexpression of GFP. To generate recombinant Ad plasmids, the shuttle vector was linearized with PmeI, mixed at a 10:1 ratio with pAdEasy and cotransfected (1 μg total) into the Escherichia coli strain BJ5138. The smaller kanamycin-resistant colonies, typically containing the large recombinant plasmids, were analyzed by restriction digests. The replication-defective vector AdRhoB, as well as the control empty vector AdeGFP (encoding enhanced GFP), were produced by transfection of 293 cymR cells with a single isolate of each recombinant Ad vector. The recombinant Ads were propagated into 293 cymR cells that repress the CMV5(CuO) promoter and expression of RhoB during packaging and expansion of the RhoB adenoviral vectors (AdRhoB). Viruses were purified by CsCl gradient centrifugation. Viral titers were determined by optical absorbance at 260 nm (1 OD=1 × 1012 physical particle ml−1).

On day 1, 1 × 105 ovarian human adenocarcinoma cells were seeded onto a six-well plate. The following day, one of the six wells was trypsinized and the cells were counted to standardize the multiplicity of infection (MOI). Transductions were performed in 5% CO2 incubators at 37 °C for 1 h using 500 μl of transduction medium (the same medium used for each cell line +2% FCS) with agitation. Pilot experiments with AdeGFP were used to estimate the optimal MOI for adenocarcinoma cells (visualization of GFP-expressing cells).

Etoposide treatment

A total of 25 × 104 cells were seeded onto six-well plates and exposed to either etoposide at the concentration specified or vehicle alone. After 36 h, either cells were stained and analyzed by flow cytometry or protein extracts were analyzed by western blotting.

Immunoblotting

Cells were seeded onto 3.5 cm dishes to 50% confluence before transduction. Total protein extracts were prepared 48 h post-transduction by lysing the cells in 150 μl of 1% Triton, 0.5% sodium dodecyl cholate, 0.1% SDS, 150 mM NaCl and 50 mM TRIS for 1 h on ice. Lysates were centrifuged at 20 000 g for 5 min to remove nuclei and precipitates. Supernatant protein concentrations were measured using the Bio-Rad protein assay with bovine serum albumin in lysis buffer as a standard. Samples containing 80 μg of total cellular protein were subjected to 10–12.5% (depending on the experiment) SDS-PAGE and transferred to a nitrocellulose membrane (GE Healthcare, Orsay, France). Equal amounts of total protein lysate were used for each blot. Membranes were blocked for 1 h at room temperature in TBST 0.1% Tween-20 5% non-fat milk. Membranes were then incubated overnight with antibodies directed against β-actin (1:10 000), RhoB (1:3000) or caspase-3 (1:200). For signal detection, the secondary HRP mouse or rabbit antibody was used at a dilution of 1:10 000. The blots were washed and developed using enhanced chemiluminescence (Amersham Biotechs, GE Healthcare, Orsay, France) according to the manufacturer's protocol and exposed to radiographic films (GE Healthcare).

Cell growth kinetics

Ovarian adenocarcinoma cells were maintained as an adherent monolayer in 25-cm2 tissue culture flasks in Dulbecco's modified Eagle's medium and 10% FCS. For vector sensitivity experiments, cells were first trypsinized and seeded onto 96-well plates (1 × 104 cells per dish) and then allowed to adhere for 24 h. Cells were then incubated with vectors for 1 h; untreated cells received identical manipulation without vector. In mock experiments, cells were treated identically with the AdeGFP vector. Cells were then washed once with phosphate-buffered saline and fresh medium was added. Cell growth was measured daily using the 3-(4,5-dimethythiazl-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)2H-tetrazolium inner salt (MTT) test.

Apoptosis evaluation

Cells were seeded onto 3.5 cm dishes to 50% confluence before transduction. Cells were transduced at an MOI of 100 with either AdeGFP or AdRhoB and left for 36 h. For cell cycle analysis, cells were fixed in cold methanol, RNase treated, and stained with propidium iodide (50 μg ml−1). Cells were analyzed for DNA content by flow cytometry (FACS-Calibur; Becton Dickinson, Pont de Claix, France). Results were expressed as percentages of elements detected in the sub-G1 peak (apoptosis). For Apo2.7 expression analysis, cells were fixed, permeabilized and analyzed by flow cytometry after staining with either isotypic control, or Apo2.7 (2.7A6A3 clone) which reacts with a 30 kDa mitochondrial membrane protein (7A6 antigen) reported as being exposed on cells undergoing apoptosis.26

All analyses were performed in duplicate.

Animals

Female Swiss athymic nude mice, 4–5 weeks old (Charles River Laboratories, Wilmington, MA, USA), were housed in filter-capped cages and kept in a sterile facility. All the mice used in this study were cared for and treated in accordance with the American Association for Cancer Research (AACR, Philadelphia, PA, USA) statement for the use of animals in cancer research. A 2-week quarantine period was imposed on all animals before starting the study.

OVCAR-3 animal model

The study used the OVCAR-3 tumor model as originally established and described previously.27 A xenograft of OVCAR-3 tumor cells (or of enhanced GFP (eGFP)-expressing OVCAR-3) was produced in the mice by i.p. implantation. The xenograft was fixed by irrigating the peritoneal cavity with normal saline and combining the wash and ascites. The cells were washed twice in phosphate-buffered saline, the pellet was resuspended and the suspension was diluted 1:3 in normal saline. Each mouse received i.p. 1 ml of the cell suspension, representing 10–12 × 106 cells.

Tumor progression experiment

The OVCAR-3 ascites were either (i) transduced, or not (control; n=5), 1 h before injection into the mice (establishment model) by AdeGFP (mock; n=5) or AdRhoB (treatment; n=5) or (ii) allowed to grow for 10 days with mice randomly assigned to different treatments, repeated every 2 days for 6 days (three treatments), of saline (control; n=5), AdeGFP i.p. (mock; n=5) or AdRhoB (treatment; n=5).

Experiments were performed 2–3 times. The efficacy of treatment was evaluated by the presence of macroscopic signs of peritoneal ascites in the animal.

Mice with peritoneal ascites were inspected once daily for assessment of overall clinical status and water and food intake. Those in extremis were killed immediately.

Fluorescence stereomicroscopy imaging

Mice were euthanized for direct internal imaging. A long midline incision was made to access the abdominal cavity. GFP fluorescence in the tumors was detected with a Leica MZFL III fluorescence stereomicroscope (Leica Microsystems, Wetzlar, Germany). High-resolution 16-bit images of 1392 × 1040 pixels were captured on a Dell PC using a thermoelectrically cooled charge-coupled device camera, model coolSNAP HQ (Roper Scientific, Evry, France). An × 8 magnification was used to visualize the tumors and the micrometastases. Selective excitation was produced with a compact arc source EL6000 (Leica Microsystems, SAS Rueil-Malmaison, France) and a GFP filter (Leica Microsystems). Color images were obtained using a Micro Color tunable RGB filter (CRI, Inc., Woburn, MA, USA). The images were processed for contrast and brightness using MetaVue 6.2 software (Universal Imaging Corp., Downingtown, PA, USA).

Statistical analysis

Results of in vitro experiments were expressed as mean±s.d. Student's two-sided t-test was used to compare values of test and control samples. P<0.05 indicated a significant difference.

Results

Overexpression of RhoB in transcomplementing or packaging cell lines (293A or 293 T cells) used to produce adenoviral or lentiviral vectors encoding RhoB induces cell death. With this in mind we cloned RhoB cDNA inside the pAdenoVATOR-CMV5(CuO)-ires GFP plasmid, which includes a CMV promoter that can be repressed by the repressor cymR. We were then able to synthesize the adenoviral vector encoding RhoB produced by cymR-expressing 293 cells (AdRhoB). These adenoviral vectors efficiently transduced different types of epithelial cells, which then expressed high amounts of ectopic RhoB as well as the fluorescent eGFP, without any complementing system. We used the mock vector encoding only eGFP as a control (AdeGFP).

AdRhoB transduction induces strong RhoB expression in ovarian cancer cell lines

Immunoblot analysis showed that OVCAR-3 or IGROV-1 ovarian adenocarcinoma cells expressed low endogenous levels of RhoB in contrast to the abundant levels expressed in HeLa cells (Figure 1a). Ovarian cancer cells were transduced with AdRhoB at different MOI (10, 50 and 100). At an MOI of 100, 100% of cells from each cell line expressed the adenoviral transgenes as assessed by visualizing eGFP expression by fluorescence microscopy (data not shown). Immunoblot analysis 48 h after transduction showed strong RhoB and eGFP overexpression in OVCAR-3 and IGROV-1-transduced cells (Figure 1b).

Figure 1
figure 1

Expression of RhoB protein. Expression of endogenous (a) and of ectopic (b) RhoB and green fluorescent protein (GFP) after transduction with AdRhoB in ovarian adenocarcinoma cells, as analyzed by western blotting.

Full size image

AdRhoB transduction inhibits ovarian cell proliferation

To evaluate the effect of RhoB ectopic expression in ovarian cancer cells, we measured cell proliferation by MTT assay at 24, 48 and 72 h post-transduction with AdRhoB or AdeGFP. We observed no notable modification in cell proliferation following AdeGFP transduction of OVCAR-3 cells or IGROV-1 cells (Figures 2a and b). In contrast, AdRhoB transduction of both OVCAR-3 and IGROV-1 cells effectively abolished cell proliferation at an MOI of 100 particles infectious per cell, within 48 h (Figures 2a and b). Moreover, as shown in Figure 2b, AdRhoB transduction of IGROV-1 at an MOI of 100 induced an important decrease in cell number, suggesting that the majority of IGROV-1 cells died after 72 h.

Figure 2
figure 2

Effect of RhoB expression on ovarian adenocarcinoma cell growth in vitro. Growth of OVCAR-3 (a) and IGROV-1 (b) control cells (WT) and after transduction at an MOI (multiplicity of infection) of 100 at day 1 with AdeGFP (GFP, green fluorescent protein) or AdRhoB (RhoB) was determined daily by colorimetric assay with MTT.

Full size image

AdRhoB transduction induces apoptotic cell death

To assess the mechanism by which RhoB induces cell death, we analyzed the effect of RhoB overexpression on cell cycle repartition by fluorescence-activated cell sorting after transduction of ovarian cells at an MOI of 100 with AdGFP or AdRhoB. As a control, we used etoposide, which is known to induce apoptosis in ovarian cancer cells.28 Both AdRhoB transduction and etoposide treatment induced a significant increase in the number of cells in the sub-G1 phase (from 18 to 24 h maximum) (Table 1).

Table 1 Percentage of apoptotic OVCAR-3 cells 18 h after either adenoviral vector transduction or etoposide exposure
Full size table

The expression of Apo2.7, a mitochondrial protein, represents an early event in apoptosis.26 We therefore used the Apo2.7 antibody to directly quantify the number of apoptotic cells by flow cytometry. At 36 h following transduction, we observed an increase in the number of positive Apo2.7 cells in both etoposide-treated cells and AdRhoB-transduced cells compared to AdGFP control cells (Table 1).

To confirm the apoptotic effect of RhoB overexpression, we assessed the expression of downstream apoptotic effectors in vitro. As shown in Figures 3a and b, 48 h after transduction, we found an increased level of cleaved caspase-3 (17 kDa form, CPP32) in etoposide-treated OVCAR-3 cells and both AdRhoB-transduced OVCAR-3 and IGROV-1 cells compared to AdGFP control cells.

Figure 3
figure 3

Immunoblot detection of cleaved caspase-3 in ovarian adenocarcinoma cells overexpressing RhoB. Relative levels of cleaved caspase-3 protein were measured by immunoblot in lysates of OVCAR-3 (a) or IGROV-1 cells (b) 48 h post-transduction with AdGFP or AdRhoB at an MOI (multiplicity of infection) of 100:1. The expression of cleaved caspase-3 protein by OVCAR-3 cells treated 36 h with increasing concentrations of etoposide is shown as a control of apoptosis induction (a). One representative experiment out of three is shown.

Full size image

Altogether these data show that overexpression of RhoB via adenoviral transduction can induce apoptosis of ovarian adenocarcinoma cells within 24 h post-transduction.

Ex vivo AdRhoB transduction of OVCAR-3 cells impairs tumor growth

We examined the effect of a single transduction of AdRhoB on the establishment of human ovarian cancer xenografts in immunodeficient mice. Following i.p. injection of 107 OVCAR-3 cells, all mice developed adenocarcinoma with a high amount of ascites. This adenocarcinoma induced the death of the mice within 4 weeks. OVCAR-3 cells from the ascites were then injected directly (control) or transduced ex vivo with AdRhoB or AdeGFP at an MOI of 100 before injection 1 h later into naïve mice. We observed no difference between the control and the AdeGFP-treated mice groups with respect to the severity of tumor progression (P=0.33). In contrast, 75% of the AdRhoB-treated mice showed no tumor development, as shown in Figure 4. Thus, a single transduction of the adenoviral vectors before tumor cell injection was sufficient to abolish tumor progression in the majority of mice.

Figure 4
figure 4

Effect of RhoB expression on ovarian adenocarcinoma cell growth in vivo. The number of ascites-free nude mice after AdRhoB-transduced OVCAR-3 peritoneal xenografts. Results represent cumulative data from three separate experiments (five mice per group).

Full size image

AdeGFP distributes among the whole peritoneal cavity

The entry of Ads into target cells represents a rate-limiting step in ectopic gene transfer and determines the therapeutic efficacy of Ad-based gene therapy, even in murine models. For this purpose, at first we examined the biodisponibility of the adenoviral vector following its injection. Here, we aimed to compare the localization of tumoral cells with that of transduced cells following injection. To assess the localization of tumoral cells, we injected mice i.p. with 107 eGFP-OVCAR-3 cells, which later developed adenocarcinoma. We assessed the localization of eGFP-OVCAR-3 cells in the peritoneal cavity at different times post i.p. injection using fluorescence stereomicroscopy imaging. As shown in Figure 5a, most of the organs inside the peritoneal cavity (ovary, uterine cord, gall bladder, spleen, retroperitoneum, intestine, parietal layer and liver) presented tumor cells after 8 days.

Figure 5
figure 5

Fluorescence analysis of tumor-bearing mice by in vivo stereomicroscopy. (a) Enhanced green fluorescent protein (eGFP) expressing OVCAR-3 cells (107) were injected intraperitoneally (i.p.) into nude mice. Mice were killed 8 or 13 days post-injection and analyzed by fluorescence stereomicroscopy. (b) Ten days after the i.p. injections of 107 parental OVCAR-3 cells (ascites), mice were injected i.p. with AdeGFP (108 particles in 500 μl saline solution). Mice were killed 36 h post-injection and analyzed by fluorescence stereomicroscopy. Representative images of eGFP fluorescence in the stomach, liver, ovary and uterus are presented. (Int, intestine; Lc, uterine left cord; liv, liver; Ov, ovary; Pa, pancreas; Pl, parietal layer; Rc, uterine right cord; retro, retro peritoneum; Sp, spleen; st, sternum; sto, stomach; Ut, uterus).

Full size image

In parallel, using fluorescence stereomicroscopy imaging, we assessed the potential of an Ad5ΔE1/ΔE3 to gene modify cells in the peritoneal cavity using a single i.p. injection of 108 AdGFP particles infectious, in ovarian cancer-bearing mice. We were able to detect gene-modified cells (that is, eGFP positive) in almost all the peritoneal cavity as shown in Figure 5b. At 36 h post i.p. injection of AdGFP particles, the stomach, liver, ovary and uterus (Figure 5b) all contained gene-modified cells. Thus, adenoviral vectors appear suitable to gene modify a high amount of cells in the peritoneal cavity. The transducable organs were also the targets of tumoral cell implantation.

Intraperitoneal AdRhoB injection inhibits OVCAR-3 tumor growth in mice

Tumor-bearing mice received three i.p. injections of 109 particles infectious of AdRhoB or AdeGFP. We observed no difference between the control and the AdeGFP-treated groups with respect to the severity of tumor progression (P=0.33). In contrast, AdRhoB treatment efficiently cured 50% of mice (P<0.001) (Figure 6). The delay of relapse and optimal survival were achieved with an AdRhoB dose of 109 viral particles, every 2 days for 1 week. Increasing the number of injections did not improve the overall survival (data not shown). This prompted us to propose that the ovarian cancer cells of the 50% of mice bearing progressive disease were resistant to RhoB-induced apoptosis. To evaluate this hypothesis, ascites from AdeGFP- or AdRhoB-treated mice showing relapse from their treatment were collected as described above. Cells were pelleted, seeded and cultured and then analyzed for the expression of eGFP and RhoB. We observed that they no longer expressed eGFP and displayed RhoB levels similar to the parental cells. AdRhoB transduction of these cells in vitro inhibited their proliferation.

Figure 6
figure 6

Kinetics of the non-ascites-bearing nude mice after OVCAR-3 peritoneal xenograft and treatment by AdRhoB. Number of the ascites-free nude mice after zOVCAR-3 peritoneal xenografts and treatment by AdRhoB (intraperitoneal route). Results represent cumulative data from two separate experiments (five mice per group).

Full size image

The 50% success rate could thus be attributed to the incapacity of the adenoviral vectors to gene modify all of the ovarian cancer cells in the peritoneal cavity.

Discussion

The current absence of treatments for ovarian cancer has driven large amounts of interest in developing new experimental therapies. The correction of specific defects responsible for the aberrant biological behavior of cancer cells represents a fascinating new approach to cancer treatment. The high frequency of p53 mutations in human cancers and the central role of p53 in regulating growth and apoptosis led researchers to initiate phase I clinical trials investigating p53 as the target for gene replacement therapy. However, these strategies are currently difficult to develop since mutated forms of p53 expressed by cancer cells are able to abolish the therapeutic effect of the wild-type p53 overexpression.18, 24

Malignant transformation is known to be a multistep process involving changes in several genes. In addition to p53, aberrations of c-erbB2, c-myc and K-ras play an important role in ovarian cancer.14, 29 Numerous studies have indicated a repression of GTPase RhoB in a significant number of cancers including ovarian cancer.7, 14, 16 Moreover, the ability of RhoB to inhibit the tumoral growth of many cancer cell lines both in vitro and in vivo6, 7 suggests that RhoB may play a tumor suppressor role.5 Here, we examined whether in vivo restoration of RhoB expression in RhoB downregulated ovarian cancer cells would reverse malignancy. We overexpressed RhoB in two ovarian cancer cell lines via AdRhoB transduction and observed a dramatic loss of tumorigenicity.

Overexpression of RhoB, following virus transduction, strongly affected the proliferation of the two cell lines in vitro as shown by the significant decrease in cell number after only 24 h of RhoB overexpression. In comparison to control cells, IGROV-1 cells appeared to be more sensitive than the OVCAR-3 cells despite the fact they expressed similar RhoB levels after adenoviral transduction. It will be interesting to examine which pathways are modified through RhoB overexpression in these two cell lines and define the signal events occurring directly downstream of RhoB expression following RhoB infection or induction.

RhoB is a small GTPase that facilitates stress-induced apoptosis in oncogenically transformed cells and limits cancer progression. Here, we have shown that overexpression of the RhoB gene in the two cell lines tested, IGROV-1 and OVCAR-3, resulted in the induction of apoptosis in vitro, as shown by the fraction of cells with sub-G1 DNA content, by the cleavage of caspase-3 and the expression of the Apo2.7 antigen. Several studies have already shown that RhoB has a pivotal role in the apoptotic response of neoplastic cells to DNA damage.7, 10, 12 However, to our knowledge this study represents the first demonstration that the overexpression of RhoB on its own using viral adenoviral transduction could induce apoptosis of cultured tumor cells.

The mechanisms used by RhoB to sensitize cells to cell death are unclear. One microarray hybridization analysis suggested that RhoB may modify the p53 response, with p53 and p53 targets identified as targets of RhoB. Moreover, p53 has previously been reported as functionally connected to the RhoGTPase network.30 We therefore compared the status of p53 expression in OVCAR-3 or IGROV-1 cells before and after AdeGFP or AdRhoB transduction. We observed neither p53 expression nor p53 phosphorylation in either of the two cell lines following RhoB overexpression, while a marked stimulation of both parameters was detected after treatment with the well-known p53 inducer, etoposide. Thus, the cytotoxic effect of RhoB overexpression does not appear to involve the p53 pathway, in contrast to the response to DNA-damaging agents.

Concerning the therapeutic use of RhoB transduction using an adenoviral vector, ovarian cancer is regarded as a particularly good candidate with the disease generally remaining confined to the abdominal cavity throughout its course, thereby enabling i.p. delivery of the therapy. This approach has the advantage of providing high viral loads close to the tumor. We first showed by fluorescence stereomicroscopy imaging that a single injection of adenoviral vectors could be sufficient to target almost all organs in the peritoneal cavity.

We observed efficient suppression of the in vivo tumorigenicity of the aggressive OVCAR-3 cell line by AdRhoB transduction in 75% of the mice. Then we have demonstrated the possibility of curing cancer by killing adenocarcinoma cells after transduction with an Ad vector encoding RhoB. This has provided the first proof-of-concept study on the use of gene transfer of RhoB as a potential treatment for ovarian cancer. However, only 50% of the mice in the present study showed a complete recovery. One theory could be that some tumor cells in AdRhoB-treated mice escaped the toxic effect of RhoB. However, this theory was denied by the fact that the cells from the ascites remained sensitive to RhoB-induced apoptosis. Thus, despite the fact that eGFP-positive cells were detected in the entire peritoneal cavity, we are unable to exclude the possibility of the Ads targeting only a fraction of cells. Indeed, the challenges facing the clinical implementation of mutation compensation approaches include the need to improve the efficiency of vector systems to increase the percentage of cells that are transduced by the therapeutic gene. Binding of Ads to the cell surface is mediated by the coxsackie Ad receptor, and internalization is achieved via interactions with integrins of the αvβ3 and αvβ5 classes. Deficiency in one or both of these membrane proteins conferring resistance to adenoviral vectors has been reported. Marked differences in coxsackie Ad receptor expression have been reported in primary cultured ovarian cancer cells.24 This underlines the rate-limiting factor of Ad delivery in ectopic gene transfer, which determines the therapeutic efficacy of Ad-based gene therapy. To improve the efficacy of adenoviral transduction, identification of the genetic and epigenetic signature of the target tumor, including the expression of cell surface molecules governing the cellular uptake of the adenoviral vector may provide the information needed to adapt a multivector treatment to each individual tumor. On the other hand, as a high percentage of ovarian cancer cells are p53 negative, oncolytic adenoviral vectors that target p53 cells (such as dl1520 (Onyx-015) containing RhoB) could be used to potentially increase the efficacy of RhoB therapy.31 Another strategy could be the use of a chimeric vector such as Δ24-RDG vector, which is an adenoviral vector deleted in amino acids 121–128 and that has been modified further to include a RDG-4C peptide into the adenoviral fiber. These vectors permit infection of cells independent of the normal coxsackie Ad receptor that is frequently expressed at very low levels on ovarian carcinoma cells.32

In conclusion, the restoration of RhoB expression in ovarian cancer is followed by induction of apoptosis in vitro and the suppression of tumorigenicity in vivo, providing evidence in favor of in vivo reactivation of the RhoB signaling pathway as a potential target for ovarian cancer therapy. Owing to the prevalence of RhoB pathway inactivation in human cancers, several pharmacological strategies aimed at restoring RhoB function must be proposed as gene therapy approaches planning to introduce a wild-type copy of the RhoB gene into tumor cells. However, as it has long been recognized that the development of invasive malignancy requires multiple genetic events with tens if not hundreds of genes that may be aberrantly expressed in malignant cells,33 it is clear that a single mutation compensation could difficultly correct the behavior of a cancer cell unless if this compensation induces the death of the cell. As a single gene mutation correction could allow also the sensitization of the cell to chemo- or radiotherapies the results presented here support efforts to treat human cancers by way of pharmacological reactivation of RhoB.