A cyclooxygenase-2 promoter-based conditionally replicating adenovirus with enhanced infectivity for treatment of ovarian adenocarcinoma


Conditionally replicating adenoviruses (CRADs) take advantage of tumor-specific characteristics for preferential replication and subsequent oncolysis of cancer cells. The antitumor effect is determined by the capability to infect tumor cells. Here, we used RGDCRADcox-2R, which features the cyclooxygenase-2 promoter for replication control and an integrin binding RGD-4C motif for enhanced infectivity of ovarian cancer cells. RGDCRADcox-2R replicated in and killed human ovarian cancer cells effectively, while the replication in nonmalignant cells was low. Importantly, the therapeutic efficacy, as evaluated in an orthotopic model of peritoneally disseminated ovarian cancer, was significantly improved and toxicity was lower than with a wild-type virus. Thus, this CRAD could be tested for treatment of ovarian cancer in humans.


Adenovirus (Ad)-mediated gene therapy has been proposed as a treatment alternative for advanced cancers refractory to conventional therapies. Ads are attractive vectors for cancer due to their unparalleled capacity for gene transfer, stability in vivo and the feasibility of high titer production. However, currently there is little evidence supporting significant clinical benefits.1 This might be due to inefficient transduction of solid tumor masses with replication-deficient Ad-vectors. To help overcome this obstacle, selectively oncolytic agents, that is, conditionally replicating adenoviruses (CRADs), have been constructed. Infection of tumor cells results in replication, oncolysis and subsequent release of the virus progeny. Normal tissue is spared due to lack of replication, which can be achieved by two strategies. Type 1 CRADs have a partial deletion in the immediately-early (E1A) or early (E1B) adenoviral genes resulting in a mutant E1 protein unable to bind cellular proteins necessary for viral cell cycle in normal cells, but not in cancer cells. Another strategy is the control of viral replication with various heterologous tumor/tissue-specific promoters, TSPs.2,3 We constructed a CRAD containing the cyclooxygenase-2 (cox-2) promoter for controlling expression of E1A.4 The expression cassette is inverted to minimize the influence by the left inverted terminal sequences and packaging signal. Cox-2 has been shown to be highly expressed in a number of epithelial tumors, including ovarian carcinoma. Furthermore, it is closely linked to carcinogenesis and progression of epithelial tumors.5 There are studies suggesting increased cox-2 gene expression after infection with various pathogens.6,7,8 However, with regard to Ad, this has not been studied. Also, an Ad featuring the cox-2 promoter has displayed specific transgene expression in cox-2-positive cancer cells, while the expression in liver and mesothelial cells was low.9,10,11 Recently, it has been demonstrated that the oncolytic potency of replicating agents is directly determined by their capability to infect target cells.12 Unfortunately, it has also been shown that expression of the Ad primary receptor, coxsackie-adenovirus receptor (CAR), is highly variable and often low on various tumor types, including ovarian cancer (reviewed in Bauerschmitz et al3). Further, there may be an inverse correlation between CAR expression and tumor grade.13 Thus, low CAR levels may hinder CRAD-mediated oncolysis. Therefore, methods to circumvent CAR-deficiency and improve cell killing have been evaluated in the context of CRADs. The concept has been validated by incorporation of an arginine–glycine–aspartic acid (RGD-4C) modification into the HI-loop of the fiber of a type 1 CRAD (Ad5-Δ24RGD).14 Subsequently, we studied the replication and oncolytic potency of this agent on ovarian cancer substrates.15 For the purpose of clinical testing with ovarian cancer patients, we have attempted to identify the best available infectivity enhanced CRAD. An RGD-4C-modified cox-2-based CRAD (RGDCRADcox-2R) was constructed and compared to Ad5-Δ24RGD, which has displayed preclinical efficacy in ovarian cancer. Another control was a wild-type Ad5, which may be a useful control for optimal replication.


RGDCRADcox-2R kills ovarian cancer cells in vitro

Monolayers of SKOV3.ip1, OV-3, OV-4, ES-2 and Hey were infected with RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt or Ad5lucRGD (Figure 1). In all cell lines, the crystal violet staining-based cell killing assay showed almost total oncolysis with Ad5-Δ24RGD (Figure 1a–e). RGDCRADcox-2R caused greater (SKOV3.ip1) or similar (OV-3, OV-4, ES-2) cell killing as Ad300wt (Figure 1a–d). Only in Hey cells did RGDCRADcox-2R show no oncolysis (Figure 1e). Ad5lucRGD was included as an E1-deleted control, and it did not cause oncolysis. Another approach for assessing cellular viability is analysis of tumor cell mitochondrial activity by MTS assay (Figure 1f–j). This quantitative cell killing confirmed the crystal violet findings in three out of five cell lines, while in OV-4 and ES-2 cell viability measured with MTS assay remained high after infection with Ad300wt and RGDCRADcox-2R (Figure1h,j). On SKOV3.ip1 and OV-3 cells, oncolysis was statistically significant with RGDCRADcox-2R in comparison to Ad5lucRGD (P=0.0054 and P<0.0001, respectively). Further, RGDCRADcox-2R killed more SKOV3.ip1 cells than Ad300wt (P=0.0104). With SKOV3.ip1 and OV-3 cells, there was no significant difference between cell killing with RGDCRADcox-2R and Ad5-Δ24RGD (P=0.1343 and 0.0827, respectively). In the other ovarian cancer lines, Ad5-Δ24RGD displayed enhanced cell killing compared to RGDCRADcox-2R (P=0.0251, P=0.0001 and P < 0.0001 for OV-4, ES-2 and Hey, respectively).

Figure 1

RGDCRADcox-2R kills ovarian cancer cells. Cells were infected with RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt and Ad5lucRGD (E1-deleted control virus). (a–e) Oncolysis was evaluated by crystal violet staining in SKOV3.ip1, OV-3, OV-4, ES-2 and Hey cells, respectively. (f–j) Cell viability was measured with MTS assay. The OD490 values of uninfected cells were set as 100%. Data are expressed as the mean±s.d.

RGDCRADcox-2R displays cox-2 expression-dependent oncolysis and low cell killing of nonmalignant mesothelial cells

To demonstrate cox-2-specific oncolysis, cox-2-positive A5499 and -negative BT4744 cell lines were analyzed for cell killing. With A549 cells, RGDCRADcox-2R caused oncolysis comparable to Ad5-Δ24RGD and Ad300wt when assessed with crystal violet staining (Figure 2a). In the MTS assay at the highest viral dose, cell viability was 22, 5.9 and 0% with RGDCRADcox-2R, Ad5-Δ24RGD and Ad300wt, respectively (Figure 2b), and oncolysis with RGDCRADcox-2R was statistically significant compared to Ad5lucRGD (P<0.0001). RGDCRADcox-2R did not kill in BT474 cells (cell viability 82%), and the difference to Ad5lucRGD was not significant (P=0.3921), while other replicating Ads displayed oncolysis (Figure 2c, d). Cell viability was 12 and 0.2% with Ad5-Δ24RGD and Ad300wt, respectively (versus RGDCRADcox-2R, 0.0037 and 0.0002, respectively).

Figure 2

RGDCRADcox-2R demonstrates cox-2 expression dependent cell killing of control cells, and low cell killing of nonmalignant mesothelial cells. Cells were infected with RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt and Ad5lucRGD (E1-deleted control virus). (a, b) Oncolysis of cox-2-positive A549 cells, included as a positive control, and (c,d) cox-2 negative control BT474 cells, was evaluated by crystal violet staining and MTS assay. (e–g) Cell viability of human mesothelial cells measured by MTS assay, mesothelial cell line AG07086A, patient mesothelial samples 1 and 2, respectively. In the MTS assay, the OD490 values of uninfected cells were set as 100%. Data is expressed as the mean±s.d.

Human mesothelial cell line AG07086A and two primary mesothelial cell samples purified from peritoneal strips were analyzed by MTS assay (Figure 2e–g). At 10 viral particles (vp)/cell, the percentage of viable cells with Ad300wt was 39, 21 and 2.2%, respectively, as compared to uninfected wells. With Ad5-Δ24RGD, viability was 24, 60 and 15%, and with RGDCRADcox-2R 104, 126 and 42%, respectively. In patient sample 2, RGDCRADcox-2R did not display any cell killing at 1 vp/cell (112%), while Ad300wt and Ad5-Δ24RGD caused cell lysis (5.8 and 76% viability, respectively) (Figure 2g). Importantly, cell killing with RGDCRADcox-2R was not significantly different from the replication-deficient Ad5lucRGD in the three nonmalignant cell samples (P=0.5266, 0.5323, 0.0572, respectively), while the other replicating agents killed these nonmalignant cells more than RGDCRADcox-2R (all P-values <0.0270).

RGDCRADcox-2R copy number in ovarian cancer primary cell spheroids

Four purified, unpassaged clinical ovarian cancer cell samples were analyzed for CRAD DNA replication (Figure 3a–d, respectively). To determine the relative increase in copies with patient samples 1–3, E4 copy number at each time point was normalized to the copy number obtained with Ad5lucRGD at 12 h (Figure 3a–c). On day 4, RGDCRADcox-2R copy number had increased 228-, 123- and 5.4-fold in patient samples 1-3, respectively, compared to 456-, 26- and 20-fold increases with Ad5-Δ24RGD, and 135-, 3.6- and 57-fold increases with Ad300wt. On day 12, RGDCRADcox-2R copy number had increased to 1374-, 333- and 6.9-fold, while these figures were 1655-, 77- and 44-fold for Ad5-Δ24RGD, and 2026-, 141- and 138-fold for Ad300wt. Thus, replication of RGDCRADcox-2R was comparable to Ad5-Δ24RGD and Ad300wt (Figure 3a–c). In patient sample 4, the DNA copy number of also Ad5lucRGD increased (Figure 3d). If E4 copies on day 6 after infection were compared to values with Ad5lucRGD at 24 h, virus copy number had increased 166-, 801-, 104- and 54-fold with Ad5lucRGD, RGDCRADcox-2R, Ad5-Δ24RGD and Ad300wt, respectively. On day 18, these values were 29-, 204-, 273- and 1342-fold increase, respectively.

Figure 3

RGDCRADcox-2R replicates in three-dimensional human ovarian primary cancer cell spheroids. Purified, unpassaged cancer cells were allowed to form spheroids, which were infected with 1000 vp/cell of RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt and Ad5lucRGD (E1-deleted control virus). Spheroids and growth medium were harvested at indicated time points, and virus copy number was measured with quantitative PCR. (a–d) The increase of virus copy number over the time period in patient samples 1–4, respectively. (e–h) Cumulative values from patients 1–4. The background values (uninfected spheroids) were subtracted.

To allow visual comparison of replication dynamics, cumulative virus copy numbers16 were calculated as a sum of all previous time points. (Figure 3e–h). RGDCRADcox-2R produced 193-, 30-, 8.8- and 1.1-fold more virus than Ad5lucRGD (patients 1–4, respectively). Ad5-Δ24RGD showed 398-, 30-, 13- and 0.7-fold total virus production compared to Ad5lucRGD. Ad300wt achieved 361-, 31-, 32- and 3.0-fold higher virus DNA production than Ad5lucRGD.

RGDCRADcox-2R in an orthotopic murine model of peritoneally disseminated ovarian cancer

Advanced carcinomatosis was allowed to develop for 10 days and then mice were treated for three consecutive days with RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt, Ad5lucRGD or no virus (Figure 4a). For RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt, Ad5lucRGD and no virus, the median survivals of mice were 107, 106, 23, 35 and 37 days, respectively. In pairwise comparisons, survival with RGDCRADcox-2R was significantly enhanced when compared to Ad300wt, Ad5lucRGD or no virus (all P<0.0001). Survival of the mice treated with RGDCRADcox-2R versus Ad5-Δ24RGD was not significantly different (P=0.9148).

Figure 4

Therapeutic effect of RGDCRADcox-2R in an animal model of peritoneally disseminated ovarian cancer. SKOV3.ip1 cells were injected i.p. into SCID mice and advanced carcinomatosis was allowed to develop for 10 days. (a) The mice received three daily i.p. injections of 1 × 108 vp of RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt or Ad5lucRGD or (b) two daily i.p. injections of 5 × 109 vp. In both experiments, RGDCRADcox-2R resulted in significantly enhanced survival (versus Ad5lucRGD or no virus, all P-values <0.0001). (c) Investigation of early mortality after injection of Ad300wt, a wild type adenovirus. Virus copy number was measured by quantitative PCR of the liver on day 19. *For comparison, liver virus copy number was measured in mice treated with a high but non-lethal i.v. dose of 5 × 1010 vp of an E1-deleted virus with an identical wild-type fiber. (d) Histopathology with H&E staining of the livers from (c) showing massive hepatic necrosis (big arrow) and vascular leakage (small arrow) in Ad300wt-treated mice. Therefore, early mortality in the Ad300wt group was probably due to constant high level virus production in the tumors, vascular dissemination and subsequent uptake by the liver.

A higher dose divided into two injections on two consecutive days was also tested (Figure 4b). For RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt, Ad5lucRGD and no virus, the median survivals of mice were 78, 109, 73, 35 and 37 days, respectively. Pairwise χ2 testing confirmed significantly improved survival with RGDCRADcox-2R compared to Ad5lucRGD or no virus (both P < 0.0001). Interestingly, the survival of mice treated with RGDCRADcox-2R, Ad5-Δ24RGD or Ad300wt did not differ significantly (RGDCRADcox-2R versus Ad5-Δ24RGD, P=0.5847; versus Ad300wt, P=0.0936).

Mice treated with the wild-type serotype 5 strain Ad300wt displayed early mortality, especially with the triple injection schema (Figure 4a). Necropsy of these mice did not reveal large tumor load or ascites. Interestingly, the livers were abnormal in visual inspection and thus virus copy number was measured. High amounts of virus were detected in the livers of prematurely dead mice (Figure 4c). When compared to livers of mice injected with a much higher dose of Ad5luc1 (a replication-deficient virus with an identical capsid) the level of virus copies was two orders of magnitude higher (Figure 4c). Furthermore, histopathology of these livers was analyzed showing massive hepatic necrosis, acute and chronic inflammation and vascular leakage, while the liver of Ad5luc1 was within normal limits (Figure 4d).


Gene therapy strategies with viral vectors have shown promise in preclinical studies, but inefficient tumor transduction in clinical trials may have limited their clinical efficacy heretofore. Therefore, CRADs have been developed to address this limitation. In this study, we evaluated the efficacy of a CRAD utilizing the cox-2 promoter for controlling E1A expression. Cox-2 is the rate-limiting enzyme in prostaglandin synthesis, and it is involved in control of inflammatory reactions. In most tissues, the activity of cox-2 promoter is low, unless it is induced by growth factors, cytokines or tumor-specific factors. Recently, it has been shown that cox-2 is expressed in many epithelial cancers, and it is related to carcinogenesis and tumor angiogenesis.5 Expression of cox-2 in ovarian cancer cell lines and, more importantly, in patient tumor specimens has been demonstrated.17 Further, malignant ascite samples of ovarian cancer patients often feature increased levels of prostaglandin E2, suggesting activation of the cox-2 promoter.17 The cox-2 promoter retains its fidelity in Ad vectors,9 and is activated in ovarian cancer cell lines and primary cancer cells.10,11 Furthermore, its activity is low in the liver,9 which is important from a safety standpoint.

As the efficacy of replicating agents is determined by their infectivity, endeavors to circumvent the frequent deficiency of CAR on clinical cancers have been evaluated. One method is genetic modification of the fiber knob with an RGD-4C motif, which allows binding to αvβ integrins, which are regularly expressed and often overexpressed on tumor vasculature and ovarian cancer cells.18,19 RGD-4C- modified Ads have been shown to achieve enhanced gene expression in and therapeutic efficacy with low-CAR targets, including ovarian cancer.18,19,20 A further advantage of this tropism modification is reduced neutralization by pre-existing neutralizing anti-adenovirus antibodies.19,20 This could be important, as most ovarian cancer patients have malignant ascites containing such antibodies, which could compromise initial infection of tumor cells and spreading of new virions.19,21 With regard to preclinical assessment of biodistribution and toxicity, preliminary results suggest that the RGD-4C modification does not affect either aspect adversely, when virus is administered i.p. to mice with ovarian carcinomatosis.20

Here, we performed a preclinical evaluation of RGDCRADcox-2R in the context of ovarian cancer. It was also compared to a type 1 CRAD, Ad5-Δ24RGD, which has shown preclinical efficacy15 and is now undergoing clinical evaluation. Ad5-Δ24RGD contains a 24-bp deletion in the E1A, in the area of the E1A protein responsible for binding retinoblastoma protein. This binding normally allows Ad to induce S-phase entry, needed for virus replication. Therefore, Ad5-Δ24RGD replicates only in cells inactive in their Rb/p16 pathway, which includes most human cancers.22

We evaluated the cell killing efficacy of RGDCRADcox-2R in five ovarian adenocarcinoma cells lines, which are shown to express cox-2 or allow cox-2-controlled transgene expression.10,11,17 In a crystal violet staining assay, RGDCRADcox-2R displayed similar or enhanced oncolysis when compared to a wild-type Ad, while Ad5-Δ24RGD achieved almost total oncolysis with all cell lines (Figure 1a–e). In a quantitative cell killing assay, the efficacy of RGDCRADcox-2R was slightly lower. In general, cell killing correlated very well with cox-2 expression. One analyzed ovarian cancer cell line, Hey, did not display oncolysis with RGDCRADcox-2R (Figure 1e, j), although it allows cox-2-mediated gene expression.11 This cell line grows very fast, which may compensate for the cell lysis induced by RGDCRADcox-2R replication. Importantly, RGDCRADcox-2R caused only minimal lysis of non-malignant mesothelial cells, while the wild-type Ad and Ad5-Δ24RGD achieved significant cell killing (Figure 2). This suggests that the cox-2 promoter retains its fidelity in a CRAD context, as has been shown earlier with E1-deleted viruses.9,10 Taken together with the earlier report of low liver toxicity,4 this could represent an important safety feature, as all published ovarian cancer trials have utilized i.p. admininstration. RGDCRADcox-2R replicated in ovarian cancer primary cell spheroids to a level comparable to wild-type Ad (Figure 3). Replication of Ad5lucRGD DNA in patient sample 4 might be due to ‘E1A-like’ activity, a previously reported phenomenon present in many cancer cells, which allows replication of Ad vector DNA in the absence of E1A.16,23

Finally, the therapeutic efficacy of the agent was evaluated in an orthotopic animal model of ovarian cancer (Figure 4). Both CRADs displayed significantly enhanced survival in comparison to controls. Although Ad5-Δ24RGD was more oncolytic in vitro, there was no significant difference between these CRADs in vivo. However, the systems analyzed are different, in vitro most viral particles are expected to enter and exit cells, while in vivo there are tissue barriers and other factors associated with the three-dimensional structure of normal organs and cancer tissues. As compared to CRADs, mice treated with wild-type Ad experienced early morbidity and mortality, which was most prominent ca. 7 days after virus injection. Livers were harvested, and they displayed massive hepatic necrosis and high amounts of viral DNA suggesting virus replication, dissemination and subsequent liver uptake and toxicity as the cause of death. The appearance of symptoms approximately 1 week after virus administration suggests that virus replication proceeded effectively in the tumor tissue, virions disseminated into the bloodstream and finally were sequestered in the liver, which eventually failed. Interestingly, no such toxicity was seen for the RGD-4C-modified CRADs. More in-depth studies are needed, but perhaps the RGD-4C-modification allowed more effective entry into tumor cells and thus less viral circulation. Further, although human Ads do not effectively replicate in mice, perhaps the degree of Ad early gene expression may, nevertheless, affect mouse toxicity. Conceivably, RGDCRADcox-2R and Ad5-Δ24RGD would have reduced production of E1A due to low cox-2 activity and an intact Rb/p16 pathway, respectively.

In conclusion, we have demonstrated that RGDCRADcox-2R allows tumor-specific replication and cell killing comparable to wild-type Ad. Thus, RGDCRADcox-2R could be an effective agent for treatment of ovarian cancer, and other tumors expressing cox-2. If proven safe in clinical studies, RGDCRADcox-2R could be useful for treatment of other tumors featuring high expression of cox-2 and low expression of CAR. In some in vitro experiments, this agent was less oncolytic than Ad5-Δ24RGD. However, RGDCRADcox-2R also caused less toxicity to nonmalignant cells. Clinical trials may ultimately determine if these features are retained in humans and which is more important – specificity or efficacy. We feel that both Ad5-Δ24RGD and RGDCRADcox-2R are promising agents, and both agents have potential benefits. The distinct approaches for regulation of replication could be useful for avoiding potential resistant clones if patients were treated with alternating cycles of the agents.

Materials and Methods

Cells and tissues

293 cells were purchased from Microbix (Toronto, Canada), while 911 cells were courtesy of Dr van der Eb (University of Leiden, The Netherlands). Lung adenocarcinoma cell line A549, breast cancer cell line BT474 and ovarian adenocarcinoma cell lines OV-3 and ES-2 were obtained from the ATCC (Manassas, VA, USA). Ovarian adenocarcinoma cell lines SKOV3.ip1, Hey and OV-4 cells were obtained from Dr Price, Dr Wolf (both MD Anderson Cancer Center, Houston, TX, USA) and Dr Eberlein (Harvard Medical School, Boston, MA, USA). Human mesothelial cell line AG07086A was obtained from the Coriell Cell Repositories (Camden, NJ, USA). All cell lines were cultured in recommended conditions.

Primary mesothelial cells were obtained by enzymatic disaggregation of fresh peritoneal strips collected from female patients undergoing intra-abdominal surgery.20 Primary ovarian adenocarcinoma cells were purified with a previously described immunomagnetic-based method24 from the malignant ascites fluid of patients undergoing a procedure for ovarian cancer. To create three-dimensional spheroids, cells were suspended in growth medium in 3% agar coated flasks.25


RGDCRADcox-2R,4 Ad5-Δ24RGD,14 E1/E3-deleted viruses Ad5lucRGD18 and Ad5luc126 have been described. Ad300wt, a wild-type human Ad5, was obtained from the ATCC. All viruses were purified on cesium chloride gradients. The vp concentration was determined at 260 nm, and standard plaque assay on 293 cells was performed to determine infectious particles. The ratio of vp/infectious particles was 48, 50, 9.3, 73 and 5.2 for RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt, Ad5lucRGD and Ad5luc1, respectively.

Cell killing assays

Cells were infected with RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt or Ad5lucRGD as reported.25 Briefly, MTS assay was performed on 96-well plates, and viruses were diluted in 50 μl of 2% growth medium. For crystal violet staining, cells were plated and infected on six-well plates, and infection volume was 500 μl. Utilizing crystal violet staining15 or MTS assay,25 oncolysis was evaluated when RGDCRADcox-2R or Ad300wt showed clear oncolysis with lowest amount of virus. The results with RGDCRADcox-2R were compared to the other groups using a two-tailed t-test (SAS, v.8.2, SAS Institute, Cary, NY, USA).

Quantitating virus replication

Primary ovarian cancer cells were purified and cultured as spheroids overnight. The next day, spheroids were infected with 1000 vp/surface cell of RGDCRADcox-2R, Ad5-Δ24RGD, Ad300wt, Ad5lucRGD, or no virus. Then, the spheroids were divided into aliquots of 105 cells in Costar® 96-well ultra low attachment plates (Corning Inc., Corning, NY, USA). Cells and growth medium were harvested together and frozen at 0.5, 1, 2, 4, 6, 8, 12 and 16 days after infection (days 2–20 from patient 4). Purification of DNA and quantitative PCR for the E4 were performed as described.25

Therapeutic ovarian cancer model

Female CB17 SCID mice (UAB CFAR SCID Mouse Core Facility) were obtained at 3–4 weeks of age and quarantined for 2 weeks. Mice were kept under pathogen-free conditions according to the American Association for Accreditation of Laboratory Animal Care guidelines. Animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee of UAB. On day 0, mice were injected i.p. with 1x107 SKOV3.ip1 cells. On days 10, 11 and 12, mice were injected i.p. with 1 × 108 vp of RGDCRADcox-2R (n=9 mice), Ad5-Δ24RGD (n=9), Ad300wt (n=9), Ad5lucRGD (n=7), or no virus (n=6) in 1 ml of Opti-MEM (Mediatech, Herndon, VA, USA). In another experiment, mice were injected i.p. on days 10 and 11 with 5 x 109 vp of RGDCRADcox-2R (n=9 mice), Ad5-Δ24RGD (n=8), Ad300wt (n=9), Ad5lucRGD (n=7) or no virus (n=6) in 1 ml of Opti-MEM. Mice were followed daily and killed when there was any evidence of pain or distress. Survival data were plotted on a Kaplan–Meier curve, and the RGDCRADcox-2R group was compared with the other groups with the χ2 test (LIFETEST procedure in SAS v.8.2). The livers of mice that died early after treatment with Ad300wt were harvested on day 19 and fixed in 10% neutral buffered formalin. Serial paraffin-embedded sections were prepared and stained with hematoxylin and eosin (H&E). Histopathology was examined in a blinded fashion by two independent pathologists. Liver DNA was purified and E4 copy number was analyzed as above. Data were obtained also with Ad5luc1, an E1/E3-deleted Ad with a wild-type fiber, using a single injection of 5 × 1010 vp.


  1. 1

    Vorburger SA, Hunt KK . Adenoviral gene therapy. Oncologist 2002; 7: 46–59.

  2. 2

    Hemminki A, Alvarez RD . Adenoviruses in oncology: a viable option? BioDrugs 2002; 16: 77–87.

  3. 3

    Bauerschmitz GJ, Barker SD, Hemminki A . Adenoviral gene therapy for cancer: from vectors to targeted and replication competent agents (Review). Int J Oncol 2002; 21: 1161–1174.

  4. 4

    Yamamoto M . Infectivity enhanced, cyclooxygenase-2 promoter-based conditionally replicative adenovirus for pancreatic cancer. Gastroenterology 2003; 125: 1203–1218.

  5. 5

    Cao Y, Prescott SM . Many actions of cyclooxygenase-2 in cellular dynamics and in cancer. J Cell Physiol 2002; 190: 279–286.

  6. 6

    Huang Q et al. The plasticity of dendritic cell responses to pathogens and their components. Science 2001; 294: 870–875.

  7. 7

    Hirschowitz E, Hidalgo G, Doherty D . Induction of cyclooxygenase-2 in non-small cell lung cancer cells by infection with ΔE1, ΔE3 recombinant adenovirus vectors. Gene Therapy 2002; 9: 81–84.

  8. 8

    Steer SA et al. Regulation of cyclooxygenase-2 expression by macrophages in response to double-stranded RNA and viral infection. J Immunol 2003; 170: 1070–1076.

  9. 9

    Yamamoto M et al. Characterization of the cyclooxygenase-2 promoter in an adenoviral vector and its application for the mitigation of toxicity in suicide gene therapy of gastrointestinal cancers. Mol Ther 2001; 3: 385–394.

  10. 10

    Casado E et al. Strategies to accomplish targeted expression of transgenes in ovarian cancer for molecular therapeutic applications. Clin Cancer Res 2001; 7: 2496–2504.

  11. 11

    Barker SD et al. The secretory leukoprotease inhibitor (SLPI) promoter for ovarian cancer gene therapy. J Gene Med 2002; 5: 300–310.

  12. 12

    Douglas JT et al. Efficient oncolysis by a replicating adenovirus (ad) in vivo is critically dependent on tumor expression of primary ad receptors. Cancer Res 2001; 61: 813–817.

  13. 13

    Okegawa T et al. The mechanism of the growth-inhibitory effect of coxsackie and adenovirus receptor (CAR) on human bladder cancer: a functional analysis of car protein structure. Cancer Res 2001; 61: 6592–6600.

  14. 14

    Suzuki K et al. A conditionally replicative adenovirus with enhanced infectivity shows improved oncolytic potency. Clin Cancer Res 2001; 7: 120–126.

  15. 15

    Bauerschmitz GJ et al. Treatment of ovarian cancer with a tropism modified oncolytic adenovirus. Cancer Res 2002; 62: 1266–1270.

  16. 16

    Lam JT et al. Replication of an integrin targeted conditionally replicating adenovirus on primary ovarian cancer spheroids. Cancer Gene Ther 2003; 10: 377–387.

  17. 17

    Denkert C et al. Expression of cyclooxygenase 2 is an independent prognostic factor in human ovarian carcinoma. Am J Pathol 2002; 160: 893–903.

  18. 18

    Dmitriev I et al. An adenovirus vector with genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism. J Virol 1998; 72: 9706–9713.

  19. 19

    Hemminki A et al. An adenovirus with enhanced infectivity mediates molecular chemotherapy of ovarian cancer cells and allows imaging of gene expression. Mol Ther 2001; 4: 223–231.

  20. 20

    Kanerva A et al. Gene transfer to ovarian cancer versus normal tissues with fiber-modified adenoviruses. Mol Ther 2002; 5: 695–704.

  21. 21

    Hemminki A et al. Serum and ascites neutralizing antibodies in ovarian cancer patients treated with intraperitoneal adenoviral gene therapy. Hum Gene Ther 2002; 13: 1505–1514.

  22. 22

    Sherr CJ . Cancer cell cycles. Science 1996; 274: 1672–1677.

  23. 23

    O'Connor RJ, Hearing P . The E4-6/7 protein functionally compensates for the loss of E1A expression in adenovirus infection. J Virol 2000; 74: 5819–5824.

  24. 24

    Barker SD et al. An immunomagnetic-based method for the purification of ovarian cancer cells from patient-derived ascites. Gynecol Oncol 2001; 82: 57–63.

  25. 25

    Kanerva A et al. Enhanced therapeutic efficacy for ovarian cancer with a serotype 3 receptor targeted oncolytic virus. Mol Ther 2003; 8: 449–458.

  26. 26

    Kanerva A et al. Targeting adenovirus to the serotype 3 receptor increases gene transfer efficiency to ovarian cancer cells. Clin Cancer Res 2002; 8: 275–280.

Download references


We thank Drs Bin Liu and Minghui Wang (Division of Human Gene Therapy, University of Alabama at Birmingham) for statistical analysis and quantitative PCR assays. This work was supported by University of Helsinki Internal Funds, the Sigrid Juselius Foundation, Finnish Cancer Society, Biocentrum Helsinki, Emil Aaltonen Foundation, Maud Kuistila Foundation, Finnish Medical Foundation, Academy of Finland, Ida Montin Foundation, Biomedicum Helsinki-Foundation, the NIH (R01 CA94084, R01 CA83821, R01 CA93796, P50 CA83591), Susan B. Komen Foundation, Deutsche Forschungsgemeinschaft (BA2076/1-2), and University of Alabama Health Services Foundation. We thank Drs H Inoue and T Tanabe for providing a plasmid phPES3 containing the cox-2 promoter.

Author information

Correspondence to A Hemminki.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Kanerva, A., Bauerschmitz, G., Yamamoto, M. et al. A cyclooxygenase-2 promoter-based conditionally replicating adenovirus with enhanced infectivity for treatment of ovarian adenocarcinoma. Gene Ther 11, 552–559 (2004) doi:10.1038/sj.gt.3302181

Download citation


  • adenovirus
  • ovarian neoplasms
  • virus replication, biological therapy

Further reading