Carrier Cell–mediated Delivery of a Replication-competent Adenovirus for Cancer Gene Therapy

Abstract

Although replication-competent viruses have been developed to treat cancers, their cytotoxic effects are insufficient, as infection is inhibited by the generation of neutralizing antibodies. To address this limitation, we developed a carrier cell system to deliver a replication-competent adenovirus. Carrier cells infected with replication-competent adenovirus were incubated with target cancer cells in a high titer of anti-adenovirus antibody. Carrier cells were injected into syngeneic subcutaneous tumors after immunization with adenovirus. Carrier cell–derived cell fragments containing viral particles were engulfed by proliferative target cancer cells. This engulfment-mediated transfer of adenovirus was not inhibited by the anti-adenovirus antibody and enabled repetitive infection. After the induction of anti-adenoviral cytotoxic T-lymphocyte (CTL) responses by immunization with adenovirus, administration of carrier cells infected with a replication-competent adenovirus induced complete tumor regression. Adenovirus-GM-CSF augmented the anti-tumor effect of carrier cells by increasing anti-adenoviral and anti-tumoral CTL responses and decreased the number of injections of carrier cells required to induce complete tumor regression. This novel carrier cell–mediated viral transfection system might prove useful in a variety of cancer therapies. Although replication-competent viruses have been developed to treat cancers, their cytotoxic effects are insufficient, as infection is inhibited by the generation of neutralizing antibodies. To address this limitation, we developed a carrier cell system to deliver a replication-competent adenovirus. Carrier cells infected with replication-competent adenovirus were incubated with target cancer cells in a high titer of anti-adenovirus antibody. Carrier cells were injected into syngeneic subcutaneous tumors after immunization with adenovirus. Carrier cell–derived cell fragments containing viral particles were engulfed by proliferative target cancer cells. This engulfment-mediated transfer of adenovirus was not inhibited by the anti-adenovirus antibody and enabled repetitive infection. After the induction of anti-adenoviral cytotoxic T-lymphocyte (CTL) responses by immunization with adenovirus, administration of carrier cells infected with a replication-competent adenovirus induced complete tumor regression. Adenovirus-GM-CSF augmented the anti-tumor effect of carrier cells by increasing anti-adenoviral and anti-tumoral CTL responses and decreased the number of injections of carrier cells required to induce complete tumor regression. This novel carrier cell–mediated viral transfection system might prove useful in a variety of cancer therapies. Although more than 300 clinical trials of cancer gene therapy have been conducted, encouraging clinical results have yet to be obtained. Recently, replication-competent viral vectors have been developed to improve anti-tumor activity. However, two major concerns over the use of such vectors remain: frequent relapse despite temporary inhibition of tumor progression1Geoerger B Grill J Opolon P Morizet J Aubert G Terrier-Lacombe MJ et al.Oncolytic activity of the E1B-55kDa-deleted adenovirus ONYX-015 is independent of cellular p53 status in human malignant glioma xenografts.Cancer Res. 2002; 62: 764-772PubMed Google Scholar and generation of high neutralizing antibody titers that subsequently inhibit repetitive viral infection.2Dematteo RP Yeh H Friscia M Caparrelli D Burke C Desai N et al.Cellular immunity delimits adenoviral therapy strategies for the treatment of neoplastic diseases.Ann Surg Oncol. 1999; 6: 88-94Crossref PubMed Scopus (28) Google Scholar Repetitive infection is difficult to achieve, although anti-CD3 antibody,2Dematteo RP Yeh H Friscia M Caparrelli D Burke C Desai N et al.Cellular immunity delimits adenoviral therapy strategies for the treatment of neoplastic diseases.Ann Surg Oncol. 1999; 6: 88-94Crossref PubMed Scopus (28) Google Scholar polyethylene glycol,3Mok H Palmer DJ Ng P Barry MA Evaluation of polyethylene glycol modification of first-generation and helper-dependent adenoviral vectors to reduce immune responses.Mol Ther. 2005; 11: 66-79Abstract Full Text Full Text PDF PubMed Scopus (218) Google Scholar liposome,4Steel JC Cavanagh HM Burton MA Kalle WH Microsphere-liposome complexes protect adenoviral vectors from neutralising antibody without losses in transfection efficiency in-vitro.J Pharm Pharmacol. 2004; 56: 1371-1378Crossref PubMed Scopus (14) Google Scholar cyclophosphamide,5Bouvet M Fang B Ekmekcioglu S Ji L Bucana CD Hamada K et al.Suppression of the immune response to an adenovirus vector and enhancement of intratumoral transgene expression by low-dose etoposide.Gene Ther. 1998; 5: 189-195Crossref PubMed Scopus (54) Google Scholar,6Jooss K Yang Y Wilson JM Cyclophosphamide diminishes inflammation and prolongs transgene expression following delivery of adenoviral vectors to mouse liver and lung.Hum Gene Ther. 1996; 7: 1555-1566Crossref PubMed Scopus (143) Google Scholar and etoposide5Bouvet M Fang B Ekmekcioglu S Ji L Bucana CD Hamada K et al.Suppression of the immune response to an adenovirus vector and enhancement of intratumoral transgene expression by low-dose etoposide.Gene Ther. 1998; 5: 189-195Crossref PubMed Scopus (54) Google Scholar have been used in a attempt to overcome the humoral immune responses to viral vectors. Non-replicative virus-transfected cells have been administered in clinical trials for glioma7Ram Z Culver KW Oshiro EM Viola JJ DeVroom HL Otto E et al.Therapy of malignant tumors by intratumoral implantation of retroviral vector-producing cells.Nat Med. 1997; 3: 1354-1361Crossref PubMed Scopus (640) Google Scholar and malignant mesothelioma,8Schwarzenberger P Harrison L Weinacker A Marrogi A Byrne P Ramesh R et al.The treatment of malignant mesothelioma with a gene modified cancer cell line: a phase I study.Hum Gene Ther. 1998; 9: 2641-2649Crossref PubMed Google Scholar as well as in a pre-clinical study of metastatic breast cancer.9Garcia-Castro J Martinez-Palacio J Lillo R Garcia-Sanchez F Alemany R Madero L et al.Tumor cells as cellular vehicles to deliver gene therapies to metastatic tumors.Cancer Gene Ther. 2005; 12: 341-349Crossref PubMed Scopus (47) Google Scholar In these studies, cells were infected with viruses to produce bystander effects by ganciclovir triphosphate and 5-fluorouracil on adjacent target cancer cells. Only two studies of replication-competent virus–infected carrier cells have been reported: one of intraperitoneal injection of PA-1 ovarian cancer cells infected with a replication-competent herpes simplex virus-1 to treat intraperitoneal ovarian cancer xenografts10Coukos G Makrigiannakis A Kang EH Caparelli D Benjamin I Kaiser LR et al.Use of carrier cells to deliver a replication-selective herpes simplex virus-1 mutant for the intraperitoneal therapy of epithelial ovarian cancer.Clin Cancer Res. 1999; 5: 1523-1537PubMed Google Scholar and one of intravenous injection of MDA-MG-231 breast cancer cells infected with wild-type adenovirus to treat metastatic foci of breast cancer xenografts in lung.9Garcia-Castro J Martinez-Palacio J Lillo R Garcia-Sanchez F Alemany R Madero L et al.Tumor cells as cellular vehicles to deliver gene therapies to metastatic tumors.Cancer Gene Ther. 2005; 12: 341-349Crossref PubMed Scopus (47) Google Scholar However, these studies, which were performed in nude mice, did not yield successful anti-tumor effects, because the PA-1 and MDA-MG-231 carrier cells did not support sufficient viral replication and, owing to their fragility, often died before contacting target cancer cells and exhibiting bystander effects. Furthermore, no studies of syngeneic mice treated with replication-competent virus–infected carrier cells have yet been performed to simulate clinical trials for cancer gene therapy, and the mechanism of the anti-tumor effects of carrier cell systems remains unclear. The IAI.3B gene was isolated using polyclonal antibodies to a high-molecular-weight fraction derived from ovarian cancer.11Campbell IG Nicolai HM Foulkes WD Senger G Stamp GW Allan G et al.A novel gene encoding a B-box protein within the BRCA1 region at 17q21.1.Hum Mol Genet. 1994; 3: 589-594Crossref PubMed Scopus (106) Google Scholar The promoter activity of the IAI.3B gene is highly specific for ovarian cancer, and a replication-competent adenovirus termed AdE3-IAI.3B, in which the E1A gene is under the control of the human IAI.3Bpromoter, replicates as efficiently as the wild-type adenovirus in ovarian cancer cells but causes re-growth after temporary inhibition of growth of ovarian cancer cell tumors.12Hamada K Kohno S Iwamoto M Yokota H Okada M Tagawa M et al.Identification of the human IAI.3B promoter element and its use in the construction of a replication-selective adenovirus for ovarian cancer therapy.Cancer Res. 2003; 63: 2506-2512PubMed Google Scholar In the study reported here, we infected human lung carcinoma A549 carrier cells with AdE3-IAI.3B, and the adenoviral particle–containing cell fragments derived from the carrier cells were engulfed by target cancer cells. This novel non-receptor-mediated adenoviral transfection system circumvented neutralization by anti-adenovirus antibody and enhanced anti-tumor activity after the induction of anti-adenoviral cytotoxic T-lymphocyte (CTL) responses by pre-immunization with adenovirus in syngeneic mice. The replicative activities of AdE3-IAI.3B in ovarian cancer cells and various other types of cells were examined by determining the production of live virus particles 48 hours after infection. The wild-type adenovirus, AdE3,replicated in all the cells tested; AdE3-IAI.3Breplicated as efficiently as AdE3 in SW626, 293, and A549 carrier cells as well as in ovarian cancer cells but not in other cells tested (P < 0.05). Adenoviruses were produced in the supernatants of SW626, 293, and A549 carrier cells better than in those of ovarian cancer cells (P < 0.01) (Figure 1a). A549 cells expressed the same degree of IAI.3B promoter activity as the ovarian cancer OCC1 cell line (A549/OCC1 promoter activity ratio = 0.84 ± 0.09; OCC1 = 1). To determine which cell line is adequate for use as a carrier cell, the therapeutic activities of AdE3-IAI.3B-infected cells were determined by measuring the anti-proliferative effects on ovarian cancer PA-1 cells (Figure 1b). Each cell line was infected with AdE3-IAI.3B for 24 hours at a multiplicity of infection (MOI) of 200, which yielded infection of all cells, and then incubated with adherent PA-1 cells for 48 hours. Of the 16 cell lines used in this study, A549, 293, and SW626 cells exhibited the most potent anti-proliferative effects on PA-1 cells. The IC50 (50% growth inhibitory concentration) of AdE3-IAI.3B-infected cells was more strongly correlated with plaque-forming unit (PFU) level in the supernatants (r = –0.6262, P < 0.01) than that in cell lysates (r = −0.5336, P < 0.05). We therefore used A549, 293, and SW626 cells as carrier cells for AdE3-IAI.3B virus for the treatment of tumors. To evaluate the anti-tumor effects of AdE3-IAI.3B-infected A549, 293, and SW626 carrier cells, subcutaneous PA-1 and RMG-1 tumors 10–15 mm in diameter were established in the flanks of nude mice. Although the growth of small tumors 5 mm in diameter in nude mice was significantly retarded by AdE3-IAI.3B until day 30 after administration, all tumors re-grew thereafter. Furthermore, administration of AdE3-IAI.3B did not eradicate large tumors 10–15 mm in diameter, even temporarily. We therefore employed a large-tumor model to evaluate the anti-tumor effects of the carrier cells. The volume of PA-1 tumors was not significantly reduced by administration of AdE3-IAI.3B compared with administration of medium alone (Figure 2b) or A549, 293, and SW626 cells alone (data not shown). However, A549, 293, and SW626 carrier cells infected with AdE3-IAI.3B completely eradicated PA-1 tumors by 81, 65, and 113 days after the injections were started, respectively, and these tumors did not relapse over the subsequent 3-month period. Administration of A549, 293, and SW626 cells alone did not suppress the growth of RMG-1 tumors (data not shown). Administration of AdE3-IAI.3B suppressed the growth in volume of RMG-1 tumors by 68% compared with administration of medium alone (P < 0.01) (Figure 2b). SW626 carrier cells suppressed the growth in volume of RMG-1 tumors by 96% compared with medium alone (P < 0.01) and eradicated three of ten tumors. A549 and 293 carrier cells completely eradicated RMG-1 tumors by 53 and 37 days after the injections, respectively. The eradicated RMG-1 tumors did not relapse over the subsequent 3-month period. To evaluate the anti-tumor effects of AdE3-IAI.3B-infected carrier cells, subcutaneous tumors 5–8 mm in diameter were established in the thighs of (C57BL/6 × C3H/He) F1 mice using cognate mouse ovarian cancer OVHM cells. Mouse cancer autografts in syngeneic mice are proliferative compared with human cancer xenografts in nude mice. Mouse cancer cells have fewer human adenovirus receptors and one or two orders of magnitude lower PFU activity of human replicative adenovirus than do human cancer cells. Administration of adenoviruses leads to strong anti-adenovirus humoral immune responses in immunocompetent mice. In fact, mouse syngeneic tumors are resistant to replication-competent viral gene therapy.13Miller CG Krummenacher C Eisenberg RJ Cohen GH Fraser NW Development of a syngeneic murine B16 cell line-derived melanoma susceptible to destruction by neuroattenuated HSV-1.Mol Ther. 2001; 3: 160-168Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar On the other hand, mouse models can mimic human cancer patients. We therefore employed a mouse syngeneic ovarian cancer model to develop a novel method of human ovarian cancer therapy. As direct intra-tumoral injection of adenoviruses or carrier cells can kill experimental mice, we mixed carrier cells with 0.2% atelocollagen. Atelocollagen prevents the diffusion of adenovirus from sites of tumor injection and prolongs the period of local retention of adenoviruses because it is liquid at 4°C and a gel at 37°C.14Sano A Maeda M Nagahara S Ochiya T Honma K Itoh H et al.Atelocollagen for protein and gene delivery.Adv Drug Deliv Rev. 2003; 55: 1651-1677Crossref PubMed Scopus (167) Google Scholar The mortality rates of mice treated with AdE3-IAI.3B at 1 × 1010 PFU, with A549 carrier cells infected with AdE3-IAI.3B at an MOI of 200, and with 0.2% atelocollagen mixed with A549 carrier cells infected with AdE3-IAI.3B at an MOI of 200 were 22 ± 10% (n = 55), 16 ± 13% (n = 62), and 5.2 ± 5.1% (n = 58), respectively. Thus, atelocollagen mixed with A549 carrier cells significantly decreased the mortality rates obtained using AdE3-IAI.3B alone (P < 0.001) and A549 carrier cells alone (P < 0.05). To evaluate the effects of anti-adenoviral cellular and humoral immunity, mice were injected six times with A549 carrier cells infected with AdE3-IAI.3B after immunization with Ad-β-gal or UV-inactivated Ad-β-gal (Figure 3a). Non-infected A549 cells alone exhibited no significant anti-tumor activity compared with non-treated medium control. Survival of control mice was not significantly different from that of mice treated with AdE3-IAI.3B alone or A549 carrier cells alone. Survival time of mice treated with A549 carrier cells after injection of Ad-β-gal was significantly longer than that of mice treated with medium control, AdE3-IAI.3B, or A549 carrier cells alone (P < 0.05). Survival time of mice treated with A549 carrier cells after injection of UV-inactivated Ad-β-gal was significantly longer than that of mice treated with A549 carrier cells after injection of Ad-β-gal (P < 0.05). To examine the effects of multiple immunizations with adenovirus, Ad-β-gal was injected into mice once, twice, or three times, and then mice were injected intra-tumorally six times with AdE3-IAI.3B-infected A549 carrier cells 3 weeks after the last injection of Ad-β-gal (Figure 3b). Survival rates did not differ significantly among the three groups. To investigate synergistic effects of granulocyte–macrophage colony-stimulating factor (GM-CSF) in treatment with carrier cells, mice were treated with A549 carrier cells infected with AdE3-IAI.3B and/or AxCAmGM-CSF and injected three times into OVHM mouse tumors after immunization of the mice with Ad-β-gal or UV-inactivated Ad-β-gal (Figure 3c). Injection of A549 carrier cells infected with AxCAmGM-CSF alone did not prolong survival. In contrast, treatment with A549 carrier cells infected with AdE3-IAI.3B alone significantly prolonged survival compared with control mice (P < 0.05). Treatment with AdE3-IAI.3B- and AxCAmGM-CSF-infected A549 carrier cells significantly prolonged survival compared with treatment with AdE3-IAI.3B-infected A549 carrier cellsafter injection of Ad-β-gal (P < 0.01) or UV-inactivated Ad-β-gal (P < 0.05). To investigate the combined effects of 293 and A549 carrier cells, A549 and/or 293 carrier cells infected with AdE3-IAI.3B and AxCAmGM-CSF were injected once intra-tumorally (Figure 3d). The survival time of mice treated with A549 or 293 carrier cells without immunization was not significantly different from that of control mice. Immunization with Ad-β-gal or UV-inactivated Ad-β-gal significantlyprolonged the survival of mice treated with A549 carrier cells (P < 0.01, P < 0.01) and 293 carrier cells (P < 0.01, P < 0.05). After immunization with Ad-β-gal or UV-inactivated Ad-β-gal, combined administration of A549 and 293 carrier cells significantly prolonged the survival of mice compared with treatment with A549 carrier cells alone (P < 0.05, P < 0.05) or 293 carrier cells alone (P < 0.05, P < 0.01). Mice that exhibited complete tumor regression were resistant to subsequent inoculation of OVHM cells. To determine anti-adenovirus antibody titers after multiple injections of adenovirus, syngeneic mice were injected with Ad-β-gal once (week 0) or three times (weeks 0, 3, and 6). Multiple injections of Ad-β-gal induced an order of magnitude higher titer of anti-adenovirus antibody at weeks 6, 9, and 12 than single injection (P < 0.05) (Figure 4a). On the other hand, anti-adenovirus antibody appeared only 6 weeks after a single injection of UV-inactivated Ad-β-gal, and titers were two orders of magnitude lower than those after a single injection of Ad-β-gal (P < 0.001). To determine which carrier cells reverse the inhibition of infection by anti-adenoviral humoral immune response, AdE3-IAI.3B-infected carrier cells were incubated with target PA-1 cells in the presence of a high titer of anti-adenovirus antibody that completely inhibited infection by AdE3-IAI.3B at an MOI of 50,000 (Figure 4b). Each type of carrier cell infected with AdE3-IAI.3B significantly inhibited PA-1 cell growth compared with control (P < 0.001). A549 carrier cells exhibited the best growth inhibitory effect; they were six times as potent in this respect as 293 carrier cells (P < 0.01). To determine whether direct intercellular contact or fragmentation of carrier cells is related to overcoming inhibition of infection by anti-adenovirus antibody, A549 carrier cells and adherent target PA-1 cells were separated using a Millicell chamber in the presence of a high titer of anti-adenovirus antibody. Cell injury by three freeze-and-thaw cycles without the Millicell chamber decreased the anti-proliferative effect of carrier cells (P < 0.0001) (Figure 4c). The small-pore-sized Millicell chamber (0.4 μm) completely blocked the anti-proliferative effect of carrier cells and frozen-and-thawed carrier cells (P < 0.0001), and the large-pore-sized Millicell chamber (12 μm) partially blocked the anti-proliferative effect of carrier cells and frozen-and-thawed carrier cells (P < 0.0001, P < 0.001). These findings suggest that cell fragments, direct intercellular contact, and living carrier cells may contribute to the anti-proliferative effects of carrier cells in the presence of a high titer of anti-adenovirus antibody. To demonstrate CTL activities in mice immunized and treated with carrier cells, OVHM and Ad-β-gal-infected OVHM cells were used as target cells in a 51Cr release assay. Immunization with Ad-β-gal or UV-inactivated Ad-β-gal alone significantly increased cytotoxic effects on Ad-β-gal-infected OVHM cells (P < 0.01) but not on uninfected OVHM cells (Figure 4d). The cytotoxic effects on Ad-β-gal-infected OVHM cells did not differ significantly between mice treated with Ad-β-gal and those treated with UV-inactivated Ad-β-gal. Injection of uninfected A549 cells did not significantly affect CTL activities in control, UV-inactivated Ad-β-gal-injected, or Ad-β-gal-injected mice. Mice that exhibited complete tumor regression with treatment with AdE3-IAI.3B-infected A549 carrier cells after Ad-β-gal immunization exhibited significantly increased cytotoxic effects on OVHM and Ad-β-gal-infected OVHM cells compared with mice immunized with Ad-β-gal alone or UV-inactivated Ad-β-gal alone (P < 0.05). Administration of AdE3-IAI.3B- and AxCAmGM-CSF-infected A549 carrier cells after Ad-β-gal immunization further increased cytotoxic effects on OVHM and Ad-β-gal-infected OVHM cells compared with administration of A549 carrier cells infected with AdE3-IAI.3B alone after Ad-β-gal immunization (P < 0.05). To investigate the mechanism of infection of carrier cells in the presence of anti-adenovirus antibody, A549 carrier cells infected with AdE3-IAI.3B were incubated with target A549 and PA-1 cells. A549 cells were mainly used as target cells because PA-1 cells were fragile and easily damaged after infection with AdE3-IAI.3B, whereas A549 cells were not. Proliferative control A549 cells exhibited cytoplasmic processes (Figure 5a). AdE3-IAI.3B- infected A549 carrier cells exhibited blebbing (Figure 5b) and delivered AdE3-IAI.3B virus particle–containing cell fragments (Figure 5c). Necrotic A549 carrier cells still contained adenoviral particles in their cytoplasm and nucleus (Figure 5d). Viral particle–containing cell fragments were captured by cytoplasmic processes (Figure 5e and f), directly infected cytoplasmic process (Figure 5f), and were engulfed by target A549 cells (Figure 5g). This engulfment was also observed in target PA-1 cells (Figure 5i). The cell fragments were engulfed only in proliferative malignant cells and not in non-proliferative normal cells. The IAI.3B promoter–introduced replication-competent adenovirus AdE3-IAI.3B replicated as efficiently as wild-type adenovirus in ovarian cancer cells expressing high levels of IAI.3B12Hamada K Kohno S Iwamoto M Yokota H Okada M Tagawa M et al.Identification of the human IAI.3B promoter element and its use in the construction of a replication-selective adenovirus for ovarian cancer therapy.Cancer Res. 2003; 63: 2506-2512PubMed Google Scholar but not in squamous cell carcinoma, glioma, or normal cells expressing low levels of IAI.3B.12Hamada K Kohno S Iwamoto M Yokota H Okada M Tagawa M et al.Identification of the human IAI.3B promoter element and its use in the construction of a replication-selective adenovirus for ovarian cancer therapy.Cancer Res. 2003; 63: 2506-2512PubMed Google Scholar Intra-tumoral administration of AdE3-IAI.3B alone yielded no prominent therapeutic effects in nude mice with large xenografts; however, injection of AdE3-IAI.3B-infected A549 or 293 carrier cells into tumors produced significant anti-tumor effects. The anti-tumor effect of carrier cells in nude mice could be due to increases in local duration of retention and local concentration of AdE3-IAI.3B by delivery from carrier cells. Direct intra-tumoral administration of replication-competent adenoviruses such as ONYX-015 did not prevent the recurrence of treated tumors, although ONYX-015 transiently eradicated tumors in nude mice.1Geoerger B Grill J Opolon P Morizet J Aubert G Terrier-Lacombe MJ et al.Oncolytic activity of the E1B-55kDa-deleted adenovirus ONYX-015 is independent of cellular p53 status in human malignant glioma xenografts.Cancer Res. 2002; 62: 764-772PubMed Google Scholar More than 90% of adenoviruses intra-tumorally injected into nude mice migrate into the systemic circulation and accumulate in the liver.6Jooss K Yang Y Wilson JM Cyclophosphamide diminishes inflammation and prolongs transgene expression following delivery of adenoviral vectors to mouse liver and lung.Hum Gene Ther. 1996; 7: 1555-1566Crossref PubMed Scopus (143) Google Scholar Replication-competent adenovirus alone is thus unable to induce complete tumor regression in nude mice, although replication-competent adenovirus–infected carrier cells can induce it. Treatment with PEGylation,15Cryole MA Chirmule N Zhang Y Wilson JM PEGylation of E1-deleted adenovirus vectors allows significant gene expression on readministration to liver.Hum Gene Ther. 2002; 13: 1887-1900Crossref PubMed Scopus (151) Google Scholar liposome encapsulation,16Yotnda P Chen DH Chiu W Piedra PA Davis A Templeton NS et al.Bilamellar cationic liposomes protect adenovectors from preexisting humoral immune responses.Mol Ther. 2002; 5: 233-241Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar cyclophosphamide,5Bouvet M Fang B Ekmekcioglu S Ji L Bucana CD Hamada K et al.Suppression of the immune response to an adenovirus vector and enhancement of intratumoral transgene expression by low-dose etoposide.Gene Ther. 1998; 5: 189-195Crossref PubMed Scopus (54) Google Scholar,6Jooss K Yang Y Wilson JM Cyclophosphamide diminishes inflammation and prolongs transgene expression following delivery of adenoviral vectors to mouse liver and lung.Hum Gene Ther. 1996; 7: 1555-1566Crossref PubMed Scopus (143) Google Scholar and etoposide5Bouvet M Fang B Ekmekcioglu S Ji L Bucana CD Hamada K et al.Suppression of the immune response to an adenovirus vector and enhancement of intratumoral transgene expression by low-dose etoposide.Gene Ther. 1998; 5: 189-195Crossref PubMed Scopus (54) Google Scholar decreased the titer of neutralizing antibody against adenoviruses and yielded partial recovery of adenoviral infectivity in target cells on the second injection. However, such modifications were ineffective in the case of third or later injections, owing to an increase in anti-adenovirus antibody titer, and therefore could not produce significant anti-tumor effects in immunocompetent animals. Moreover, murine cancer cells are resistant to replication-competent viral gene therapy.13Miller CG Krummenacher C Eisenberg RJ Cohen GH Fraser NW Development of a syngeneic murine B16 cell line-derived melanoma susceptible to destruction by neuroattenuated HSV-1.Mol Ther. 2001; 3: 160-168Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar However, our carrier cell system completely eradicated murine tumors in syngeneic mice and produced significant anti-tumor effects despite a high titer of anti-adenovirus antibody. Electron-microscopic analysis demonstrated that A549 carrier cells bearing replication-competent adenoviruses delivered adenoviral particles through their cell fragments to target cells. Importantly, engulfment of cell fragments was observed only in proliferative malignant cells and not in non-proliferative normal cells. Although previous studies have shown that apoptotic cells are engulfed in fragments by dividing cancer cells but not by non-dividing normal cells,17Simamura E Hirai KI Shimada H Koyama J Apoptosis and epithelial phagocytosis in mitomycin C-treated human pulmonary adenocarcinoma A549 cells.Tissue Cell. 2001; 33: 161-168Crossref PubMed Scopus (17) Google Scholar this is to our knowledge the first demonstration of engulfment of adenovirus-infected cell fragments by dividing cancer cells. Carrier cell–mediated adenoviral infection consequently produced CTL responses to target cells as well as adenoviruses. Furthermore, mice with complete tumor regression obtained protective immunity. Replication-competent virus–infected carrier cell systems have previously been reported only in replication-competent herpes simplex virus-1 mutant–infected PA-1 cells10Coukos G Makrigiannakis A Kang EH Caparelli D Benjamin I Kaiser LR et al.Use of carrier cells to deliver a replication-selective herpes simplex virus-1 mutant for the intraperitoneal therapy of epithelial ovarian cancer.Clin Cancer Res. 1999; 5: 1523-1537PubMed Google Scholar and in wild-type adenovirus–infected MDA-MG-231 cells.9Garcia-Castro J Martinez-Palacio J Lillo R Garcia-Sanchez F Alemany R Madero L et al.Tumor cells as cellular vehicles to deliver gene therapies to metastatic tumors.Cancer Gene Ther. 2005; 12: 341-349Crossref PubMed Scopus (47) Google Scholar Since herpes simplex virus has a virion envelope that fuses with cell membrane and releases capsid into the cytoplasm, whereas adenovirus has no virion envelope,18Roizman B Knipe DM Herpes simplex viruses and their replication.Fields Virology. Lippincott Williams and Wilkins, Philadelphia, PA.2001: 2399-2459Google Scholar the method for cell fragment–mediated transfer of adenovirus we have developed is novel, is independent of adenovirus receptors, and is not blocked by anti-adenovirus antibody. A549 and 293 cells have been used as adenovirus-producing cells,19Imler JL Chartier C Dreyer D Dieterle A Sainte-Marie M Faure T et al.Novel complementation cell lines derived from human lung carcinoma A549 cells support the growth of E1-deleted adenovirus vectors.Gene Ther. 1996; 3: 75-84PubMed Google Scholar and our study also demonstrated that they produced and secreted large amounts of replication-competent adenovirus. Combined use of A549 and 293 carrier cells also yielded significant anti-tumor effects superior to those of either carrier cell population alone. Notably, as adenovirus begins to replicate in 6 and 12 hours following infection with 293 and A549 cells, respectively, 293 cells can support viral replication earlier than A549 cells.19Imler JL Chartier C Dreyer D Dieterle A Sainte-Marie M Faure T et al.Novel complementation cell lines derived from human lung carcinoma A549 cells support the growth of E1-deleted adenovirus vectors.Gene Ther. 1996; 3: 75-84PubMed Google Scholar GM-CSF is a cytokine with multiple biological effects, including enhancement of the cytotoxicity of CTLs. Clinical trials have revealed clear anti-tumor and immuno-modulating effects of GM-CSF gene therapy in treating renal cell carcinoma, melanoma, lung cancer, bladder cancer, and hematological malignancies.20Qin H Chatterjee SK Cancer gene therapy using tumor cells infected with recombinant vaccinia virus expressing GM-CSF.Hum Gene Ther. 1996; 7: 1853-1860Crossref PubMed Scopus (55) Google Scholar,21Mahvi DM Sondel PM Yang NS Albertini MR Schiller JH Hank J et al.Phase I/IB study of immunization with autologous tumor cells transfected with the GM-CSF gene by particle-mediated transfer in patients with melanoma or sarcoma.Hum Gene Ther. 1997; 8: 875-891Crossref PubMed Scopus (50) Google Scholar,22Shi FS Weber S Gan J Rakhmilevich AL Mahvi DM Granulocyte-macrophage colony-stimulating factor (GM-CSF) secreted by cDNA-transfected tumor cells induces a more potent antitumor response than exogenous GM-CSF.Cancer Gene Ther. 1999; 6: 81-88Crossref PubMed Scopus (42) Google Scholar GM-CSF-secreting tumor cells have been found to stimulate potent, specific, and long-lasting anti-tumor immunity, partly because GM-CSF stimulates the proliferation and differentiation of antigen-presenting cells.23Armstrong CA Botella R Galloway TH Murray N Kramp JM Song IS et al.Antitumor effects of granulocyte-macrophage colony-stimulating factor production by melanoma cells.Cancer Res. 1996; 56: 2191-2198PubMed Google Scholar,24Mastrangelo MJ Maguire Jr., HC Eisenlohr LC Laughlin CE Monken CE McCue PA et al.Intratumoral recombinant GM-CSF-encoding virus as gene therapy in patients with cutaneous melanoma.Cancer Gene Ther. 1999; 6: 409-422Crossref PubMed Scopus (290) Google Scholar Although the anti-tumor effects of GM-CSF alone were insufficient to eradicate tumors, our carrier cell experiments demonstrated that animals could be completely cured with injection of carrier cells co-infected with AdE3-IAI.3B and AxCA mGM-CSF after immunization with adenovirus. Of the 32 protocols for ovarian cancer gene therapy authorized by the Recombinant DNA Advisory Committee in the United States, only two have yielded partial responses in trials. This poor outcome could be due to intra-abdominal injection of vector, as almost all injected vector moves rapidly into the systemic circulation through the peritoneum. Since residual tumor in the peritoneal cavity greater than 2 cm in diameter is indicative of a poor prognosis in patients with ovarian cancer, direct intra-tumoral injection can improve the rate of cure. In this sense, a laparoscopic approach may be extremely useful for intra-abdominal ovarian tumor, although use of multiple intra-tumoral injections should be avoided to ensure patient compliance. Our carrier cell system, which consists of a single injection of A549 and 293 carrier cells infected with replication-competent adenovirus and adenovirus-GM-CSF, and which resulted in complete cure of animals, is thus a promising and clinically feasible means of treating ovarian cancer. Cell lines and adenoviruses. Human ovarian teratocarcinoma PA-1 cells; ovarian clear cell carcinoma RMG-1 and MH cells; ovarian adenocarcinoma OCC1, 420, and OVCAR3 cells; non–small cell lung carcinoma A549 cells; cervical squamous cell carcinoma ME-180, C33A, and HT-3 cells; glioma U87 cells; colon adenocarcinoma SW626 (ovarian metastasis) and HT-29 cells; normal keratinocyte K42 and skin fibroblast F27 cells; and murine ovarian carcinoma OVHM cells were cultured as described previously.12Hamada K Kohno S Iwamoto M Yokota H Okada M Tagawa M et al.Identification of the human IAI.3B promoter element and its use in the construction of a replication-selective adenovirus for ovarian cancer therapy.Cancer Res. 2003; 63: 2506-2512PubMed Google Scholar Construction, purification, and plaque assay of adenoviruses were performed as described previously.12Hamada K Kohno S Iwamoto M Yokota H Okada M Tagawa M et al.Identification of the human IAI.3B promoter element and its use in the construction of a replication-selective adenovirus for ovarian cancer therapy.Cancer Res. 2003; 63: 2506-2512PubMed Google Scholar Cell count assay. PA-1 cells were plated at a density of 10,000 cells/well in 12-well plates, cultured for 48 hours with cells infected with AdE3-IAI.3B for 24 hours, and then counted. PA-1 cells were plated at a density of 10,000 cells/well in 6-well plates, cultured for 10 days in high titer (×6,000) of anti-adenovirus antibody (Takeda Pharmaceutical, Japan) with carrier cells infected with AdE3-IAI.3B for 24 hours, and then counted. A Millicell filter chamber of 0.4-μm or 12-μm pore size (Millipore, Bedford, MA) was used to separate PA-1 target cells from A549 carrier cells. Cytotoxic assay. Splenic lymphocytes were isolated from tumor-bearing female (C57BL/6 × C3H/He) F1 mice (CLEA Japan) 3 weeks after injection of phosphate-buffered saline, Ad-β-gal (1 × 1010 PFU), or UV-inactivated Ad-β-gal (1 × 107 PFU), and from mice with complete tumor regression after injection of A549 carrier cells infected with AdE3-IAI.3B and/or AxCAmGM-CSF after pre-immunization. CTL activity was determined by 51Cr release assay as described previously.5Bouvet M Fang B Ekmekcioglu S Ji L Bucana CD Hamada K et al.Suppression of the immune response to an adenovirus vector and enhancement of intratumoral transgene expression by low-dose etoposide.Gene Ther. 1998; 5: 189-195Crossref PubMed Scopus (54) Google Scholar Inhibition of subcutaneous ovarian tumor growth in nude mice. PA-1 or RMG-1 cells (1 × 107) were injected into female nude (nu/nu) mice (CLEA Japan) (n = 10). A549, SW626, or 293 carrier cells were infected for 24 hours with AdE3-IAI.3B at an MOI of 200, 250, and 5, respectively. A549 and SW626 carrier cells were irradiated at 200 and 40 Gy, respectively. Then medium alone, AdE3-IAI.3B (1 × 1010 PFU), A549, SW626, or 293 carrier cells (5 × 106 cells/mouse) were injected into tumors 10–15 mm in diameter on days 0, 1, 2, 3, 4, and 5. Inhibition of subcutaneous ovarian tumor growth in syngeneic mice. Murine OVHM cells (1 × 106) were injected into female (C57BL/6 × C3H/He) F1 mice (CLEA Japan) (n = 10). Medium alone, AdE3-IAI.3B (1 × 1010 PFU), or A549 carrier cells (5 × 106) infected with AdE3-IAI.3B at an MOI of 200 were injected into tumors 5–8 mm in diameter on days 0, 1, and 2 or on days 0, 1, 2, 3, 4, and 5. A549 carrier cells (5 × 106) infected with AxCAmGM-CSF at an MOI of 100, or AdE3-IAI.3B at an MOI of 200 and AxCAmGM-CSF at an MOI of 10, were injected on days 0, 1, and 2. A549 carrier cells (5 × 106) infected with AdE3-IAI.3B at an MOI of 200 and AxCAmGM-CSF at an MOI of 10, 293 carrier cells (5 × 106) infected with AdE3-IAI.3B at an MOI of 5 and AxCAmGM-CSF at an MOI of 1, or A549 carrier cells (2.5 × 106) infected with AdE3-IAI.3B at an MOI of 200 and AxCAmGM-CSF at an MOI of 10 combined with 293 carrier cells (2.5 × 106) infected with AdE3-IAI.3B at an MOI of 5 and AxCAmGM-CSF at an MOI of 1 were injected on day 0. A549 and/or 293 carrier cells were infected with each virus for 24 hours and mixed with 0.2% atelocollagen before injection. A549 carrier cells were irradiated at 200 Gy before injection. Mice were pre-immunized with Ad-β-gal (1 × 1010 PFU) on day –21, days –21 and –42, or days –21, –42, and –63, or UV-inactivated Ad-β-gal (1 × 107 PFU) on day –21. Anti-adenovirus antibody assay. Serum samples were collected from female (C57BL/6 × C3H/He) F1 mice (n = 5) 3, 6, 9, and 12 weeks after a single injection of Ad-β-gal (1 × 1010 PFU) or UV-inactivated Ad-β-gal (1 × 107 PFU), and 3 weeks after the first, second, and third injections of Ad-β-gal (1 × 1010 PFU) and 6 weeks after the third injection of Ad-β-gal (1 × 1010 PFU). Ad-β-gal was injected every 3 weeks. Anti-adenovirus antibodies were assayed for prevention of β-gal expression in A549 cells as described previously.12 Electron microscopy. A549 carrier cells were infected with AdE3-IAI.3B at an MOI of 100 for 24 hours and incubated with A549 or PA-1 target cells in anti-adenovirus antibody–containing medium for 24 hours. A549 carrier cells with A549 or PA-1 target cells were fixed with 2% glutaraldehyde and scraped. Carrier and target cells were further fixed with 2% buffered-osmium tetroxide for 2 hours and embedded in epon epoxy resin. Thin sections were stained with uranyl acetate and lead citrate and examined. Statistical analysis. Values are mean ± SD and were examined using the unpaired t-test, Welch test, and regression analysis. Survival data were plotted on Kaplan-Meier curves and examined with the log-rank test using the LIFETEST procedure. Findings of P < 0.05 were considered significant. This research was supported by a grant-in-aid from the Ministry of Education, Science, Sports, and Culture, Japan, and by the Integrated Center for Science, Ehime University. We thank K. Oka for preparing culture medium and M. Nose and S. Hirose for comments and discussion. The authors have no conflicting financial interests.