Dendritic cells (DCs) are potent antigen-presenting cells that play a critical role in the initiation of antitumor immune responses. In this study, we show that genetic modifications of a murine epidermis-derived DC line and primary bone marrow–derived DCs to express a model antigen β-galactosidase (βgal) can be achieved through the use of a replication-deficient, recombinant adenovirus vector, and that the modified DCs are capable of eliciting antigen-specific, MHC-restricted CTL responses. Importantly, using a murine metastatic lung tumor model with syngeneic colon carcinoma cells expressing βgal, we show that immunization of mice with the genetically modified DC line or bone marrow DCs confers potent protection against a lethal tumor challenge, as well as suppression of preestablished tumors, resulting in a significant survival advantage. We conclude that genetic modification of DCs to express antigens that are also expressed in tumors can lead to antigen-specific, antitumor killer cells, with a concomitant resistance to tumor challenge and a decrease in the size of existing tumors.

Vascular endothelial growth factor (VEGF), a potent angiogenic mediator, is overexpressed in most solid tumors. On the basis of the knowledge that solid tumor growth beyond a small volume is critically dependent on angiogenesis, and that adenovirus (Ad) vectors can mediate efficient in vivo gene transfer and expression, we hypothesized that Ad-mediated transfer of a secreted form of the extracellular domain of the flt-1 VEGF receptor (Adsflt) would suppress tumor growth on a regional basis. To evaluate this concept, three tumor models were examined using a murine colon carcinoma cell line and syngeneic BALB/c mice. First, mice with preestablished splenic CT26.CL25 tumors and liver metastases were given Adsflt on AdNull intravenously and, after 15 days, spleens and livers were harvested to quantify tumor burden. Adslft-treated animals had minimal residual splenic tumors and liver metastases; in contrast, control animals had bulky splenic tumors and extensive liver metastases (p < 0.003). Second, mice with preestablished lung metastases showed a significant reduction in pulmonary metastases with regionally administered Adslft (intratracheal, p < 0.02) but not when the vector was systemically administered (intravenous, p > 0.9). Finally, mice with primary subcutaneous tumors treated with intratumoral administration of Adslft showed significant tumor suppression (p < 0.05) not observed in AdNull-treated mice or mice given Adslft intravenously (p > 0.3). We conclude that Ad-mediated in vivo regional delivery of a secreted form of the extracellular domain of the flt-1 VEGF receptor can effectively inhibit regional tumor growth, a strategy that may provide a means to control tumor growth within the treated organ without the risk of systemic antiangiogenesis.

Although replication-deficient adenovirus (Ad) vectors are efficient vehicles for in vivo gene transfer, persistence of expression of the Ad genome is limited in immunocompetent hosts by cellular immunity directed against the gene product of the vector. While most attention has been focused on cytotoxic T lymphocytes (CTL) directed against the low-level early and late Ad gene expression in the Ad vector-infected target cells, significant cellular immunity is likely also directed against the product of heterologous transgenes. To evaluate this concept, in vivo generation of CTL was evaluated in C57B1/6 and BALB/c mice with Ad vectors expressing a variety of heterologous transgenes, including Escherichia coli chloramphenicol acetyl transferase (CAT), beta-galactosidase (beta-Gal), cytosine deaminase, and human thrombopoietin (hTPO), with an Ad vector expressing no transgene ("null") as a control. Following intravenous administration of Ad vectors, spleen cells were harvested 2 weeks later, stimulated for 5 days with syngeneic cells infected with various Ad vectors, and then evaluated for CTL activity using 51Cr-release from syngeneic Ad vector-infected targets. In all cases, CTL directed against the heterologous transgene products was observed, although there were differences in the amounts of transgene-specific CTL. CTL directed against the transgene were also observed with other routes of administration, including intratracheal, subcutaneous, and intraperitoneal administration. These observations suggest that inclusion of a heterologous transgene in Ad vectors enhances the elimination of vector-infected cells, a circumstance that will be partially circumvented using autologous genes. For some applications, specific immune responses to products of transgenes delivered by Ad vectors might be exploited for therapeutic purposes.