Cancer immunotherapy aims at reversing the immunosuppressive tumor environment and enhancing anti-tumor immunity. This thesis comprises studies on murine models for prostate (TRAMP-C2) and bladder (MB49) cancer with the aim to explore if the introduction of an adenoviral vector expressing CD40 ligand (AdCD40L) can induce anti-tumor immune responses. We show in subcutaneous mouse models that AdCD40L treatment suppresses tumor growth. Bladder cancer is known to secrete immunosuppressive IL-10 which may inhibit T cell function. We show that introducing AdCD40L into mouse bladder tumors inhibits IL-10 production and reverses immunosuppression. AdCD40L-transduced mouse prostate cancer cells showed caspase activation and reduced cell viability. Vaccination with CD40L-modified prostate cancer cells induces anti-tumor responses and protects mice against rechallenge with native TRAMP-C2 cells. In order to enhance AdCD40L therapy, we explored the possibility of combining it with the histone deacetylase inhibitor FK228, also known as depsipeptide. We show that FK228 upregulates coxsackie and adenovirus receptor expression and thereby enhances adenoviral-mediated CD40L expression in both murine and human prostate cancer cells. Increasing amounts of FK228 or AdCD40L reduces prostate cancer cell viability, while the combined treatment gives at least an additive therapeutic effect. Moreover, we show that AdCD40L transduction of prostate cancer cells induces endogenous CD40 expression and sensitize them for CD40L-mediated therapy. In order to conduct prostate-specific gene therapy, prostate-specific promoters can be used to drive transgene expression. However, there are no reports on prostate-specific promoters that are transcriptionally active in mouse cells. Here we show that by using the two-step transcription activation system (TSTA), we can enhance the activity of a recombinant human promoter sequence and obtain activity in mouse prostate cancer cells as well. This finding paves the way for future studies of prostate-specific gene therapy in immunocompetent mouse models.

Each year liver cancer kills more than half a million people, making it the third leading cause of cancer-related death worldwide. Annual incidence continues to rise steadily, both domestically and globally, increasing the burden of this disease. Advancements in the ability to obtain detailed molecular profiles of tumors have led to the successful development of targeted therapies for a number of different cancers. Unfortunately, however, the molecular pathogenesis of liver cancer is poorly understood relative to many other types of malignancies. Thus, the identification of factors contributing to the development and progression of liver tumors is a major goal of current research. In pursuit of this goal, I have utilized the Sleeping Beauty (SB) transposon system as a tool for forward genetic mutagenesis screening in mice. The SB system recapitulates the kinetics of spontaneous tumor development in humans by providing a stepwise accumulation of mutations. Micro-evolutionary processes within a developing tumor lead to the selective expansion of cells harboring mutations that confer some kind of selective advantage. By identifying the most prevalent mutation events within a specific tumor type across a large number of independent samples, a list of genes implicated as being involved in tumorigenesis can be generated. Using this approach, the Dlk1-Dio3 imprinted domain was identified as a site of frequent mutation in SB-induced hepatocellular carcinomas (HCCs). I discovered that the mechanistic basis for recurrent selection of transposon insertion within this domain in liver tumors involved activated expression of Retrotransposon-like 1 (Rtl1). I also found that RTL1 activation is a common event in human HCC, suggesting that it could potentially be beneficial as a therapeutic target in a subset of patients. Etiological factors related to liver cancer development are varied, but are linked by the fact that each provides a chronic liver injury stimulus that promotes the development of hepatic fibrosis. In fact, ˜ 90% of human HCC occurs in this context, and yet the majority of mouse liver cancer models fail to account for this important environmental component of the disease. I have conducted a screen for genetic drivers of liver cancer in the presence or absence of hepatic fibrosis. Comparison of mutation profiles between fibrotic and non-fibrotic tumors revealed largely non-overlapping sets of candidate genes, indicative of a differential selective pressure for mutations depending on the fibrotic context of the liver. Driver mutations identified preferentially in the presence of liver fibrosis have a high likelihood of relevance to human disease, given the similarities in environmental context and kinetics of mutation acquisition. Consistent with this idea, multiple genes with well-established roles in human HCC were found to be preferentially mutated in SB-induced tumors developed in a fibrotic liver. Before a candidate cancer gene identified in an animal model system can have an impact on human disease, its proposed role in tumorigenesis must be validated. Existing techniques for validation of putative liver cancer genes suffer from significant limitations including high cost, low throughput, and a level of complexity that prohibits widespread utilization. I have contributed to the generation of a novel tool for in vivo validation of candidate genes that is not subject to these limitations. By combining elements of recombinant adenoviral vectors and the piggyBac transposition system, we have generated a highly flexible gene delivery system with significant advantages over existing techniques. The Ad-PB system has broad accessibility and applicability, making it a valuable tool for advancing efforts to improve cancer therapies.

Adenovirus-mediated gene therapy is a potential complement to standard cancer treatments. Advantages are that vectors can be used to target tumors and that replicating viruses lead to increased therapeutic dosage. In this thesis, an oncolytic serotype 5 adenovirus (Ad5), Ad[i/PPT-E1A, E3], was developed where viral replication is controlled by the insulator-shielded (i) prostate-specific PPT promoter. The adenoviral E3 region was inserted for its immune regulatory and lysis functions. Ad[i/PPT-E1A, E3] had improved cytotoxic abilities both in vitro and in a prostate cancer xenograft mouse model compared to a virus lacking the E3 region. To further improve adenoviral vectors, the histone deacetylase inhibitor (HDACi) FK228 was studied. FK228 has been proposed to enhance the effect of adenoviral therapy by upregulation of CAR, the primary receptor for Ad5 infection. In the present study, we observed that FK228 promotes transgene expression even better when administered after viral transduction, indicating a post-transductional enhancement of transgene expression. Another interesting finding was that FK228 reduced transgene expression from the PPT promoter in the prostate cancer cell line LNCaP. This is explained by the fact that different HDACi have the ability to provoke a neuroendocrine phenotype of LNCaP. A potential drawback with adenoviral gene therapy is the rapid clearance of the virus from the circulation. Viral particles have been coated with polyethylene glycol (PEG) to evade immune recognition, a strategy that works well in mouse models. However, less is known about the effects of adenoviral PEGylation in human blood. We have studied cell interactions and immune responses to PEGylated and uncoated Ad5 vectors in human whole blood using a blood loop model with constant blood flow. Limited effects of PEGylation were observed in human blood, which were associated with the neutralizing ability of the donor blood. An important finding that donors with high neutralizing ability in whole blood do not necessarily have neutralizing antibodies against the virus strongly implies that neutralization should be measured in whole blood.