Background

Cancer afflicts all classes of society, but incidence rates increase with age. Thus, with increasing lifespan in developed countries, cancer poses an increasing burden on healthcare systems. Most investigations rely on cancer cell lines, which have defined phenotypes and genotypes, are simple to maintain in the laboratory, and usually straightforward methods are established to manipulate them (e.g., introduce genes, silence genes…). However, cells do not exist as individuals in the body, they are an integral part of tissues, which are vascularized and composed of various cell types that make cell-cell contacts and exchange signaling molecules. Thus, tissues are evidently more representative and relevant models than individual cells to study cell biology and pathologies, such as cancer.

Although several factors are involved in cancer development, such as unrepaired DNA damages, reactive oxygen species, metabolism, genomic instability, the principal model remains intact. Specifically, tumor suppressors are either silenced or mutated while oncoproteins get activated, leading to uncontrolled proliferation and cancer development.

While various cellular proliferation assays (cell counting, transformation, clonogenic, live cell imaging…) and gene expression manipulation methods are well established for cell lines, comparably reliable methods are lacking in the nascent field of ex vivo tissue models. We thus established a robust method for lentiviral transduction and inducible gene expression in ex vivo prostate tissues.

Prostate cancer (PC) is the most common UK male cancer. Most patients respond to androgen ablation therapy that targets the androgen receptor (AR), but patients invariably progress to castrate resistant PC (CRPC) for which there is no curative treatment available and where the prognosis is extremely poor (Feldman et al., 2001). Research investigating novel therapeutics and drug efficacy in PC is hampered by the lack of reliable models that reproduce the patients’ disease. Few PC cell lines are available and they only represent very late disease. Established, well-characterized cell lines (LNCaP, VCaP, DUCaP, CWR22R, PC3 and DU145) were developed from metastasis with AR mutation, amplification and AR-variant expression or even AR loss, but are not representative of early disease presentation for the majority of men where AR gene is generally unaltered (Reference 1). Additionally, cells are cultured on plastic as 2D monolayers of a homogeneous cell population and therefore lack tissue architecture and interactions of distinct cell types. The importance of the tumor microenvironment is also increasingly acknowledged with not only epithelium, but also stromal components (not represented in cell lines), which contribute to PC and CRPC development. Clinical value of using genetically engineered mouse models has the caveat that mice fail to develop PC and their prostate anatomy is dramatically different from the human prostate. Moreover, animal xenograft experiments suffer similar limitations as cell line-based investigations and are further limited by most xenografts being subcutaneously implanted and do not accurately represent complex, heterogeneous tumors. Not surprisingly therefore, there is high variability in the correlation between xenograft studies and clinical trials reported in PC (Centenera et al., 2012 and 2013). These factors substantially contribute towards the very low success rate observed in phase III clinical trials of novel PC therapeutics (significantly less than the general cancer average of 5%). Identification of new, more clinically relevant models of PC that are specific to individual PC patients would allow to improve this statistic and to address an unmet need. Thus, we have developed an ex vivo prostate tissue model to assess proliferation, but also oncogenic and tumor suppressive potential of exogenously expressed genes. These cultures can be generated from patients at all stages of disease and maintain tissue architecture, comprising both stroma and epithelial compartments, thereby retaining crucial contributory autocrine and paracrine signaling interactions.