The goals of our project are to isolate prostate epithelial populations, including prostate stem cells, and to evaluate the tumorigenic and the metastatic capacity of such populations upon transformation by deletion of the tumor suppressor, PTEN, either alone or in combination with alteration of the p53 tumor suppressor. PTEN deletion or inactivation is common in human prostate cancer, and PTEN deletion in mouse prostate epithelium has been shown to mimic human prostate cancer in the production of adenocarcinoma with a luminal phenotype, which is capable of giving rise to androgen independent secondary tumors. The addition of p53 deletion has been shown to result in more highly penetrant and rapidly evolving disease as a result of decreased cellular senescence. Because PTEN deletion is an embryonic lethal event, loss of PTEN has been engineered to be specific for prostate epithelium, using CRE-dependent excision. We have established appropriate strains of mice carrying genetically modified loci and validated methods to obtain specific PTEN and/or p53 excision both in vivo and in vitro. Existing mouse models of prostate cancer demonstrate relatively little metastasis and usually to lymph nodes only. Possible explanations for this observation are, 1) polyclonal tumor formation in transgenic models leads to mortality due to tumor burden before the conditions for metastatic spread occur, or 2) the wrong precursor population is transformed, or 3) transforming events are missing that lead to the metastatic phenotype. An approach that allows the above possibilities to be systematically investigated is based upon orthotopic transplantation of selected tumor initiating cells. We have successfully transplanted prostate epithelial cells into the dorsal prostate of the mouse and longitudinally followed the progression of the transplanted cells using bioluminescent imaging. Orthotopic transplantation will be used to evaluate tumorigenic and metastatic potential in various prostate epithelial populations, separated on the basis of cell surface markers and transformed with defined genetic events. The specificity of Pten and P53 deletion in prostate epithelial cells was confirmed using immunohistochemical analysis of donor prostate tissue. A thorough pathological analysis of donor genotype mice has been accomplished, and the conclusion of these studies is that Pb-Cre driven Pten and P53 deletion results in highly invasive prostate AC. In addition, a low frequency of prostatic urothelial carcinoma, adenosquamous carcinoma, and sarcomatoid carcinoma is observed. Donor mice develop diffuse prostatic intraepithelial neoplasia (PIN) in all prostatic lobes, with high grade PIN appearing as early as 10 weeks. Adenocarcinoma (AC) can be detected by 12 weeks. Most donor genotype mice die of urinary tract outflow obstruction as early as 3.5 months and usually by 5-6 months with 100% mortality by 7 months (n=60 mice analyzed). AC ranges from a well differentiated glandular pattern to a more solid undifferentiated pattern. Tumors have a prominent fibrovascular stroma with moderate multifocal lymphoplasmacytic infiltrates. PIN/AC appears more severe in the proximal portions of the prostate. The most distal portions of the dorsal prostate rarely progress beyond PIN, while AC is often seen in older mice in the distal anterior and ventral prostates. The AC is highly invasive with tumor cells invading adjacent stroma, nerves, blood vessels, and lymphatics. Disseminated tumor cells in regional lymph nodes have been observed following identification with pan-cytokeratin staining. Similarly, rare single prostate tumor cells can be detected within capillaries in the lung, and occasionally a few tumor cells can be observed extending into the adjacent pulmonary parenchyma. There is no evidence of micrometastasis in the period subsequent to AC appearance and prior to death. Thus, we believe this is a desirable model to interrogate cooperating genetic aberrations leading to metastasis. Immunohistochemical analyses of tumors using defined cell lineage markers shows that Pten/P53 deletion displays a similar profile to Pten deletion alone but with more rapid and extensive development of AC. PIN is comprised of a proliferation of CK5+ basal cells along the basement membrane and CK8+ luminal cells at the edge of the lumen. CK8+/Synaptophysin(Syn)+ and CK5+/CK8+ intermediate cells are also expanded in suprabasal layers. P63+ cells also increase with PIN yet are mostly lost with the transition to carcinoma. With the progression to AC a variety of tumor cell immunophenotypes are observed, including CK5+/CK8-, CK5-/CK8+, CK5+/CK8+, CK8+/Syn+, and probably CK5+/CK8+/syn+ cells (Fig2A,B&C). Bcl-2 is detected in occasional carcinoma cells, but not in control mouse prostate epithelial cells. We conclude that there is a proliferation of stem/progenitor cells in Pten-/-P53-/-prostate epithelium. Because Pten/P53 deletion may occur independently in distinct progenitors, questions raised by the presence of the various intermediate populations are their differentiation potential and precursor-progeny relationship. Our main approach relies upon the orthotopic injection of tumor cells into the dorsal prostate as a means to provide the appropriate microenvironment for tumor growth and evolution as well as the appropriate anatomic location for dissemination of AC. For the purpose of evaluating secondary tissue-specific growth, we are supplementing orthotopic assays with subcutaneous and intratibial sites. A surgical microscope and a surgical micromanipulator are employed to allow injection into the dorsal prostate, a region of the prostate that requires a significant amount of manipulation to access. Initial experiments were evaluated using BLI the day after injection to assay successful implantation and possible leakage. Comparing different volumes indicated that a 10-20 microl volume is optimal because lower volumes have given inconsistent day 1 takes. We have not observed tumor formation to date with cell inoculums less than 100,000 cells, although animals injected with fewer cells are still being observed. Leakage can be sensitively imaged as a signal in the scrotum of the mouse, the point of accumulation for tumor cells that have leaked into the peritoneal cavity. A high percentage of injections display no obvious leakage (approximately 90%). Currently, recipient mice are assayed for BL soon after surgery (1-2 weeks) to confirm the success and accuracy of injection and then imaged subsequently at varying intervals ranging from weekly to monthly to assess tumor growth. Using fractionation approaches, we have found that AC and adenosquamous carcinoma arise from populations enriched for progenitor and basal cells. Of interest, the adenosquamous carcinoma has been observed to metastasize from the orthotopic tumor to the lungs and to lymph nodes at a low frequency.
