TISSUE AND CELL-CELL INTERACTIONS IN PROSTATE-CANCER PROGRESSION

Prostate cancer is unique among human cancers in the wide discrepancy between the high prevalence of histologic changes recognizable as canc er (latent cancer) and the much lower prevalence of clinically recogni zable disease (clinical cancer). Latent or incidental prostate cancer is often found at autopsy, affecting nearly 33% of all men older than 50 years, yet the number of new cases of clinically manifest prostate cancer represents only a small fraction of this latent disease. This d iscrepancy indicates that life-threatening prostate cancer results fro m progression of only a subset of the latent and clinically insignific ant lesions and not the progression of all small tumors. Therefore, we must understand the underlying molecular and cellular determinants of prostate cancer progression. The role of tissue and cell-cell interac tions in prostate cancer progression is complex. During early stages o f progression, relatively normally organized stroma can suppress discr ete steps in carcinogenesis. This phenomenon is exemplified by studies using the mouse prostate reconstitution (MPR) model system, where inv estigators showed that intrinsic properties of the mesenchyme can modu late the conversion of oncogene-initiated benign hyperplastic prostati c epithelium to the malignant phenotype. When the ras and myc oncogene s were introduced into both the mesenchymal and epithelial compartment s of the microdissected urogenital sinus, poorly differentiated prosta te cancer was produced at a high frequency (more than 90%) in inbred C 57BL/6 mice. In contrast, benign prostatic hyperplasia converted to ca ncer at a low frequency (less than 10%) in BALB/c MPR mice under simil ar conditions. Heterologous MPR composed of BALB/c mesenchyme and C57B L/6 epithelium or vice versa showed that intrinsic properties of BALB/ c mesenchyme can arrest the progression of ras + myc-initiated C57BL/6 epithelium from benign hyperplasia to cancer. Further progression of prostate cancer beyond its earliest forms of the malignant phenotype l eads to increasingly disorganized stroma, where inappropriate stromal cell-cancer cell interactions can occur. Such interactions may include endothelial cell-cancer cell, lymphocyte-cancer cell, and neuron-canc er cell interactions. In addition to stromal cell-cancer cell interact ions, the direct interaction of cancer cells with other cancer cells m ay lead to the establishment of important selection criteria whereby s ubsets of prostate cancer cells progress to metastatic disease. Unders tanding these various cell-cell interactions requires additional exper imental studies that address the molecular and cellular basis of prost ate cancer progression using appropriate model systems that contain th e critical cell types. Recently we modified the MPR model system to in volve the use of p53 ''knock-out'' urogenital sinus tissue as a target for oncogene-initiated prostate cancer. Using this protocol, we produ ced metastatic prostate cancer from normal cells in vivo. Importantly, the organ specificity of the metastatic disease closely resembles tha t in humans, and thus we can study genetic and phenotypic variability throughout the course of progression. In addition, because we use reco mbinant retroviruses to transduce the initiating oncogenes, we can tra ck clonal progression and determine common progenitors of organ-specif ic metastatic disease. This model should prove useful in dissecting im portant tissue and cell-cell interactions in prostate cancer progressi on.