Purpose: We have shown that the phenotypically undifferentiated (PSA(-/lo)) prostate cancer cell population harbors long-term self-renewing cancer stem cells (CSC) that resist castration, and a subset of the cells within the PSA(-/lo) population bearing the ALDH(hi)CD44(+)alpha 2 beta 1(+) phenotype (Triple Marker(+)/TM+) is capable of robustly initiating xenograft tumors in castrated mice. The goal of the current project is to further characterize the biologic properties of TM+ prostate cancer cell population, particularly in the context of initiating and propagating castration-resistant prostate cancer (CRPC). Experimental Design: The in vivo CSC activities were measured by limiting-dilution serial tumor transplantation assays in both androgen-dependent and androgen-independent prostate cancer xenograft models. In vitro clonal, clonogenic, and sphere-formation assays were conducted in cells purified from xenograft and patient tumors. qPCR, Western blot, lentiviral-mediated gene knockdown, and human microRNA arrays were performed for mechanistic studies. Results: By focusing on the LAPC9 model, we show that the TM+ cells are CSCs with both tumor-initiating and tumor-propagating abilities for CRPC. Moreover, primary patient samples have TM+ cells, which possess CSC activities in "castrated" culture conditions. Mechanistically, we find that (i) the phenotypic markers are causally involved in CRPC development; (ii) the TM+ cells preferentially express castration resistance and stem cell-associated molecules that regulate their CSC characteristics; and (iii) the TM+ cells possess distinct microRNA expression profiles and miR-499-5p functions as an oncomir. Conclusions: Our results define the TM+ prostate cancer cells as a population of preexistent stem-like cancer cells that can both mediate and propagate CRP Candhighlight the TM+ cell population as a therapeutic target. (C) 2016 AACR.