Abstract 3063: Stem cell lineage hierarchy by keratin profiling in normal human prostate epithelial cells and prostate cancer

Background: Both mechanical and nonmechanical functions of keratins have been implicated in stem cells and cancers. Using a sphere-based label retention assay, we recently isolated and characterized prostate stem cells and cancer stem-like cells from mixed progenitor populations (PMID 28651114), identifying keratin 13 (KRT13) as a specific prostate stem cell marker regulating self-renewal. Herein we utilize detailed keratin profiles to further clarify the human prostate epithelial lineage hierarchy and identify prostate cancer stem-like cells. Methods and Results: Primary prostate epithelial cells were 3D cultured (5 days) to form prostaspheres (PS), followed by PS-based long-term label retention and FACS to separate stem cells from progenitors. In normal prostate tissues from three healthy donors, RNA-seq revealed enrichment of KRT13, 23, 80, 78, 86 and 4 in label-retaining prostate stem cells while KRT6, 17, 14, 5, 8, 18 and P63 were enriched in nonretaining progenitors. We next used Fluidigm C1 captured single-cell RNA-seq and identified three major clusters in the label-retaining stem cell population; Cluster I represents quiescent stem cells (KRT13, 23, 80, 78, 4 enriched) and Clusters II and III contain active stem cells and bipotent progenitors, respectively (KRT16, 17, 6 enriched). GSEA analysis found stem cell and cancer-related pathways enrichment in Cluster I. Three additional clusters were identified in nonretaining progenitor cells, with Cluster IV representing unipotent basal progenitor cells (KRT5, 14, 6, 16 enriched) and Clusters V and VI early- and late-stage unipotent luminal progenitors (KRT8, 18, 10 enriched). Cancer stem-like cells were similarly isolated from three prostate cancer specimens and RNA-seq with MetaCore pathway analysis found enrichment of cytoskeleton remodeling keratin filaments. Interestingly, in addition to normal stem cell keratins (KRT13, 23, 80, 78, 4), other keratins (KRT10, 19, 6, 75, 16, 79, 3, 82) were enriched in cancer stem-like cells. Surprisingly, stem-like cells from patient-matched benign regions revealed a similar keratin profile, suggesting a cancer field effect for stem-like cell populations. Conclusion: Taken together, using gene profiling with an emphasis on keratin patterns, we have delineated the lineage hierarchy of human prostate stem cells originating from the activation of quiescent stem cells to bipotent progenitors that give rise to unipotent basal and luminal progenitor cells. We have identi