Discussion
Epithelial proliferation is concentrated at the tips of the developing prostate ducts and early studies focused on the duct tip as the probable site of progenitor/stem cells (Kinbara et al., 1996). Subsequent studies examining label retention and regenerative capacity implicated the proximal duct as the reservoir of stem/progenitor cells in the adult prostate (Tsujimura et al., 2002, Kinbara et al., 1996, Burger et al., 2005 and Xin et al., 2005). Further studies sorting for putative stem cell markers and testing for regenerative potential confirmed a relative abundance of stem cells in the proximal duct as compared to the intermediate and distal duct segments (Evans and Chandler, 1987, Lawson et al., 2007, Burger et al., 2005 and Xin et al., 2005). Labeling with either BrdU or GFP at the onset of ductal budding yielded a small population of labeled cells in the adult prostate concentrated in the proximal ducts near their urethral origin. In our study, we also observed label-retaining stromal cells localized to the stromal compartment surrounding the proximal ducts. This could be coincidence, but it does suggest the possibility that slow-cycling epithelial and stromal cells are co-localized within a specific niche in the adult gland and share regulatory signaling mechanisms (Goto et al., 2006, Potten and Loeffler, 1990 and Lavker and Sun, 2000). The most striking findings were robust proliferation of these cells at E16, the preponderance of AR gene expression and the proliferative response to castration.
The ducts of the adult mouse prostate are lined by a pseudostratified epithelium composed of basal cells, luminal cells and rare neuroendocrine cells. The identity and location of stem cells within this epithelial layer are still a matter of debate. At one time basal cells were widely believed to contain progenitor cells capable of differentiating into basal, luminal and neuroendocrine cells (Collins et al., 2001 and Wang et al., 2001). This view has been challenged by recent observations suggesting that stem cells may also reside in the luminal cell layer (Slack, 2000). Single cells co-expressing the markers used in our studies (Lin− Sca-1+CD133+CD44+CD117+) have been shown capable of regenerating a fully differentiated prostate epithelium (Leong et al., 2008). We found that GFP label retaining cells were enriched for co-expression of Lin− Sca-1+CD133+CD44+CD117+. While only a fraction of label retaining cells co-expressed the four stem cell markers, they accounted for approximately a quarter of all the cells co-expressing these markers. This was unexpected and suggests that initiation of prostate ductal budding is associated with a uniquely robust proliferation of epithelial stem/progenitor cells. This is to our knowledge the first evidence that initiation of prostate organogenesis is accompanied by a burst of proliferation among cells that will become a tissue specific stem or progenitor cells in the adult organ. We speculate that this proliferative burst potentially creates a window of vulnerability of these reserve cells to mutation or imprinting changes that could predispose to neoplasia in the adult. Whether a similar burst of proliferation among stem/progenitor cells in other developing organs remains to be determined.
AR is present in the nucleus of most epithelial and stromal cells of the intact adult prostate. Luminal cells are predominantly, if not exclusively, androgen positive whereas only half of all basal cells are (Wang et al., 2009). AR in the stroma mediates paracrine stimulation of epithelial proliferation while AR in epithelial cells stimulates luminal cell differentiation and protein synthesis. It has been assumed that prostate stem cells lack AR (Mirosevich et al., 1999); however, there is some evidence to challenge this view as it applies to the human prostate (Oldridge et al., 2012) and the mouse prostate (Slack, 2000). We found that most BrdU label retaining epithelial cells in the adult prostate were AR positive. More striking was the observation that four-marker positive cells are nearly all AR positive. Insofar as previous studies showed that 14 in 97 four-marker positive cells exhibit the regenerative capacity bona fide stem cells in a tissue recombination assay (Leong et al., 2008), our data suggests that at least some, if not all, stem cells are AR positive.
AR has been shown to exert a growth inhibitory effect in luminal cells. In transgenic mice lacking epithelial AR, epithelial cells are less differentiated and hyper-proliferative (Heer, 2011). It is possible that selective proliferation of slow-cycling and 4-marker cells after castration reflects a release of androgen-mediated inhibition of proliferation of cells that do not require androgen for survival. There is precedent for this in the breast where estrogen suppresses stem cell proliferation (Wu et al., 2007). An alternative explanation for proliferation and migration of LRCs and 4-marker cells following castration would be the response to injury. Progenitor cells play a primary role in the regenerative response to injury in a variety of adult tissues (Simões et al., 2011, Amcheslavsky et al., 2009, Imitola et al., 2004, Brunt et al., 2010 and Zhang et al., 2000) and it has been postulated that factors released by injured tissues stimulate stem cell proliferation and attract stem cells to the sites of injury (Crosnier et al., 2006 and Inui and Sakaguchi, 1992). The effects of castration on the adult prostate are a combination of ischemic injury due to vascular disruption, epithelial apoptosis and acute inflammation (Bodine, 1995, Hayek et al., 1999 and Kerr and Searle, 1973). Whether a direct response to decreased testosterone levels or in response to castration-induced injury, our observation may be relevant to the behavior of tumor stem cells in prostate cancer. To the extent that tumor stem cells phenocopy the features of normal adult stem/progenitor cells, our findings suggest that treatment of human prostate cancer with androgen deprivation could inadvertently produce an expansion of tumor stem cells.
