The prostate gland exhibits a remarkable capacity to regenerate after successive cycles of castration and restoration of normal androgen levels, which has been attributed to the existence of a small population of hormone-insensitive cells with stem cell properties. Such cells are highly relevant to our understanding of normal prostate development, carcinogenesis and response to castration. The purification of these putative stem cells, to the exclusion of the majority (>99%) of other cell types, has, however, long remained an experimental challenge. The paper by Leong and colleagues,1 which describes the regeneration of vestigial prostate glands in mice from single epithelial cell implants, is, therefore, of considerable interest to both basic and clinical scientists.
This topic is not new: Isaacs and Coffey2 postulated the existence of androgen-independent tissue stem cells in the prostate more than 20 years ago, having observed regeneration of rat prostate after repeated cycles of androgen deprivation and replacement. Their model is generally accepted to hold not only for normal prostate, but also for hormone-naive prostate cancer in the early stages of castration therapy. The possibility of a long-lived amplifying cell population as the source of regeneration cannot, however, be ruled out, even by the current study,1 which addresses the nature of the normal prostate stem cell more precisely than ever before.
The authors initially determined gene expression levels within the different lobes and regions of the mouse prostate, focusing on candidate stem cell markers—in particular CD117 (also known as c-kit, stem cell factor receptor).1 Their findings offer support for the model that stem cell populations are enriched within the proximal region of the mouse prostate,3 and the findings by our laboratory that human prostate stem cells are located within the basal compartment of the gland.4, 5
As predicted by a model of stem cell regeneration after castration, the authors demonstrated an increase in the number of cells expressing CD117, CK14, and CD44 (but not the luminal marker CD24) in castrated mice. This change was reversed after hormone replacement, presumably owing to cell differentiation and expansion. To test the regenerative potential of the CD117+ population, Leong et al.1 sorted the CD117+ and CD117- fractions (from adult male C57BL/6 mouse donors) and implanted these (in a complex with rat embryonic urogenital sinus mesenchymal cells) under the renal capsule of athymic nu/nu mouse hosts (Figure 1). The CD117+ cells, but not CD117- cells, could generate a prostate in this setting. Histological examination of the CD117+ grafts demonstrated a branching morphology with epithelial tubules composed of basal, luminal and neuroendocrine lineages.
Figure 1 | The single-cell transplantation procedure.
Adult C57BL/6 mouse prostates were harvested, dissociated, lineage depleted, and stained with propidium iodide and antibodies against cell-surface markers. Single viable (propidium iodide-) Lin-Sca-1+CD133+CD44+CD117+ cells were sorted into individual wells of a plate containing collagen solution. Following collagen gelation, rat stromal cells were added to each well. Gels were grafted under the renal capsule of host athymic nu/nu mice, along with a subcutaneous slow-release testosterone pellet. Abbreviation: FACS, fluorescence activated cell sorting. Permission obtained from Macmillan Publishers © Leong, K. G. et al. Nature 456, 804–808 (2008).
The 'gold standard' of stem cell function is the ability to self-renew, which is required for maintenance of tissue integrity over a lifetime. To evaluate the self-renewal capacity of CD117+ cells and to provide an objective indicator of the frequency of stem cell activity in the CD117+ population, limiting dilution analysis was performed in vivo. The data indicated that around 10% of single Lin-Sca-1+CD133+CD44+CD117+ cells had prostate stem cell function; that is, they could generate a prostate when individually transplanted under the renal capsule of athymic nu/nu male mouse hosts. Furthermore, secondary and tertiary transplants of CD117+ cells, but not CD117- cells, also gave rise to morphologically and immunohistochemically defined prostate acini.
Leong et al.1 also investigated a functional role for the CD117 antigen in prostate stem cells by administering anti-CD117 blocking antibody to castrated C57BL/6 mice, and then assessing prostate regeneration after hormone replacement. Impairment of CD117 signaling led to in vivo inhibition of prostate regeneration, accompanied by a decrease in levels of cycling cells and an increased basal : luminal cell ratio. This observation raises the intriguing possibility that regeneration was mediated by growth factors secreted under hormonal influence from the prostate stromal cells, whose critical role in prostate tissue (and tumor) development was recently highlighted by Niu et al.6 However, as Leong et al.1 note, despite the smaller prostate size seen in CD117+/- heterozygotes (Figure 2), only studies that selectively knock out CD117 gene expression in stem cells will conclusively demonstrate a functional requirement for CD117 in prostate development.
Figure 2 | Multiple prostate epithelial cell types derived from a stem cell transplant.
The images show immunohistochemical staining of sections of grafts for basal cell (CK14; green) and luminal cell (CK18; red) cytokeratins. Nuclei are stained with DAPI (blue). Of particular interest is the difference in size and complexity of the glands derived from a heterozygous mutant CD117 mouse (right panel) compared with the wild-type homozygote, which indicates a functional role for CD117 expression in mouse prostate gland development. Permission obtained from Macmillan Publishers © Leong, K. G. et al. Nature 456, 804–808 (2008).
One evident shortcoming in the conclusions of Leong et al.1 concerns the established stem cell markers CD133 (also known as prominin-1) and Sca-1. As they correctly comment, both markers are expressed by a subpopulation of stromal cells in the prostate (as indeed is CD117, whose expression is detectable in both stromal and interstitial cells of the prostate7). Such cells must be depleted before candidate stem cells are selected, using a combination of surface markers to identify epithelial stem cells. Moreover, extrapolation of results from mice to humans is also potentially problematic, as CD133 seems to be expressed by 30% of prostate epithelial cells in mice, compared with only a rare population (<0.1%) of basal epithelial cells in the human prostate.4, 5 Such differences might be attributable to antibody specificity. The mouse ortholog of human CD133 is expressed in a broad range of adult epithelial cells, including the stem and progenitor population. However, the specificity of the widely used antibody against human CD133 is dependent upon antigen glycosylation, which is only seen on the surface of more primitive cell types.8
Another controversy, which the study did not address, is the role of p63—a 'master controller' of epithelial lineage development—in prostate epithelium.9 Whereas Leong et al.1 detected p63 in the proximal basal epithelium and found a number of p63-expressing cells, which co-localize with CD117+ cells in human benign prostatic hyperplasia, the precise functional role of p63 remains undetermined.
Importantly, while the subrenal capsule of the immunocompromised mouse host might be the site of choice for future human xenograft experiments, the results of grafting candidate mouse cells into the prostate of syngeneic mice—a more natural microenvironment—will be awaited with great interest.
This study adds significantly to the body of evidence on stem cell biology and prostate development. It definitively shows that prostate epithelial stem cells are not directly dependent upon androgens for survival. Rather surprisingly, however, Leong et al.1 did not report data on androgen receptor expression in either the Lin-Sca-1+CD133+CD44+CD117+ cells or their differentiated progeny. If the androgen receptor was not expressed in the mouse stem cells (as has been shown in human prostate epithelial stem cells4), then hormonal control must have been modulated by the powerful embryonic differentiating stimulus of the embryonic urogenital sinus mesenchyme (an analog of the androgen-receptor-expressing stroma in adult prostate tissues). Thus, treatments for hormone-sensitive disease should also be assessed for their effect on the accompanying stroma, rather than just on the epithelial components of prostate cancer.6
As a final thought, if stem cells are indeed involved in prostate carcinogenesis—consistent with their long lifespan and capacity for regeneration, as demonstrated by Leong et al.1—then the development of prostate cancer therapies on the basis of detectable tumor regression will result in agents that only kill differentiated, cycling tumor cells while sparing the rarer, quiescent cancer stem cell population.10 Leong et al.1 have confirmed that the phenotype of the normal stem cell is quite distinct from that of the majority of luminal cells in the prostate. If cancer-initiating cells have similar developmental origins to normal stem cells, then they are certainly likely to be resistant to castration-based therapies. The development of new cancer therapies that target stem cells will require carefully defined clinical end points and trial criteria, as any agent directed against stem cells will potentially kill <0.1% of the total cells in a tumor mass. The optimum therapeutic strategy is most likely to involve a combination of anti-stem-cell therapy and established agents directed against the androgen-sensitive cells in the tumor mass.
If the 1990s was the decade of molecular genetics, which culminated in the sequencing of the human genome, then we are now well into the decade of the stem cell. For urologists, the existence of stem cells in the prostate, as convincingly demonstrated by Leong et al.,1 will require a reassessment of the 'prostate facts' learned in medical school. Not all prostate cells, whether normal or cancerous, are directly sensitive to androgens. This knowledge should have a profound effect on how we apply hormone-based strategies for prevention and treatment of prostatic disease.
2
1-integrin expression