Normal Prostate Stem/Progenitor Cells and Tumorigenesis

Adult prostate, unlike blood, skin or gut, is not a rapidly renewing organ. Nevertheless, the mouse prostate (Figure 1A) undergoes atrophy upon removing androgen by castration and can regenerate to the original size upon re-administration of androgen. Remarkably, this regression-regeneration cycle can be repeated for up to 30 times attesting the presence of stem cells (SCs). The mouse prostate has four anatomical lobes termed anterior prostate (AP; also called coagulation gland), dorsal and lateral prostate (DP and LP, respectively; sometimes called DLP), and ventral prostate (VP) (Figure 1A). The mouse prostatic tubules have a proximal to distal orientation in relation to the urethra (Figure 1B). There are multiple populations of murine prostatic stem/progenitor cell populations that have been reported (reviewed in Laffin & Tang, 2010).

Dr. E. Wilson’s group performed painstaking BrdU pulse-chase experiments in the mouse prostate subjected to multiple rounds of castration-regeneration cycles and their results have shown that the long-term label-retaining cells (LT-LRCs) are localized preferentially to the proximal region (i.e., close to the urethra) of the prostatic tubules (Figure. 1B). Subsequent work from her group demonstrated that the proximally located prostate LT-LRCs are kept more quiescent possible by TGFβ. Her group further demonstrated that prostatic epithelial cells possessing high Aldefluor activity (i.e., ALDH+) are enriched in stem cells.

A commonly employed experimental approach to demonstrate the stem cell activity in the prostate is the tissue recombination (TR) coupled with kidney capsule (KC) transplantation, or TR/KC, assays, in which candidate prostate epithelial cells are ‘recombined’, in collagen gels, with rat or mouse urogenital sinus mesenchyme (UGSM) pieces or cells. The recombinants are transplanted under the KC and the outgrowths analyzed, weeks to months later, for different lineages of cells based on the expression of various phenotypic markers. Using the TR/KC assays, the Wilson and Witte labs show that the mouse prostatic SCs in the proximal regions are enriched in the Sca-1+ cell population. The Witte group further demonstrates that regenerative prostatic SCs are basally located and can be prospectively purified using Sca-1+, CD49f (integrin α6)+, and Trop2+ phenotype and these cells preferentially express Bmi-1 (Figure 1C). Late 2008, Dr. WQ Gao’s lab reported Lin-Sca-1+CD133+CD44+CD117+ adult mouse prostate SCs localized in the basal cell layer, from which a single cell can regenerate a prostatic outgrowth in renal capsule (Figure 1C). The relationship between these basally localized prostate SC populations (i.e., CD49f+Trop2hi vs. c-Kit+) remains unclear except that both seem to be Sca-1+.

The TR/KC assays, though useful and powerful in revealing the potential SC activities of the candidate cell populations, resemble a wound healing response and may over-estimate SC frequency due to confounding factors in the UGSM and from the host. In contrast, in vivo lineage-tracing studies using various gene promoters driving reporters such as β-gal or GFP, can circumvent such technical issues and provide critical information on cell lineage development in the organism. Late 2009, Dr. M. Shen’s group employed a lineage tracing strategy to uncover a luminal prostate SC population called CARNs (castration resistant and expressing Nkx-3.1) that manifest themselves upon extended castration (Figure 1C). The phenotypes of CARNs in intact prostates remain unclear and the relationship between CARNs and basal SCs is also unclear at the moment. Recently, Dr. Z. Wang’s lab and Dr. L. Xin’s group have provided evidence that the adult mouse prostate contains self-sustaining progenitors in both luminal and basal layers that can self-renew and regenerate luminal and basal cells, respectively.

Taken together, all these recent studies, especially those supported by in vivo lineage tracing, indicate that the prostatic SCs likely localize in both basal and luminal cell layers. Importantly, both basal and luminal progenitors seem to be able to function as the cells-of-origin for PCa.

Our lab recently generated a transgenic (Tg) mouse line in which human 15-lipoxygenase 2 (i.e., 15-LOX2), which metabolizes arachidonic acid to produce a fatty acid molecule (15[S]-HETE) and is normally expressed, most prominently, in the luminal cell layer of human prostatic glands but lost in 70% human prostate cancer (HPCa), is directed to the mouse prostate using the ARR2Pb promoter. Surprisingly, the 15-LOX2 Tg prostate is hyperplastic with increased numbers of both luminal and basal cells. Strikingly, several mouse prostate basal/SC markers, including cytokeratin 5, p63, Sca-1, Trop2, and Nkx3.1, are significantly upregulated in the 15-LOX2 Tg prostate (Suraneni et al., Oncogene, 2010), linking 15-LOX2-induced prostate hyperplasia to prostatic SCs.

Our lab has also established a novel tetracycline-regulated bigenic animal model that can allow us to prospectively purify out LT-LRCs (long-term LRCs) LIVE to study their potential SC properties, their relationship to the reported mouse prostate SC populations, and their potential involvement in tumorigenesis.

Much less is known about the SCs and their lineage development in the human prostate. Human prostatic glands have well-demarcated basal and luminal epithelial cells that express distinct markers (Figure 2). For example, the human prostatic luminal cells express PSA (prostate-specific antigen, which is human-specific), AR, (androgen receptor), PAP (prostate acid phosphatase), cytokeratins 8 and 18 (CK8/CK18), CD57, and 15-LOX2 whereas the basal cells express CK5/14, p63, Bcl-2, and hTERT (Figure 2). It is interesting to note that most of the basal cell-restricted molecules are well-known SC regulators. In support, several basally localized subpopulations of cells, e .g., those expressing CD44, integrin α2β1 and α6, Trop2, CD133, and ABCG2, have been reported to possess stem/progenitor cell properties (Figure 2). Our lab, over the years, has studied and made extensive use of primary normal human prostate (NHP) epithelial cells isolated from normal/benign human prostates and cultured in serum/androgen-free medium. Our work has defined these cells as CD44+α2β1+hTERT+p63+15LOX2-AR-PSA- progenitor cells (Tang et al., JBC, 2002; Bhatia et al., JBC, 2003; Tang et al., Oncogene, 2004; Bhatia et al., Oncogene, 2005; Bhatia et al., JBC, 2008). hTERT overexpression and p16-shRNA lead to NHP cell immortalization and, importantly, such immortalized NHP progenitors are karyotypically normal, exhibit a global SC gene-expression profile, have an unlimited proliferative capacity and retain tri-potential differentiation capacity in vivo to regenerate normal prostatic glands at ≤1,000 cells (Bhatia et al., JBC., 2008). Of interest, senescent NHP cells, like senescent fibroblasts, can promote tumorigenesis through cell fusion (Bhatia et al., Int. J. Cancer, 2008). For more readings on NHP stem/progenitor cells, please see: Tang et al., Mol. Carcinogenesis, 2007; Honorio et al., 2009; Jeter & Tang, 2010.

It is our hope that studies on normal prostate SCs will shed important light on their potential involvement in prostate tumorigenesis although we consider the cell-of-origin for human PCa a ‘chicken-and-egg’ question that may defy a definitive answer. Consequently, our lab has been focusing on identifying, in established prostate tumors, stem-like PCa cells or PCSCs.