Castration triggers growth of previously static androgen-independent lesions in the transgenic adenocarcinoma of the mouse prostate (TRAMP) model

MALAT1 Fusions and Basal Cells Contribute to Primary Resistance against Androgen Receptor Inhibition in TRAMP Mice

Simple Summary

We deeply characterized a frequently used mouse model of prostate cancer and found cellular and molecular regulators of resistance against antihormonal treatment, such as basal cell function and MALAT1 gene fusions. As these mechanisms also occur in human disease, our findings highlight the importance of this model for human cancer and may be helpful for future research focusing on overcoming antihormonal treatment resistance.

Abstract

Targeting testosterone signaling through androgen deprivation therapy (ADT) or antiandrogen treatment is the standard of care for advanced prostate cancer (PCa). Although the large majority of patients initially respond to ADT and/or androgen receptor (AR) blockade, most patients suffering from advanced PCa will experience disease progression. We sought to investigate drivers of primary resistance against antiandrogen treatment in the TRAMP mouse model, an SV-40 t-antigen driven model exhibiting aggressive variants of prostate cancer, castration resistance, and neuroendocrine differentiation upon antihormonal treatment. We isolated primary tumor cell suspensions from adult male TRAMP mice and subjected them to organoid culture. Basal and non-basal cell populations were characterized by RNA sequencing, Western blotting, and quantitative real-time PCR. Furthermore, effects of androgen withdrawal and enzalutamide treatment were studied. Basal and luminal TRAMP cells exhibited distinct molecular signatures and gave rise to organoids with distinct phenotypes. TRAMP cells exhibited primary resistance against antiandrogen treatment. This was more pronounced in basal cell-derived TRAMP organoids when compared to luminal cell-derived organoids. Furthermore, we found MALAT1 gene fusions to be drivers of antiandrogen resistance in TRAMP mice through regulation of AR. Summarizing, TRAMP tumor cells exhibited primary resistance towards androgen inhibition enhanced through basal cell function and MALAT1 gene fusions.

Wnt/Beta-Catenin Signaling and Prostate Cancer Therapy Resistance

Metastatic or locally advanced prostate cancer (PCa) is typically treated with androgen deprivation therapy (ADT). Initially, PCa responds to the treatment and regresses. However, PCa almost always develops resistance to androgen deprivation and progresses to castrate-resistant prostate cancer (CRPCa), a currently incurable form of PCa. Wnt/β-Catenin signaling is frequently activated in late stage PCa and contributes to the development of therapy resistance. Although activating mutations in the Wnt/β-Catenin pathway are not common in primary PCa, this signaling cascade can be activated through other mechanisms in late stage PCa, including cross talk with other signaling pathways, growth factors and cytokines produced by the damaged tumor microenvironment, release of the co-activator β-Catenin from sequestration after inhibition of androgen receptor (AR) signaling, altered expression of Wnt ligands and factors that modulate the Wnt signaling, and therapy-induced cellular senescence. Research from genetically engineered mouse models indicates that activation of Wnt/β-Catenin signaling in the prostate is oncogenic, enables castrate-resistant PCa growth, induces an epithelial-to-mesenchymal transition (EMT), promotes neuroendocrine (NE) differentiation, and confers stem cell-like features to PCa cells. These important roles of Wnt/β-Catenin signaling in PCa progression underscore the need for the development of drugs targeting this pathway to treat therapy-resistant PCa.

Genetic deletion of osteopontin in TRAMP mice skews prostate carcinogenesis from adenocarcinoma to aggressive human-like neuroendocrine cancers

Osteopontin (OPN) is a secreted glycoprotein, that belongs to the non-structural extracellular matrix (ECM), and its over expression in human prostate cancer has been associated with disease progression, androgen independence and metastatic ability. Nevertheless, the pathophysiology of OPN in prostate tumorigenesis has never been studied. We crossed TRansgenic Adenocarcinoma of the Mouse Prostate (TRAMP) mice with OPN deficient (OPN-/-) mice and followed tumor onset and progression in these double mutants. Ultrasound examination detected the early onset of a rapidly growing, homogeneous and spherical tumor in about 60% of OPN-/- TRAMP mice. Such neoplasms seldom occurred in parental TRAMP mice otherwise prone to adenocarcinomas and were characterized for being androgen receptor negative, highly proliferative and endowed with neuroendocrine (NE) features. Gene expression profiling showed up-regulation of genes involved in tumor progression, cell cycle and neuronal differentiation in OPN-deficient versus wild type TRAMP tumors. Down-regulated genes included key genes of TGFa pathway, including SMAD3 and Filamin, which were confirmed at the protein level. Furthermore, NE genes and particularly those characterizing early prostatic lesions of OPN-deficient mice were found to correlate with those of human prostate NE tumours. These data underscore a novel role of OPN in the early stages of prostate cancer growth, protecting against the development of aggressive NE tumors.

Interaction of the Androgen Receptor, ETV1, and PTEN Pathways in Mouse Prostate Varies with Pathological Stage and Predicts Cancer Progression

To examine the impact of common somatic mutations in prostate cancer (PCa) on androgen receptor (AR) signaling, mouse models were designed to perturb sequentially the AR, ETV1, and PTEN pathways. Mice with "humanized" AR (hAR) alleles that modified AR transcriptional strength by varying polyglutamine tract (Q-tract) length were crossed with mice expressing a prostate-specific, AR-responsive ETV1 transgene (ETV1(Tg)). While hAR allele did not grossly affect ETV1-induced neoplasia, ETV1 strongly antagonized global AR regulation and repressed critical androgen-induced differentiation and tumor suppressor genes, such as Nkx3-1 and Hoxb13. When Pten was varied to determine its impact on disease progression, mice lacking one Pten allele (Pten(+/-) ) developed more frequent prostatic intraepithelial neoplasia (PIN). Yet, only those with the ETV1 transgene progressed to invasive adenocarcinoma. Furthermore, progression was more frequent with the short Q-tract (stronger) AR, suggesting that the AR, ETV1, and PTEN pathways cooperate in aggressive disease. On the Pten(+/-) background, ETV1 had markedly less effect on AR target genes. However, a strong inflammatory gene expression signature, notably upregulation of Cxcl16, was induced by ETV1. Comparison of mouse and human patient data stratified by the presence of E26 transformation-specific ETS fusion genes highlighted additional factors, some not previously associated with prostate cancer but for which targeted therapies are in development for other diseases. In sum, concerted use of these mouse models illuminates the complex interplay of AR, ETV1, and PTEN pathways in pre-cancerous neoplasia and early tumorigenesis, disease stages difficult to analyze in man.

Lobe-specific lineages of carcinogenesis in the transgenic adenocarcinoma of mouse prostate and their responses to chemopreventive selenium

BACKGROUND

The transgenic adenocarcinoma of mouse prostate (TRAMP) model is by far the most practical transgenic model for preclinical prostate cancer chemoprevention studies. It is critical to characterize the prostate lobe-specificity of lesion lineages to consolidate the advantages of this model and minimize its limitations for chemoprevention studies.

METHODS

We dissected dorsolateral (DLP), ventral (VP), and anterior prostate (AP) lobes, and macroscopic tumors from 90 male C57BL/6J TRAMP mice at 22-24 weeks of age (WOA) and analyzed lesions by histological, biochemical and proteomic approaches. To determine whether methylseleninic acid (MSeA) led to a deletion of initiated cells, we gave oral MSeA to TRAMP mice from 5 to 23 WOA or from 5 to 15 WOA and analyzed lesions at 23 WOA.

RESULTS

All tumors (n = 18) were T-antigen(+), synaptophysin (SYP)(+), androgen-receptor(-), and E-cadherin(-) poorly differentiated neuroendocrine carcinomas (NE-Ca). They were traceable most frequently to VP (66.7%) and rarely to DLP (11.1%) and AP (5.6%) with an estimated life-time incidence of 1 out of 3 mice. In DLP, epithelial lesions ranged from mild-to-severe atypical hyperplasia, with T-antigen(+), SYP(-), androgen-receptor(+), and E-cadherin(+). Proteomic profiling revealed many molecular differences between VP and DLP. In MSeA experiment, 6 out of 19 (31.5%) mice developed NE-Ca in the control group, only 2 in each MSeA group of 17-18 mice (11.1-11.8%) bore a detectable NE-Ca.

CONCLUSION

The C57BL/6J TRAMP mouse represents at least two lineages of prostate carcinogenesis. Chemoprevention studies should incorporate this knowledge for efficacy assessment and molecular target validations.

Animal models relevant to human prostate carcinogenesis underlining the critical implication of prostatic stem/progenitor cells

Recent development of animal models relevant to human prostate cancer (PC) etiopathogenesis has provided important information on the specific functions provided by key gene products altered during disease initiation and progression to locally invasive, metastatic and hormone-refractory stages. Especially, the characterization of transgenic mouse models has indicated that the inactivation of distinct tumor suppressor proteins such as phosphatase tensin homolog deleted on chromosome 10 (PTEN), Nkx3.1, p27(KIP1), p53 and retinoblastoma (pRb) may cooperate for the malignant transformation of prostatic stem/progenitor cells into PC stem/progenitor cells and tumor development and metastases. Moreover, the sustained activation of diverse oncogenic signaling elements, including epidermal growth factor receptor (EGFR), sonic hedgehog, Wnt/β-catenin, c-Myc, Akt and nuclear factor-kappaB (NF-κB) also may contribute to the acquisition of more aggressive and hormone-refractory phenotypes by PC stem/progenitor cells and their progenies during disease progression. Importantly, it has also been shown that an enrichment of PC stem/progenitor cells expressing stem cell-like markers may occur after androgen deprivation therapy and docetaxel treatment in the transgenic mouse models of PC suggesting the critical implication of these immature PC cells in treatment resistance, tumor re-growth and disease recurrence. Of clinical interest, the molecular targeting of distinct gene products altered in PC cells by using different dietary compounds has also been shown to counteract PC initiation and progression in animal models supporting their potential use as chemopreventive or chemotherapeutic agents for eradicating the total tumor cell mass, improving current anti-hormonal and chemotherapies and preventing disease relapse.

Differential androgen receptor signals in different cells explain why androgen-deprivation therapy of prostate cancer fails

Prostate cancer is one of the major causes of cancer-related death in the western world. Androgen-deprivation therapy (ADT) for the suppression of androgens binding to the androgen receptor (AR) has been the norm of prostate cancer treatment. Despite early success to suppress prostate tumor growth, ADT eventually fails leading to recurrent tumor growth in a hormone-refractory manner, even though AR remains to function in hormone-refractory prostate cancer. Interestingly, some prostate cancer survivors who received androgen replacement therapy had improved quality of life without adverse effect on their cancer progression. These contrasting clinical data suggest that differential androgen/AR signals in individual cells of prostate tumors can exist in the same or different patients, and may be used to explain why ADT of prostate cancer fails. Such a hypothesis is supported by the results obtained from transgenic mice with selective knockout of AR in prostatic stromal vs epithelial cells and orthotopic transplants of various human prostate cancer cell lines with AR over-expression or knockout. These studies concluded that AR functions as a stimulator for prostate cancer proliferation and metastasis in stromal cells, as a survival factor of prostatic cancer epithelial luminal cells, and as a suppressor for prostate cancer basal intermediate cell growth and metastasis. These dual yet opposite functions of the stromal and epithelial AR may challenge the current ADT to battle prostate cancer and should be taken into consideration when developing new AR-targeting therapies in selective prostate cancer cells.

Differential requirement for focal adhesion kinase signaling in cancer progression in the transgenic adenocarcinoma of mouse prostate model

Increasing evidence indicates that adhesion signaling plays an important role in the tumor microenvironment, contributing to cancer progression, invasion, and metastasis. Focal adhesion kinase (FAK) is a nonreceptor protein tyrosine kinase that regulates adhesion-dependent cell signaling and has been implicated in mediating steps in cancer progression and metastasis in many human cancers, including prostate. We have investigated the role of FAK in the appearance of adenocarcinoma (atypical epithelial hyperplasia of T antigen) and neuroendocrine carcinomas in the transgenic adenocarcinoma of mouse prostate (TRAMP) model using either Cre-mediated recombination to genetically ablate FAK expression or pharmacologic inhibition of FAK activity with the small-molecule inhibitor, PF-562,271. We provide evidence that loss of FAK or its inhibition with PF-562,271 does not alter the progression to adenocarcinoma. However, continued FAK expression (and activity) is essential for the androgen-independent formation of neuroendocrine carcinoma. These data indicate that integrin signaling through FAK is an important component of cancer progression in the TRAMP model and suggest that treatment modalities targeting FAK may be an appropriate strategy for patients with castrate-resistant cancer.

Early and delayed castrations confer a similar survival advantage in TRAMP mice

The most appropriate time to introduce androgen deprivation therapy for prostate cancer remains controversial. Our aim was to evaluate the effects of early versus delayed surgical castration on prostate cancer progression and survival in the transgenic adenocarcinoma of the mouse prostate (TRAMP) model. TRAMP mice were randomly divided into three groups: the early castration group (on which castration was performed at the age of 4 weeks), the delayed castration group (on which castration was performed when abdominal tumours could be palpated), and the sham-castrated group. Mice were monitored daily throughout their lives until cancer-related death or the development of an obviously moribund appearance, at which time the individual mouse was killed. Androgen receptor expression in prostate tumours was also evaluated. The results shows that the average lifespan in early castration, delayed castration and sham-castrated groups were 54.1 weeks, 59.9 weeks and 39.1 weeks, respectively. Both early castration and delayed castration conferred a statistically significant survival advantage when compared with the sham-castrated group (P<0.001). However, the difference in lifespan between the early castration group and the delayed castration group was not statistically significant (P=0.85). The increase in lifespan in the TRAMP mice that received either early or delayed castration correlated with lower G/B value (genitourinary tract weight/body weight) at death than the sham-castrated mice. In conclusion, early and delayed castrations in TRAMP mice prolonged survival to a similar extent. This finding may provide a guide for clinical practice in prostate cancer therapy.

Profiling human androgen receptor mutations reveals treatment effects in a mouse model of prostate cancer

Gain-of-function mutations in the androgen receptor (AR) are found in prostate cancer and are implicated in the failure of hormone therapy. Most studies have emphasized the ligand-binding domain (LBD) where mutations can create promiscuous receptors, but mutations in the NH(2)-terminal transactivation domain have also been found. To assess AR alteration as a mechanism of treatment resistance, a mouse model (h/mAR-TRAMP) was used in which the murine AR coding region is replaced by human sequence and prostate cancer initiated by a transgenic oncogene. Mice received either no treatment, androgen depletion by castration, or treatment with antiandrogens, and 20 AR transcripts were sequenced per end-stage tumor. All tumors expressed several mutant alleles, although most mutations were low frequency. Some mutations that occurred multiple times within the population were differentially located dependent on treatment. Mutations in castrated or antiandrogen-treated mice were widely dispersed but with a prominent cluster in the LBD (amino acids 736-771), whereas changes in intact mice centered near the NH(2)-terminal polymorphic glutamine tract. Functional characterization of selected LBD mutant alleles showed diverse effects on AR activity, with about half of the mutations reducing transactivation in vitro. One receptor, AR-R753Q, behaved in a cell- and promoter-dependent manner, although as a germ-line mutation it causes androgen insensitivity syndrome. This suggests that alleles that are loss of function during development may still activate a subset of AR targets to become gain of function in tumorigenesis. Mutant ARs may thus use multiple mechanisms to evade cancer treatment.

Divergent effects of castration on prostate cancer in TRAMP mice: possible implications for therapy

PURPOSE

Divergent responses to androgen deprivation have been found in patients and in animal models of prostate cancer. The molecular basis for these different outcomes is unknown. Our aim was to identify the molecular responses of prostate cancer with divergent outcomes to androgen deprivation in TRAMP mice.

EXPERIMENTAL DESIGN

Castrated and noncastrated B6xFVB TRAMP mice were evaluated for survival, tumor development, pathology, and expressions of specific proteins at different time points.

RESULTS

TRAMP mice responded differentially to androgen deprivation. In the majority, primary tumors regressed after castration (positive response), whereas in others the tumors grew even more aggressively than in the noncastrated mice (negative response). Mice with regressed tumors had the highest survival rates. Androgen receptor was elevated in all tumors from castrated mice despite significant differences in tumor sizes. In positively responding tumors, expressions of Bcl-2 and Grp78 were greatly increased by 10 weeks after castration, whereas expressions of Bax, Bcl-xl, SV40 T antigen, and c-myc were lower. These tumors also showed a reduction in proliferating cells compared with noncastrates and negatively responding tumors. Most of these changes disappeared 20 weeks after castration, by which time there was an increase in the size of primary tumors as well as in distant metastasis.

CONCLUSIONS

In TRAMP prostate cancer that responded positively to castration, different expression patterns of proteins involved in cellular apoptosis, stress, and proliferation occur approximately 10 weeks after castration. This may be an optimal time for targeting Bcl-2, and perhaps Grp78, to enhance the antitumor effects of androgen deprivation.

Effects of castration on the development of prostate adenocarcinoma from its precursor HGPIN and on the occurrence of androgen-independent, poorly differentiated carcinoma in TRAMP mice

Androgen ablation is thought to exert selective pressure for the development of androgen-independent forms of prostate cancer. This study was set up to investigate the effects of surgical castration on the development of prostate adenocarcinoma (ADC) from its precursor (high-grade prostate intraepithelial neoplasia (HGPIN)) and on the occurrence of androgen-independent, poorly differentiated carcinoma (PDC) in (C57Bl/6 transgenic adenocarcinoma of mouse prostate) TRAMP mice. It was found that castration cures HGPIN and ADC and prevents their further occurrence and growth, but has no effect on PDC. This indicates that in this model, PDC is not the progression of ADC favoured by androgen ablation and that its initiating cells are different from those of HGPIN and ADC.

Androgen receptor variants and prostate cancer in humanized AR mice

Androgen, acting via the androgen receptor (AR), is central to male development, differentiation and hormone-dependent diseases such as prostate cancer. AR is actively involved in the initiation of prostate cancer, the transition to androgen independence, and many mechanisms of resistance to therapy. To examine genetic variation of AR in cancer, we created mice by germ-line gene targeting in which human AR sequence replaces that of the mouse. Since shorter length of a polymorphic N-terminal glutamine (Q) tract has been linked to prostate cancer risk, we introduced alleles with 12, 21 or 48 Qs to test this association. The three "humanized" AR mouse strains (h/mAR) are normal physiologically, as well as by cellular and molecular criteria, although slight differences are detected in AR target gene expression, correlating inversely with Q tract length. However, distinct allele-dependent differences in tumorigenesis are evident when these mice are crossed to a transgenic prostate cancer model. Remarkably, Q tract variation also differentially impacts disease progression following androgen depletion. This finding emphasizes the importance of AR function in androgen-independent as well as androgen-dependent disease. These mice provide a novel genetic paradigm in which to dissect opposing functions of AR in tumor suppression versus oncogenesis.

Origin of androgen-insensitive poorly differentiated tumors in the transgenic adenocarcinoma of mouse prostate model

Following castration, the transgenic adenocarcinoma of mouse prostate (TRAMP) model demonstrates rapid development of SV40-Tag-driven poorly differentiated tumors that express neuroendocrine cell markers. The cell population dynamics within the prostates of castrated TRAMP mice were characterized by analyzing the incorporation of 5-bromodeoxyuridine (BrdUrd) and the expression of SV40-Tag, synaptophysin, and androgen receptor (AR). Fourteen days postcastration, the remaining epithelial cells and adenocarcinoma cells were nonproliferative and lacked detectable SV40-Tag or synaptophysin expression. In contrast, morphologically distinct intraglandular foci were identified which expressed SV40-Tag, synaptophysin, and Ki67, but that lacked AR expression. These proliferative SV40-Tag and synaptophysin-expressing intraglandular foci were associated with the rare BrdUrd-retaining cells. These foci expanded rapidly in the postcastration prostate environment, in contrast to the AR- and SV40-Tag-expressing adenocarcinoma cells that lost SV40-Tag expression and underwent apoptosis after castration. Intraglandular foci of synaptophysin-expressing cells were also observed in the prostates of intact TRAMP mice at a comparable frequency; however, they did not progress to rapidly expanding tumors until much later in the life of the mice. This suggests that the foci of neuroendocrine-like cells that express SV40-Tag and synaptophysin, but lack AR, arise independent of androgen-deprivation and represent the source of the poorly differentiated tumors that are the lethal phenotype in the TRAMP model.

Glutamine tract length of human androgen receptors affects hormone-dependent and -independent prostate cancer in mice

The androgen receptor (AR) is involved in the initiation and progression of prostate cancer and its transition to androgen independence. Genetic variation in AR may contribute to disease risk and has been studied for a polymorphic N-terminal glutamine (Q) tract that shows population heterogeneity. While the length of this tract is known to affect AR in vitro, association with disease is complicated by genetic and environmental factors that have led to discordant epidemiological findings. To clarify the effect of Q tract polymorphism on prostate cancer, we created mice bearing humanized AR genes (h/mAr) varying in Q tract length. ARs with short Q tracts (12Q), which are transcriptionally more active, induce earlier disease in the transgene-induced TRAMP prostate cancer model than alleles with median (21Q) or long (48Q) tracts. Disease length varies within each genotype, with greater differentiation and AR expression in slower growing tumors. Remarkably, following androgen ablation, Q tract length has effects that are also allele-dependent and in directions opposite to those in hormone intact mice. Differences in AR activity conferred by Q tract length thus appear to direct distinct pathways of androgen-independent as well as androgen-dependent progression, and highlight substantial risk that may be associated with alterations in the androgen axis. This AR allelic series in humanized mice provides an experimental paradigm to dissect the role of AR in prostate cancer initiation and progression, to model response to treatment and to test therapies targeted specifically to the human AR.

5-aza-2'-deoxycytidine delays androgen-independent disease and improves survival in the transgenic adenocarcinoma of the mouse prostate mouse model of prostate cancer

PURPOSE

We have previously shown that 5-aza-2'-deoxycytidine (5-aza) is an effective chemopreventive agent capable of preventing early disease progression in the transgenic adenocarcinoma of the mouse prostate (TRAMP) model. The purpose of this study was to determine the effect of 5-aza on preexisting TRAMP prostate cancers and prevention of androgen-independent prostate cancer.

EXPERIMENTAL DESIGN

TRAMP mice with established prostate cancers were treated with 5-aza, castration, castration + 5-aza, or vehicle control (PBS). One cohort of 22 mice per treatment was euthanized after 10 weeks of treatment, whereas a second cohort of 14 mice per group was followed until death to determine survival. Histologic sections of prostate, pelvic lymph nodes, lung, and liver were blinded and analyzed by a certified genitourinary pathologist (K.J.W.).

RESULTS

Combined treatment (castration + 5-aza) provided significant survival benefits over either single treatment (combined versus castration P = 0.029, combined versus 5-aza P = 0.036). At 24 weeks of age, 86% of mice within the PBS cohort exhibited histologic evidence of prostate cancer, whereas only 47% of the combined cohort exhibited malignant disease (P < 0.0001). Additionally, whereas 43% of the PBS treatment group exhibited lymph node metastases, these were only observed in 21% of the combined treatment mice.

CONCLUSIONS

This is the first study to examine the effect of 5-aza and combined castration + 5-aza on preexisting prostate cancer in an animal model. Based on these preclinical findings, we suggest that 5-aza treatment may prolong the time to an androgen-independent status and thus survival in a hormone-deprived setting in prostate cancer.

Breast cancer resistance protein-mediated efflux of androgen in putative benign and malignant prostate stem cells

Malignantly transformed stem cells represent a potential common nidus for the primary cancer and the recurrent cancer that arises after treatment failure. Putative prostate stem cells and prostate tumor stem cells in benign and malignant human prostate tissue, in primary human prostate xenografts, and in the transgenic adenocarcinoma of the mouse prostate (TRAMP) mouse model of prostate cancer, are defined by expression of breast cancer resistance protein (BCRP), a marker of pluripotent hematopoietic, muscle, and neural stem cells, and by an absence of androgen receptor (AR) protein. Inhibition of BCRP-mediated efflux of dihydrotestosterone by novobiocin or fumitremorgin C in a rat prostate progenitor cell line that expresses BCRP and AR mRNAs, but minimal AR protein, results in stabilization and nuclear translocation of AR protein, providing a mechanism for lack of AR protein in BCRP-expressing stem cells. In both benign and malignant human prostate tissue, the rare epithelial cells that express BCRP and lack AR protein are localized in the basal cell compartment, survive androgen deprivation, and maintain proliferative potential in the hypoxic, androgen-deprived prostate. Putative prostate tumor stem cells that express BCRP but not AR protein in TRAMP are the source of a BCRP-negative and AR-negative, Foxa2- and SV40Tag-expressing, transit amplifying compartment that progresses to the poorly differentiated carcinomas that arise rapidly after castration. Therefore, BCRP expression isolates prostate stem/tumor stem cells from the prostate tissue microenvironment through constitutive efflux of androgen, protecting the putative tumor stem cells from androgen deprivation, hypoxia, or adjuvant chemotherapy, and providing the nidus for recurrent prostate cancer.