Most prostate tumors rely on male hormones – called androgens – to survive. Aggressive prostate cancer is often treated with drugs that block androgens, which usually cause the prostate tumors to shrink. One class of the drugs works by binding to and inactivating the androgen receptor protein on prostate cancer cells. However, aggressive prostate tumors can often become resistant to these anti-androgen therapies.

It is not clear where the resistant cancer cells come from. In 2009, researchers showed that the normal prostate contains some cells that appear to be independent of androgens. A subset of these cells – also known as CARNs – can act as stem or progenitor cells that can repair the prostate after injury. These normal androgen-independent cells can also be the cells from which prostate tumors arise. Here, Chua et al. – including one of the researchers from the 2009 study – investigated how these CARN cells behave when the androgen receptor is deleted.

When the androgen receptor was genetically removed in CARN cells of otherwise healthy mice, the behavior of CARN cells was unaffected. When the androgen receptor was deleted together with a protein that normally suppresses the formation of tumors, it protected the mice from prostate cancer. However, Chua et al. also observed that deleting the androgen receptor could not prevent the tumor from growing when two cancer-causing mutations were present. These tumors were similar to human prostate tumors that are resistant to anti-androgen therapy.

Since CARN cells may also exist in humans, this new way of making prostate cancers in mice may be used to study how these resistances arise in patients. A better understanding of how prostate tumors develop might lead to new treatments in which the androgen receptor is blocked in combination with other new protein targets.

Elucidating the cell type(s) of origin of cancer and the molecular drivers of tumor initiation is of fundamental importance in understanding the basis of distinct tumor subtypes as well as differences in treatment response and patient outcomes (Blanpain, 2013; Rycaj and Tang, 2015; Shibata and Shen, 2013; Visvader, 2011). Furthermore, since cancer often originates from stem cells and/or lineage-restricted progenitor cells, the identification of stem/progenitor cells is of considerable significance. In the case of the prostate, however, both the specific identity of stem/progenitor cells as well as cell types of origin for cancer have remained unclear (Lee and Shen, 2015; Wang and Shen, 2011; Xin, 2013).

In the normal prostate epithelium, there are three primary cell types, corresponding to secretory luminal cells, an underlying layer of basal cells, and rare neuroendocrine cells (Shen and Abate-Shen, 2010; Toivanen and Shen, 2017). Lineage-tracing studies have shown that both luminal and basal cells are mostly lineage-restricted (unipotent) in the normal adult mouse prostate as well as during androgen-mediated prostate regeneration (Choi et al., 2012; Liu et al., 2011; Lu et al., 2013; Wang et al., 2013). In addition, cells within the basal compartment possess stem/progenitor properties in a range of ex vivo assays as well as during inflammation and wound repair (Goldstein et al., 2008; Höfner et al., 2015; Kwon et al., 2014; Lawson et al., 2007; Toivanen et al., 2016; Wang et al., 2013). However, recent studies have shown that luminal cells can also display stem/progenitor properties in specific in vivo and ex vivo contexts (Chua et al., 2014; Karthaus et al., 2014; Kwon et al., 2016; Wang et al., 2009). Furthermore, there is now considerable evidence supporting a luminal origin for prostate cancer, both in mouse models (Wang et al., 2009; Wang et al., 2014) as well as in human tissues (Gurel et al., 2008; Meeker et al., 2002).

Androgen receptor (AR) plays a central role in many aspects of normal prostate development as well as prostate cancer progression (Cunha et al., 2004; Toivanen and Shen, 2017; Watson et al., 2015). In the prostate epithelium of adult hormonally intact mice, AR is primarily expressed by luminal cells, but is also found in a subset of basal cells (Lee et al., 2012; Mirosevich et al., 1999; Xie et al., 2017). Several studies have shown that conditional deletion of AR in the adult prostate epithelium results in a short-term increase in proliferation of luminal cells (Wu et al., 2007; Xie et al., 2017; Zhang et al., 2016a), indicating a role for AR in normal prostate homeostasis. Importantly, AR can act as a master regulator of prostate epithelial specification in a fibroblast reprogramming assay (Talos et al., 2017).

In the context of prostate cancer, tumor recurrence after androgen-deprivation therapy is due to the emergence of castration-resistant prostate cancer (CRPC), which is associated with increased AR activity that can be targeted by second-generation anti-androgen therapies (Watson et al., 2015). However, treatment failure following such anti-androgen therapies is frequently associated with the appearance of AR-negative tumor cells, which are typically associated with highly aggressive lethal disease (Beltran et al., 2014; Vlachostergios et al., 2017; Watson et al., 2015). In some cases, this AR-negative CRPC contain large regions displaying a neuroendocrine phenotype (CRPC-NE) (Beltran et al., 2016, 2014; Ku et al., 2017; Mu et al., 2017; Zou et al., 2017).

Previous work from our laboratory has identified CARNs as a luminal stem/progenitor cell within the androgen-deprived normal mouse prostate epithelium that is also a cell of origin for prostate cancer (Wang et al., 2009). Following androgen administration to induce prostate regeneration, CARNs can generate both luminal and basal progeny in vivo, as well as in renal grafting and organoid assays (Chua et al., 2014; Wang et al., 2009). Although CARNs express AR (Wang et al., 2009), it has been unclear whether AR is required for any or all the progenitor properties of CARNs, and whether the intrinsic castration-resistance of untransformed CARNs might resemble the castration-resistance of tumor cells in CRPC. Below, we show that the progenitor properties of CARNs are largely unaffected by loss of AR, whereas their ability to serve as cells of origin for prostate cancer are altered by AR deletion in a context-dependent manner. Notably, cell lines derived from AR-deleted CARNs have molecular profiles that resemble those for CRPC, and AR-deleted CARNs can serve as a cell of origin for focal neuroendocrine differentiation in a novel mouse model of AR-negative prostate cancer.

To investigate whether the stem/progenitor properties of CARNs are dependent upon AR function, we have used an inducible targeting approach in genetically engineered mice. For this purpose, we used mice carrying a conditional allele of Ar (De Gendt et al., 2004) together with the inducible Nkx3.1^CreERT2 driver (Wang et al., 2009) and the R26R-YFP reporter to visualize cells and their progeny in which Cre-mediated recombination has taken place (Srinivas et al., 2001); as Ar is an X-linked gene, deletion of a single allele in males is sufficient to confer a hemizygous null phenotype. Since CARNs are Nkx3.1-expressing cells found under androgen-deprived conditions, we castrated adult male mice carrying the Cre driver and reporter alleles, followed by tamoxifen induction to induce Cre-mediated activity specifically in CARNs (Figure 1A).

Figure 1

Download asset Open asset

Figure 1—source data 1

Quantitation of CARNs and AR-deleted CARNs in vivo.

https://doi.org/10.7554/eLife.28768.004

Download elife-28768-fig1-data1-v3.docx

Using this strategy, we compared the properties of CARNs in Nkx3.1^CreERT2/+; R26R-YFP/+ mice, which we denote as ‘control’ mice, with those in Nkx3.1^CreERT2/+; Ar^flox/Y; R26R-YFP/+ mice, which we denote as ‘AR-deleted’ mice. We found that the percentage of lineage-marked YFP-positive cells, corresponding to CARNs, was not significantly different (p=0.51) between the control (0.36 ± 0.17%, n = 5 mice) and AR-deleted mice (0.31 ± 0.06%, n = 5 mice) (Figure 1B,C). Notably, we found that 87.1% of the YFP-positive cells in Nkx3.1^CreERT2/+; Ar^flox/Y; R26R-YFP/+ mice (n = 344/395 cells in four mice) were AR-negative, indicating that AR deletion occurred with high efficiency. Furthermore, these YFP-positive cells expressed the luminal markers cytokeratins 8 and 18 (CK8 and CK18), but not cytokeratin 5 (CK5) and p63, indicating that AR deletion does not alter the luminal phenotype of CARNs (Figure 1D). These findings indicate that AR deletion does not affect the frequency or luminal properties of CARNs.

To investigate the progenitor properties of AR-deleted CARNs, we examined their ability to generate progeny during androgen-mediated regeneration. We implanted subcutaneous mini-osmotic pumps containing testosterone into control Nkx3.1^CreERT2/+; R26R-YFP/+ mice as well as Nkx3.1^CreERT2/+; Ar^flox/Y; R26R-YFP/+ mice, followed by tissue harvest at 4, 7, 14, and 28 days later; the final 28-day time point corresponds to a fully regenerated prostate (Figure 2A). We found that the YFP-marked cells and cell clusters were similar in the control and AR-deleted prostates at 4 and 7 days after testosterone administration (Figure 2B,C). However, at 14 and 28 days, the control prostates contained many YFP-expressing cell clusters with more than 4 cells, whereas the prostates with AR-deleted CARNs mostly contained YFP-expressing single cells or doublets (Figure 2B,C).

Figure 2

Download asset Open asset

Figure 2—source data 1

Quantitation of YFP⁺ cells during regeneration.

https://doi.org/10.7554/eLife.28768.006

Download elife-28768-fig2-data1-v3.docx

To compare the proliferative ability of control and AR-deleted CARNs and their progeny, we pursued BrdU pulse-chase experiments during prostate regeneration. We performed castration and tamoxifen administration on control and AR-deleted mice, followed by androgen-mediated regeneration for 28 days, with administration of daily doses of BrdU either from days 1 through 4 of regeneration or from days 11 through 14 (Figure 3A,B). When BrdU was administered from days 1 through 4 of regeneration, we could readily detect BrdU⁺YFP⁺ cells in the control prostates (50.9 ± 11.8%, n = 3 mice) as well as AR-deleted prostates (62.9 ± 14.9%, n = 3 mice) (Figure 3C,E). In contrast, when BrdU was administered from days 11 through 14, we could only detect BrdU⁺YFP⁺ cells in the control prostates (11.1 ± 6.2%, n = 3 mice), but not in the AR-deleted prostates (0%, n = 3 mice) (Figure 3D,F). This difference suggests that AR-deleted CARNs and/or their progeny have a defect in proliferation during later stages of regeneration, consistent with the analysis of YFP⁺ cluster size (Figure 2B).

Figure 3

Download asset Open asset

Figure 3—source data 1

Quantitation of BrdU incorporation and renal grafting data.

https://doi.org/10.7554/eLife.28768.008

Download elife-28768-fig3-data1-v3.docx

Notably, although YFP-expressing basal cells could be readily identified in both control and AR-deleted prostates, there was an increase in the percentage of basal cells within the YFP⁺ population in the AR-deleted mice (Figure 2D). This difference was evident using either the basal marker CK5 (2.1% CK5⁺AR⁺YFP⁺ versus 19.2% CK5⁺AR^–YFP⁺) or p63 (3.5% p63⁺AR⁺YFP⁺ versus 14.6% p63⁺AR^–YFP⁺) (Figure 2D). These findings indicate that AR-deleted CARNs favor generation of basal progeny and/or that there is decreased proliferation or survival of luminal progeny during regeneration.

As a further test of the progenitor properties of AR-deleted CARNs, we examined their ability to generate prostate ducts in a tissue recombination/renal grafting assay. Previously, we had shown that single CARNs were capable of generating ducts in this assay (Wang et al., 2009). We isolated YFP-positive cells from control and AR-deleted mice that had undergone castration and tamoxifen induction, and recombined 10 YFP-positive cells together with 2.5 × 10⁵ rat embryonic urogenital mesenchyme cells, followed by renal grafting (Figure 3G). We found that both control and AR-deleted CARNs could generate prostate ducts (Figure 3H), but that the AR-deleted CARN-s were significantly less efficient (12.5% graft efficiency, n = 16) compared to the control CARNs (p=0.003; 68.8% graft efficiency, n = 16), consistent with a proliferation defect in the AR-deleted CARNs.

Based on these findings, we further investigated the properties of CARNs and AR-deleted CARNs by establishing adherent cell lines. Using a novel method based on conditions that we previously established for culture of prostate organoids (Chua et al., 2014), we successfully generated adherent cell lines from single YFP⁺ cells isolated from castrated and tamoxifen-treated Nkx3.1^CreERT2/+; Ar^flox/Y; R26R-YFP/+ mice. Genotyping of the resulting lines led to identification of Ar-positive (non-recombined allele) and Ar-negative (recombined allele) lines, which we term APCA and ADCA (Ar-Positive CArn-derived and Ar-Deleted CArn-derived) lines. These cell lines could be propagated as adherent cells in the presence of Matrigel and DHT. Under these conditions, we found that the APCA (n = 2) and ADCA (n = 2) lines were morphologically indistinguishable (Figure 4A). These cell lines were comprised of a mixture of cells expressing basal (CK5) or luminal (CK8) markers or both, as well as Foxa1, an epithelial marker that encodes a transcriptional partner of AR (Gao et al., 2003; He et al., 2010) (Figure 4A). Furthermore, both the APCA and ADCA lines showed robust proliferation at similar levels, as demonstrated by Ki67 immunostaining, CellTiter-Glo assays, and colony formation in the presence or absence of DHT (Figure 4A–C).

Figure 4 with 1 supplement see all

Download asset Open asset

Figure 4—source data 1

Epithelial cell lines established from mouse and human prostate tissue.

https://doi.org/10.7554/eLife.28768.011

Download elife-28768-fig4-data1-v3.docx

To determine the relative efficiency of forming APCA and ADCA lines from AR⁺ and AR^– CARNs, respectively, we sorted 60 single YFP⁺ cells from castrated and tamoxifen-treated Nkx3.1^CreERT2/+; Ar^flox/Y; R26R-YFP/+ mice into individual wells of a 96-well plate. We found that six YFP⁺ cells gave rise to adherent lines, with four of these corresponding to AR⁺ lines that had failed to undergo Cre-mediated recombination of the conditional Ar allele, and two lines corresponding to AR^– lines. After correcting for the 87.1% efficiency of recombination of the AR-floxed allele in vivo, these data indicate that the relative plating efficiency for the AR^– CARNs compared to AR⁺ CARNs is 7.4%, consistent with the decreased grafting efficiency of AR^– CARNs.

Notably, we were also able to use this methodology to establish 14 primary human prostate epithelial cell lines from benign prostatectomy specimens at high efficiency. Similar to the mouse APCA cell lines, these HPE (Human Prostate Epithelial) cell lines are propagated as adherent cells in the presence of Matrigel and DHT. All these lines display similar marker phenotypes, expressing basal and luminal markers as well as AR and PSA, and are highly proliferative (Figure 4—figure supplement 1).

To assess the ability of the APCA and ADCA cell lines to reconstitute prostate ducts, we performed tissue recombination assays by combining 1 × 10⁵ cells with rat urogenital mesenchyme followed by renal grafting. We found that the APCA lines could generate prostate ducts (n = 10 grafts with two lines; 100% efficiency), some with evidence of secretions, whereas the ADCA lines (n = 6 grafts with one line; 67% efficiency) generated ducts that lacked prostate secretions (Figure 4D). Next, we tested the role of AR in this tissue reconstitution assay by treating the mice grafted with APCA cells (n = 12 grafts with two lines) with tamoxifen at 7 weeks after grafting in order to induce Ar deletion. We found that tamoxifen treatment resulted in grafts containing prostate ducts composed of a mixture of AR-positive and negative cells, but with a decreased efficiency of graft formation relative to the same APCA lines in the absence of tamoxifen (42% versus 100% efficiency) (Figure 4D). Taken together, these results show that AR deletion decreases the efficiency of prostate duct formation by CARN-derived cells, consistent with the results obtained using AR-deleted CARNs (Figure 3H). Notably, since ADCA cells do not display a growth disadvantage relative to APCA cells in culture, this difference in duct formation is likely to be due to a non-cell-autonomous effect mediated by the urogenital mesenchyme in grafts.

To examine the molecular basis for differences between the ADCA and APCA lines (n = 2 lines each), we performed RNA-sequencing followed by bioinformatic analyses. Based on the RNA expression profiling data, we constructed a differential expression signature comparing ADCA cells to APCA cells. Using the resulting ADCA signature to examine pathway enrichment by Gene Set Enrichment Analysis (GSEA) (Subramanian et al., 2005), we found up-regulation of gene sets involved in DNA replication and repair pathways, as well as cell cycle and apoptosis (Figure 5A), suggesting that cellular proliferation and survival are affected by AR deletion. We also compared the ADCA signature with a signature defined between expression profiles of AR-null and AR-positive mouse prostate luminal cells (Xie et al., 2017) and found enrichment for up-regulated genes (Figure 5B). Next, we performed a cross-species comparison of the ADCA signature with a signature defined between profiles of human prostate luminal and basal epithelial cells (Zhang et al., 2016b) and found that there was no significant enrichment in either tail (Figure 5C), indicating that AR deletion does not drive APCA cells towards a specific lineage. Furthermore, we performed GSEA comparisons of the ADCA signature with several signatures obtained from analyses of human prostate cancer progression. In particular, we observed enrichment for up-regulated genes when compared to a signature of CRPC from Best and colleagues (Best et al., 2005), as well as to a signature of metastatic CRPC from Stanbrough and colleagues (Stanbrough et al., 2006) (Figure 5D,E). Moreover, we observed a strong enrichment when compared to a signature from Beltran and colleagues (Beltran et al., 2016) defined between CRPC with neuroendocrine differentiation (CRPC-NE) and non-neuroendocrine CRPC (Figure 5F), consistent with our observation of pathway enrichment for gene sets corresponding to axon guidance and small-cell lung cancer (Figure 5A).

Figure 5

Download asset Open asset

Finally, we tested the ability of AR-deleted CARNs to serve as a cell of origin for prostate cancer, based on the previous finding that prostate cancer can initiate from CARNs after specific deletion of Pten and androgen-mediated regeneration (Wang et al., 2009). We used Nkx3.1^CreERT2/+; Pten^flox/flox; R26R-YFP/+ controls (which we term NP-CARN) and Nkx3.1^CreERT2/+; Pten^flox/flox; Ar^flox/Y; R26R-YFP/+ mice (NPA-CARN) in an experimental paradigm involving castration, tamoxifen-treatment, and androgen-mediated regeneration for one month. We found that AR deletion resulted in a significant difference between the NP-CARN and NPA-CARN phenotypes, as the NP-CARN controls displayed high-grade prostatic intraepithelial neoplasia (PIN), whereas the NPA-CARN prostates showed a weak phenotype corresponding to diffuse hyperplasia with mild inflammation and increased apoptosis, (Figure 6A). The NPA-CARN prostates contained YFP-positive cells that also expressed phospho-Akt (pAkt), indicating successful deletion of Pten, but these cells were only found as solitary or as small clusters, unlike the large clusters of YFP⁺pAkt⁺ cells observed in the control NP prostates (Figure 6A). Furthermore, the NPA-CARN prostates displayed a decreased proliferative index relative to NP-CARN (2.7%, n = 3 vs. 9.2%, n = 3), as well as increased apoptosis (2.6%, n = 3 vs. 0.7%, n = 3) (Figure 6B). Taken together, findings indicate that AR is required for tumor initiation following Pten deletion in CARNs.

Figure 6

Download asset Open asset

Figure 6—source data 1

Tumor phenotypes and marker quantitation.

https://doi.org/10.7554/eLife.28768.014

Download elife-28768-fig6-data1-v3.docx

In contrast, AR deletion did not affect tumor initiation following combined deletion of Pten and activation of the oncogenic Kras^G12D allele. Using a similar protocol for castration, tamoxifen-treatment, and androgen-mediated regeneration, we compared the phenotypes of Nkx3.1^CreERT2/+; Pten^flox/flox; Kras^LSL-G12D/+; R26R-YFP/+ controls (NPK-CARN) and Nkx3.1^CreERT2/+; Pten^flox/flox; Kras^LSL-G12D/+; Ar^flox/Y; R26R-YFP/+ mice (NPKA-CARN). In both genotypes, deletion of Pten and activation of oncogenic Kras resulted in formation of tumors with large clusters of YFP⁺ cells that express pAkt and Ras (Figure 6C). Furthermore, both NPK-CARN and NPKA-CARN tumors displayed high proliferative indices (20%, n = 3 vs. 19%, n = 3) and low frequencies of apoptosis (0.9%, n = 3 vs. 0.8%, n = 3) (Figure 6D). Notably, we observed an important difference between the NPK-CARN and NPKA-CARN tumors, as all the NPKA-CARN tumors contained a low but variable percentage of synaptophysin-positive neuroendocrine cells among total epithelial cells (0.7%, n = 3), which were never observed in the NPK-CARN controls (0%, n = 3) (Figure 6E,F). We also observed rare cells in all three NPKA-CARN tumors that expressed other neuroendocrine markers such as Chromogranin A, Foxa2, and Aurora kinase A (Figure 6E). Since the synaptophysin-postive cells co-expressed YFP (Figure 6E), we conclude that transformed AR-negative CARNs can give rise to neuroendocrine cells.

Taken together, our analyses have defined specific roles for AR in regulating the progenitor properties of CARNs, and indicate that the intrinsic castration-resistance of CARNs is independent of AR function. We find that targeted deletion of AR does not affect the percentage of CARNs, their luminal marker expression, or their ability to generate basal cells during androgen-mediated regeneration. However, there are fewer luminal progeny from AR-deleted CARNs during regeneration in vivo, and there is a decreased efficiency of prostate duct formation by both AR-deleted CARNs and ADCA cells in renal grafts. Thus, AR deletion in CARNs may primarily affect the proliferation and/or survival of their luminal progeny in vivo, although an effect on CARNs themselves cannot be excluded.

Interestingly, our results suggest potential roles of the stroma in modulating the proliferation of CARNs and/or their luminal progeny. Notably, BrdU incorporation assays reveal a proliferation defect of AR-deleted CARNs during later stages of regeneration but not during early regeneration. One possible explanation is that AR activity may cell-autonomously regulate the proliferation of luminal progeny of CARNs; alternatively, however, stromal remodeling during later stages of regeneration may alter non-cell autonomous signals that regulate luminal proliferation. Furthermore, since ADCA cells do not display a growth defect in culture, their decreased efficiency of prostate duct formation in grafts is likely due to a non-cell-autonomous inhibitory effect from the stroma.

Our study has also yielded interesting insights into differences between CARNs and other luminal epithelial cells. While this manuscript was in preparation, another study also investigated the requirements of AR in CARNs, and reported that AR-deleted CARNs completely failed to generate progeny during regeneration (Xie et al., 2017). This apparent discrepancy may be partially explained by our observation that AR-deleted CARNs can still generate basal progeny, and by the failure of progeny from AR-deleted CARNs to proliferate at later stages of androgen-mediated regeneration. However, we concur that CARNs require AR function to generate viable luminal progeny, which is not the case for most luminal cells during homeostasis or regeneration (Xie et al., 2017; Zhang et al., 2016a). Furthermore, the decreased proliferation of AR-deleted CARNs during regeneration contrasts with the transient increase in luminal proliferation observed after inducible AR deletion in the adult prostate epithelium, which is also a non-cell-autonomous effect mediated by the stroma (Zhang et al., 2016a). Together with our previous finding that CARNs display increased organoid formation efficiency relative to other luminal cells (Chua et al., 2014), these findings support the identification of CARNs as a distinct luminal population with stem/progenitor properties, and highlight the complexity of AR functions in the epithelial and stromal compartments.

In addition, we note that Xie and colleagues reported that Pten deletion in CARNs resulted in tumor formation after regeneration (Xie et al., 2017), unlike the absence of tumors that we observe in NPA-CARN mice. At present, the basis for this discrepancy remains unclear. Our finding that AR deletion results in failure of tumor formation following Pten inactivation could be due to differences between CARNs and bulk luminal cells and/or to differences due to Pten loss in the regressed versus hormonally intact prostate epithelium. In principle, these possibilities could potentially be distinguished using inducible Cre-drivers to delete Pten in bulk luminal cells in regressed versus hormonally intact mice.

Our findings on CARNs in mouse models may be of potential relevance for human prostate biology and cancer. Although CARNs are defined in the regressed prostate epithelium, and our in vivo studies involve manipulations performed after castration in mice, there is evidence that CARN-like cells exist in the human prostate from tissue-slice culture experiments (Zhao et al., 2010), as well as from analyses of prostate tumors after androgen-deprivation (Germann et al., 2012). However, it is less clear whether multipotent luminal progenitors can be identified in the context of the hormonally intact human prostate. Previous lineage-reconstruction studies using patterns of mitochondrial DNA mutations have indicated the existence of multipotent epithelial progenitors (Blackwood et al., 2011; Gaisa et al., 2011), and recent work has provided evidence for multipotent basal progenitors localized to the most proximal region of the prostate as well as more distally located unipotent luminal progenitors (Moad et al., 2017). Notably, ex vivo studies of human prostate organoids have demonstrated the existence of bipotential luminal progenitors (Karthaus et al., 2014). Thus, we believe that current data favor a general similarity of epithelial lineage relationships in the two species, suggesting that findings deduced from analyses of mice may be translatable to the human prostate.

The ability of CARNs to retain progenitor properties even in the absence of AR raises the possibility that CARNs represent a cell of origin for prostate cancers that are particularly susceptible to develop castration-resistance. Notably, under conditions of AR down-regulation, such as those that may occur during aging or inflammation, CARNs that lack tumor suppressors such as PTEN may represent a latent target for subsequent oncogenic events that can confer tumor growth, such as those activating the ERK MAP kinase pathway. Interestingly, our bioinformatic analyses of the ADCA cell line signature shows enrichment with castration-resistance signatures based on expression data from human prostate cancer patients (Best and Stanbrough signatures), consistent with increasing evidence supporting AR-independent mechanisms of castration-resistance (Beltran et al., 2014; Vlachostergios et al., 2017; Watson et al., 2015). In addition, the observed enrichment with the Beltran CRPC-NE signature suggests a similarity in gene expression programs with advanced cancers that lack AR activity, as neuroendocrine differentiation in prostate tumors is associated with loss of AR expression (Beltran et al., 2011). Notably, consistent with a role for AR loss in the emergence of neuroendocrine phenotypes, tumors in NPKA-CARN mice can display focal neuroendocrine differentiation, which has also been recently described in other mouse models of advanced prostate cancer (Ku et al., 2017; Zou et al., 2017).

In this regard, we note that the NP-CARN and NPK-CARN mice develop tumor phenotypes similar to those in NP and NPK mice, which have the same genotypes but whose tumors are induced by the Nkx3.1^CreERT2 driver in hormonally intact adult prostate (Aytes et al., 2013; Floc'h et al., 2012). Interestingly, NP tumors are initially castration-sensitive (Floc'h et al., 2012), consistent with the inability of NPA-CARN mice to develop tumors, whereas NPK tumors are castration-resistant (Aytes et al., 2013), consistent with the phenotype of NPKA-CARN tumors. The molecular basis for this switch is currently unclear, but it is conceivable that it involves ETS family transcription factors, which are known to interact with AR to positively and negatively modulate its activity (Baena et al., 2013; Bose et al., 2017; Chen et al., 2013); interestingly, ETV4 is up-regulated in NPK tumors and may be involved in this switch (Aytes et al., 2013). However, the focal neuroendocrine differentiation observed in NPKA-CARN tumors suggests that oncogenic transformation of AR-deleted CARNs can also facilitate transdifferentiation of luminal cells to neuroendocrine fates, as we have demonstrated for a Pten and Trp53 mutant mouse model (NPp53) after anti-androgen treatment (Zou et al., 2017).

Finally, since tumors initiated from CARNs following combined Pten deletion and Kras activation are at least partially independent of AR from their outset, it is conceivable that such tumors are intrinsically more resistant to second-generation anti-androgen therapies. Interestingly, recent studies have also identified distinct castration-resistant progenitors that express Bmi1 (CARBs) that are cells of origin for prostate cancer (Yoo et al., 2016). The development of targeted therapies directed at molecular features of CARNs and/or other castration-resistant luminal cells may therefore be relevant for successful combination with anti-androgen therapies.

1. 1. Chua
   2. CW
   3. Li BI
   4. Mitrofanova
   5. A
   6. Shen MM

   (2017) Differential requirements of androgen receptor in luminal progenitors during prostate regeneration and tumor initiation (APCA and ADCA lines RNASeq)

   GSE99233.

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE99233

1. 1. Best CJ
   2. Gillespie JW
   3. Yi Y
   4. Chandramouli GV
   5. Perlmutter MA
   6. Gathright Y
   7. Erickson HS
   8. Georgevich L
   9. Tangrea MA
   10. Duray PH
   11. González S
   12. Velasco A
   13. Linehan WM
   14. Matusik RJ
   15. Price DK
   16. Figg WD
   17. Emmert-Buck MR
   18. Chuaqui RF

   (2005) Molecular alterations in primary prostate cancer after androgen ablation therapy

   Publicly available at the NCBI Gene Expression Omnibus (accession no: GSE2443).

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE2443
2. 1. Stanbrough M
   2. Bubley GJ
   3. Ross K
   4. Golub TR
   5. Rubin MA
   6. Penning TM
   7. Febbo PG
   8. Balk SP

   (2006) Increased expression of genes converting adrenal androgens to testosterone in androgen-independent prostate cancer

   Publicly available at the NCBI Gene Expression Omnibus (accession no: GSE32269).

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE32269
3. 1. Beltran H
   2. Prandi D
   3. Mosquera JM
   4. Benelli M
   5. Puca L
   6. Cyrta J
   7. Marotz C
   8. Giannopoulou E
   9. Chakravarthi BV
   10. Varambally S
   11. Tomlins SA
   12. Nanus DM
   13. Tagawa ST
   14. Van Allen EM
   15. Elemento O
   16. Sboner A
   17. Garraway LA9
   18. Rubin MA
   19. Demichelis F

   (2016) Divergent clonal evolution of castration-resistant neuroendocrine prostate cancer

   Publicly available at dbGaP (accession no. dbGaP phs000909.v1.p1).

   https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000909.v1.p1
4. 1. Zhang D
   2. Park D
   3. Zhong Y
   4. Lu Y
   5. Rycaj K
   6. Gong S
   7. Chen X
   8. Liu X
   9. Chao HP
   10. Whitney P
   11. Calhoun-Davis T
   12. Takata Y
   13. Shen J
   14. Iyer VR
   15. Tang DG

   (2016) Stem cell and neurogenic gene-expression profiles link prostate basal cells to aggressive prostate cancer

   Publicly available at the NCBI Gene Expression Omnibus (accession no: GSE67070).

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE67070
5. 1. Xie Q
   2. Liu Y
   3. Cai T
   4. Horton C
   5. Stefanson J
   6. Wang ZA

   (2017) Dissecting cell-type-specific roles of androgen receptor in prostate homeostasis and regeneration through lineage tracing

   Publicly available at the NCBI Gene Expression Omnibus (accession no: GSE76724).

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE76724

1. 1. Aytes A
   2. Mitrofanova A
   3. Kinkade CW
   4. Lefebvre C
   5. Lei M
   6. Phelan V
   7. LeKaye HC
   8. Koutcher JA
   9. Cardiff RD
   10. Califano A
   11. Shen MM
   12. Abate-Shen C

   (2013) ETV4 promotes metastasis in response to activation of PI3-kinase and Ras signaling in a mouse model of advanced prostate cancer

   PNAS 110:E3506–E3515.

   https://doi.org/10.1073/pnas.1303558110

   • PubMed
   • Google Scholar
2. 1. Baena E
   2. Shao Z
   3. Linn DE
   4. Glass K
   5. Hamblen MJ
   6. Fujiwara Y
   7. Kim J
   8. Nguyen M
   9. Zhang X
   10. Godinho FJ
   11. Bronson RT
   12. Mucci LA
   13. Loda M
   14. Yuan GC
   15. Orkin SH
   16. Li Z

   (2013) ETV1 directs androgen metabolism and confers aggressive prostate cancer in targeted mice and patients

   Genes & Development 27:683–698.

   https://doi.org/10.1101/gad.211011.112

   • PubMed
   • Google Scholar
3. 1. Banach-Petrosky W
   2. Jessen WJ
   3. Ouyang X
   4. Gao H
   5. Rao J
   6. Quinn J
   7. Aronow BJ
   8. Abate-Shen C

   (2007) Prolonged exposure to reduced levels of androgen accelerates prostate cancer progression in Nkx3.1; Pten mutant mice

   Cancer Research 67:9089–9096.

   https://doi.org/10.1158/0008-5472.CAN-07-2887

   • PubMed
   • Google Scholar
4. 1. Beltran H
   2. Prandi D
   3. Mosquera JM
   4. Benelli M
   5. Puca L
   6. Cyrta J
   7. Marotz C
   8. Giannopoulou E
   9. Chakravarthi BV
   10. Varambally S
   11. Tomlins SA
   12. Nanus DM
   13. Tagawa ST
   14. Van Allen EM
   15. Elemento O
   16. Sboner A
   17. Garraway LA
   18. Rubin MA
   19. Demichelis F

   (2016) Divergent clonal evolution of castration-resistant neuroendocrine prostate cancer

   Nature Medicine 22:298–305.

   https://doi.org/10.1038/nm.4045

   • PubMed
   • Google Scholar
5. 1. Beltran H
   2. Rickman DS
   3. Park K
   4. Chae SS
   5. Sboner A
   6. MacDonald TY
   7. Wang Y
   8. Sheikh KL
   9. Terry S
   10. Tagawa ST
   11. Dhir R
   12. Nelson JB
   13. de la Taille A
   14. Allory Y
   15. Gerstein MB
   16. Perner S
   17. Pienta KJ
   18. Chinnaiyan AM
   19. Wang Y
   20. Collins CC
   21. Gleave ME
   22. Demichelis F
   23. Nanus DM
   24. Rubin MA

   (2011) Molecular characterization of neuroendocrine prostate cancer and identification of new drug targets

   Cancer Discovery 1:487–495.

   https://doi.org/10.1158/2159-8290.CD-11-0130

   • PubMed
   • Google Scholar
6. 1. Beltran H
   2. Tomlins S
   3. Aparicio A
   4. Arora V
   5. Rickman D
   6. Ayala G
   7. Huang J
   8. True L
   9. Gleave ME
   10. Soule H
   11. Logothetis C
   12. Rubin MA

   (2014) Aggressive variants of castration-resistant prostate cancer

   Clinical Cancer Research 20:2846–2850.

   https://doi.org/10.1158/1078-0432.CCR-13-3309

   • PubMed
   • Google Scholar
7. 1. Best CJ
   2. Gillespie JW
   3. Yi Y
   4. Chandramouli GV
   5. Perlmutter MA
   6. Gathright Y
   7. Erickson HS
   8. Georgevich L
   9. Tangrea MA
   10. Duray PH
   11. González S
   12. Velasco A
   13. Linehan WM
   14. Matusik RJ
   15. Price DK
   16. Figg WD
   17. Emmert-Buck MR
   18. Chuaqui RF

   (2005) Molecular alterations in primary prostate cancer after androgen ablation therapy

   Clinical Cancer Research 11:6823–6834.

   https://doi.org/10.1158/1078-0432.CCR-05-0585

   • PubMed
   • Google Scholar
8. 1. Blackwood JK
   2. Williamson SC
   3. Greaves LC
   4. Wilson L
   5. Rigas AC
   6. Sandher R
   7. Pickard RS
   8. Robson CN
   9. Turnbull DM
   10. Taylor RW
   11. Heer R

   (2011) In situ lineage tracking of human prostatic epithelial stem cell fate reveals a common clonal origin for basal and luminal cells

   The Journal of Pathology 225:181–188.

   https://doi.org/10.1002/path.2965

   • PubMed
   • Google Scholar
9. 1. Blanpain C

   (2013) Tracing the cellular origin of cancer

   Nature Cell Biology 15:126–134.

   https://doi.org/10.1038/ncb2657

   • PubMed
   • Google Scholar
10. 1. Bose R
    2. Karthaus WR
    3. Armenia J
    4. Abida W
    5. Iaquinta PJ
    6. Zhang Z
    7. Wongvipat J
    8. Wasmuth EV
    9. Shah N
    10. Sullivan PS
    11. Doran MG
    12. Wang P
    13. Patruno A
    14. Zhao Y
    15. Zheng D
    16. Schultz N
    17. Sawyers CL
    18. International SU2C/PCF Prostate Cancer Dream Team

    (2017) ERF mutations reveal a balance of ETS factors controlling prostate oncogenesis

    Nature 546:671–675.

    https://doi.org/10.1038/nature22820

    • PubMed
    • Google Scholar
11. 1. Chen Y
    2. Chi P
    3. Rockowitz S
    4. Iaquinta PJ
    5. Shamu T
    6. Shukla S
    7. Gao D
    8. Sirota I
    9. Carver BS
    10. Wongvipat J
    11. Scher HI
    12. Zheng D
    13. Sawyers CL

    (2013) ETS factors reprogram the androgen receptor cistrome and prime prostate tumorigenesis in response to PTEN loss

    Nature Medicine 19:1023–1029.

    https://doi.org/10.1038/nm.3216

    • PubMed
    • Google Scholar
12. 1. Choi N
    2. Zhang B
    3. Zhang L
    4. Ittmann M
    5. Xin L

    (2012) Adult murine prostate basal and luminal cells are self-sustained lineages that can both serve as targets for prostate cancer initiation

    Cancer Cell 21:253–265.

    https://doi.org/10.1016/j.ccr.2012.01.005

    • PubMed
    • Google Scholar
13. 1. Chua CW
    2. Shibata M
    3. Lei M
    4. Toivanen R
    5. Barlow LJ
    6. Bergren SK
    7. Badani KK
    8. McKiernan JM
    9. Benson MC
    10. Hibshoosh H
    11. Shen MM

    (2014) Single luminal epithelial progenitors can generate prostate organoids in culture

    Nature Cell Biology 16:951–961.

    https://doi.org/10.1038/ncb3047

    • PubMed
    • Google Scholar
14. 1. Cunha GR
    2. Ricke W
    3. Thomson A
    4. Marker PC
    5. Risbridger G
    6. Hayward SW
    7. Wang YZ
    8. Donjacour AA
    9. Kurita T

    (2004) Hormonal, cellular, and molecular regulation of normal and neoplastic prostatic development

    The Journal of Steroid Biochemistry and Molecular Biology 92:221–236.

    https://doi.org/10.1016/j.jsbmb.2004.10.017

    • PubMed
    • Google Scholar
15. 1. De Gendt K
    2. Swinnen JV
    3. Saunders PT
    4. Schoonjans L
    5. Dewerchin M
    6. Devos A
    7. Tan K
    8. Atanassova N
    9. Claessens F
    10. Lécureuil C
    11. Heyns W
    12. Carmeliet P
    13. Guillou F
    14. Sharpe RM
    15. Verhoeven G

    (2004) A Sertoli cell-selective knockout of the androgen receptor causes spermatogenic arrest in meiosis

    PNAS 101:1327–1332.

    https://doi.org/10.1073/pnas.0308114100

    • PubMed
    • Google Scholar
16. 1. Fabregat A
    2. Sidiropoulos K
    3. Garapati P
    4. Gillespie M
    5. Hausmann K
    6. Haw R
    7. Jassal B
    8. Jupe S
    9. Korninger F
    10. McKay S
    11. Matthews L
    12. May B
    13. Milacic M
    14. Rothfels K
    15. Shamovsky V
    16. Webber M
    17. Weiser J
    18. Williams M
    19. Wu G
    20. Stein L
    21. Hermjakob H
    22. D'Eustachio P

    (2016) The Reactome pathway Knowledgebase

    Nucleic Acids Research 44:D481–D487.

    https://doi.org/10.1093/nar/gkv1351

    • PubMed
    • Google Scholar
17. 1. Floc'h N
    2. Kinkade CW
    3. Kobayashi T
    4. Aytes A
    5. Lefebvre C
    6. Mitrofanova A
    7. Cardiff RD
    8. Califano A
    9. Shen MM
    10. Abate-Shen C

    (2012) Dual targeting of the Akt/mTOR signaling pathway inhibits castration-resistant prostate cancer in a genetically engineered mouse model

    Cancer Research 72:4483–4493.

    https://doi.org/10.1158/0008-5472.CAN-12-0283

    • PubMed
    • Google Scholar
18. 1. Gaisa NT
    2. Graham TA
    3. McDonald SA
    4. Poulsom R
    5. Heidenreich A
    6. Jakse G
    7. Knuechel R
    8. Wright NA

    (2011) Clonal architecture of human prostatic epithelium in benign and malignant conditions

    The Journal of Pathology 225:172–180.

    https://doi.org/10.1002/path.2959

    • PubMed
    • Google Scholar
19. 1. Gao N
    2. Zhang J
    3. Rao MA
    4. Case TC
    5. Mirosevich J
    6. Wang Y
    7. Jin R
    8. Gupta A
    9. Rennie PS
    10. Matusik RJ

    (2003) The role of hepatocyte nuclear factor-3 alpha (Forkhead Box A1) and androgen receptor in transcriptional regulation of prostatic genes

    Molecular Endocrinology 17:1484–1507.

    https://doi.org/10.1210/me.2003-0020

    • PubMed
    • Google Scholar
20. 1. Germann M
    2. Wetterwald A
    3. Guzmán-Ramirez N
    4. van der Pluijm G
    5. Culig Z
    6. Cecchini MG
    7. Williams ED
    8. Thalmann GN

    (2012) Stem-like cells with luminal progenitor phenotype survive castration in human prostate cancer

    Stem Cells 30:1076–1086.

    https://doi.org/10.1002/stem.1087

    • PubMed
    • Google Scholar
21. 1. Goldstein AS
    2. Lawson DA
    3. Cheng D
    4. Sun W
    5. Garraway IP
    6. Witte ON

    (2008) Trop2 identifies a subpopulation of murine and human prostate basal cells with stem cell characteristics

    PNAS 105:20882–20887.

    https://doi.org/10.1073/pnas.0811411106

    • PubMed
    • Google Scholar
22. 1. Gurel B
    2. Iwata T
    3. Koh CM
    4. Jenkins RB
    5. Lan F
    6. Van Dang C
    7. Hicks JL
    8. Morgan J
    9. Cornish TC
    10. Sutcliffe S
    11. Isaacs WB
    12. Luo J
    13. De Marzo AM

    (2008) Nuclear MYC protein overexpression is an early alteration in human prostate carcinogenesis

    Modern Pathology 21:1156–1167.

    https://doi.org/10.1038/modpathol.2008.111

    • PubMed
    • Google Scholar
23. 1. He HH
    2. Meyer CA
    3. Shin H
    4. Bailey ST
    5. Wei G
    6. Wang Q
    7. Zhang Y
    8. Xu K
    9. Ni M
    10. Lupien M
    11. Mieczkowski P
    12. Lieb JD
    13. Zhao K
    14. Brown M
    15. Liu XS

    (2010) Nucleosome dynamics define transcriptional enhancers

    Nature Genetics 42:343–347.

    https://doi.org/10.1038/ng.545

    • PubMed
    • Google Scholar
24. 1. Höfner T
    2. Eisen C
    3. Klein C
    4. Rigo-Watermeier T
    5. Goeppinger SM
    6. Jauch A
    7. Schoell B
    8. Vogel V
    9. Noll E
    10. Weichert W
    11. Baccelli I
    12. Schillert A
    13. Wagner S
    14. Pahernik S
    15. Sprick MR
    16. Trumpp A

    (2015) Defined conditions for the isolation and expansion of basal prostate progenitor cells of mouse and human origin

    Stem Cell Reports 4:503–518.

    https://doi.org/10.1016/j.stemcr.2015.01.015

    • PubMed
    • Google Scholar
25. 1. Jackson EL
    2. Willis N
    3. Mercer K
    4. Bronson RT
    5. Crowley D
    6. Montoya R
    7. Jacks T
    8. Tuveson DA

    (2001) Analysis of lung tumor initiation and progression using conditional expression of oncogenic K-ras

    Genes & Development 15:3243–3248.

    https://doi.org/10.1101/gad.943001

    • PubMed
    • Google Scholar
26. 1. Karthaus WR
    2. Iaquinta PJ
    3. Drost J
    4. Gracanin A
    5. van Boxtel R
    6. Wongvipat J
    7. Dowling CM
    8. Gao D
    9. Begthel H
    10. Sachs N
    11. Vries RGJ
    12. Cuppen E
    13. Chen Y
    14. Sawyers CL
    15. Clevers HC

    (2014) Identification of multipotent luminal progenitor cells in human prostate organoid cultures

    Cell 159:163–175.

    https://doi.org/10.1016/j.cell.2014.08.017

    • PubMed
    • Google Scholar
27. 1. Ku SY
    2. Rosario S
    3. Wang Y
    4. Mu P
    5. Seshadri M
    6. Goodrich ZW
    7. Goodrich MM
    8. Labbé DP
    9. Gomez EC
    10. Wang J
    11. Long HW
    12. Xu B
    13. Brown M
    14. Loda M
    15. Sawyers CL
    16. Ellis L
    17. Goodrich DW

    (2017) Rb1 and Trp53 cooperate to suppress prostate cancer lineage plasticity, metastasis, and antiandrogen resistance

    Science 355:78–83.

    https://doi.org/10.1126/science.aah4199

    • PubMed
    • Google Scholar
28. 1. Kwon OJ
    2. Zhang L
    3. Ittmann MM
    4. Xin L

    (2014) Prostatic inflammation enhances basal-to-luminal differentiation and accelerates initiation of prostate cancer with a basal cell origin

    PNAS 111:E592–E600.

    https://doi.org/10.1073/pnas.1318157111

    • PubMed
    • Google Scholar
29. 1. Kwon OJ
    2. Zhang L
    3. Xin L

    (2016) Stem cell antigen-1 identifies a distinct androgen-independent murine prostatic luminal cell lineage with bipotent potential

    Stem Cells 34:191–202.

    https://doi.org/10.1002/stem.2217

    • PubMed
    • Google Scholar
30. 1. Lawson DA
    2. Xin L
    3. Lukacs RU
    4. Cheng D
    5. Witte ON

    (2007) Isolation and functional characterization of murine prostate stem cells

    PNAS 104:181–186.

    https://doi.org/10.1073/pnas.0609684104

    • PubMed
    • Google Scholar
31. 1. Lee SH
    2. Shen MM

    (2015) Cell types of origin for prostate cancer

    Current Opinion in Cell Biology 37:35–41.

    https://doi.org/10.1016/j.ceb.2015.10.002

    • PubMed
    • Google Scholar
32. 1. Lee SO
    2. Tian J
    3. Huang CK
    4. Ma Z
    5. Lai KP
    6. Hsiao H
    7. Jiang M
    8. Yeh S
    9. Chang C

    (2012) Suppressor role of androgen receptor in proliferation of prostate basal epithelial and progenitor cells

    Journal of Endocrinology 213:173–182.

    https://doi.org/10.1530/JOE-11-0474

    • PubMed
    • Google Scholar
33. 1. Lesche R
    2. Groszer M
    3. Gao J
    4. Wang Y
    5. Messing A
    6. Sun H
    7. Liu X
    8. Wu H

    (2002) Cre/loxP-mediated inactivation of the murine Pten tumor suppressor gene

    Genesis 32:148–149.

    https://doi.org/10.1002/gene.10036

    • PubMed
    • Google Scholar
34. 1. Liu J
    2. Pascal LE
    3. Isharwal S
    4. Metzger D
    5. Ramos Garcia R
    6. Pilch J
    7. Kasper S
    8. Williams K
    9. Basse PH
    10. Nelson JB
    11. Chambon P
    12. Wang Z

    (2011) Regenerated luminal epithelial cells are derived from preexisting luminal epithelial cells in adult mouse prostate

    Molecular Endocrinology 25:1849–1857.

    https://doi.org/10.1210/me.2011-1081

    • PubMed
    • Google Scholar
35. 1. Lu TL
    2. Huang YF
    3. You LR
    4. Chao NC
    5. Su FY
    6. Chang JL
    7. Chen CM

    (2013) Conditionally ablated Pten in prostate basal cells promotes basal-to-luminal differentiation and causes invasive prostate cancer in mice

    The American Journal of Pathology 182:975–991.

    https://doi.org/10.1016/j.ajpath.2012.11.025

    • PubMed
    • Google Scholar
36. 1. Meeker AK
    2. Hicks JL
    3. Platz EA
    4. March GE
    5. Bennett CJ
    6. Delannoy MJ
    7. De Marzo AM

    (2002)

    Telomere shortening is an early somatic DNA alteration in human prostate tumorigenesis

    Cancer Research 62:6405–6409.

    • PubMed
    • Google Scholar
37. 1. Mirosevich J
    2. Bentel JM
    3. Zeps N
    4. Redmond SL
    5. D'Antuono MF
    6. Dawkins HJ

    (1999) Androgen receptor expression of proliferating basal and luminal cells in adult murine ventral prostate

    Journal of Endocrinology 162:341–350.

    https://doi.org/10.1677/joe.0.1620341

    • PubMed
    • Google Scholar
38. 1. Moad M
    2. Hannezo E
    3. Buczacki SJ
    4. Wilson L
    5. El-Sherif A
    6. Sims D
    7. Pickard R
    8. Wright NA
    9. Williamson SC
    10. Turnbull DM
    11. Taylor RW
    12. Greaves L
    13. Robson CN
    14. Simons BD
    15. Heer R

    (2017) Multipotent basal stem cells, maintained in localized proximal niches, support directed long-ranging epithelial flows in human prostates

    Cell Reports 20:1609–1622.

    https://doi.org/10.1016/j.celrep.2017.07.061

    • PubMed
    • Google Scholar
39. 1. Mu P
    2. Zhang Z
    3. Benelli M
    4. Karthaus WR
    5. Hoover E
    6. Chen CC
    7. Wongvipat J
    8. Ku SY
    9. Gao D
    10. Cao Z
    11. Shah N
    12. Adams EJ
    13. Abida W
    14. Watson PA
    15. Prandi D
    16. Huang CH
    17. de Stanchina E
    18. Lowe SW
    19. Ellis L
    20. Beltran H
    21. Rubin MA
    22. Goodrich DW
    23. Demichelis F
    24. Sawyers CL

    (2017) SOX2 promotes lineage plasticity and antiandrogen resistance in TP53- and RB1-deficient prostate cancer

    Science 355:84–88.

    https://doi.org/10.1126/science.aah4307

    • PubMed
    • Google Scholar
40. 1. Ogata H
    2. Goto S
    3. Sato K
    4. Fujibuchi W
    5. Bono H
    6. Kanehisa M

    (1999) KEGG: kyoto encyclopedia of genes and genomes

    Nucleic Acids Research 27:29–34.

    https://doi.org/10.1093/nar/27.1.29

    • PubMed
    • Google Scholar
41. 1. Park JH
    2. Walls JE
    3. Galvez JJ
    4. Kim M
    5. Abate-Shen C
    6. Shen MM
    7. Cardiff RD

    (2002) Prostatic intraepithelial neoplasia in genetically engineered mice

    The American Journal of Pathology 161:727–735.

    https://doi.org/10.1016/S0002-9440(10)64228-9

    • PubMed
    • Google Scholar
42. 1. Rycaj K
    2. Tang DG

    (2015) Cell-of-origin of cancer versus cancer stem cells: assays and interpretations

    Cancer Research 75:4003–4011.

    https://doi.org/10.1158/0008-5472.CAN-15-0798

    • PubMed
    • Google Scholar
43. 1. Shen MM
    2. Abate-Shen C

    (2010) Molecular genetics of prostate cancer: new prospects for old challenges

    Genes & Development 24:1967–2000.

    https://doi.org/10.1101/gad.1965810

    • PubMed
    • Google Scholar
44. 1. Shibata M
    2. Shen MM

    (2013) The roots of cancer: stem cells and the basis for tumor heterogeneity

    BioEssays 35:253–260.

    https://doi.org/10.1002/bies.201200101

    • PubMed
    • Google Scholar
45. 1. Srinivas S
    2. Watanabe T
    3. Lin CS
    4. William CM
    5. Tanabe Y
    6. Jessell TM
    7. Costantini F

    (2001) Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus

    BMC Developmental Biology 1:4.

    https://doi.org/10.1186/1471-213X-1-4

    • PubMed
    • Google Scholar
46. 1. Stanbrough M
    2. Bubley GJ
    3. Ross K
    4. Golub TR
    5. Rubin MA
    6. Penning TM
    7. Febbo PG
    8. Balk SP

    (2006) Increased expression of genes converting adrenal androgens to testosterone in androgen-independent prostate cancer

    Cancer Research 66:2815–2825.

    https://doi.org/10.1158/0008-5472.CAN-05-4000

    • PubMed
    • Google Scholar
47. 1. Subramanian A
    2. Tamayo P
    3. Mootha VK
    4. Mukherjee S
    5. Ebert BL
    6. Gillette MA
    7. Paulovich A
    8. Pomeroy SL
    9. Golub TR
    10. Lander ES
    11. Mesirov JP

    (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles

    PNAS 102:15545–15550.

    https://doi.org/10.1073/pnas.0506580102

    • PubMed
    • Google Scholar
48. 1. Talos F
    2. Mitrofanova A
    3. Bergren SK
    4. Califano A
    5. Shen MM

    (2017) A computational systems approach identifies synergistic specification genes that facilitate lineage conversion to prostate tissue

    Nature Communications 8:14662.

    https://doi.org/10.1038/ncomms14662

    • PubMed
    • Google Scholar
49. 1. Toivanen R
    2. Mohan A
    3. Shen MM

    (2016) Basal progenitors contribute to repair of the prostate epithelium following induced luminal anoikis

    Stem Cell Reports 6:660–667.

    https://doi.org/10.1016/j.stemcr.2016.03.007

    • PubMed
    • Google Scholar
50. 1. Toivanen R
    2. Shen MM

    (2017) Prostate organogenesis: tissue induction, hormonal regulation and cell type specification

    Development 144:1382–1398.

    https://doi.org/10.1242/dev.148270

    • PubMed
    • Google Scholar
51. 1. Visvader JE

    (2011) Cells of origin in cancer

    Nature 469:314–322.

    https://doi.org/10.1038/nature09781

    • PubMed
    • Google Scholar
52. 1. Vlachostergios PJ
    2. Puca L
    3. Beltran H

    (2017) Emerging variants of castration-resistant prostate cancer

    Current Oncology Reports 19:32.

    https://doi.org/10.1007/s11912-017-0593-6

    • PubMed
    • Google Scholar
53. 1. Wang X
    2. Kruithof-de Julio M
    3. Economides KD
    4. Walker D
    5. Yu H
    6. Halili MV
    7. Hu YP
    8. Price SM
    9. Abate-Shen C
    10. Shen MM

    (2009) A luminal epithelial stem cell that is a cell of origin for prostate cancer

    Nature 461:495–500.

    https://doi.org/10.1038/nature08361

    • PubMed
    • Google Scholar
54. 1. Wang ZA
    2. Mitrofanova A
    3. Bergren SK
    4. Abate-Shen C
    5. Cardiff RD
    6. Califano A
    7. Shen MM

    (2013) Lineage analysis of basal epithelial cells reveals their unexpected plasticity and supports a cell-of-origin model for prostate cancer heterogeneity

    Nature Cell Biology 15:274–283.

    https://doi.org/10.1038/ncb2697

    • PubMed
    • Google Scholar
55. 1. Wang ZA
    2. Shen MM

    (2011) Revisiting the concept of cancer stem cells in prostate cancer

    Oncogene 30:1261–1271.

    https://doi.org/10.1038/onc.2010.530

    • PubMed
    • Google Scholar
56. 1. Wang ZA
    2. Toivanen R
    3. Bergren SK
    4. Chambon P
    5. Shen MM

    (2014) Luminal cells are favored as the cell of origin for prostate cancer

    Cell Reports 8:1339–1346.

    https://doi.org/10.1016/j.celrep.2014.08.002

    • PubMed
    • Google Scholar
57. 1. Watson PA
    2. Arora VK
    3. Sawyers CL

    (2015) Emerging mechanisms of resistance to androgen receptor inhibitors in prostate cancer

    Nature Reviews Cancer 15:701–711.

    https://doi.org/10.1038/nrc4016

    • PubMed
    • Google Scholar
58. 1. Wu CT
    2. Altuwaijri S
    3. Ricke WA
    4. Huang SP
    5. Yeh S
    6. Zhang C
    7. Niu Y
    8. Tsai MY
    9. Chang C

    (2007) Increased prostate cell proliferation and loss of cell differentiation in mice lacking prostate epithelial androgen receptor

    PNAS 104:12679–12684.

    https://doi.org/10.1073/pnas.0704940104

    • PubMed
    • Google Scholar
59. 1. Xie Q
    2. Liu Y
    3. Cai T
    4. Horton C
    5. Stefanson J
    6. Wang ZA

    (2017) Dissecting cell-type-specific roles of androgen receptor in prostate homeostasis and regeneration through lineage tracing

    Nature Communications 8:14284.

    https://doi.org/10.1038/ncomms14284

    • PubMed
    • Google Scholar
60. 1. Xin L

    (2013) Cells of origin for cancer: an updated view from prostate cancer

    Oncogene 32:3655–3663.

    https://doi.org/10.1038/onc.2012.541

    • PubMed
    • Google Scholar
61. 1. Yeh S
    2. Hu YC
    3. Wang PH
    4. Xie C
    5. Xu Q
    6. Tsai MY
    7. Dong Z
    8. Wang RS
    9. Lee TH
    10. Chang C

    (2003) Abnormal mammary gland development and growth retardation in female mice and MCF7 breast cancer cells lacking androgen receptor

    The Journal of Experimental Medicine 198:1899–1908.

    https://doi.org/10.1084/jem.20031233

    • PubMed
    • Google Scholar
62. 1. Yoo YA
    2. Roh M
    3. Naseem AF
    4. Lysy B
    5. Desouki MM
    6. Unno K
    7. Abdulkadir SA

    (2016) Bmi1 marks distinct castration-resistant luminal progenitor cells competent for prostate regeneration and tumour initiation

    Nature Communications 7:12943.

    https://doi.org/10.1038/ncomms12943

    • PubMed
    • Google Scholar
63. 1. Zhang B
    2. Kwon OJ
    3. Henry G
    4. Malewska A
    5. Wei X
    6. Zhang L
    7. Brinkley W
    8. Zhang Y
    9. Castro PD
    10. Titus M
    11. Chen R
    12. Sayeeduddin M
    13. Raj GV
    14. Mauck R
    15. Roehrborn C
    16. Creighton CJ
    17. Strand DW
    18. Ittmann MM
    19. Xin L

    (2016a) Non-Cell-autonomous regulation of prostate epithelial homeostasis by androgen receptor

    Molecular Cell 63:976–989.

    https://doi.org/10.1016/j.molcel.2016.07.025

    • PubMed
    • Google Scholar
64. 1. Zhang D
    2. Park D
    3. Zhong Y
    4. Lu Y
    5. Rycaj K
    6. Gong S
    7. Chen X
    8. Liu X
    9. Chao HP
    10. Whitney P
    11. Calhoun-Davis T
    12. Takata Y
    13. Shen J
    14. Iyer VR
    15. Tang DG

    (2016b) Stem cell and neurogenic gene-expression profiles link prostate basal cells to aggressive prostate cancer

    Nature Communications 7:10798.

    https://doi.org/10.1038/ncomms10798

    • PubMed
    • Google Scholar
65. 1. Zhao H
    2. Nolley R
    3. Chen Z
    4. Peehl DM

    (2010) Tissue slice grafts: an in vivo model of human prostate androgen signaling

    The American Journal of Pathology 177:229–239.

    https://doi.org/10.2353/ajpath.2010.090821

    • PubMed
    • Google Scholar
66. 1. Zou M
    2. Toivanen R
    3. Mitrofanova A
    4. Floch N
    5. Hayati S
    6. Sun Y
    7. Le Magnen C
    8. Chester D
    9. Mostaghel EA
    10. Califano A
    11. Rubin MA
    12. Shen MM
    13. Abate-Shen C

    (2017) Transdifferentiation as a mechanism of treatment resistance in a mouse model of castration-resistant prostate cancer

    Cancer Discovery 7:736–749.

    https://doi.org/10.1158/2159-8290.CD-16-1174

    • PubMed
    • Google Scholar

Author details

1. Chee Wai Chua

   1. Department of Medicine, Columbia University Medical Center, New York, United States
   2. Department of Genetics and Development, Columbia University Medical Center, New York, United States
   3. Department of Urology, Columbia University Medical Center, New York, United States
   4. Department of Systems Biology, Columbia University Medical Center, New York, United States
   5. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States

   Present address

   1. State Key Laboratory of Oncogenes and Related Genes, Renji-Med X Clinical Stem Cell Research Center, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China
   2. Department of Urology, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China

   Contribution

   Conceptualization, Data curation, Formal analysis, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing—original draft

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9634-7265

2. Nusrat J Epsi

   1. Department of Health Informatics, Rutgers School of Health Professions, Rutgers, The State University of New Jersey, Newark, United States
   2. Rutgers Biomedical and Health Sciences, Rutgers, The State University of New Jersey, Newark, United States

   Contribution

   Formal analysis, Investigation, Visualization, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-5363-075X

3. Eva Y Leung

   1. Department of Medicine, Columbia University Medical Center, New York, United States
   2. Department of Genetics and Development, Columbia University Medical Center, New York, United States
   3. Department of Urology, Columbia University Medical Center, New York, United States
   4. Department of Systems Biology, Columbia University Medical Center, New York, United States
   5. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States

   Contribution

   Formal analysis, Investigation, Visualization, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1291-6625

4. Shouhong Xuan

   1. Department of Medicine, Columbia University Medical Center, New York, United States
   2. Department of Genetics and Development, Columbia University Medical Center, New York, United States
   3. Department of Urology, Columbia University Medical Center, New York, United States
   4. Department of Systems Biology, Columbia University Medical Center, New York, United States
   5. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States

   Contribution

   Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0571-7855

5. Ming Lei

   1. Department of Medicine, Columbia University Medical Center, New York, United States
   2. Department of Genetics and Development, Columbia University Medical Center, New York, United States
   3. Department of Urology, Columbia University Medical Center, New York, United States
   4. Department of Systems Biology, Columbia University Medical Center, New York, United States
   5. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States

   Contribution

   Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

6. Bo I Li

   1. Department of Medicine, Columbia University Medical Center, New York, United States
   2. Department of Genetics and Development, Columbia University Medical Center, New York, United States
   3. Department of Urology, Columbia University Medical Center, New York, United States
   4. Department of Systems Biology, Columbia University Medical Center, New York, United States
   5. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States

   Contribution

   Data curation, Formal analysis, Funding acquisition, Investigation, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-4104-4179

7. Sarah K Bergren

   1. Department of Medicine, Columbia University Medical Center, New York, United States
   2. Department of Genetics and Development, Columbia University Medical Center, New York, United States
   3. Department of Urology, Columbia University Medical Center, New York, United States
   4. Department of Systems Biology, Columbia University Medical Center, New York, United States
   5. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States

   Contribution

   Formal analysis, Supervision, Investigation, Methodology

   Competing interests

   No competing interests declared

8. Hanina Hibshoosh

   1. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States
   2. Department of Pathology and Cell Biology, Columbia University Medical Center, New York, United States

   Contribution

   Resources, Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

9. Antonina Mitrofanova

   1. Department of Health Informatics, Rutgers School of Health Professions, Rutgers, The State University of New Jersey, Newark, United States
   2. Rutgers Biomedical and Health Sciences, Rutgers, The State University of New Jersey, Newark, United States

   Contribution

   Data curation, Software, Formal analysis, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Writing—original draft, Project administration

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0671-6512

10. Michael M Shen

    1. Department of Medicine, Columbia University Medical Center, New York, United States
    2. Department of Genetics and Development, Columbia University Medical Center, New York, United States
    3. Department of Urology, Columbia University Medical Center, New York, United States
    4. Department of Systems Biology, Columbia University Medical Center, New York, United States
    5. Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, New York, United States

    Contribution

    Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Investigation, Visualization, Writing—original draft, Project administration, Writing—review and editing

    For correspondence

    mshen@columbia.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-4042-1657