Sphere-Formation Assay: Three-Dimensional in vitro Culturing of Prostate Cancer Stem/Progenitor Sphere-Forming Cells

Cancer Stem Cells (CSCs) are a sub-population of cells, identified in most tumors, responsible for the initiation, recurrence, metastatic potential, and resistance of different malignancies. In prostate cancer (PCa), CSCs were identified and thought to be responsible for the generation of the lethal subtype, commonly known as Castration-Resistant Prostate Cancer (CRPC). In vitro models to investigate the properties of CSCs in PCa are highly required. Sphere-formation assay is an in vitro method commonly used to identify CSCs and study their properties. Here, we report the detailed methodology on how to generate and propagate spheres from PCa cell lines and from murine prostate tissue. This model is based on the ability of stem cells to grow in non-adherent serum-free gel matrix. We also describe how to use these spheres in histological and immuno-fluorescent staining assays to assess the differentiation potential of the CSCs. Our results show the sphere-formation Assay (SFA) as a reliable in vitro assay to assess the presence and self-renewal ability of CSCs in different PCa models. This platform presents a useful tool to evaluate the effect of conventional or novel agents on the initiation and self-renewing properties of different tumors. The effects can be directly evaluated through assessment of the sphere-forming efficiency (SFE) over five generations or other downstream assays such as immuno-histochemical analysis of the generated spheres.

INTRODUCTION Stem cells are undifferentiated, self-renewing cells that provide the source of all types of specialized cells in the body, from embryonic development (embryonic stem cells) and throughout adulthood (tissue-specific adult stem cells). These cells are characterized by their self-renewal ability, developmental potency, and their ability to differentiate into different downstream cell lineages.
While the zygote, blastomeres, and extraembryonic tissues are totipotent, i.e., able to differentiate into all tissues of the three germ layers, adult stem cells are multipotent cells committed to a specialized lineage. These can include neural stem cells, giving rise to neurons and glial cells; and hematopoietic cells which give rise to different types of blood cells (1). A similar subgroup of self-renewing undifferentiated cells was identified within tumors, showing interesting regenerating ability post-therapy. Based on these findings, the concept of cancer stem cells (CSCs) was developed as a subpopulation of tumorigenic cells, capable of initiating and driving tumorigenesis. The identification of the first CSC in acute myeloid leukemia (AML) in 1994 (2) has given way to potential isolation of similar tissue-specific CSCs and progenitor cells from other tumors (3).
The CSC model proposes a hierarchical organization whereby tumor growth is dependent on CSCs, a presumably small population, that have self-renewal ability and differentiation potential (4)(5)(6), thereby giving rise to more differentiated (tumors are mostly anaplastic rather than differentiated) tumor cells, similar to the role of stem cells in normal tissue (7). Another important feature of CSCs is their resistance to cytotoxic chemotherapy and ionizing radiation (8,9). Some theories have attributed their resistance to their presumably slow cell cycle and overexpression of efflux pumps (10), suggesting that such treatment may instead enrich the CSC sub-population within each tumor (8,9,11). Of high clinical relevance becomes the development of novel therapies, specifically targeting CSCs, potentially able to eliminate the regenerating capacity of the tumor.
Cancer stem cells have been identified in many of the solid tumors including brain, breast, prostate, colon, lung, and others. The properties and role of prostate cancer (PCa) stem cells is an active field of research where further investigation is needed to assess the cellular expression of Androgen Receptor (AR) by prostate CSCs, and confirm their ability to give rise to the metastatic form of PCa, namely CRPC (12)(13)(14)(15)(16)(17)(18)(19). Therefore, in vitro assays that favor the growth and propagation of CSCs is essential to enable their molecular/cellular characterization. Lately, it has been shown that CSCs have the ability to form multicellular three-dimensional (3D) spheres in vitro when grown in non-adherent serum-free conditions (20,21). Such 3D cultures allow the growth and propagation of CSCs, as well as evaluating the potential use of various conventional and novel drugs to target these tumor-initiating cells (21,22). However, most of the currently used protocols for 3D culturing of tumor spheroids in suspension exhibit forced floating and hanging drop approaches for screening of drugs (21,(23)(24)(25), which display several limitations and challenges pertaining Abbreviations: CSCs, cancer stem cells; PCa, prostate cancer; CRPC, castrationresistant prostate cancer; SFA, sphere-formation assay; SFE, sphere-forming efficiency; AML, acute myeloid leukemia; AR, androgen receptor; 3D, threedimensional; IACUC, Institutional Animal Care and Use Committee; UGS, urogenital system; DMEM, Dulbecco's Modified Eagle's Medium; FBS, fetal bovine serum; PrEGM, prostate epithelial cell growth media; PFA, paraformaldehyde; BSA, bovine serum albumin; NGS, normal goat serum; DEPC, diethyl pyrocarbonate; CK5, cytokeratin 5; CK8, cytokeratin 8; CK14, cytokeratin 14; Ecad, E-cadherin; EMT, epithelial-mesenchymal transition; BLE, berberis libanotica ehrenb. efficient assessment of the number and size of cultured spheres, as they are mobile and can merge with one another (24,26).
The time, cost, and technical challenge of performing selfrenewal in vivo studies highlight the need to develop alternative methods. Hence, in vitro sphere-forming assays have been established to investigate PCa (27,28), similar to those developed to study the nervous system (29) and mammary glands (30). Spheres with self-renewing properties formed in a 3D culture matrix which resembles the native microenvironment can be generated from human and mouse prostate epithelial cells. The sphere formation assay (SFA) provides a useful tool to assess the stem cells' population residing in tumors or cancerous cell lines and screen for drugs specifically targeting CSCs.
Here, we report the methodology for generating and propagating prostate spheres (prostatospheres) from murine prostate tissue and from human and murine PCa cell lines. This method has been previously used to generate prostate spheres from primary murine PCa cells (31)(32)(33)(34)(35)(36), as well as human and murine-derived PCa cell lines (32,35,37

Stepwise Procedures
This study was carried out in accordance with the recommendations of the NIH Guide and the American University of Beirut Guidelines for Use and Care of Animals. The protocol was approved by the Institutional Animal Care and Utilization Committee of the American University of Beirut. For sphere-formation assay, prostate cells can be prepared from murine prostate tissues and from human and murine PCa cell lines.

Isolation of primary prostate cells from murine prostate tissues (Timing ∼ 2 h)
1.1. Sacrifice the male mouse (8-12 weeks old) using isoflurane inhalation followed by cervical dislocation. In this case, we pass the mixture through a nylon mesh filter of a 40 µm pore size into a 50 mL conical tube and then transfer the mixture to a 15 mL conical tube and continue as mentioned above. 1.8. Make sure cells are properly dissociated and of appropriate morphology. Spin down the cells at 200 x g for 5 min, aspirate the media, wash the cells once with 3 mL of PrEGM. Resuspend the pellet in 1 mL of PrEGM and count the cells using a hemocytometer and trypan blue (to differentiate living cells from dead ones). Note: the previous steps discussing the isolation of primary prostate cells from murine prostate tissue shall be done as fast as possible, to preserve survival of these cells. is a gel-like substance that tends to become very loose and to break whenever plated in the center of the well (especially that spheres are left in culture for up to 2 weeks). Therefore, this is significantly avoided by plating the Matrigel TM matrix around the rim. Third, it confines the spheres' growth to a defined area in the well that makes the task of counting spheres much easier and more precise. In addition, it insures the distribution of cells in a single-cell format, avoiding clumping and overgrowth. Note: Potential pitfall: As a possible alternative, cells could be plated in the middle of the well (and not all over the well to minimize the use of Matrigel TM , which is another reason why plating around the rim is preferred). Although cells might be crowded in that way, but still possible to do. In this case, and if opted to plate in the middle, semi-automated counting could be achieved via acquisition of images along the z-axis and in a tile scan manner to cover the whole area and the different optical sections i.e., thickness of Matrigel TM . Note: Troubleshooting: Matrigel TM cell suspensions must be followed up daily to check up over spheres growth and formation. In general, spheres will start to appear by day three or four. Each cell line exhibits a distinct pattern in the process of spheres formation and growth, which must be monitored to optimize the followup schedule. It is very important to periodically acquire bright field images, using an inverted light microscope, of developing spheres to assess for any changes in morphology. Note: if the user is interested in assessing the effects of certain drugs on the sphere forming ability, a proper amount of the drug must be used in specific wells and compared to control wells with the appropriate media. We and others have shown the effects of many drugs, targeting specific signaling pathways, on sphere formation ability of presumably putative cancer stem/progenitor cells (31,32,34,36,(41)(42)(43)(44). There are many potential strategies of assessing the effect of a drug or a novel therapeutic agent in an SFA setting as outlined in Figure 1. A continuous treatment over five generations can indicate whether the effect of a drug is cumulative or not. Alternatively, we can establish and propagate spheres over five generations without treatments to enrich for cancer stem/ progenitor cells, followed by treatments at G5 to assess if a drug can target this enriched cell population or not. . This is a very important step as to verifying the extensive self-renewal ability of stem cells (45). Note: Troubleshooting: the SFE of each cell line has to be calculated through manual counting of the spheres, as stated in section 3.8. The number of sphere-forming units (SFUs) is dynamic and might change from one generation to another and from one cell type to another throughout the 5 generations starting from the same amount of cells at the start point (22,31,32,(34)(35)(36)(37). This is used to assess the self-renewal ability of the CSCs.

Isolation of cells from prostate cancer cell lines (Timing
Each sphere presumably originated from a single stem/progenitor cell and therefore, the differentiation potential within spheres should be tested. To do so, spheres could be collected at any generation (preferably at each generation), and could be subjected to immunohistochemistry and immunofluorescence analysis as explained before (31,32) and as shown in Figure 3. These methods can be used to further assess the effects of used treatments on differentiation pathways of targeted tumors/cells. Furthermore, total protein and RNA can be extracted from treated and non-treated FIGURE 1 | Schematic illustrating strategy of drug treatments in sphere-formation assay. After isolating single cell suspension from PCa tissues or PCa cell lines, drugs can be added to every generation in the sphere-formation assay (A), or drugs can be added to the first generation of spheres only (G1D0) and then spheres can be serially propagated to investigate whether the effect is permanent or reversal (B), or spheres can be formed and serially passaged so that you have extensively grew and enriched for stem cells, and then drugs are added at G5D0 to potentially target those cells (C). The sphere-formation efficiency has to be calculated for each generation to assess self-renewal ability of sphere-forming cells. At each generation, spheres could be processed for immune-histochemical analysis to check for differentiation markers, or proteins and total RNA could also be extracted to assess for differentiation.   (47)] was also detected, besides expression of the stem cell marker CD49f and SCA-1, which have been shown to identify putative prostate stem-like cells (48,49). The nuclei were stained with anti-fade reagent Fluorogel II with DAPI. Scale bars = 100 µm. Representative confocal microscopy images were acquired using the 63x oil objective and images were processed using the Zeiss ZEN 2012 image-analysis software. Microscopic analysis was performed using Zeiss LSM 710 laser scanning confocal microscope (Zeiss). DAPI, 40,6-diamidino-2-phenylindole.
spheres at any generation and subjected to WB and qRT-PCR analyses of the stemness and differentiation markers (41).

Immunofluorescence staining (Timing ∼ 2 days)
5.1. Spheres are grown in 35-mm glass bottom culture plates with 10-mm microwell in Matrigel TM -containing media following the same protocol as described above. 5.2. Using a pasteur pipette, aspirate the media from the center not to disturb the Matrigel TM . Freezing Medium (stored at −80 • C). 6.3. Fasten the embedded block to the block holder in the cryostat at −20 • C. 6.4. Section the frozen spheres block into desired thickness (typically 5 to 10 microns). 6.5. Mount the sphere sections on chilled, pre-cleaned, uncoated microscope glass slides (Sections can be stored in a sealed slide box at −80 • C for later use). 6.6. Apply staining methods as described above.

ANTICIPATED RESULTS
In this manuscript, we successfully generated prostate spheres from different human PCa cell lines (22RV1, RWPE1, PC3, and DU145). Using the above protocol, we also generated prostate spheres from primary single cell suspensions obtained from wt and Pten −/− TP53 −/− mouse prostate tissues as demonstrated in Abou-Kheir et al. (31,32), Agarwal et al. (36), and Daoud et al. (37), and from our previously established murine PCa cells, which represent androgen-dependent PCa and CRPC, namely PLum-AD and PLum-AI, respectively, as in Daoud et al. (37). In addition, TMPRSS2-driven ERG expression has been shown to increase the self-renewal and expand the numbers of clonogenic self-renewing CRPC subpopulation progenitors, as assayed by in vitro prostatospheres formation assay in Casey et al. (34). Representative images of the spheres generated from different cell lines, as well as primary murine prostate epithelial cells are presented in Figure 2.
Single cell suspensions of prostate cells either from mouse (Figure 2, upper panel) or human cell lines (Figure 2, bottom panel) were mixed with Matrigel TM (1:1) and then plated at a density of 5,000 cells/well in a 24-well plate. Media was changed every 2-3 days and bright field images were taken at days 8-12. Spheres were successfully propagated for at least 5 generations to assess the self-renewal ability of the stem/progenitor cells population. Single cells were monitored for single sphere formation over several days. The use of Matrigel TM semisolid matrix (or other semisolid matrices, as suggested in the procedure) allows seeded cells to remain embedded in place and keeps them less likely to migrate and move within the matrix. This is practically useful to overcome several limitations seen in prostate spheroids cultured in suspension, which tend to coalesce, merge, and therefore require very low plating density (as low as 1 cell per well), to form clonal colonies. Louis et al. (39) mirrored this issue in neurospheres formed in suspension as compared to others formed in semisolid collagen-based matrix, stating similar advantages to the applied 3D-culture (39). Hereby, we present a high-resolution live imaging movie of individual wt murine PCa cells embedded in a semisolid Matrigel TM -based 3D culturing system, showing their behavior in forming spheres, each of which is derived from a single cell (Video S1).
The number of spheres were counted and compared to the initial number of seeded cells. The percentage of sphere-forming units (SFU) within each cell line, i.e., the ratio of average number of prostatospheres to the initial number of seeded cells, is representative of the stem/progenitor cells subpopulation in culture. The average SFU seen using primary cells taken from murine prostate is 0.5%, while the average SFU using prostate cell lines is 5-10%, as seen in El-Merahbi et al. (35). The average SFU of each cell line is consistent over the 5 generations, which marks the ability of these cells to self-renew.
In our previous work on wt and Pten −/− TP53 −/− primary prostate cells (32) and murine PCa cell lines PLum (37), generated spheres were characterized through immunofluorescent staining, showing the expression of lineage markers of basal (CK5), luminal (CK8) and neuroendocrine cells (β3 tubulin) of prostate tissue. In this manuscript, the generated spheres from human PCa cell lines were stained for epithelial cell markers (CK5, CK8, CK14, and β3 tubulin), as well as stem cell markers (p63, SOX2, CD49f, and SCA-1). Figure 3 shows selective immunofluorescence images of stained prostatospheres derived from primary wt murine prostate cells. Interestingly, the spheres generated from those cells have self-renewal ability when subjected to a minimum of five propagations, and after staining of the prostatospheres at G0 and G5, no significance difference was detected in the expression levels (31,32,36,37).
Due to the heterogenic character of PCa cell lines, and the lack of universal markers for CSC in prostate tumors, generated spheres from all human and murine PCa cell lines were stained for an array of CSC markers (50) including: CD44, SOX2, CD117 (c-kit), and CD49f ( Figure S1). The levels of expression of different CSC markers (CD44, CD133, SSEA4, c-kit, NKx3.1, OCT-4, CD49f, and CD24) was further assessed by measuring their respective mRNA levels using qRT-PCR, on the spheres derived from human and murine PCa cell lines. As expected, each cell line presented a selective combination of stem cell markers: for example, PC3-derived spheres showed relatively high expression of CD44, CD133, and SSEA4 (compared to nonsignificant expression of other markers), while DU45-derived spheres showed an increased expression of c-kit (CD117), NKx3.1, and OCT4 ( Figure S2, Table S1).
This assay can further be used as a functional reporter of the progenitor activity of different cell lines, as well as the differentiation and self-renewal ability of the stem/progenitor cell population (27). This property provides a platform to test the effects of traditional and novel therapeutic agents on PCa cell lines. We have previously used this assay to compare the effects of selective inhibitors of the AKT/mTOR pathway (Rapamycin and Triciribine), as well as the downstream signaling of the androgen receptor (Nilutamide and bicalutamide) on wild-type and Pten −/− TP53 −/− mouse prostate spheres (31). The effect of these drugs was monitored through measuring the sphere-forming efficiency (SFE) and the average volume of the formed spheres; which were both decreased compared to non-treated control prostatospheres. Similarly, we studied the effects of Ehrenb extracts on the sphere-formation ability of PCa cell lines DU45 and PC3. The decreased ability to form spheres on G1, and the limited self-renewal ability during propagation of these spheres over 5 generations as outlined in Figure 1, correlated with the ability of this extract to limit cellular growth, migration, and proliferation on 2D-functional assays (35).

DISCUSSION
In this manuscript, we outline a protocol to generate and propagate prostate spheres (prostatospheres) from prostate tissues and human and murine PCa cell lines. This assay depends on the ability of stem cells to survive in a 3D culture matrix (for example, Matrigel TM ) in low or serum-free medium while differentiated cells grow in adherent monolayer cultures and FBS-containing medium. Primary plating of single-cells suspension in a 3D matrix doesn't confirm the existence of an enriched stem cell sub-population, which is characterized by its self-renewal ability and differentiation potential. Multiple and serial propagations of the spheres are needed to confirm the existence of CSCs within a tumor or a cell line, dependent on finding a relatively constant SFU/SFE within each generation. The issue of self-renewal represents a central feature of CSCs and is tightly associated with their pathologic ability to regenerate tumors after treatment. It represents the ability of these cells to reproduce indefinitely, while maintaining its multipotent ability to differentiate. 45 discussed this issue in terms of propagation and self-renewal of neuronal adult stem cells, cultured as neurospheres in suspension. (45) They properly argue that non-stem progenitor cells retain the ability of selfrenewal and can indeed propagate into secondary or tertiary spheres (G2 and G3, respectively). Theoretically, prostate stem cells shall maintain self-renewal ability indefinitely. Due to technical limitations faced in vitro, propagating prostate cellderived spheres over five generations (or more) can be practical to isolate and propagate prostate CSCs; the sphere-forming ability of progenitor cells decreases subsequently, while that of prostate CSCs is maintained.
The clonal origin of the formed spheres is yet another important aspect of this assay. Previous work showed the tendency of neurospheres to coalesce and merge when cultured in suspension. In order to generate clonal colonies using cells grown in suspension to culture neurospheres, deposition of a single cell per well presented the golden criteria to reach that goal (39). Previous studies have successfully cultured prostatospheres in serum-free suspension culture, starting from PCa cell lines (51,52). However, this issue hasn't been properly addressed. The protocol proposed herein suggests growing of cells in semi-solid matrix, to avoid their migration and aggregation, and overcome the technical burden of culturing non-adherent spheres starting from single cell per well. Discrete spheres were grown and monitored starting from a single cell suspension (Video S1). Matrigel TM provides an enriched matrix that harbors different glycoproteins and growth factors seen in the basement membrane, including collagen IV, laminin, and Fibroblast Growth Factor; it provides a semi-solid matrix able of simulating the rich extracellular medium of cells in vivo (53). The use of Matrigel TM has been further implicated in long term organoids culture, another 3D-culturing modality that closely mirrors the in vivo prostate of murine and human derived cells (54). Moreover, Matrigel TM has been used to facilitate the establishment of tumorigenic Xenografts derived from human cell lines in immunocompromised mice (53). A previous study found that the 3D solid spheres of mammary glands, cultured in Matrigel TM , could be used for repopulating gland-free mammary fat pads in mice reaching an engraftment frequency of 71% (55). This means that the spheres formed can be used for tissue transplants and that this assay could be further developed in the future to be applicable on human tissue grafts. Interestingly, it was also found that the sphere assay could be used to compare the cellular mechanisms of normal and malignant cells through deciphering active pathways and differentiation patterns. It was found that the spheres derived from normal lung tissue, cultured in Matrigel TM , were able to form a lumen while the spheres derived from cell lines and tumor biopsies of lung cancer were not (56), showing a potential use of the sphere-formation assay to compare the underlying cellular processes governing CSCs and those in normal tissue stem cells, and their pathophysiological implications on progression of the disease. Another study, by Hur et al. found that hematospheres could be isolated from blood and cultured in Matrigel TM to allow the formation of a network of vessels (57), proposing potential therapeutic implications to treat coronary artery occlusions or atherosclerotic arteries, by vessel replacement therapies. Because the use of SFA has been recently implemented on assessing stem cell-like properties in tumors, there has not been much focus on drug treatments targeting these self-renewing cells. However, there have been various studies that focused on testing particular drugs on generated spheres. SFA was used, along with different assays, to show that Nigericin, an antibiotic that suppresses Golgi function in eukaryotic cells, suppresses colorectal cancer metastasis by inhibiting epithelialmesenchymal transition (EMT) (58). The drug Metformin was shown to inhibit the growth of thyroid carcinoma cells, suppress the self-renewing property of the thyroid CSCs and act as an enhancing supplement to chemotherapeutic agents (43). Furthermore, Eckol has been proven to suppress the maintenance of stem cell properties and malignancies in glioma stem-like cells, hence suggesting a reduced chance of cancer recurrence (42). Also, spheres from esophageal tumor origin have been found to involve stem-like cells with elevated aldehyde dehydrogenase activity (59). Berberis libanotica Ehrenb (BLE) extracts as well were shown to have high therapeutic potential in targeting PCa and eradicating the self-renewal ability of highly resistant CSCs (35). All the mentioned examples have used SFA to assess the effect of drugs on the CSCs or stem cell-like populations in tumors; these studies primarily relied on two main criteria in their assessment: average volume of the generated spheres and the average number of sphere-forming units (SFU). Yet, most studies focus on testing particular drugs on generated spheres for single generation, without assessing the effect on spheres propagated for several generations, which possess enriched stem cell/progenitor properties.
To emphasize the importance of the use of SFA in stem cell and cancer research, we have put together a detailed protocol of how to conduct a semisolid Matrigel TM -based sphere-formation assay in the absence or presence of drugs, consecutively or alternatively, in the context of PCa stem cells. The advantage of using this sphere-formation in vitro assay is to isolate, propagate, purify, and amplify specific populations of prostatic normal and CSCs. It enables studying stem cells at different stages of their formation (at different generations) and detecting markers of their signaling pathways. It also solely depends on the functional intrinsic property of stem cells in forming a complex structure in a 3D environment, namely self-renewal ability and differentiation potential.

AUTHOR CONTRIBUTIONS
HB, KC, RC, OH, AM, and FB contributed to the project design and execution of experiments, analysis of results, and writing of manuscript. AE-H, DM, Y-NL, GD, and WA-K contributed to overlooking and following up with experiments, result analysis and manuscript proofreading. Y-NL, GD, and WA-K contributed to project design, result analysis, manuscript writing and proofreading. All authors critically revised and edited the manuscript and approved the final draft.

FUNDING
This study was supported by funding from the Medical Practice Plan (MPP) at AUB-FM (Grant #320080). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

ACKNOWLEDGMENTS
We would like to thank all members in the Abou-Kheir's Laboratory (The WAK Lab) for their help on this work. In addition, we would like to thank all members of the core facilities in the DTS Building at the American University of Beirut (AUB) for their help and support. Also, we would like to thank Ross Lake Center for Cancer Research at the National Cancer Institute for helping with the acquisition of the movie (Video S1).

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc. 2018.00347/full#supplementary-material Figure S1 | Immunophenotyping of spheres derived from human and murine prostate cancer cell lines. Immunofluorescent images of PC3, DU145, 22RV1, RWPE-1, PLum-AI, PLum-AD derived prostatospheres, stained for selected cancer stem cell markers including CD44, SOX2, CD117, and CD49f (white arrows). Those cancer stem cell markers have been shown to identify putative prostate stem-like cells (47,49,60,61). CD49f stain (white arrows) is displayed in the middle of some prostatospheres in addition to the cells that express this marker at the surface. Noteworthy, in some cases, CD49f is ubiquitously expressed at the surface of the spheres (as in RWPE-1, DU145, and PC3), which is due to the presence of basal epithelial cells that are mostly found at the outer layers of the sphere. CD44, on the other hand, is a widely used cell surface marker that is expressed in many cells including stem cells (62) and was found to be homogeneously expressed across the different prostatosphere cells (as shown in the images to the left). The nuclei were stained with anti-fade reagent Fluorogel II with DAPI. Representative confocal microscopy images were acquired using the 63x oil objective and images were processed using the Zeiss ZEN 2012 image-analysis software. Microscopic analysis was performed using Zeiss LSM 710 laser scanning confocal microscope (Zeiss). Scale bar = 20 µm.  Table S1 for primers sequence). Video S1 | Recorded growth of prostatospheres in Matrigel TM matrix. This high-resolution movie shows the growth of single cell suspensions of wt murine prostate into individual spheres. This technique overcomes the limitations of culturing prostate spheroids in suspension, by limiting the migration and aggregation of generated spheres. This time-lapse movie of prostate sphere-formation assay was captured hourly over 100 h using an Olympus Viva View FL Incubator Microscope with a 40x objective (Olympus).