Stromal Gli signaling regulates the activity and differentiation of prostate stem and progenitor cells

Interactions between cells in the stroma and epithelium facilitate prostate stem cell activity and tissue regeneration capacity. Numerous molecular signal transduction pathways, including the induction of sonic hedgehog (Shh) to activate the Gli transcription factors, are known to mediate the cross-talk of these two cellular compartments. However, the details of how these signaling pathways regulate prostate stem and progenitor cell activity remain elusive. Here we demonstrate that, although cell-autonomous epithelial Shh-Gli signaling is essential to determine the expression levels of basal cell markers and the renewal potential of epithelial stem and progenitor cells, stromal Gli signaling regulates prostate stem and progenitor cell activity by increasing the number and size of prostate spheroids in vitro. Blockade of stromal Gli signaling also inhibited prostate tissue regeneration in vivo. The inhibition of stromal Gli signaling suppressed the differentiation of basal and progenitor cells to luminal cells and limited prostate tubule secretory capability. Additionally, stromal cells were able to compensate for the deficiency of epithelial Shh signaling in prostate tissue regeneration. Mechanistically, suppression of Gli signaling increased the signaling factor transforming growth factor β (TGFβ) in stromal cells. Elevation of exogenous TGFβ1 levels inhibited prostate spheroid formation, suggesting that a stromal Gli–TGFβ signaling axis regulates the activity of epithelial progenitor cells. Our study illustrates that Gli signaling regulates epithelial stem cell activity and renewal potential in both epithelial and stromal compartments.

Interactions between cells in the stroma and epithelium facilitate prostate stem cell activity and tissue regeneration capacity. Numerous molecular signal transduction pathways, including the induction of sonic hedgehog (Shh) to activate the Gli transcription factors, are known to mediate the cross-talk of these two cellular compartments. However, the details of how these signaling pathways regulate prostate stem and progenitor cell activity remain elusive. Here we demonstrate that, although cell-autonomous epithelial Shh-Gli signaling is essential to determine the expression levels of basal cell markers and the renewal potential of epithelial stem and progenitor cells, stromal Gli signaling regulates prostate stem and progenitor cell activity by increasing the number and size of prostate spheroids in vitro. Blockade of stromal Gli signaling also inhibited prostate tissue regeneration in vivo. The inhibition of stromal Gli signaling suppressed the differentiation of basal and progenitor cells to luminal cells and limited prostate tubule secretory capability. Additionally, stromal cells were able to compensate for the deficiency of epithelial Shh signaling in prostate tissue regeneration. Mechanistically, suppression of Gli signaling increased the signaling factor transforming growth factor ␤ (TGF␤) in stromal cells. Elevation of exogenous TGF␤1 levels inhibited prostate spheroid formation, suggesting that a stromal Gli-TGF␤ signaling axis regulates the activity of epithelial progenitor cells. Our study illustrates that Gli signaling regulates epithelial stem cell activity and renewal potential in both epithelial and stromal compartments.
Stromal-epithelial cell interactions are essential for prostate development and adult prostate tissue regeneration (1). The prostate gland is composed of numerous connected tubules and the surrounding stromal microenvironment. Each tubule con-sists of three types of prostate epithelial cells (PrECs), 3 including secretory luminal cells, basal cells, and neuroendocrine cells (2,3). Luminal cells typically expressing cytokeratin (CK) 8 or 18 locate at the apical region of the epithelium, are sensitive to androgen stimulation, and produce secretory proteins. Basal cells expressing CK5 or p63 reside underneath the luminal cells and attach to the basement lamina (4). Both luminal and basal cells in the epithelial compartment contain stem/progenitor cells, which are capable of self-sustentation in tissue regeneration (5).
PrECs are surrounded by the stromal microenvironment, providing an important niche to nurse epithelial progenitor cells (6). The stromal compartment comprises a variety of cell types, including smooth muscle cells, subepithelial cells, wrapping cells, interstitial fibroblasts, and others (3,7). Stromal cells secrete important signaling factors to stimulate prostate development and adult prostate tissue regeneration (1,8). These paracrine signaling factors include stromal androgen, fibroblast growth factor (FGF), TGF␤, bone morphogenetic protein (BMP), Sonic hedgehog (Shh), and others; they induce prostatic secretion in luminal cells and maintain the self-renewal capacity of prostate stem/progenitor cells (9 -12). Shh-Gli signaling exists in both prostate basal cells and stromal cells (7). However, how this signaling pathway controls the renewal capacity of stem/progenitor cells through the stromal-epithelial interaction remains elusive.
Shh-Gli signaling is an important signal transduction pathway in regulating the normal development of multiple organs, including growth of prostatic tissues and differentiation of prostate epithelia (13,14). It has been reported that Shh-Gli signaling facilitates prostate branching morphogenesis through regulation of hepatocyte growth factor (15). In mammalian cells, the Gli family consists of three members (Gli1, Gli2, and Gli3). Although both Gli2 and Gli3 have C-terminal transcrip-tional activation and N-terminal transcriptional repression domains, Gli1 contains only a C-terminal transcriptional activation domain. Shh signaling is primarily mediated through Gli2 and Gli3 (16). Gli3T is a mutant missing the C-terminal transcriptional activation domain and has been used as a constitutive transcriptional repressor of Gli signaling (17). In the Shh-dependent canonical pathway, the binding of Shh to a 12-transmembrane receptor, Patched (Ptc/Ptch/Ptch1), induces trans-localization of Gli2/3 to the nucleus and regulates expression of downstream target genes such as Gli1, Bcl2, Ptch1, and others (16,18,19).
We have shown previously that epithelial Gli signaling regulates the expression of p63 and plays an essential role in the maintenance of the homeostasis and renewal potential of prostate stem/progenitor cells (20). Dysregulation of Gli signaling by oncogenic events, such as the synergy of Kras and androgen receptor (AR), promote the pathological expansion of basal/ progenitor cells (20,21). In this study, we demonstrate that the Shh-Gli signaling axis mediates the interaction of stromalepithelial cells in prostate tissue regeneration under physiological conditions. Particularly, Gli signaling in stromal cells regulates the activity of prostate epithelial stem/progenitor cells and promotes prostate spheroid formation in vitro and tissue regeneration in vivo. Mechanistically, stromal Gli signaling regulates TGF␤1/2 expression, and exogenous TGF␤1 inhibits prostate spheroid formation. Our study emphasizes the importance of stromal Gli signaling in the regulation of prostate stem cell activity and tissue regeneration capacity.

Cell autonomous Shh-Gli signaling regulates the renewal potential of prostate progenitor cells
We have demonstrated previously that epithelial Gli signaling is essential in regulating the renewal potential of prostate stem/progenitor cells (20). To further examine the role of the cell-autonomous Shh-Gli signaling axis in controlling prostate stem cell activity, lentiviral vectors overexpressing pre-Shh(WT) or the pre-Shh(C25S) mutant were constructed (Fig. S1, A and B). Shh(WT) undergoes both cholesterol modification at the C terminus and palmitoylation at the N terminus (Fig. 1A). Mutation of the cysteine to serine blocked palmitoylation of Shh without affecting its maturation (Fig. 1, A and B, and Fig. S1C), indicating that both pre-Shh(WT) and pre-Shh(C25S) were processed to form mature Shh protein (19 kDa).
The self-renewal potential of prostate stem/progenitor cells can be examined by prostate sphere formation (Fig. 1C) (22). We examined whether the cell-autonomous Shh-Gli signaling axis regulates prostate stem cell activity by sphere formation assay. Overexpression of Gli3T and knockdown of Gli1 and Gli2 were confirmed by down-regulation of Ptc1 and Bcl2 or Gli1/2 expression, respectively (Fig. S2). As reported previously (20), overexpression of Gli3T, a dominant-negative repressor of Gli signaling, in PrECs significantly inhibited primary prostate sphere formation (Fig. 1D). Although overexpression of Shh(C25S) inhibited the number of primary prostate spheres, overexpression of Shh(WT), shRNA-Gli1, or shRNA-Gli2 resulted in no change in sphere number (Fig. 1D). However, the number of secondary spheres was significantly inhibited by overexpression of Gli3T, Shh(C25S), or shRNA-Gli2 but not shRNA-Gli1 (Fig. 1E). The data indicate that suppression of Shh-Gli2/3 signaling by loss of Shh palmitoylation, knockdown of Gli2, or overexpression of the repressor Gli3T inhibits the renewal activity of prostate stem cells.
P63, CK5, and CK14 are well-established markers of prostate basal/progenitor cells, and CK8 and CK18 are known luminal markers. We have shown previously that overexpression of Gli3T and shRNA-Gli2 inhibited p63 expression (20). Therefore, we analyzed whether Shh(WT) or Shh(C25S) regulates the levels of these proteins in PEB cells, a cell line isolated from mouse prostate basal cells (23). Although overexpression of Shh(WT) in PEB cells did not further increase the expression levels of basal or luminal markers, overexpression of Gli3T down-regulated expression levels of p63, CK5, CK14, and AR but not CK8 or CK18. Similarly, Shh(C25S) decreased the levels of p63, CK5, and CK14 but not CK8, CK18, or AR (Fig.  1F). Collectively, the data suggest that the cell-autonomous Shh-Gli signaling axis regulates the renewal potential of prostate spheroids by potential regulation of basal or stem/ progenitor cells.

The presence of mesenchymal stromal cells compensates for the deficiency of epithelial Shh signaling in prostate tissue regeneration
The interactions between stromal and epithelial cells is essential in prostate tissue regeneration (1). Epithelial stem/ progenitor cell activity was impaired by overexpression of Shh(C25S) or Gli3T. We examined whether WT mesenchymal cells could compensate for the deficiency of cell-autonomous Shh signaling in epithelial cells. In contrast to the inhibition of primary prostate sphere formation by Shh(C25S) or Gli3T (Fig.  1D), epithelial cells overexpressing Shh(WT) or Shh(C25S) showed no difference in the size of regenerated prostate tissue in the presence of WT urogenital sinus mesenchyme (UGSM) cells ( Fig. 2A). RFP-expressing tubules indicated that the tubules were infected with a control vector, Shh(WT), or Shh(C25S). Histologically, regenerated prostate tubules showed no difference in expression levels of CK5/CK8 or p63 (Fig. 2B). Similarly, regenerated prostate tissues derived from overexpression of Gli3T in epithelial cells showed a similar tubule structure as the control vector (Fig. S3). The data suggest that impaired epithelial Shh signaling could be rescued by the presence of normal mesenchymal cells.

Mesenchymal stromal cells enhance prostate spheroid formation and elevate epithelial Gli and p63 expression
To evaluate the contribution of stromal cells to the regulation of prostate epithelial stem/progenitor cell activity, an in vitro stromal cell sphere co-culture assay was developed in which the formation of prostate spheres is under the stimulation of mesenchymal stromal cells (Fig. 3A). The number and size of prostate spheroids were significantly increased when PrECs were co-cultured with UGSM cells (Fig. 3, B and C). Additionally, stromal cells significantly increased the number The expression of pre-Shh protein (ϳ50 kDa) or mature protein (19 kDa) was detected by immunoblotting (right panel), lysates were subjected to click chemistry, and palmitoylated-proteins were detected by fluorescence (left panel). C, schematic of primary and secondary spheroid formation. Prostate tissues were isolated from BL6 mice and dissociated into single cells. PrECs were transduced with pre-Shh(WT), pre-Shh(C25S), Gli3T, or shRNA-Gli1/2. Only PrECs with stem cell activity (indicated in orange) could form primary spheroids after embedding within Matrigel. Primary spheres were digested into spheroid cells and embedded in Matrigel to form secondary spheres. D, primary PrECs transduced with control vector, pre-Shh(WT), pre-Shh(C25S), Gli3T, shRNA-Gli1, or shRNA-Gli2 were mixed with Matrigel and plated in separate wells in a 12-well plate. Sphere number was counted after 10 days of incubation. E, the primary spheres were dissociated and replated in the same way as described in D to generate secondary spheres. F, PEB cells were transduced with control vector, pre-Shh(WT), pre-Shh(C25S), or Gli3T (positive control). The protein lysates were analyzed for p63, CK5, CK14, AR, CK8, CK18, Gli3T, Shh (mature form), and tubulin expression by immunoblotting. *, p Ͻ 0.05; ***, p Ͻ 0.001; N.S., not significant.

Role of Gli signaling in prostate tissue development
of prostate spheroid aggregates. The aggregates contained two to six spheroids, with some aggregates having more than six spheroids (Fig. 3B).
Further analysis showed that the expression levels of progenitor and basal markers (p63, CK5, or CK14), AR, and luminal markers (CK8 or CK18) were increased in the PrECsϩUGSM group compared with the PrECs alone group (Fig. 3D). The expression levels of Gli1, Gli2, and Gli3 in prostate spheres grown with or without the stimulation of UGSM cells were also compared. Although Gli1 was barely detected in spheres (data not shown), the levels of Gli2 and Gli3 were significantly elevated in the PrECsϩUGSM group compared with the PrECs alone group (Fig. 3D). Additionally, the numbers of single spheres and spheroid aggregates derived from the PrECsϩUGSM group were significantly higher than those from PrECs alone in the secondary spheroid passage, suggesting that UGSM cells enhance sphere renewal potential (Fig. 3E). Collectively, the data suggest that mesenchy- . Arrows indicate p63 ϩ cells. Of note, the RFP signal is usually bleached after antigen retrieval so that the staining for CK5 (red) will not interfere with the RFP in tissues (see also Fig. S7).

Role of Gli signaling in prostate tissue development
mal stromal cells enhance the activity of prostate stem/ progenitor cells, which is associated with epithelial Gli signaling.

Stromal Gli signaling promotes prostate stem/progenitor activity
We studied the contribution of stromal Gli signaling in supporting epithelial stem/progenitor cell activity. To exclude the endogenous stromal cells from the PrEC preparation, prostate basal cells were isolated based on Lin Ϫ CD49f ϩ Sca1 ϩ markers (4) (Fig. 4, A-C). Additionally, UGSM cells were transduced with Gli3T, shRNA-Gli1, or shRNA-Gli2 by lentiviral infection (Fig. 4A). The number of spheres increased more than 2-fold in the basal cellsϩUGSM or UGSM vector groups compared with basal cells alone or basal cellsϩ3T3 cells (Fig. 4D). The size of spheres increased in the presence of UGSM, UGSMϩGli3T/ shRNA-Gli2/shRNA-Gli1, or 3T3 cells compared with basal cells alone (Fig. 4D). The UGSM-induced sphere number was significantly inhibited by overexpression of Gli3T or shRNA-Gli2 but not by shRNA-Gli1 (Fig. 4D). Additionally, the number of spheres and spheroid aggregates significantly decreased in the PrECsϩUGSM-Gli3T group compared with the PrECsϩ UGSM control group (Fig. 4E), suggesting that overexpression of Gli3T inhibited spheroid renewal potential. The data suggest

. Mesenchymal stromal cells enhance prostate spheroid formation and increase the expression levels of Gli transcription activators and basal and luminal markers of prostate spheres.
A, schematic of the UGSM-sphere co-culture assay. Primary PrECs were isolated from prostate tissue. UGSM cells were isolated from E16.5 mouse embryos. PrECs were mixed with 5000 UGSM cells and Matrigel and plated in the rim of a 12-well plate, and another 5000 UGSM cells were inoculated in the center of a well. Of note, only PrECs with stem cell activity (indicated in orange) are able to form spheres. UGSM cells remained as cells and were not able to form spheres in the co-culture assay. B and C, the numbers of single spheres and spheroid aggregates (two, three, four, five, six, and more than six spheroid aggregates, B) and size of single spheres (C) were recorded on day 10. Representative images are shown. The presence of UGSM cells increased the size and number of prostate spheres and spheroid aggregates. Scale bar is 100 m. D, prostate spheres grown with or without UGSM cells were collected. After UGSM cells were removed from the mixture of spheres and UGSM cells by filtering through 40-m mesh, spheres from both groups were collected and compared. Proteins were extracted from the isolated prostate spheroids (UGSM cells were excluded) for analysis of Gli2/3 and GAPDH expression. Cell lysate derived from embryonic stem cells was used as a positive control for the expression of Gli2 and Gli3. Expression levels of AR, p63, CK8, CK18, CK5, CK14, and GAPDH in two sets of experiments are shown. E, prostate spheres from the PrECs and PrECsϩUGSM groups were dissociated into single cells. 6000 dissociated cells were mixed with Matrigel and evaluated for secondary sphere formation. The number of single spheres (labeled 1 on the x axis) and spheroid aggregates (with two, three, or four spheres) were recorded. *, p Ͻ 0.05; ***, p Ͻ 0.001; ND, not detected. were recorded on day 10, and sphere images were taken (bottom panel). Scale bar is 100 m. E, the prostate spheroids derived from the UGSM-control or UGSM-Gli3T groups were dissociated into single cells. Cells were mixed with Matrigel and subjected to the secondary sphere assay. The x axis shows single spheres (labeled 1) or spheroid aggregates (spheroid aggregates with two, three, or four spheres). The y axis represents the ratio of sphere formation. 1000:1 means 1000 epithelial cells forming one prostate sphere. F, schematic of the co-culture of Lin Ϫ CD49f ϩ Sca1 ϩ basal cells with adult stromal cells. Adult stromal cells were isolated based on the Lin Ϫ CD49f Ϫ Sca1 ϩ population as shown in C. The stromal cells were grown in the same medium as used for growing UGSM cells. The adult stromal cells were transduced with control vector (FUCRW) or Gli3T by lentiviral infection. Lin Ϫ CD49f ϩ Sca1 ϩ basal cells were mixed with adult stromal cells in the co-culture assay. G, sphere number and size were recorded on day 10, and sphere images were taken. **, p Ͻ 0.01; ***, p Ͻ 0.001; N.S., not significant; ND, not detected.

Role of Gli signaling in prostate tissue development
that the expression of stromal Gli2/3 promotes epithelial stem cell activity.
Next, we further examined whether sphere formation could also be regulated by adult prostate stromal cells. Primary adult stromal cells were isolated based on Lin Ϫ CD49f Ϫ Sca1 ϩ markers (Fig. 4, B and C). The isolated adult stromal cells grew similarly as UGSM cells. Stromal cells were transduced with control or Gli3T by lentiviral infection (Fig. 4F). Similar to the UGSM sphere assay (Fig. 4D), adult stromal cells promoted the number and size of prostate spheres. The number, but not the size, of prostate spheres was inhibited by the suppression of Gli signaling through overexpression of Gli3T (Fig. 4G). Collectively, the data indicate that adult stromal Gli signaling regulates epithelial stem cells as well.

Blockade of stromal Gli signaling inhibits prostate tissue regeneration in vivo
We examined the contribution of stromal Gli signaling in prostate tissue regeneration in vivo. To exclude the potential of endogenous stromal contribution in tissue regeneration, basal cells were isolated based on Lin Ϫ CD49f ϩ Sca1 ϩ markers (Fig. 5, A-C). The isolated cells were mixed with or without UGSMexpressing control vector or Gli3T (Fig. 5A) and subjected to a prostate tissue regeneration assay. The number of regenerated tubules was significantly elevated in regenerated tissue derived from the basal cellsϩUGSM group compared with the basal cells alone group (Fig. 5, D and E, and Fig. S4). However, it was significantly inhibited in the basal cellsϩUGSM-Gli3T group compared with the basal cellsϩUGSM group (Fig. 5, D and E). Additionally, the secretion (stained as a pinkish color by H&E staining) was largely not visible in the lumen of tubules in the basal cellsϩUGSM-Gli3T group compared with those in the basal cellsϩUGSM group (Fig. 5D, b and c).
We characterized cell types in the regenerated tubules from the different groups. The number of p63 ϩ and/or CK5 ϩ cells in the tubules derived from the basal cells alone or basal cellsϩ UGSM-Gli3T group were significantly elevated compared with the basal cellsϩUGSM group (Fig. 5, D, F, and H, and Fig. S4). In contrast, the number of CK8 ϩ cells in the tubules derived from the basal cells alone or basal cellsϩUGSM-Gli3T groups were significantly lower (Fig. 5, D and G, and Fig. S4). Collectively, the data indicate that blockade of Gli signaling in UGSM cells inhibited the tissue regeneration potential and the differentiation of prostate basal/progenitor cells into luminal cells.
Similarly, by using PrECs (unsorted cells, endogenous stromal cells not excluded) for prostate tissue regeneration, the number of tubules and branching morphogenesis of the regenerated tubules were also inhibited in the PrECsϩUGSM-Gli3T group compared with the PrECsϩUGSM control group (Fig. S5).

Mesenchymal stromal Gli signaling regulates TGF␤ expression levels to affect stem cell activity
We investigated how blockade of stromal Shh-Gli signaling regulates other signaling pathways that are related to stem cell activity in stromal cells. Overexpression of Gli3T or knockdown of Gli3 was confirmed by decreased mRNA levels of Gli downstream genes such as Gli1 and Bcl2 (Fig. 6, A and C).
Although overexpression of Gli3T resulted in a significant increase in TGF␤2 levels in UGSM cells (Fig. 6B), shRNA-Gli3 increased the expression levels of TGF␤1 and TGF␤R1/2/3 (Fig. 6D). The difference in regulation of TGF␤1 or TGF␤2 expression levels might be due to a broader effect of Gli3T that could block both Gli1/2 or Gli3 signaling, considering they have different downstream targets (24,25). Interference of Gli signaling in UGSM cells also affected AR, Notch, and BMP expression (Fig. S6).
We examined how exogenous TGF␤1 affects the activity of prostate progenitor cells using the sphere formation assay. An increase in TGF␤1 in the medium inhibited the number (Fig.  6E) and size (Fig. 6F) of prostate spheres and spheroid aggregates (Fig. 6G) stimulated by UGSM cells. A high concentration of TGF␤1 (10 ng/ml) completely inhibited UGSM-stimulated prostate spheroid formation (Fig. 6E). The results were consistent with the inhibitory effect of sphere formation in vitro (Fig.  4, D and G) and tubule formation in vivo (Fig. 5D). The data indicate that stromal Shh-Gli signaling modulates the TGF␤ signaling pathway in stromal cells, thereby regulating epithelial stem cell activity.

Discussion
Our study demonstrates that cell-autonomous Shh-Gli signaling in the epithelial compartment dictates the renewal potential of prostate stem/progenitor cells. Prostate basal cells are a major pool of stem/progenitor cells in prostate tissue (4). The cells maintain high expression levels of Shh and Gli3 (7). Blockade of Shh signaling by overexpression of the Shh(C25S) mutant inhibits expression of p63 or CK5, a marker of progenitor/basal cells in prostate tissue (26,27), and prostate spheroid formation. This result is consistent with our previous experimental evidence showing that overexpression of Gli3T, a dominant-negative suppressor of Gli signaling, inhibits the renewal capacity of prostate progenitor cells (20). Our studies illustrate that the Shh-Gli2/3-p63 signal transduction pathway is essential for maintaining homeostasis of prostate stem cells in the epithelial compartment (Fig. 7).
Our study also illustrates that stromal Gli signaling regulates epithelial stem cell activity (Fig. 7). Stromal-epithelial cell interactions are important for prostate development and adult prostate regeneration (1). We show that mesenchymal stromal cells elevate expression of Gli2/3 and p63/CK5 in prostate spheroids from the co-culture assay. The epithelial CK8/CK18 levels are also elevated through stimulation of stromal cells, suggesting that stromal cells promote the differentiation process as well. Blockade of stromal Gli signaling inhibits the number of spheres in vitro and regenerated prostate tubules in vivo. Our data support Shh-Gli signaling as an important pathway that facilitates cross-talk of stromal cells with epithelial basal/ progenitor cells in prostate development and tumor initiation (7). In particular, stromal Gli2/3-TGF␤1/2 signaling promotes the activity of prostate stem/progenitor cells. In contrast to previous reports (28), the experimental design of this study allowed us to clearly dissect the contribution of Shh signaling in either the epithelial or stromal compartment toward the regulation of prostate tissue regeneration. Similar to other epithelial stem Role of Gli signaling in prostate tissue development cells, the microenvironment is essential for promoting the renewal activity of prostate progenitor cells (6,29).
Prostate epithelium and reactive stromal cells co-evolve to regulate prostate development (3). TGF␤ and hedgehog signal-ing are two well-characterized pathways in embryonic development (30). Numerous studies have reported that TGF␤ regulates hedgehog signaling and Gli expression. Reciprocally, Gli signaling positively regulates TGF␤ levels (24,31). Our data   PCR (B and D). E and F, prostate tissues were dissociated into PrECs. PrECs alone or PrECs mixed with UGSM cells in the co-culture sphere assay were grown in prostate sphere growth medium with control or 1, 5, or 10 ng/ml of TGF␤1. The number (E) and size (F) of spheres were counted on day 10. G, the number of single spheres (labeled 1 on the x axis) and spheroid aggregates (aggregates with two, three, four, five, or six spheres) in PrECs cultured with UGSM with/without 1 ng/ml of TGF␤1 were recorded. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; ND, not detected.

Role of Gli signaling in prostate tissue development
reveal transcriptional regulation of TGF␤ by Gli signaling in stromal cells. Blockade of Gli signaling by overexpression of Gli3T inhibits stroma-stimulated basal cell differentiation in prostate tissue regeneration. Particularly stromal Gli signaling regulates numerous signaling routes, including TGF␤ expression levels, which inhibits the differentiation of basal/progenitor cells (p63 ϩ or CK5 ϩ ) to CK8 ϩ luminal cells. As a result, it further controls the fate of epithelial stem cells.
Blockade of Gli signaling by Gli3T (Fig. 6, A and B) or shRNA-Gli3 (Fig. 6, C and D) results in an increase in TGF␤2 and TGF␤1 levels in UGSM cells, respectively. The differential effect of Gli3T and shRNA-Gli3 on the regulation of TGF␤ isoforms might be due to the different genetic approaches for suppression of Gli signaling. Gli3T is a constitutive repressor form of Gli3 that antagonizes, in a dominant fashion, the transcription of Gli factors, including Gli1 and Gli2. Overexpression of Gli3T leading to elevation of TGF␤2 expression has also been reported for pancreatic cancer cells (32) or with activation of Smoothened (24). On the other hand, microarray analysis suggests that Gli3 alone regulates TGF␤1 expression levels in thymocytes at the embryonic day 18.5 (E18.5) (33). Gli3 is bifunctional, meaning that it acts as a transcriptional activator as Gli3A or as a repressor as Gli3R. Gli3 is mostly expressed in the fetal stage (34) and is predominantly the Gli3R isoform in stromal cells (35). In contrast to Gli3T, shRNA-Gli3 potentially suppresses Gli3R levels, leading to up-regulation of TGF␤1 in UGSM cells (Fig. 6, C and D).
Targeting palmitoylation of Shh is a therapeutic approach for inhibition of Shh-Gli signaling. Our studies demonstrate that loss of Shh palmitoylation inhibits the renewal potential of prostate epithelial stem/progenitor cells. Although palmitoylation is not required for the autocleavage of pre-Shh protein to mature Shh (36), it is essential for the formation of the multimeric form of Shh to perform its physiological activity, such as the regulation of embryonic development (37,38). Shh-Gli signaling is associated with numerous types of cancer, including prostate cancer (39). Up-regulation of Shh in prostate cancer cells modulates the tumor microenvironment and facilitates osteoblastic cells in metastatic prostate cancer (40). The majority of inhibitors targeting Shh-Gli signaling are based on blockade of Shh downstream signaling, including Smo, Ptch, or Gli transcriptional activators (41,42). Therefore, targeting palmitoylation of Shh provides a direct therapeutic approach to inhibit the ligand activity. Palmitoylation of Shh is catalyzed by Hhat, which transfers palmitoyl-CoA to the N-terminal cysteine of Shh (36). It has been shown that mutation of Hhat, which disrupts the palmitoylation of Shh, interferes with testicular organogenesis (43). Recently identified inhibitors, such as RU-SKI 39/41/43/50, show promising inhibitory effects on Shh signaling (44,45). Our results provide biological evidence that,  (20). Additionally, epithelial stem cell activity is influenced by stromal Gli signaling. Blocking stromal Gli signaling (by overexpression of Gli3T or shRNA-Gli3) increases the expression levels of TGF␤1 or TGF␤2 and other signaling. Elevation of TGF␤ inhibits proliferation and/or differentiation of progenitor cells, subsequently inhibiting prostate tissue regeneration or spheroid formation. As a result, stromal Gli signaling, through regulation of TGF␤ levels, contributes to the differentiation and activity of epithelial stem cells.

Role of Gli signaling in prostate tissue development
although inhibition of Shh palmitoylation suppresses cell-autonomous Shh-Gli signaling, the normal stromal microenvironment is sufficient to overcome the deficit of epithelial Shh-Gli signaling. Therefore, the efficacy of the inhibitors in targeting stromal Gli signaling should also be investigated.

Plasmid and lentiviral production
The ORF of mouse Shh was PCR-amplified from the parental vector, pBS-Shh(WT) (Addgene, plasmid 13999) (primers are listed in Table S1) and inserted into the XbaI site of the FUCRW vector. The C25S mutation was introduced by sequential PCR site-directed mutagenesis technique. In the first round of the PCR reaction, pre-mShh-5Ј(XbaI)/pre-mShh-C25S-R and pre-mShh-C25S-F/pre-mShh-3Ј(XbaI) fragments were amplified individually with pBS-mShh as the template. In the second PCR, full-length Shh-C25S was amplified by the two flanking primers, pre-mShh-5Ј(XbaI) and pre-mShh-3Ј(XbaI), with the two fragments generated in the first PCR as the template. PCR was performed in a 20-l reaction mixture containing 100 ng of DNA template and 20 ng of each primer. The thermocycler steps were 2 min at 95°C, 30 cycles of 30 s (s) at 95°C, 30 s at 58°C, and 1 min extension at 72°C, followed by a final extension at 72°C for 5 min. The PCR products were gel-purified, digested by XbaI, and inserted into the XbaI site of the FUCRW vector. Because all lentiviral vectors were derived from the FUCRW parental vector, they carry an RFP marker under the cytomegalovirus promoter. Lentivirus production and infection were performed as described previously (21). All procedures followed the safety guidelines and regulations of the University of Georgia.

Isolation of primary prostate epithelial cells and prostate basal cells
Whole prostate tissues were isolated from five C57BL/6J mice (2 months old) and minced. The minced tissues were further digested by collagenase in DMEM for at least 1 h at 37°C. The collagenase-digested cell suspension was further digested with 0.05% trypsin for 5 min to dissociate cell clusters into single cells (22). After washing with DMEM to remove trypsin, the dissociated primary PrECs were resuspended in 1 ml of PrEGM medium.
For the isolation of prostate basal cells, the above PrECs were sorted based on Lin Ϫ Sca1 ϩ CD49f ϩ markers as described previously (4). In brief, 10 l of cell suspension (around 5 ϫ 10 4 cells) was aliquoted into four tubes, each containing 0.5 ml of PrEGM medium. One microliter of CD49f-PE (0.2 mg/ml), 1 l of Sca-1-APC (0.2 mg/ml), 1 l of Lin-FITC containing a mixture of CD45 (0.17 mg/ml), CD31 (0.17 mg/ml), and Ter119 (0.17 mg/ml) or none were added to each tube, respectively. The samples were used for a gating control in FACS. Additionally, a mixture of 5 l of CD49f-PE (0.2 mg/ml), 4 l Sca-1-APC (0.2 mg/ml), and 5 l of Lin-FITC containing a mixture of CD45 (0.17 mg/ml), CD31 (0.17 mg/ml), and Ter119 (0.17 mg/ml) was added to the parental tubes. All tubes were incubated on ice for 30 min. Stained cells were subjected to cell sorting by the MoFlo XDP (Beckman Coulter). Basal cells (Lin Ϫ CD49f ϩ Scal ϩ ) were collected and counted for the sphere formation assay.

Primary sphere formation from PrECs and secondary sphere formation from dissociated primary spheroid cells (without stromal cells)
Isolated primary PrECs from BL6 mice as described above were counted and seeded at 5 ϫ 10 5 cells/well in a 12-well plate. A lentivirus carrying the control vector or overexpression of Shh(WT), Shh(C25S), Gli3T, shRNA-Gli1, or shRNA-Gli2 was added to dissociated primary prostate cells with a multiplicity of infection of 10 -20, respectively. After 2 h of spin infection (at 1500 rpm), the lentivirus was removed from each well. The transduced cells were resuspended from the well, washed twice, and finally resuspended in 50 l of PrEGM medium (Lonza, catalog no. CC-3166). Fifty microliters of cell suspension was mixed with 50 l of Matrigel and plated around the rims of the wells in a 12-well plate. The rims of the wells allowed the Matrigel to create a 3D space for PrECs to grow as prostate spheres. After the cell-Matrigel mixture solidified at 37°C for 20 min, 1 ml of PrEGM was added. The sphere number and size were recorded after 10 days of incubation.
For the experiments of sphere formation in the presence of TGF␤1 ligand, isolated primary PrECs from BL6 mice as described above were counted. 7.5 ϫ 10 3 cells were resuspended in 50 l of PrEGM medium, mixed with 50 l of Matrigel, and plated around the rims of the wells in a 12-well plate. When solidified at 37°C for 20 min, 1 ml of PrEGM was added. After overnight incubation, the medium was replaced with fresh PrEGM medium containing 0, 1, 5, or 10 ng/ml of TGF␤1 (a gift from Dr. Peter Sun's laboratory). The number and size of spheres were recorded at day 10.
For the secondary spheres, dispase was added to digest the Matrigel matrix. The primary spheres were collected and treated sequentially with collagenase and trypsin as described above. Digested cells were passed through a 22-gauge syringe 3 times and filtered with a 40-m cell strainer. Cells were then resuspended in 400 l of PrEGM medium and sorted for RFPpositive cells by flow cytometry. Fifty microliters of 8 ϫ 10 3 RFP ϩ cells in PrEGM were mixed with 50 l of Matrigel and reseeded in a well of a 12-well plate. The number of spheres was counted after 10 days of incubation. Growth medium and Matrigel conditions were the same as for the primary sphere assay (22).

Sphere-stromal cell co-culture assay
In this assay, spheres were formed under the induction of UGSM cells or adult stromal cells. Of note, UGSM or stromal cells alone did not form spheres in the Matrigel in the co-culture assay.
For sphere-adult stromal cells co-culture, UGSM cells isolated as described above were replaced with adult stromal cells to perform similar experiments. Adult stromal cells were isolated based on Lin Ϫ CD49f Ϫ Scal ϩ . The isolated cells were grown in the same medium as UGSM cells (DMEM, 5% Nu-serum, 5% FBS, 0.05% insulin, 0.01 M dihydrotestosterone, and penicillin-streptomycin) and could be passaged in cell culture for three to five passages. Adult stromal cells were transduced with control vector, Gli3T, shRNA-Gli1, or shRNA-Gli2 by lentiviral infection depending on the experimental settings. The number of spheres was counted 10 days later.

Real-time PCR
UGSM cells were infected with a lentivirus depending on the experimental setting and cultured for 5 days. Cells were washed with PBS for RNA or protein extraction. Total RNA was isolated using the RNeasy Kit (Qiagen) following the protocol of the manufacturer. Complementary DNA was reverse-transcribed from 1.5 g of total RNA in a 20-l reaction with a high-capacity cDNA reverse transcription kit (Life Technologies). The reverse transcription products were diluted 30 times with distilled H 2 O, and 2 l was used as template for each realtime PCR reaction. The reactions were performed using PerfeCTa SYBR Green FastMix (Quanta Biosciences). The thermal cycling conditions were composed of an initial denaturation step at 95°C for 1 min, 40 cycles at 95°C for 10 s, and 60°C for 50 s. The experiments were carried out in triplicate. Relative quantification of -fold changes of gene expression was obtained by the 2 Ϫ⌬⌬Ct method, with GAPDH as the internal reference gene. The sequences of primers used for RT-PCR are listed in Table S1.

Click chemistry for the detection of Shh palmitoylation
PEB or 293T cells infected with a virus carrying a control vector, Shh(WT), or Shh(C25S) were grown in 6-cm dishes to 80 -90% confluence. To metabolically label palmitoylated proteins, cells were further cultured in medium containing 2% BSA (fatty acid-free) and 50 M palmitic acid azide (15-azidopentadecanoic acid) (Thermo Fisher, C10265) for 293T cells or 50 M 17-octadecynoic acid (Cayman, 90270) for PEB cells for 24 h. Cells were washed twice with PBS, and proteins were extracted with lysis buffer (100 mM sodium phosphate (pH 7.5), 150 mM NaCl, and 1% Nonidet P-40) containing protease and protein thioesterase inhibitors (0.2 mM hexadecylsulfonyl fluoride and 10 M palmostatin B). To perform click chemistry reactions, 50 g of cell lysate was incubated with reaction buffer (0.2 mM TAMRA-azide, 5 mM sodium L-ascorbate, 1 mM BTTP, and 2 mM CuSO 4 ) at a 1:1 ratio (v/v) for 1 h at room temperature in the dark. 4ϫ SDS gel loading buffer containing 150 mM ␤-mercaptoethanol was added, and samples were heated for 10 min at 70°C and separated by SDS-PAGE. After gel electrophoresis, the gel was soaked in a fixation solution (40% methanol, 10% acetic acid, and 50% water) for 1 h, followed by rinsing with deionized water (three times, 5 min each time) at room temperature. The fluorescence signal was detected using a Typhoon TRIOϩ variable mode imager (GE Healthcare). The image was analyzed with ImageQuant (GE Healthcare). The expression of Shh was also confirmed by Western blotting, and ␥-tubulin was used as the loading control.

Prostate regeneration assay and immunohistochemistry
For the prostate regeneration assay, primary prostate cells were isolated from 8-to 12 week-old BL6 male mice. The isolated primary prostate cells were transduced with control vector, Shh(WT), or Shh(C25S) by lentiviral infection. Basal cells were also isolated based on Lin Ϫ CD49f ϩ Sca1 ϩ from primary prostate cells as described above. The transduced cells (2 ϫ 10 5 cells/graft), including control vector or Gli3T, isolated basal cells, or unsorted cells, were combined with UGSM cells (2-3 ϫ 10 5 cells/graft), followed by 25 l of collagen type I (adjusted to pH 7.0). For Shh-induced transformation, UGSM cells were transduced with control, Shh(WT), or Shh(C25S) by lentiviral infection. The transduced UGSM cells were combined with PrECs. After overnight incubation, grafts were implanted under the kidney capsule in CB.17 SCID/SCID (SCID) mice by survival surgery. After 8 weeks, regenerated prostate tissues were obtained for histological and immunohistochemistry analysis.
All animals were maintained and used according to the surgical and experimental procedures of protocol A2013-03-008, approved by the Institutional Animal Care and Use Committee of the University of Georgia.
Formalin-fixed/paraffin-embedded specimens were sectioned at 4-m thickness and mounted on positively charged slides. Sections were stained with H&E and immunohisto-Role of Gli signaling in prostate tissue development chemistry (IHC) analysis was performed as described previously (21,46,47).

Statistical analysis
Prism software was used to carry out statistical analysis. The data are presented as mean Ϯ S.E. and were analyzed by Student's t test. All t tests were performed at the two-sided 0.05 level for significance.