Inhibition of Epithelial Ductal Branching in the Prostate by Sonic Hedgehog Is Indirectly Mediated by Stromal Cells*

Sonic hedgehog (Shh), a vertebrate homologue of the Drosophila segment-polarity gene hedgehog, has been reported to play an important role during normal development of various tissues. Abnormal activities of Shh signaling pathway have been implicated in tumorigenesis such as basal cell carcinomas and medulloblastomas. Here we show that Shh signaling negatively regulates prostatic epithelial ductal morphogenesis. In organotypic cultures of developing rat prostates, Shh inhibited cell proliferation and promoted differentiation of luminal epithelial cells. The expression pattern of Shh and its receptors suggests a paracrine mechanism of action. The Shh receptors Ptc1 ( P atched 1 ) and Ptc2 were found to be expressed in prostatic stromal cells adjacent to the epithelium, where Shh itself was produced. This paracrine model was confirmed by co-culturing the developing prostate in the presence of stromal cells transfected with a vector expressing a constitutively active form of Smoothened, the real effector of the Shh signaling pathway. Furthermore, expression of activin A and TGF-β1 that were shown previously to inhibit prostatic epithelial branching was up-regulated following Shh treatment in the organotypic cultures. Taken together, these results suggest that Shh negatively regulates prostatic ductal branching indirectly by acting on the surrounding stromal cells, at least partly via up-regulating expression of activin A and TGF-β1.

The prostate is derived from the embryonic urogenital sinus under the influence of testosterone levels and interactions between the mesenchyme and epithelial cells during embryogenesis (1,2). The initial prostatic primodium is formed around birth in the rodent. The prostatic epithelial buds continue to undergo extensive ductal outgrowth and branching into the surrounding mesenchyme during the first 3 weeks after birth (3,4). In the mature prostate, there are essentially two major types of epithelial cells within the prostatic epithelium: the luminal and basal cells. They can be distinguished based on their location, morphology, function, and expression profile of specific cytokeratins (5). The luminal cells located in the apical region of the epithelium are generally believed to be terminally differentiated cells that are androgen-dependent and produce secretory proteins (5). The basal cells residing between the luminal cells and the underlying basement membrane express cytokeratin 14 or p63 (6) and are thought to serve as progenitor cells (7). Currently, the molecular mechanisms that control prostatic ductal morphogenesis and luminal epithelial cell differentiation are still poorly understood.
Among several molecules that might be involved in regulation of prostatic ductal branching (reviewed in Ref. 8), sonic hedgehog (Shh) 1 is a good candidate. Originally identified as a homologue of the segment-polarity gene hedgehog of the Drosophila (9), Shh has been reported to be expressed at numerous sites of epithelial and mesenchymal cell lineages during development (10,11). It serves as a morphogen to exert its biological functions (12), including dorsoventral patterning of body axis, specification of neuronal and oligodendrocytes cell fates, cell proliferation, cell differentiation, axonal outgrowth, and cell survival (13,14). Both Drosophila and mouse genetics show that the seven transmembrane protein, Smo (Smoothened) is required for Shh signaling (15)(16)(17). Ptc (Patched) appears to negatively regulate Smo in the absence of Shh (18 -20). Even though it is widely accepted that Shh binds to Ptc (21,22) with high affinity, it is unclear how Shh binding results in the activation of its downstream target Smo, as well as the subsequent signaling pathway. The original model suggested that Ptc binds directly to Smo and represses its activity in the absence of Shh. Upon Shh binding the normal inhibition by Ptc is released, and Smo initiates signaling. However, recent studies in Drosophila have suggested that Ptc may not repress Smo activity through a direct interaction but rather that Ptc inhibits Smo activity from a distance (23)(24)(25)(26)(27), possibly through the regulation of vesicular trafficking. The main target of Shh activity is a family of zinc finger transcription factors known as Gli or Ci in the fly. In vertebrates there are three Ci orthologues: Gli1, Gli2, and Gli3 (9). Disruption or mutation of the Shh pathway results in developmental disorders (28 -30) and are highly associated with several human diseases including basal cell carcinoma (31)(32)(33) and medulloblastomas (34). Plant alkaloid cyclopamine, an antagonist on Shh signaling by blocking the activity of Smo (35,36), has shown promises in preclinical models of medulloblastomas (34) and could be proven useful in the treatment of Shh-associated tumors.
Although a recent study suggested that Shh signaling might be involved in the initial morphogenesis of the prostate during embryogenesis (37), prostatic development and maturation mainly occurs during the postnatal period (3)(4)(5). Regulation of epithelial ductal branching and differentiation of luminal epithelial cells by Shh after the prostatic primodium is formed around birth is not clearly addressed. In the present experiments, we focused on a possible role of Shh during postnatal prostatic ductal morphogenesis and luminal epithelial cell differentiation. We determined the cellular expression pattern of Shh and its receptors, Ptc1 and Ptc2, and studied functions of Shh during postnatal prostatic development by comparing Shh with testosterone and cyclopamine in postnatal prostate organ cultures. We found that Shh inhibited epithelial ductal branching of the prostate. Shh appeared to act indirectly through stromal cells, because Ptc1 and Ptc2 were selectively expressed in the stromal cells. More importantly, co-culturing developing prostate with stromal cells overexpressing a constitutively active form of Smo (SmoM2) (33) mimicked the effects of Shh on prostatic branching morphogenesis. This effect led to an upregulation of activin A and TGF-␤1, which are shown previously (38 -40) to be expressed in the stromal cells and inhibit prostatic epithelial branching. Therefore, we propose that Shh regulates prostatic branching morphogenesis indirectly via the stroma by at least in part up-regulating expression of activin A and TGF-␤1.

Organotypic Cultures, Cell Transfection, and RNA Preparation-
The ventral prostate whole mounts were dissected from postnatal day 2 (P2) rats and plated on cell culture inserts (8-m pore size; BD Biosciences) individually in serum-free medium. Serum-free medium was made of Dulbecco's modified Eagle's medium/F-12 plus serum-free supplement (I-1884; Sigma), 1% bovine serum albumin, 2 mM glutamine, 5 mg/ml glucose, 25 ng/ml fungizone, and 100 units/ml penicillin and 100 mg/ml streptomycin, modified as described previously (41). Shh (400 nM; Genentech), cyclopamine (5 M; kind gift of Dr. W. Gaffield at Western Regional Research Center, United States Department of Agriculture, Albany, CA), and testosterone (10 -8 M; Sigma) was added to the medium at the beginning of the cultures. The medium was changed every other day. For some cultures, BrdUrd was introduced into the medium for 17 h before the cultures were fixed on the fourth day. In some experiments, individual P2 rat prostate was cut into 16 pieces, mixed with stromal cells transfected either with a control vector or a vector expressing a constitutively active form of Smo (SmoM2) or Gli1 when they were plated together in the cell culture inserts. Stromal cells were isolated from E17 rat urogenital sinus by tearing away urogenital epithelium following incubation in 0.2 M EDTA in Hanks'-buffered saline as described previously (42). Cell transfection was done 1 day before the co-culture. The cultures were fixed at various time points in 4% paraformaldehyde in phosphate buffer (0.1 M, pH 7.4) for 30 min at room temperature before they were processed for immunocytochemistry (43). Cell transfection was performed using GenePorter, essentially in a similar manner as described previously (44,45). Total RNA was isolated from four cultures of each of the experimental groups 2 days after treatment with Shh using RNeasy columns (Qiagen), treated with DNase I (Roche Molecular Biochemicals) for 15 min at room temperature, and cleaned using RNeasy columns as described (44).
Immunocytochemistry, Image Acquisition, Cell Counts, and Data Analysis-For BrdUrd and cytokeratin 14 double immunocytochemistry, fixed cultures were treated with 2 N HCl for 30 min at room temperature. After washes, the preparations were then incubated with a mixture of anti-BrdUrd antibody (1:40; BD Biosciences) and anticytokeratin 14 (1:10,000; Berkeley Antibody Company) in phosphatebuffered saline containing 3% normal donkey serum and 0.2% Triton overnight at 4°C. Positive staining was visualized by incubation with FITC-conjugated donkey anti-mouse and Texas red-conjugated donkey anti-rabbit secondary antibodies as described previously (44). Some of the cultures were labeled with anti-Shh monoclonal antibody (5E1, 1:1000; Developmental Studies Hybridoma Bank) or double stained with anti-p63 antibody (1:200 dilution; Santa Cruz Biotechnology) mediated via FITC-conjugated secondary antibody and 7AAD (Molecular Probes). Cultures were mounted in Fluoromount-G (Southern Biotechnology) containing counter staining dye 4Ј,6-diamidino-2-phenylindole (Sigma) and viewed using a Zeiss Axiophot epifluorescent microscope. Images were captured with Compix imaging systems using a cooled RGB CCD camera and analyzed using Adobe PhotoShop 4.0. Sections labeled with anti-p63 antibody (Santa Cruz Biotechnology) and 7AAD (Molecular Probes) were viewed using a Bio-Rad MRC-600 confocal microscope.
Cell counts were performed on the digital images captured from the sections of the cultures. For BrdUrd labeling index, BrdUrd-labeled cells versus the total number of epithelial cells in given epithelial ducts (cytokeratin 14 highlighted region) or a given area of stromal area were counted against DAPI counter-stained nuclei. Approximately 1,500 cells were counted from randomly selected 15-20 sections from each of the experimental groups. Data were then normalized to the control cultures and are expressed as mean Ϯ S.D. For luminal cell differentiation analysis, p63 negative cells and total epithelial cells (7AADpositive) in randomly selected epithelial ducts were counted from five to six sections from each of the experimental groups, normalized to the control group, and expressed as mean Ϯ S.D. Two-way, unpaired t test was used for statistical analysis.
Gene Expression Analysis by in Situ Hybridization and PCR-[ 33 P]UTP-labeled sense and antisense riboprobes were generated from PCR products cloned into transcription vectors (46). The murine Ptc1 and Ptc2 cDNA, cloned in pBluescript, (Stratagene, La Jolla, CA) spanned from nucleotides 3518 to 4289 of MMU46155 (for Ptc1) and from nucleotides 883 to 1309 of AJ133485 (for Ptc2), respectively. Formalin-fixed, paraffin-embedded, dissected mouse prostatic tissue (postnatal days 7 and 16 and adult) was sectioned and deparaffinized. The in situ hybridization was performed as described previously (44). Tissue pre-treatments consisted of a proteinase K digestion (4 g/ml) for 30 min at 37°C, a 0.5ϫ SSC wash for 10 min at room temperature, and dehydration prior to application of the pre-hybridization solution. Slides were processed for autoradiography and exposed for 4 weeks before developing. Sense control probes revealed no hybridization above background (not shown).

Shh Inhibits Prostatic Epithelial Ductal Branching and
Growth-When whole mount ventral prostate organs were isolated from postnatal day 2 rats and maintained in serum-free medium (see "Materials and Methods"), epithelial ductal branching and expansion continued (Fig. 1A), although at a slower pace than in the cultures where testosterone was included (Fig. 1D). These observations are in agreement with previous work by Cunha and co-workers (47). However, a significant inhibition of branching morphogenesis was observed in cultures treated with 400 nM Shh (Fig. 1B). Lower concentrations of Shh (200 nM) also led to clear, though less profound, growth inhibition (data not shown). In contrast, cultures grown in the presence of an Shh pathway antagonist, cyclopamine at 2.5 or 5 M, showed enhanced epithelial growth (Fig. 1C). This inhibition was further confirmed by immunocytochemistry of cultures with an anti-cytokeratin 14 antibody that highlights the epithelial area in the ductal units (5). As shown in Fig. 1F, the number of epithelial ductal tips was much lower in cultures where Shh was present than in control cultures (Fig. 1E). On the other hand, the tips of the epithelial ducts in the cultures containing cyclopamine were enlarged (Fig. 1G). Similar growth-inhibiting effects were also observed when the tissues were cultured in the presence of both testosterone and Shh (data not shown), suggesting that Shh can override the stimulating effects of testosterone on epithelial cell growth. In addition, when both Shh and cyclopamine were added together to the culture, the growth pattern was the same as in cultures treated with cyclopamine alone (not shown), indicating that the concentration of cyclopamine used (5 M) is sufficient to block both exogenous and endogenous Shh activity.
The concentrations of Shh used in the present experiments (200 and 400 nM) are similar to previous studies on induction of midbrain dopaminergic neurons (48). To further rule out the possibility that the growth-inhibiting effects of Shh we observed might be because of any potential general toxicity of Shh, we did cell death analysis by performing TUNEL staining on the sections of the cultures from all four experimental groups. Only negligible number of TUNEL-positive cells were observed in any of the cultures grown in the presence or absence of Shh at 400 nM, cyclopamine at 5 M, or testosterone at 10 -8 M. However, a dramatic elevation in the number of TUNEL-positive cells was seen (92.0 Ϯ 60.9, n ϭ 4) in sections of the cultures grown in the presence of 10 M cyclopamine, suggesting that at higher concentrations, cyclopamine might be toxic.
Shh Inhibits Cell Proliferation in the Prostate-To examine whether Shh influences prostatic morphogenesis by affecting cell proliferation, we performed BrdUrd immunocytochemistry with all four groups of cultures. We found that BrdUrd labeling index was greatly reduced in both epithelial and stromal areas of the Shh-treated cultures (Fig. 2, C and D versus A and B). In contrast, BrdUrd labeling index was elevated in the cultures treated with cyclopamine (Fig. 2, E and F). Testosterone, a potent mitogen used as a positive control, showed the highest BrdUrd labeling index (Fig. 2, G and H). Cell counts of the number of BrdUrd-labeled cells versus total number of epithelial cells in a given ductal unit or total stromal cells in a given stromal area provided quantitative support to our observations. The BrdUrd labeling index in the epithelium is as follows: control, 80.3 Ϯ 21.4; Shh-treated, 48.6 Ϯ 18.6; cyclopamine-treated, 93.1 Ϯ 11.8; testosterone-treated, 111.44 Ϯ 16.8 (p Ͻ 0.01, between Shh-treated and control or between testosterone-treated and control; p Ͻ 0.05 between cyclopamimetreated and control). Data were normalized to the control group and are plotted in Fig. 3.
Shh Promotes Terminal Differentiation of Luminal Epithelial Cells-By postnatal day 5 of normal development, some of the basal cells that are normally cytokeratin 14/p63-positive start to become postmitotic and differentiate into cytokeratin 14/p63-negative luminal cells (5). To determine whether luminal cell differentiation is also affected by Shh, cultures were maintained for 7 days in serum-free medium and were doublelabeled with anti-p63 antibody that highlights basal epithelial cells and 7AAD that stains nuclei of all cells. We found that the proportion of luminal cells among the total epithelial cell population within an individual ductal unit was significantly higher in the Shh-treated cultures (Fig. 4B) as compared with control cultures (Fig. 4A). Consistent with this observation, the proportion of luminal cells in ductal units was significantly lower in the cultures treated with cyclopamine (Fig. 4C). Tes- tosterone showed the highest percentage as it is expected to promote luminal cell differentiation in addition to cell proliferation (47). The ratio of luminal to total epithelial cells from all four types of cultures are shown in Fig. 4E (p Ͻ 0.05, between Shh-treated and control; p Ͻ 0.01 between control and cyclopamine or testosterone-treated). These observations suggest that Shh appears to regulate prostatic morphogenesis not only by inhibiting proliferation of progenitor cells but also by promoting terminal differentiation of luminal cells. Cell counts of luminal cells per ductal unit revealed an absolute increase in luminal cells (control, 9.2 Ϯ 1.1; Shh-treated, 15.2 Ϯ 1.9; mean Ϯ S.E., n ϭ 35, p Ͻ 0.01) and provided further support for this model. Therefore, Shh pushes the progenitor cells out of cell cycle earlier than scheduled and induces a precocious differentiation into luminal epithelial cells. (22,50) have demonstrated that Shh exerts its biological function through its receptors Ptc1 and Ptc2. Ptc genes also constitute good reporters for Shh pathway activation as they are both up-regulated upon Shh signaling. Therefore, determination of the expression patterns of Ptc1 and Ptc2 in the prostatic tissue would help us to understand the mechanism of action by Shh and in particular the cell type targeted by Shh activity. We carried out in situ hybridization on tissue sections prepared from mouse prostates at different developmental stages. Although the sense probes did not show any specific labeling (data not shown), signals for Ptc1 and Ptc2 were specifically detected in stromal cells that are adjacent to the epithelium at P7 (Fig. 5, A-D). Expression of Ptc1 and Ptc2 became much lower in P16 prostates, and no detectable signal was seen in adult prostate (data not shown), suggesting a down-regulation of the two receptors as the prostate matures.

Cellular Expression Patterns of Shh Receptor and Shh-Previous studies
To find out where Shh is produced, we performed immunocytochemistry of the P2 prostate cultures using an anti-Shh antibody. As shown in Fig. 5E, immunoreactivity was seen confined to the epithelium of the prostate. These results indicate that Shh is produced by epithelial cells whereas its receptors are expressed by the stromal cells.

Co-culturing Stromal Cells Transfected with a Constitutively Active Form of Smo Leads to a Growth Inhibition in the Developing Prostatic Epithelium-The expression patterns of Ptc1
and Ptc2 in stromal cells and the presence of Shh in the epithelium of the prostate suggest a paracrine mechanism by which Shh might act on stromal cells that then influence epithelial ductal branching and growth. To provide further experimental evidence for this model, we cultured developing prostates that were minced into pieces together with stromal cells that expressed a constitutively active form of Smo (SmoM2) (33) or a control vector. In the control cultures, prostatic glandular ductal structures were well formed resembling intact prostate tissue (Fig. 6A). However, cultures maintained in the presence of SmoM2-transfected stromal cells showed an inhibition of epithelial branching similar to what we observed in Shh-treated cultures (see Fig. 6B and Fig. 1B). Activation of the Shh signaling pathway by transfecting the stromal cells with a downstream effector gene, Gli1, which up-regulated expression of Ptc1 (data not shown), led to similar results (Fig. 6C). To visualize more clearly the epithelial buds of these cultures, we immunostained them with an anti-cytokeratin 14 antibody. As shown in Fig. 6, D-F, the number of epithelial buds was greatly reduced in the cultures in which stromal cells were transfected with SmoM2 or Gli1. These findings support the notion that Shh inhibits prostatic epithelial branching by acting on the stromal cells. Consistent with this notion, we observed that Shh inhibited proliferation of purified human or rodent stromal cells but not that of human LNCaP and PC3 prostatic cancer cell lines. 2 Ptc1 and Ptc2 Expression Was Up-regulated in Developing Mouse Prostates Treated with Shh-Ptc1 and Ptc2 are known downstream targets of the Shh pathways whose expression is up-regulated upon Shh signaling. To confirm that the biological effects we observed in these cultures were elicited by Shh action on stromal cells, we examined the expression levels of Ptc1 and Ptc2 in the cultures following Shh treatment by quantitative RT-PCR. As shown in Fig. 7, the expression levels of Ptc1 and Ptc2 were elevated significantly after Shh treatment (p Ͻ 0.01, n ϭ 4 for each group).
Sonic Hedgehog Up-regulates Expression of Activin A and TGF-␤1-As an attempt to understand how Shh exerts its inhibitory effects on epithelial branching morphogenesis via stromal cells, we focused our attention on TGF-␤ superfamily members including activin A, TGF-␤1, BMP4, and BMP7 as these molecules are potential candidates that mediate the effects of Shh in other tissue types (for reviews see Refs. 9 and 51). Furthermore they are reported to be expressed by prostatic stromal cells and are capable of inhibiting prostatic epithelial cell growth (38 -40, 52, 53). We performed TaqMan RT-PCR analysis to determine whether Shh treatment alters expression of these TGF-␤ superfamily molecules in the cultures. As shown in Fig. 8, cultures treated with Shh significantly upregulated expression of activin A and TGF-␤1 as compared with the control cultures (p Ͻ 0.01, n ϭ 4 for each group; see Fig. 8). In contrast, expression of BMP4 and BMP7 remained essentially unchanged following Shh treatment (p Ͼ 0.05).

DISCUSSION
In the present study, we demonstrated that Shh regulates prostatic morphogenesis by inhibiting epithelial ductal branching. Such inhibition was achieved by negative regulation of cell proliferation of both epithelial cells and underlying stromal cells and by promotion of terminal differentiation of the luminal epithelial cells. We also showed that the stromal cells, but not the epithelial cells, express the receptors for Shh, Ptc1, and Ptc2. More importantly, co-culturing the developing prostate with stromal cells transfected with SmoM2 mimicked the growth-inhibitory effects of Shh. These two lines of evidence strongly suggest that Shh acts on stromal cells to indirectly regulate the proliferation and differentiation of prostatic epithelial cells.
Activation  A and B) and Ptc2 (C and D) in P7 mouse prostates. Shown are hematoxylin-eosin brightfield (A and C) and corresponding darkfield (B and D) for each image. Strong expression of Ptc1 is detected in the stroma (arrows) adjacent to epithelial buds of the developing prostate. The epithelium is negative (arrowheads). The expression pattern of Ptc2 is similar to that of Ptc1, although the expression intensity of Ptc2 appears weaker. Ptc2 is also detected in the stroma (arrows) whereas the associated epithelium (arrowhead) is negative. E, Shh immunocytochemistry of a section from a P2 rat ventral prostate maintained in culture for 4 days indicates that Shh immunoreactivity is seen in the epithelium (epi) but not in the stroma (str). Scale bar, 100 m. effects that depend upon the tissue types. In the skin (31-33), cerebellum (34,54,55), and kidney (56), Shh appears to stimulate cell proliferation and to inhibit cell differentiation. However, in some other tissues, Shh inhibits cell proliferation and induces terminal differentiation of specific cell types. Our findings in the prostate seem to be consistent with studies on gastric gland cells, pancreatic cells, midbrain dopaminergic neurons, and spinal cord motoneurons. For example, activation of Shh signaling inhibits gastric gland cell proliferation and morphogenesis (57). Shh has also been shown to indirectly inhibit expansion of pancreas mass by suppressing within the endoderm the expression of PDX-1, a transcription factor that is necessary for the formation of the pancreas (58 -60). In the developing nervous system, Shh acts as a morphogen and induces the differentiation of dopaminergic neurons (48), serotonergic (61), and motor neurons (62). However, our observations are different from what was reported recently by Bushman and co-workers (37) in the prostate. Their results suggest that Shh enhances epithelial cell proliferation and facilitates the initial ductal budding based on experiments using Shh neutralizing antibody. There could be several explanations for the discrepancy. One difference is that we studied prostate development at a postnatal stage when major prostatic ductal branching and cell differentiation occur whereas Bushman and co-workers (37) focused on the initial prostatic budding process from the embryonic urogenital sinus before birth. They combined the beads pre-absorbed with Shh neutralizing antibody with E15 urogenital sinus tissue, which is the primodium that will give rise to the prostate before they grafted the tissue-bead mixture under the renal capsule, and they observed an impairment of prostatic ductal morphogenesis 7-10 days post-grafting (37). It is possible that the tissue responsiveness to Shh may change as development proceeds. Such stage-specific responsiveness of Shh has been described during normal development of avian skin (63). Alternatively, localized expression of Shh in the distal aspect of the epithelial bud (64) may yield different outcomes as compared with ubiquitous administration of Shh. Ubiquitous overexpression of Shh in the lung epithelium leads to an increase in mesenchyme versus epithelium (65,66). Such increased mesenchymal cell mass results in defects in the branching morphogenesis (66,67) as epithelial outgrowth is inhibited everywhere along the bud. On the other hand, loss of Shh signaling can also cause defects in epithelial branching (66) in the form of enlarged epithelial buds as epithelial outgrowth is initiated ubiquitously. Despite the difference in the effects of Shh at various developmental stages, activation of Shh signaling occurs in mesenchymal/ stromal cells in both cases. While we demonstrated by in situ hybridization that Shh receptors are expressed in the stromal cells, Lamm et al. (64) looked at another downstream target of the pathway, Gli1, and found it up-regulated by RT-PCR in mesenchymal cells. In addition, our tissue recombination experiments using SmoM2-or Gli1-transfected stromal cells demonstrated that activation of the Shh signaling pathway in stromal cells show similar growth-inhibiting effects on epithelial cells as soluble Shh. Similar paracrine action mechanism of epithelial cell-derived Shh has been observed in the developing mouse kidney and Xenopus small intestine (68). During development of the metanephric kidney, Shh is expressed in the ureteric epithelium, and its receptor Ptc1 is expressed in the surrounding mesenchymal cells (56). During organogenesis of the Xenopus midgut, the effectors of hedgehog signaling are localized to the gut mesoderm whereas hedgehog is produced in the endoderm (68). Activation of hedgehog signaling leads to an inhibition on the transition of midgut endoderm into intestinal epithelium (68). FIG. 7. Shh elicits an up-regulation of Ptc1 and Ptc2 expression. TaqMan RT-PCR analysis indicates that addition of Shh into the culture results in an up-regulation of the receptors of Shh, Ptc1, and Ptc2. This experiment confirms that the biological effects of Shh we observed in the cultures are indeed attributable to the activation of Shh signaling. RNAs were extracted from the cultures 2 days following treatment with Shh, cyclopamine, or testosterone.

FIG. 8. Shh increases expression of activin A and TGF-␤1 in prostatic organ cultures.
TaqMan RT-PCR analysis indicates that addition of Shh into the culture results in an up-regulation of activin A and TGF-␤1 but not BMP4 and BMP7. RNAs were extracted from the cultures 3 days following treatment of the cultures with Shh. The probe and primers for activin A were designed based on the sequence of the gene encoding inhibin ␤A chain as activin A is a homodimer with two ␤A chains. Other activin molecules and inhibin contain one of the other ␤B, ␤C, ␤E, or ␣ subunits. After normalization to Gapdh, the data were further normalized to the cultures without Shh (Untreated). The values represent -fold increases for each of the genes examined in Shh-treated cultures.
Stromal influence on prostatic epithelial cell proliferation and differentiation during embryogenesis and tumorigenesis has been demonstrated previously (69). Tissue recombination studies show that urogenital sinus mesenchymal cells can induce epithelial cells from non-prostatic tissues to become prostatic epithelial cells (42,70), and such induction is determined by the amount of mesenchymal tissue present (71,72). Stromaepithelial cell co-culture experiments also indicate that although carcinoma-associated stromal cells enhances prostatic carcinogenesis (72,73), normal stromal cells are capable of inhibiting prostatic carcinogenesis by inducing differentiation and inhibiting the proliferation of the epithelium (52,74,75). Consistent with this notion, a reduction in the number of normal stromal smooth muscle cells is closely associated with progression of prostatic carcinogenesis (76). Although the molecular mechanisms through which stroma modulates the epithelial cell phenotype are still unknown, a few well characterized signaling pathways, such as steroid hormones and specific growth factors, may contribute to the paracrine regulation of epithelial cell growth and differentiation by stromal cells. For example, experiments using androgen-insensitive urothelium, prepared from the testicular feminized (Tfm/y) mice, elegantly demonstrate that androgens exerts their effects by acting through androgen receptors in the urogenital sinus mesenchymal cells to stimulate epithelial proliferation, prostatic ductal branching morphogenesis, and columnar cytodifferentiation (5,77). There is also evidence of a link between elevated serum levels of insulin-like growth factor-I (IGF-I) and an increased risk of prostate cancer (78). In normal prostate IGFs are produced only by stromal cells whereas normal epithelial cells express the IGF receptors and synthesize specific IGF-binding proteins that modulate effects of IGFs (79).
Although the exact mechanism by which Shh signaling inhibits prostatic epithelial cell growth is not yet fully understood, our quantitative PCR analysis suggests that Shh action on the stromal cells enhances the expression of the inhibitory growth factors activin A and TGF-␤1, which may, in turn, act on epithelial cells. Consistent with this notion, conditioned media collected from primary stromal cell cultures inhibit prostatic epithelial cell proliferation (52). Activin A (38, 39) and TGF-␤1 (40) have been shown to exert inhibitory effects on epithelial cell growth and branching morphogenesis and are found to be expressed in the prostatic stromal cells. In this regard, members of the TGF-␤ superfamily have been demonstrated to mediate the effects of Shh in several developmental systems (for reviews, see Refs. 9 and 51). For example, TGF-␤2, a very close TGF-␤1 family member, has been shown to mediate the effects of Shh on hypertrophic differentiation and parathyroid hormone-related peptide expression in the perichondrium of embryonic mouse metatarsal bones grown in organ cultures (80). Loss of responsiveness of prostatic epithelial cancer cells to Shh might be attributable to the facts that expression of TGF-␤ receptors is dramatically down-regulated in these cancer cells (49,81) and that stromal cells are absent in the prostatic cancer cell cultures. On other hand, it should be noted that BMP4, a vertebrate homolog of Drosophila Decapentaplegic, which acts as a downstream target of Shh in other tissue types (for reviews, see Refs. 9 and 51), is found to be expressed in the prostate and shown to inhibit prostatic epithelial branching (53), but its expression is not regulated by Shh based on our RT-PCT analysis (Fig. 8). Considered together, these data suggest that Shh might influence epithelial branching morphogenesis by at least in part up-regulating expression of activin A and TGF-␤1. As demonstrated in the developing lung (65, 66) and midgut (68), our findings support the notion that appropriate regulation of Shh signaling in the stroma plays a critical role in stromal-epithelial cell interactions that determines proper epithelial ductal branching and differentiation in the prostate.