|Home | About | Journals | Submit | Contact Us | Français|
While prostate gland development is dependent on androgens, other hormones including retinoids and estrogens can influence this process. Brief exposure to high-dose estrogen during the neonatal period in rats leads to permanent, lobe-specific aberrations in the prostate gland, a phenomenon referred to as developmental estrogenization. We have previously shown that this response is mediated through alterations in steroid receptor expression; however, further downstream mechanisms remain unclear. Herein, we examined Sonic hedgehog (Shh)-patched (ptc)-gli in the developing rat prostate gland, its role in branching morphogenesis, and the effects of neonatal estrogens on its expression and localization to determine whether a disturbance in this signaling pathway is involved in mediating the estrogenized phenotype. Shh was expressed in epithelial cells at the distal tips of elongating ducts in discreet, heterogeneous foci, while ptc and gli1–3 were expressed in the adjacent mesenchymal cells in the developing gland. The addition of Shh protein to cultured neonatal prostates reduced ductal growth and branching, decreased Fgf10 transcript, and increased Bmp4 expression in the adjacent mesenchyme. Shh-induced growth suppression was reversed by exogenous Fgf10, but not noggin, indicating that Fgf10 suppression is the proximate cause of the growth inhibition. A model is proposed to show how highly localized Shh expression along with regulation of downstream morphogens participates in dichotomous branching during prostate morphogenesis. Neonatal exposure to high-dose estradiol suppressed Shh, ptc, gli1, and gli3 expressions and concomitantly blocked ductal branching in the dorsal and lateral prostate lobes specifically. In contrast, ventral lobe branching and Shh-ptc-gli expression were minimally affected by estrogen exposure. Organ culture studies with lateral prostates confirmed that estradiol suppressed Shh-ptc-gli expression directly at the prostatic level. Taken together, the present findings indicate that lobe-specific decreases in Shh-ptc-gli expression are involved in mediating estradiol-induced suppression of dorsal and lateral lobe ductal growth and branching during prostate morphogenesis.
Prostate gland development is dependent on androgens which stimulate ductal outgrowth, branching morphogenesis, cellular differentiation, and onset of secretory activity (George and Peterson, 1988; Siiteri and Wilson, 1974). In addition, there is evidence that other hormones including estrogens and retinoids can influence these processes, although their exact role in normal prostatic development is less clear (Aboseif et al., 1997; Price, 1963; Seo et al., 1997). In humans, prostate morphogenesis occurs entirely in utero under the influence of testosterone secreted from the fetal testes (Lowsley, 1912; Shapiro, 1990). There is also clear indication that rising maternal estrogens during the third trimester have a direct effect on the human prostate since the epithelium develops marked squamous metaplasia which sloughs at birth when maternal estrogen levels abruptly fall (Brody and Goldman, 1940; Zondek and Zondek, 1975). Furthermore, maternal exposure to pharmacological levels of diethylstilbestrol (DES) has been shown to induce prostatic abnormalities in human offspring (Driscoll and Taylor, 1980). Consequently, it has been proposed that excessive estrogenization during prostatic development may contribute to the high incidence of benign prostatic hyperplasia (BPH) and prostatic carcinoma currently observed in the aging male population (Santti et al., 1994).
Prostate gland development in the rodent is initiated late in gestation as buds emerge from the urogenital sinus (UGS), and in contrast to humans, the gland undergoes extensive branching morphogenesis and cellular differentiation during the postnatal period (Hayashi et al., 1991). Thus, the neonatal rodent prostate gland has emerged as a useful model for fetal prostate development in humans. Brief exposure of rodents to estrogens during the neonatal period has been shown to have permanent, irreversible, and dose-dependent effects on the prostate gland’s morphology, cellular organization, and function (Prins, 1992; Prins and Birch, 1995; Pylkkanen et al., 1993; Rajfer and Coffey, 1979; vom Saal et al., 1997). If estrogenic exposures are high, the permanent imprints include reduced prostatic growth, epithelial differentiation defects, altered secretory function, and aging-associated dysplasia similar to prostatic intraepithelial neoplasia or PIN (Naslund and Coffey, 1986; Prins, 1992, 1997; Prins et al., 1993; Rajfer and Coffey, 1979). In the rat model, the responses are lobe-specific with differentiation abnormalities and adult-onset dysplasia occurring with the highest frequency and severity in the ventral prostate lobe (VP), while the dorsal (DP) and lateral (LP) lobes show greater inhibition of ductal branching and complexity (Prins, 1992, 1997). This process, referred to as estrogenic imprinting or developmental estrogenization, is used as a model to evaluate the role of exogenous and endogenous estrogens as a potential predisposing factor for prostate diseases later in life.
The mechanism of developmental estrogenization of the prostate gland is not well understood. Previous studies from our laboratory using estrogen receptor (ER) knockout mice for ERα and ERβ determined that the effects are mediated through stromal ERα which is transiently up-regulated following neonatal estrogen exposure (Prins and Birch, 1997). Furthermore, the expression of several steroid receptors which mediate steroid action in the developing prostate is drastically altered both temporally and quantitatively by neonatal estrogens. Most notably, androgen receptor (AR) is markedly down-regulated, while ERα, progesterone receptor (PR), and retinoic acid receptors (RARα, RARβ, and RXRα) are significantly up-regulated in a cell-specific manner (Prins and Birch, 1995, 1997; Prins et al., 1998, 2001a,b, 2002; Pu et al., 2003; Woodham et al., 2003). In addition, retinoid metabolizing enzymes and retinoid levels are affected by estrogenization in a lobe-specific manner (Prins et al., 2002; Pu et al., 2003). The net effect of these alterations is that the developing prostate is no longer under predominant androgen regulation through AR but is rather driven by estrogens, progesterone, and retinoid signals. We currently hypothesize that critical developmental genes which dictate prostate morphogenesis may be downstream targets of steroid action in the prostate gland. Furthermore, we predict that the expression levels and patterns of these steroid-regulated developmental genes will be disturbed by the prostatic steroid receptor shift as a result of neonatal estrogenic exposure. In the present study, we test the hypothesis that the Shh-ptc-gli signaling pathway may be altered in response to prostatic estrogenization and that this may, in part, mediate specific aspects of that phenotype.
Sonic hedgehog (Shh) is a secreted glycoprotein produced by epithelial cells at mesenchymal interfaces in developing tissues where it is involved in determination of cell fate, proliferation, and embryonic patterning (see review by Ingham and McMahon, 2001). Shh is expressed in a spatially defined manner in developing glands, most commonly at the apical edge of outgrowing ducts. This secreted morphogen binds to membrane-bound patched (ptc) receptors on adjacent mesenchymal cells, thus establishing epithelial–mesenchymal cross-talk. Liganding of ptc by Shh relieves its inhibition on smoothened (smo) which allows for activation of Gli transcription factors that directly mediate Shh’s effects. In vertebrates, there are three known Gli transcripts; gli1, gli2, and gli3 which have both redundant and unique actions. Importantly, gli1 and gli2 are transcriptional activators, while gli3 is believed to be a transcriptional repressor (Meyer and Roelink, 2003) which permits tight regulation of Shh actions. Both short-range and long-range actions of Shh have been described which differ as a function of concentration gradients (Gritli-Linde et al., 2001). In most structures, Shh is considered to be a critical regulatory morphogen since it regulates the expression of other secreted morphogens and homeobox genes (Chuang and McMahon, 2003; Haraguchi et al., 2001; Perriton et al., 2002; Roberts et al., 1995, 1998; Schneider et al., 2000). It has also been shown to induce ptc and gli1 expression, thus establishing an autoregulatory loop (Marigo and Babin, 1996).
Recently, the spatiotemporal expression of Shh-ptc-gli1 was characterized for the murine prostate gland (Lamm et al., 2002), while Shh-ptc mRNA expression was temporally described for the rat ventral prostate (Freestone et al., 2003). Similar to the lung, Shh was localized to the distal tips of the outgrowing murine prostatic buds, while ptc and gli1 were expressed in the adjacent mesenchyme. Expression of these signaling molecules was highest in the perinatal UGS complex and levels declined as prostate morphogenesis proceeded. Since studies with Shh null mice demonstrated that prostatic budding could be initiated by testosterone in the absence of Shh, it does not appear to be necessary for prostatic induction (Berman et al., 2004; Freestone et al., 2003). Nonetheless, inhibition of Shh action in UGS and prostate organ cultures by cyclopamine, an inhibitor of smo, demonstrated that Shh plays an important role in prostate ductal outgrowth, patterning, and epithelial differentiation during glandular morphogenesis (Berman et al., 2004; Freestone et al., 2003; Lamm et al., 2002). The exact nature of that role, however, remains unclear.
Tight regulation of developmental gene expression in specific spatiotemporal patterns is critical for normal development of all structures. However, little is known about the mechanisms involved in activating Shh gene expression in general and in the prostate specifically. There is evidence that spatiotemporal expression of Shh can be regulated by steroids in several organs. Retinoids have been shown to regulate Shh expression in the developing lungs, central nervous system, wing buds, and skeleton (Cardoso et al., 1996; Helms et al., 1994, 1997; Suzuki et al., 1999; Tamura et al., 1997). While testosterone is not required for Shh expression in the UGS, it can up-regulate transcript levels of Shh (Freestone et al., 2003; Podlasek et al., 1999). Furthermore, the localization of Shh expression in the mouse prostate is dramatically influenced by dihydrotestosterone, suggesting that androgens may regulate a critical male-specific spatial expression pattern (Lamm et al., 2002). It is therefore possible that a shift in steroidal regulation of prostate morphogenesis from androgen dominance to estrogen, progesterone, and/or retinoid dominance, as occurs following neonatal estrogen exposure, may alter the expression of Shh and/or its cognate signaling molecules.
Since a staged spatiotemporal accounting of Shh, ptc, and gli1–3 expression is not complete for the rat prostate lobes, we first present a detailed characterization of these molecules during the active phases of rat prostate branching morphogenesis. The potential role for Shh during prostate development was next examined in a neonatal prostate organ culture system using exogenous Shh or anti-Shh antibodies with subsequent analysis of branching morphogenesis, prostatic Fgf10, and Bmp4 expression. We then sought to determine whether Shh, ptc, and/or gli1, gli2, and gli3 are regulated by estrogens in the developing prostate gland and whether a disturbance of this pathway is involved in mediating the neonatal estrogenization of the prostate. Our findings indicate that Shh-ptc-gli expression is suppressed by estradiol in a lobe-specific manner and suggest that this may play a role in reduced ductal outgrowth and branching defects in the dorsolateral prostate.
All rats were handled in accordance with the principles and procedures of the Guiding Principles for the Care and Use of Animal Research, and the experiments were approved by the Institutional Animal Care Committee. Timed pregnant female Sprague–Dawley rats were purchased from Zivic-Miller (Pittsburgh, PA), housed individually in a temperature-controlled (21°C) and light-controlled (14 h L–10 h D) room and fed standard Purina rat chow (Ralston-Purina, St. Louis, MO) ad libitum. The day of birth was designated as day 0. All males from a single mother were assigned to one of two groups and treated on postnatal days (pnd) 0, 3, and 5 with subcutaneous injections of either 25 µg 17β-estradiol-3-benzoate (Sigma–Aldrich Chemical Co., St. Louis, MO) in 25 µl of peanut oil (Arachis sp.) or oil alone as controls. Pups from both treatment groups were killed by decapitation on pnd 1, 3, 6, 10, 30, or 90, and the UGS-prostate complexes were dissected for subsequent analysis. Thus, pups killed on pnd 1 and 3 were exposed to a single dose of estradiol on day 0, while offspring killed on pnd 6 and later were exposed three times to estradiol. After completion of morphogenesis (pnd 30), prostatic ductal complexity was examined in each treatment group. The LPs were separated into LP1 and LP2 ductal arrays and microdissected in 0.5% collagenase IV. The number of terminal ductal tips was counted manually under the microscope for each LP1 and LP2, and data were analyzed by Student t test. Due to complexity of the VP at pnd 30, accurate ductal tip and branch point counting for this lobe were not possible.
Since the in vivo estrogenic effects on Shh-ptc-gli expression were concentrated in the LP, in vitro experiments were performed with this lobe to determine if the effect was mediated directly at the prostatic level. LPs were isolated on pnd 0 and cultured on Millicell-CM filters (Millipore Corp., Bedford, MA) floating in 2-ml medium in a six-well plates. One side of the LP from each animal was cultured in basal organ culture medium (BOCM), while the contralateral lobe from each animal was cultured in BOCM with 20 µM 17β-estradiol (Sigma–Aldrich). Previous in vitro studies over a fivefold log dose range had established that this concentration produced strong but not complete growth inhibition of the prostate lobes (Putz and Prins, 2002). The BOCM consisted of DMEM/F-12 (Invitrogen/GIBCO) containing 10−8 M testosterone (Sigma–Aldrich), 50 µg/ml gentamycin, and 1 × insulin–transferrin–selenium (Invitrogen). The LPs (n = 10) were cultured in a 5% CO2 incubator for 6 days with medium changed every 48 h. Photographs were captured with a Burle video camera and Snappy 3.0 software to monitor growth and determine 2-D area. At pnd 6, the prostate compartment was used for RNA isolation and realtime RT–PCR.
To examine the effects of exogenous Shh on prostate branching, VPs and LPs were isolated on pnd 0 and separately cultured as described above for 4–6 days in BOCM with 2 µg/ml mouse Shh protein (R&D Systems) or 2 µg/ml BSA (contralateral lobes). This was replicated on six separate sets of tissue. To determine the effect of exogenous Shh on gene expression, eight additional paired cultures were terminated after 18 h (i.e., before observable growth alterations), RNA was isolated, and Fgf10 and Bmp4 mRNA were quantitated by real-time RT–PCR. In other cultures, VPs from pnd 0 pups were cultured for 4–6 days in the presence of (1) BOCM alone, (2) BOCM plus 2 µg/ml Shh, (3) BOCM plus 2 µg/ml Shh and 0.5 µg/ml Fgf10 (R&D Systems), or (4) BOCM plus 2 µg/ml Shh and 1 µg/ml noggin (R&D Systems), a Bmp4 antagonist (n = 3–4 for each group). Further assessment of the effect of Shh on Fgf10 expression was made by culture of pnd 0 UGS/prostate complexes in anti-Shh antibody (5 µg/ml 5E1; Developmental Studies Hybridoma Bank, University of Iowa) or control IgG (n = 3) for 24 h followed by whole-mount in situ hybridization (ISH) for Fgf10 mRNA.
To examine the effect of localized Shh on prostatic branching and Fgf10 expression, SHH beads were employed. Affi-Gel blue beads (150-µM diameter; Bio-Rad) were washed in PBS, incubated in 1 mg/ml human SHH protein (kindly provided by Curis, Inc., Cambridge, MA) at 37°C for 1 h and stored at 4°C for up to 1 week. Control beads were incubated with an equal concentration of BSA. Paired pnd 0 VPs and LPs received SHH beads positioned in the proximal or distal mesenchyme with BSA beads placed in a corresponding region of the contralateral lobe. The lobes were cultured for 4 days in BOCM, after which three from each group were analyzed for Fgf10 mRNA by whole-mount ISH.
The UGS-prostatic complexes were fixed in 4% paraformaldehyde, dehydrated, and digested with proteinase K. Following prehybridization, tissues were hybridized overnight at 60°C with 0.5–0.6 µg/ml digoxigenin-labeled RNA probes, washed at high and low stringency, incubated overnight at 4°C in antidigoxygenin alkaline phosphatase-conjugated antiserum (Roche, Indianapolis, IN), and color-reacted with NBT and BCIP (Roche). To allow for temporal and treatment comparisons, pnd 1, 3, and 6 prostatic complexes from in vivo control and estrogenized rats were processed together, and direct comparisons were made within each run. A minimum of four separate ISH assays were performed for each separate gene from in vivo experiments. For in vitro studies, all tissues from individual experiments were processed together to allow direct comparisons of signal intensity. The prostates were photographed with a Zeiss AxioCam color digital camera using AxioVision 2.0.5 software. To identify cellular localization of gene expression, ISH-stained tissues were cross-sectioned at 10 µm.
Cloned cDNA for mouse Shh (0.9 kb) was kindly provided by Dr. Liang Ma (Tulane University, LA), and cloned cDNAs for mouse gli1 (1.6 kb), mouse gli2 (1.0 kb), and mouse gli3 (2.4 kb) were kindly provided by Dr. A. Joyner (Skirball Institute of Biomolecular Medicine, NY). The rat Patched1 (ptc1) and rat Fgf10 templates were prepared by TA cloning a 660 and a 498 bp PCR fragment, respectively, into PCR II vector. Digoxigenin-labeled RNA probes were prepared by in vitro transcription using appropriate RNA polymerases (DIG RNA labeling kit, Roche).
Ptc protein and p63, a basal cell marker, were localized by immunocytochemistry as described (Prins et al., 1991). Briefly, frozen sections were fixed in 2% paraformaldehyde, blocked with 2% serum, incubated overnight at 4°C with anti-ptc antibody (1 µg/ml G-19; Santa Cruz Biotechnology, Santa Cruz, CA) or anti-p63 antibody (0.4 µg/ml H-137; Santa Cruz), reacted with biotinylated anti-IgG (Vector Laboratories, Inc., Burlingame, CA), and detected with avidin-biotin peroxidase (ABC-Elite, Vector Labs) using diaminobenzidine tetrachloride as a chromagen. For controls, normal goat or rabbit IgG (Vector) was substituted for primary antibody on separate sections.
Two procedures were used for RNA extraction and reverse transcription (RT) depending upon tissue volume. A standard assay for pnd 6–90 VP involved RNA extraction with Trizol (Invitrogen, Carlsbad, CA), DNase I digestion (Roche), and RT with AMV at 42°C for 60 min using the RT System (Promega, Madison, WI). A microassay for smaller tissues (pnd 1–6 VP, pnd 6–10 LP and DP) used RNeasy Kit (Qiagen, Valencia, CA) for RNA extraction, On-Column DNase I Digestion, and RT with MMLV at 37°C for 60 min using First Strand cDNA Synthesis Kit (Fermentas Inc., Hanover, MD). Random primers were used for reverse transcription.
The exon spanning primers and dual-labeled probe sets used for PCR are shown in Table 1. For dual-labeled probes, the 5′-reporters were FAM for Shh, gli1, Fgf10, and Bmp4; Hex for ptc, gli2, and RPL19; and Texas red for gli3, and the 3′ was labeled with black hole quencher. Plasmids containing each DNA sequence (Shh, ptc1, gli1, gli2, gli3, Fgf10, Bmp4, and RPL19) were cloned with TOPO TA cloning kit (Invitrogen) and used for standard curves in each reaction to directly quantitate target DNA levels. Ribosomal protein L19 (RPL19) was quantitated and served as an internal reference for normalization. Direct comparisons of RPL19 per unit total RNA revealed no effect of estrogen treatment in developing prostates. Real-time PCR was performed in duplex with Platinum qPCR Supermixture-UDG (Invitrogen) using the iCycler (Biorad, Hercules, CA). Reaction conditions were optimized for single gene and multiplex PCR, and the cycle conditions were 95°C for 3 min and 40 cycles of 95°C for 15 s and 60°C for 30 s. Optical data obtained by real-time PCR were analyzed with the manufacturer’s software (iCycle Optical System Interface Version 3.0). Each assay was repeated three to ten times using different tissues. Statistical analysis involved ANOVA and two-tailed Student t test (Sigma Plot, Version 8.02, SPSS Inc., Chicago, IL).
Whole-mount ISH combined with real-time RT–PCR was used to determine the spatiotemporal expression patterns of Shh, ptc, and gli1, 2, and 3 in the rat prostate gland during the active period of branching morphogenesis. There were no noticeable differences between the VP, DP, or LP with regards to expression patterns of these hedgehog signaling molecules. Prostatic budding is initiated at fetal day 18.5 of a 21-day gestation in the rat when UGS epithelial cells penetrate into the surrounding UGS mesenchyme in the dorsal, lateral, and ventral directions. At fetal day 20 (E20), the earliest time point examined in the present study, there was robust Shh mRNA expression in the prostatic buds in all three lobes with the greatest concentration in the central-to-distal aspects of these protruding ducts (Fig. 1A). By pnd 1 (i.e., 2 days later), Shh transcript levels had declined relative to the intense expression observed at fetal day 20, and expression was now restricted to the distal region of the elongating ducts (Figs. 1B, F) where it remained for the remainder of morphogenesis (Fig. 1C). Cross-sectional analysis demonstrated that Shh mRNA was confined to the epithelial cells at the leading edge of the prostatic buds and was absent in the proximal-to-central ducts (Fig. 1D). Importantly, the Shh expression pattern at the distal tips was not evenly distributed among all epithelial cells but rather was heterogeneous as early as pnd 1, a pattern which persisted as ducts elongated and branched. This is shown in Figs. 1F – H where distinct Shh expression foci are observed with adjacent regions of lower expression levels. By pnd 6, there was a noticeable decline in overall Shh signal intensity by whole-mount ISH (Fig. 1C), and this was confirmed by RT–PCR which showed a significant reduction in Shh transcript levels at pnd 6 as compared to pnd 1 (Fig. 2A, oil controls in hatched bars). By day 30, when prostatic branching is complete, Shh transcript levels reached a nadir where they remained through adulthood (Fig. 2B, hatched bars).
Ptc, the transmembrane receptor for Shh, was localized to the distal periductal mesenchymal cells in the developing prostate gland of the newborn rat, thus establishing an epithelial to stromal cell signaling pathway (Fig. 3A). As the ducts elongated and branched, the highest concentration of ptc mRNA was visualized in the distal ducts, and an expression boundary was observed between the proximal and central ducts (Fig. 3C). Cross-sectional analysis of ISH stained tissues revealed the strongest ptc mRNA signal in mesenchymal cells immediately adjacent to outgrowing ducts (Fig. 3E). A reduced intensity signal was also noted in distal epithelial cells (Fig. 3E); however, there was no epithelial signal for ptc mRNA in the proximocentral ducts (Fig. 3F). This localization pattern was confirmed by immunocytochemistry which revealed ptc protein in periductal mesenchymal and epithelial cells in the distal ducts (Fig. 3G), whereas in the proximocentral regions, ptc protein was present at low levels in the mesenchyme alone (Fig. 3H). By pnd 6, there was a significant decline in ptc mRNA levels observed both by whole-mount ISH as well as by RT–PCR (Fig. 2A, hatched bars) and levels continued to decline to a nadir in adulthood (Fig. 2B).
The three downstream gli transcription factors exhibited similar spatiotemporal expression patterns during prostatic development with some subtle differences (Fig. 4). Strong expression of gli1, 2, and 3 mRNA by periductal mesenchymal cells was observed on the day of birth. As the ducts elongated and branched, strong gli1, 2, and 3 expression was retained at the leading distal tips (Fig. 4); however, proximodistal variations were observed between them. Gli1 was expressed in the central and distal periductal mesenchyme (Figs. 4A and B), gli2 expression was retained along the entire ductal length with an increasing proximodistal gradient (Figs. 4D and E), and gli3 localized to the distal tip exclusively with a broader mesenchymal expression than observed for gli1 and gli2 (Figs. 4G and H). While cross-sectional analysis of the ISH-stained tissues revealed strong mesenchymal localization for all three glis, weaker stain was also observed in the distal tip epithelial cells for gli1 and gli3, but not for gli2 (Figs. 4C, F, and I). These subtle differences may have physiologic relevance, since gli1 and 2 are believed to be activators of Shh targets whereas gli3 is generally believed to be a repressor of downstream genes (Ingham and McMahon, 2001). Similar to Shh and ptc expression, levels of gli1, 2, and 3 mRNA significantly declined by pnd 6 and reached nadirs at day 30 where they remained through adulthood (Fig. 2, hatched bars).
In contrast to a stimulatory role for Shh in branching morphogenesis of other structures (Ingham and McMahon, 2001; Pepicelli et al., 1998), several recent studies have shown that Shh protein addition to cultured prostates inhibits ductal growth, while Shh antagonists increase branching, suggesting an inhibitory role for Shh in prostate morphogenesis (Berman et al., 2004; Freestone et al., 2003; Wang et al., 2003). To understand the mechanism of this effect, experiments were undertaken using the cultured neonatal rat prostate as a model system. Shh was applied either globally to the medium of cultured VPs or LPs or locally via placement of SHH-soaked beads in proximal or distal mesenchyme, and growth was monitored over several days. As was previously observed, the addition of Shh protein to culture medium markedly suppressed ductal elongation and branching morphogenesis throughout the prostate lobes (Fig. 5A). Immunocytochemistry for p63, a basal cell marker, revealed that despite the reduction in growth and branching, differentiation of the epithelial cells into a bilayer of luminal and basal cells was not affected by exogenous Shh (Figs. 5D – E). Local application of SHH beads within the distal mesenchyme of either the VP or LP resulted in inhibition of ductal elongation and branching in the immediate vicinity of the bead, while distant ducts remained unaffected (Fig. 5B). In contrast, SHH beads in the proximal mesenchyme did not influence subsequent duct elongation or branching (Fig. 5C) which suggested that the ability of Shh to affect branching required a distinct local environment.
One possibility is that Shh affects the expression of local growth factors necessary for prostatic growth and branching. To address this, the expression of Fgf10 and Bmp4 mRNA was assessed as a function of Shh manipulation since (1) both of these mesenchymal genes are known downstream targets of Shh in other systems (Haraguchi et al., 2001), (2) Fgf10 is expressed in the distal, but not proximal mesenchyme in the developing prostate (Thomson and Cunha, 1999), (3) Fgf10 plays an stimulatory role in prostatic branching morphogenesis (Donjacour et al., 2003), and (4) Bmp4 plays an inhibitory role in prostatic ductal outgrowth and branching (Lamm et al., 2001). Using RT–PCR, Fgf10 and Bmp4 transcripts were quantitated in VPs and LPs cultured with Shh protein in the medium. Since this culture system results in a marked shift in the prostatic epithelial–stromal ratio, gene expression was assessed before observable growth alterations. In both lobes, Shh protein significantly down-regulated Fgf10 expression and increased Bmp4 expression within 18 h, suggesting that the growth inhibitory effects of Shh may be mediated through alterations in the expression of these morphogens (Fig. 6A). To address this directly, replacement experiments were conducted with exogenous Fgf10 protein or noggin, a Bmp4 antagonist, added to prostates cultured with Shh protein for 4–6 days. As shown in Fig. 6B, Shh protein suppressed ductal elongation and branching of the VP, and this was largely overridden by exogenous Fgf10. In contrast, antagonism of Bmp4 with noggin was unable to reverse the growth suppression by Shh protein (Fig. 6C). Prostates cultured with either Fgf10 or noggin alone in the presence of 10 nM testosterone exhibited a modest increase in prostate size and branching (data not shown); however, this was not significant when measured by 2-D analysis. As an alternate approach to demonstrate Fgf10 regulation by Shh, UGS-prostatic complexes were removed on pnd 0 and cultured for 24 h with anti-Shh or control antibody (n = 3). The expression of ptc, which is directly stimulated by Shh, was repressed which confirms that the antibodies functionally blocked Shh action (Fig. 7A). Importantly, Fgf10 expression was markedly increased in Shh-inhibited prostates when compared to controls as determined by wmISH (Fig. 7B). Taken together, these data indicate that the prostatic growth suppression by Shh is mediated through suppression of mesenchymal Fgf10.
Since Shh is exclusively expressed in distal ductal tips of the branching prostate, we examined the local regulation of Fgf10 expression by SHH beads implanted along the distal mesenchyme. Prostatic-UGS complexes were removed on pnd 0; SHH or BSA beads were placed along the LP and VP surfaces and cultured for 24 h. Whole-mount ISH for Fgf10 revealed a strong reduction in mesenchymal Fgf10 mRNA expression in explants with SHH beads as compared to the BSA controls (Figs. 7C – D). Similarly, in VPs cultured for 3 days with SHH beads implanted in the distal mesenchyme, Fgf10 expression was reduced in the immediate vicinity of the SHH beads where ductal growth and branching were inhibited (Figs. 7E – F). In contrast, Fgf10 expression was unaffected at a distance from the beads where branching was normal as well as in the immediate vicinity of BSA beads (Fig. 7E). These data indicate that Shh at distinct distal sites causes local down-regulation of Fgf10 expression in the adjacent mesenchyme.
Exposure to estradiol in vivo on pnd 0, 3, and 5 reduced Shh-ptc and gli mRNA expression in the DP and LP. In contrast, expression of these signaling molecules in the VP was not significantly affected. As early as pnd 1, wmISH consistently revealed strong suppression of Shh transcript in the LP and DP regions when tissues from control and estrogen-treated rats were processed together (Fig. 8A, top), while Shh expression was similar between the two groups in the VP (Fig. 8A, bottom). This suppression of Shh expression in the dorsolateral prostate persisted through pnd 6 (Fig. 8B). Furthermore, real-time RT–PCR of Shh mRNA levels in day 6 and 10 LP and DP revealed a significant reduction following estradiol exposure in those lobes specifically (Fig. 9), while no statistical difference was noted in the VP between the treatment groups at any time point (Fig. 2). Similar to Shh, estradiol reduced ptc expression in the LP and DP as early as pnd 1 (Fig. 8C), and this effect persisted through pnd 6 for the LP and pnd 10 for the DP (Figs. 8D and and9).9). In contrast, VP ptc expression was not affected by estradiol exposure at any time point (Figs. 2A and B; 8C and D). Likewise, gli1 and gli3 expressions were suppressed in the LP and DP through pnd 10 following neonatal estradiol exposure (Figs. 8E and F; ;9),9), while expression in the VP was not changed (Figs. 2 and 8E and F). Notably, gli2 expression was not affected by this steroid in any of the prostate lobes (Figs. 2 and and9).9). It is noteworthy that estrogen did not affect the localization of the hedgehog signaling molecules in the prostate lobes.
To determine whether the effects of estrogen on prostatic Shh-ptc-gli expression were direct prostatic effects or indirectly mediated through systemic alterations following estrogen treatment, an organ culture system was employed for the developing LP. Previous in vitro studies with VPs have shown a recapitulation of several, but not all, of the estrogen effects, indicating that estrogen has direct effects at the level of the prostate gland (Jarred et al., 2000). In the present study, LPs were removed on pnd 0 and grown in vitro with testosterone or testosterone plus 20 µM estradiol (contralateral lobe). As shown in Fig. 10A, LP growth and branching in vitro were markedly suppressed by estradiol. Measurement of Shh transcript levels in the cultured LPs by real-time RT–PCR revealed lower levels of Shh mRNA following in vitro estradiol exposure (Fig. 10B); however, this trend was not statistically significant (P = 0.09). In contrast, ptc, gli1, gli2, and gli3 mRNA were all significantly reduced by estradiol in vitro to the full extent as observed in vivo (Fig. 10B). Together, these findings indicate that the estrogenic effects are mediated directly at the prostatic level.
Our previous studies have shown that neonatal exposure to estrogens interrupts branching morphogenesis and cellular differentiation in the developing prostate and that these phenomena have a degree of lobe specificity to them (Prins, 1992, 1997). While all lobes show a similar overall reduction in growth and adult size, branching is most restricted in the LP. In light of the above lobe-specific findings with regards to Shh-ptc-gli expression, we further characterized the estrogen-induced branching deficits by microdissecting the lobes in collagenase on pnd 30 (when branching is normally completed) and counting the number of terminal ductal tips in the lateral LP1 and LP2 ductal systems. Fig. 11A shows a single LP from a control animal where the elongated LP1 ducts and bushy LP2 ducts are visualized with multiple branch points and terminal ductal tips or acini. In contrast, the dorsolateral-UGS complex from an estrogenized rat clearly shows the lack of branching and absence of acini in the dorsal and lateral ducts (Fig. 11B). Comparison of the number of terminal ductal tips or acini in the lateral LP1 and LP2 lobes of the control and estrogenized rats revealed a significant difference between the groups (Fig. 11E). Furthermore, the number of LP1 tips in estrogenized rats was 6.7 ± 0.3 which is the exact number of main LP1 ducts found in a control LP, indicating a complete inhibition of branching in that lateral region as a result of high-dose estrogen exposure. While the number of tips was greatly reduced in the estrogenized LP2, there was visual and numerical evidence that some branching occurred in that region, albeit minimal. In contrast, the VP of the estrogenized prostate (Fig. 11D), although significantly reduced in size as compared to the control ventral lobe (Fig. 11C), exhibited a complex branching pattern, so much so that an accurate counting of the branch points and terminal tips was not possible.
The present study provides a clear spatiotemporal picture of the expression patterns of the Shh-ptc-gli signaling molecules in the separate rat prostate lobes during development. The patterns described are in close agreement with similar findings for the mouse prostate gland (Lamm et al., 2002; Podlasek et al., 1999) and the developing rat ventral lobe (Freestone et al., 2003). Overall, the expression of the hedgehog signaling molecules is high at birth when budding is underway and rapidly declines over the next several days when morphogenesis is complete. Whole-mount ISH of the UGS-prostatic complex from fetal day 20 to pnd 6 showed that, during the initial budding phase, Shh has a broad epithelial expression along the ductal length which rapidly transitions into a distal tip epithelial expression pattern as the ducts elongate and branch. Furthermore, the expression of Shh at the tips is not uniform but rather appears heterogeneous with foci of higher Shh expression at specific sites which may allow for highly localized Shh actions at those sites. The highest concentration of ptc and gli1–3 transcripts was localized to the condensed mesenchymal cells adjacent to the elongating epithelial ducts in the distal regions, which in total provides for the greatest Shh signal being transmitted at the distal aspect of the developing gland. Similar patterning for Shh-ptc has been observed in several developing structures with axis polarity including the lungs (Helms et al., 1997; Pepicelli et al., 1998), external genitalia (Haraguchi et al., 2001), and limbs, (Yang et al., 1997) suggesting a common mechanistic role for this paracrine signal. It is noteworthy that ptc receptor and gli1 and gli3 transcription factors were also localized, albeit at lower expression levels, in the distal tip epithelial cells but were not expressed in the proximal or central ductal epithelial cells. A low level of ptc was previously detected by RT– PCR in isolated epithelial cells of the mouse prostate (Lamm et al., 2002). Thus, in addition to paracrine signaling, Shh may have a role in directly activating genes in the distal epithelial cells in an autocrine manner.
The highly localized expression of Shh at the leading edge of the outgrowing ducts is undoubtedly of critical importance in its functional role. Several other developmental genes, including Nkx3.1 (Bhatia-Gaur et al., 1999; Prins et al., 2001a,b), Hoxb13 (Sreenath et al., 1999), and Bmp7 (Huang et al., 2003), have been localized to the prostatic distal tip epithelium, while Fgf10 and Bmp4 have been localized to the distal mesenchyme (Huang et al., 2004; Thomson and Cunha, 1999) making this region reminiscent of the distal signaling center described for lungs (Cardoso, 2000; Weaver et al., 1999). The lung distal signaling center is believed to be involved in cell fate specification along the proximal–distal axis of the respiratory tract, and it is proposed that cells at the distal tip, exposed to high concentrations of morphogens, assume a distal phenotype, whereas cells at a distance assume or retain a proximal character. It is well recognized that prostatic ducts have morphologic and functional differences along the proximal–distal axis, although the molecular determinants of this polarity have been unclear (Prins et al., 1992). We propose that a distal signaling center exists within the rodent prostate gland with Shh-ptc-gli playing a central role in determining polarity along the proximal–distal ductal axis.
Unlike Shh whose expression is strictly limited to the distal epithelium, ptc, gli2, and, to a degree, gli1 are expressed along the ductal length with decreasing expression towards the proximal duct. This continued proximal–central duct expression as the ducts elongate and branch may allow for long-range Shh actions along the ductal length. Recent studies (Gritli-Linde et al., 2001) demonstrated that Shh protein associates with GAGs and can be detected over long distances in the basement membrane, raising the possibility of graded Shh effects along the ductal length. Long-range as well as short-range actions of Shh have been described in other systems, and these differ based on its concentration gradient as well as the presence or absence of specific transcription factors in the target cells (Ingham and McMahon, 2001). Thus, it is noteworthy that the activating gli1 and gli2 are expressed along the ductal length, while gli3, a repressor of Shh targets, is expressed only in the distal mesenchyme. These differences may lead to distinctive molecular responses to Shh along the ductal length as opposed to the distal signaling center.
The precise roles of Shh in prostatic development are not well clarified at the present time. Although initially thought to be essential for prostatic budding (Podlasek et al., 1999), recent studies with Shh null mice demonstrated that Shh is not required for prostatic initiation (Berman et al., 2004; Freestone et al., 2003). Experiments to address the role of Shh in prostatic ductal growth and branching have produced conflicting data. Implantation of anti-SHH antibody beads in the UGS inhibited prostatic ductal outgrowth (Podlasek et al., 1999), while cyclopamine, an Shh antagonist, reduced distal epithelial cell proliferation (Freestone et al., 2003) which suggests a stimulatory role for Shh in prostatic growth. In contrast, addition of Shh to the culture medium reduced prostatic ductal length and branching, while cyclopamine had the opposite effect, suggesting an inhibitory role for Shh in these processes (Berman et al., 2004; Freestone et al., 2003; Wang et al., 2003). The present findings help to explain the growth suppressive effects of exogenous Shh in the prostate by demonstrating that decreased Fgf10 expression following Shh application is the proximate cause of growth and branching suppression. Ductal elongation and branching are known to be driven by Fgf10 in the prostate gland as well as other structures (Bellusci et al., 1997; Donjacour et al., 2003; Hoffman et al., 2002; Thomson and Cunha, 1999). In the present study, the addition of Shh to prostate cultures rapidly down-regulated prostatic Fgf10 transcript levels, while anti-Shh antibodies increased Fgf10 expression throughout the distal mesenchyme which indicates that Fgf10 is downstream of Shh in the prostate gland. Importantly, the Shh-induced growth suppression was reversed by concomitant addition of Fgf10 to the prostate cultures. Shh addition to the cultures also resulted in increased expression of prostatic Bmp4, a secreted morphogen known to restrict ductal outgrowth (Lamm et al., 2001). However, the addition of noggin, a Bmp4 antagonist, did not override the Shh-induced growth inhibition which indicates that the elevation in Bmp4 does not by itself mediate the Shh effects. Overexpression of Shh has been shown to suppress hair follicle morphogenesis as opposed to its normal stimulatory role which emphasizes that the exact timing, location, and concentration of Shh affect the resultant phenotype (Ellis et al., 2003). Taken together, we propose that the reported effects of ubiquitous administration of Shh or Shh antagonists to cultured prostates are a result of global alterations in Fgf10 expression throughout the prostatic mesenchyme.
In normal developing tissues including the prostate, the spatial expression of developmental genes is critical in dictating the final phenotype. While evaluating the effects of regional Shh application using SHH beads, we observed that growth and branching inhibition was dependent on the location of the beads. Thus, SHH beads implanted proximally, where Fgf10 is not expressed, had little effect, while similar beads implanted in the distal mesenchyme locally down-regulated mesenchymal Fgf10 and concomitantly inhibited ductal elongation and branching in the immediate vicinity. This is significant in light of our present results on the regionalized expression pattern of prostatic Shh. Not only is Shh highly restricted to the distal tip of elongating ducts, its expression appears heterogeneous within the distal signaling center with focal areas of higher and lower expression. We offer a model in Fig. 12 for the potential involvement of Shh in dichotomous branching of the prostate ducts which incorporates the highly localized, focal Shh expression pattern along with the findings that Shh suppresses Fgf10 and enhances Bmp4 expression in adjacent mesenchymal cells in the distal prostate gland. We propose that as the prostatic ducts elongate and make contact with Fgf10-expressing mesenchymal cells, focal expression of Shh at the distal tips will locally down-regulate Fgf10 expression in the mesenchyme immediately adjacent to the Shh foci. This localized reduction in Fgf10 will create expression domains with relatively higher Fgf10 levels lateral to the Shh foci (Fig. 12b). Higher Fgf10 activation of FgfRiiib on adjacent epithelial cells will increase epithelial proliferation at the site of each lateral Fgf10 subdomain (Fig. 12c). Disparate epithelial proliferation rates will result in the sprouting of two buds on each side of the Shh foci, thus initiating a branchpoint (Fig. 12d). Contributing to this, mesenchymal Bmp4 expression will be locally up-regulated adjacent to the Shh foci, and this will inhibit ductal outgrowth at that restricted Shh/Bmp4 domain. In this manner, Shh, Fgf10, and Bmp4 may work in synchrony to initiate and maintain dichotomous branching at the distal aspects of elongating ducts. Importantly, interruption of this signaling network by altered expression of Shh and/or Fgf10 or their cognate receptors would result in growth and branching abnormalities. It is noteworthy that a similar hypothesis has been ventured for the role of Shh with Fgf10 in branching morphogenesis of the lung (Acosta et al., 2001; Weaver et al., 2000).
A major aim of the present study was to determine if neonatal estrogenic exposure affected prostatic Shh-ptc-gli expression and whether this contributed to the estrogenized phenotype in the developing and adult prostate. The estrogenized rat prostate can be described as a proximalized phenotype in which the morphologic and differentiation characteristics of the proximal ducts extend far into the glandular structure at the expense of a central–distal phenotype, which is suppressed (Prins et al., 2001b). For the dorsolateral prostate, this proximalized phenotype includes branching deficiencies where the proximal ducts fail to branch in the distal regions. Since it has been demonstrated that disruption of the lung distal signaling center disrupts distal pattern formation and leads to proximalized phenotypes in the pulmonary system (Cardoso, 2000; Weaver et al., 1999), we suspected that the proximalized DLP phenotype was related to disruption of distal signaling center genes that are specifically involved in distal branching. The present results support this hypothesis by showing a lobe-specific effect of estrogens on both Shh-ptc-gli expression and ductal outgrowth and branching. Following in vivo exposure to estrogens, Shh, ptc, gli1, and gli3 were significantly reduced in the DP and LP within 24 h, while the VP appeared unaffected. Furthermore, we observed that the main ducts of the DLP were stunted and exhibited marked branching deficiencies, while the VP exhibited similar branching complexity in the estrogenized and control rats. It is important to recognize that other factors are affected by estrogens besides the Shh-ptc-gli system which could account for the reduced branching independent of Shh. However, taken together with the model proposed above for the role of Shh in dichotomous branching, the present results strongly implicate the lobe-specific suppression of the Shh-ptc-gli pathway in mediating the suppression of ductal outgrowth and branching in the dorsal and lateral lobes.
It is of note that estradiol does not affect the localization of Shh expression in the prostate. This is relevant since Lamm et al. (2002) have shown that DHT shifts Shh expression in the female UGS from the cranial margin to the tips of the nascent buds; that is, DHT activates a male-specific spatial pattern. Apparently, estradiol does not counteract this effect. It is also important to point out that estrogenization of the prostate inhibits epithelial proliferation and stimulates mesenchymal proliferation leading to a relative increase in prostatic stromal mass (Prins, 1992). However, this shift in cell ratios cannot account for the decreased ptc and gli expression in the present study since the relative number of stromal cells increases following estrogenization yet mesenchymal ptc and gli decrease.
The organ culture studies demonstrated that estradiol is capable of suppressing LP Shh-ptc-gli expression in vitro which indicates that the estrogenic effects are mediated directly at the prostatic level. However, it is noteworthy that the estrogenic suppression of Shh mRNA was not as marked in vitro as observed in vivo which implicates the additional involvement of a systemic factor in lowering prostatic Shh levels. One potential candidate is the decreased circulating androgens as a result of neonatal estrogen exposure (Corbier et al., 1992; Haavisto et al., 2001; Prins, 1992). Although androgens are not required for Shh expression in the UGS, they have been shown to increase levels of Shh transcript (Freestone et al., 2003; Lamm et al., 2002) which raises the possibility that androgens have a supplemental role in maintaining normal Shh expression. We propose that a combination of reduced circulating androgens together with direct effects of estrogen may result in the in vivo suppression of DP and LP Shh levels. We also predict that the direct prostatic effects of estrogen on epithelial Shh expression are mediated through paracrine factors since ERα, which initiates the events, is expressed in periductal mesenchymal cells (Prins and Birch, 1997; Prins et al., 2001a,b). In contrast to Shh, estrogenic regulation of ptc and gli1–3 was fully recapitulated in vitro, indicating that their suppression is mediated entirely at the prostatic level. It is of interest that ptc and glis are markedly down-regulated by estradiol in vitro when Shh levels are only marginally suppressed which suggests that their reduced levels are not secondary to reduced Shh expression but may be direct targets of estrogen action through the up-regulated ERα in the mesenchyme. Alternatively, estrogens may indirectly function at the prostatic level through retinoids via RAR/RXRs (Prins et al., 2002; Pu et al., 2003) or through the ERα-induced progesterone receptor in the mesenchymal cells (Sabharwal et al., 2000).
The molecular basis for lobe-specific effects of estrogen on Shh and its downstream signaling molecules is not clear at present. Our previous studies have shown similar ERα localization and up-regulation following neonatal estrogen exposure in all prostate lobes; thus, the presence or absence of the initiating receptor is not a factor in lobe specificity (Prins and Birch, 1997). Likewise, AR is immediately down-regulated to a similar degree in all lobes following neonatal estrogenization (Prins and Birch, 1995). Recently, we observed a lobe-specific expression pattern for RXRγ, retinaldehyde dehydrogenase 2 (RALDH2), and CYP 26 which form and degrade intracellular retinoic acid, respectively, as well as cellular retinoic acid binding proteins, CRABP 1 and 2 which sequester and regulate retinoic acid availability within cells (Pu et al., 2003). Transcript levels for each of these molecules were high in the developing DP and LP but were minimally expressed in the VP, suggesting a greater developmental role for retinoids in the DLP. Furthermore, neonatal estrogens drastically reduced RALDH2 and CYP26 transcripts and suppressed CRABP 1 and 2 levels in the DP and LP specifically which may compromise retinoid availability. This is highly significant since retinoic acids have been shown to be critical regulators of Shh expression in multiple developing structures (Cardoso et al., 1996; Goyette et al., 2000; Helms et al., 1994, 1997; Niederreither et al., 2002). Future studies are aimed at examining the potential lobe-specific retinoid regulation of Shh-ptc-gli in the DLP.
In closing, lobe-specific responses to neonatal estrogens have been previously described, but their molecular basis has been unclear. While the VP is the site of epithelial differentiation defects and adult-onset dysplasia, the LP epithelium shows minimal differentiation abnormalities but major branching deficiencies (Prins, 1997). These distinctive regional responses to the same initial compound suggest that they are mediated through separate pathways. Previous studies have shown that permanent reductions in AR and ERβ are specific to the VP and are thought to contribute to adult dysplasia in that lobe (Prins, 1992; Prins et al., 1993, 1998). Since the present data indicate that the Shh-ptc-gli pathway participates in prostatic branching morphogenesis through localized regulation of other critical morphogens, it is likely that the regional suppression of the Shh-ptc-gli signaling system in the DP and LP contributes to the lobe-specific growth deficits in that prostatic region.
The authors would like to thank Dr. Liang Ma at Tulane University for assistance in establishing whole-mount ISH and for generously providing the Shh cDNA plasmid, Dr. Alexander Joyner at New York University for generously providing the mouse cDNA plasmids for gli1, gli2, and gli3, Lynn Birch for technical assistance, and Dr. Oliver Putz for graphic renditions of our data. Work was supported by NIH grant DK-40890.