BMP a Putative Regulator of Epithelial Homeostasis in the Human Prostate Is a Potent Inhibitor of Prostate Cancer Bone Metastasis in Vivo

【摘要】  Bone morphogenic protein 7 (BMP7) counteracts physiological epithelial-to-mesenchymal transition, a process that is indicative of epithelial plasticity. Because epithelial-to-mesenchymal transition is involved in cancer, we investigated whether BMP7 plays a role in prostate cancer growth and metastasis. BMP7 expression in laser-microdissected primary human prostate cancer tissue was strongly down-regulated compared with normal prostate luminal epithelium. Furthermore, BMP7 expression in prostate cancer cell lines was inversely related to tumorigenic and metastatic potential in vivo and significantly correlated to E-cadherin/vimentin ratios. Exogenous addition of BMP7 to human prostate cancer cells dose-dependently inhibited transforming growth factor ß-induced activation of nuclear Smad3/4 complexes via ALK5 and induced E-cadherin expression. Moreover, BMP7-induced activation of nuclear Smad1/4/5 signaling transduced via BMP type I receptors was synergistically stimulated in the presence of transforming growth factor ß, a growth factor that is enriched in the bone microenvironment. Daily BMP7 administration to nude mice inhibited the growth of cancer cells in bone. In contrast, no significant growth inhibitory effect of BMP7 was observed in intraprostatic xenografts. Collectively, our observations suggest that BMP7 controls and preserves the epithelial phenotype in the human prostate and underscore a decisive role of the tumor microenvironment in mediating the therapeutic response of BMP7. Thus, BMP7 can still counteract the epithelial-to-mesenchymal transition process in the metastatic tumor, positioning BMP7 as a novel therapeutic molecule for treatment of metastatic bone disease.
--------------------------------------------------------------------------------
The phenotypic changes of increased motility and invasiveness of carcinoma cells are reminiscent of the epithelial-to-mesenchymal transition (EMT) that occurs during embryonic development and morphogenesis.1-3 During EMT, epithelial cell layers lose polarity and cell-cell contact, undergo a dramatic remodeling of the cytoskeleton, express mesenchymal components, and manifest a migratory phenotype by resolving cell-matrix contacts and digesting adjacent basal lamina.4,5 Whereas in the developing embryo EMT is a prerequisite for the formation of various tissues and organs, EMT in postnatal life is required for repair and remodeling of numerous tissues or organs.1,6
Plasticity resulting from cells shifting between epithelial and mesenchymal phenotypes is discernible either by EMT or the reverse process of mesenchymal-to-epithelial transition (MET).2 Both processes have emerged as a fundamental principle for reprogramming of gene transcription and as a major determinant of stem cell fate in development and in adult tissue homeostasis. The process of EMT is currently considered a potential mechanism of the disease progression in malignant and fibrotic disorders.1 An example of the latter is the persistence of a pathogenic insult of the kidney that causes disruption of the tubular basement membrane and eventually may lead to renal fibrosis due to transition of tubular epithelial cells into a migratory mesenchymal phenotype (EMT-derived fibroblasts in the interstitium).6
In epithelial cells transforming growth factor ß (TGFß) has been identified as one of the main inducers of EMT during development and in fibrotic disorders,5-10 although another member of the TGFß superfamily, bone morphogenetic protein 7 (BMP7) or osteogenic protein-1, is involved in the maintenance of the epithelial phenotype by induction of MET.11-14 Zeisberg et al13,14 demonstrated that BMP7 was capable of counteracting TGFß-induced EMT and could reverse chronic renal injury.
Prostate cancer is the second most frequently diagnosed cancer and the second leading cause of cancer death in the Western male population. Evidence is mounting that during carcinogenesis, developing prostate cancer cells acquire mesenchymal (and even osteoblastic) characteristics and migratory features concomitant with a loss of epithelial characteristics like E-cadherin expression.15,16 In cancer progression, "oncogenic" EMT refers to clusters of malignant cells that lose epithelial characteristics and acquire self-sustained migratory and highly invasive phenotypes.5 We hypothesized that the Gleason histological grading of prostate cancer parallels EMT starting from "well-differentiated" cells with well-defined boundaries that still resemble healthy prostate epithelium (Gleason patterns 1 and 2), via infiltrative cancer cells with less-defined boundaries with extensions into adjacent non-neoplastic prostate tissue (Gleason pattern 3), toward "poorly differentiated" highly migratory prostate cancer cells that have lost many epithelial characteristics and have acquired an invasive metastatic and mesenchymal phenotype (Gleason patterns 4 and 5).
Several defects in the canonical TGFß signaling pathway have been reported in prostate cancer17,18 and may contribute to the development and progression of prostate cancer as alterations in TGFß and TGFß-receptor 2 expression are associated with poor clinical outcome.19 A growing number of in vivo studies demonstrate that inhibitors of TGFß or TGFß receptors may reduce the metastatic and/or invasive properties of a variety of experimental cancers by preventing EMT pathways,5,20,21 thus underscoring the importance of TGFß in oncogenic EMT associated with cancer progression.
This study was designed to investigate whether tumorigenicity and invasive behavior are associated with modulated BMP7 expression in clinical prostate cancer specimens and in human prostate cancer cell lines with different tumorigenic potential. Furthermore, we studied if systemic administration of BMP7 affects the growth of human prostate cancer cells in orthotopic tumor and bone metastasis models by whole-body bioluminescent imaging (BLI). We present evidence for cross talk between BMP7 and TGFß signaling in the regulation of EMT in prostate cancer, and we identified BMP7 as a potential therapy for metastatic bone disease.

【关键词】  putative regulator epithelial homeostasis prostate inhibitor prostate metastasis

Materials and Methods

Tissue Sampling

Radical prostatectomy specimens were obtained at the Department of Urology of the University of Bern, Switzerland. Written informed consent was obtained from all patients, and tissue sampling was approved by the local ethical committee. Within 15 minutes of surgical excision, samples of prostate cancer tissue or noncancerous tissue of the prostate were taken and either snap-frozen or immersed in RNAlater (Qiagen, Basel, Switzerland).22 Tissue adjacent to the respective samples was processed for paraffin embedding and served as histological control. Histological diagnosis and grading were performed by a pathologist (R.M.).

Laser Capture Microdissection (LCM)

Six-µm cryosections were treated according to the HistoGene LCM frozen section staining kit (Arcturus, Bucher Biotech, Basel, Switzerland).22,23 The P.A.L.M. Micro-Beam system (P.A.L.M. Microlaser Technologies, Bernried, Germany) was used to excise approximately 1000 pure epithelial cells of noncancerous prostate (NP) and prostate cancer tissue in each prostate. Total RNA was extracted with the Pico Pure RNA Isolation Kit (Arcturus).

Real-Time Polymerase Chain Reaction (PCR) Analysis

Total RNA (see above) was reverse transcribed using random primers in the presence of RNase inhibitor (Roche Diagnostics, Rotkreuz, Switzerland). Real-time PCR was performed with exon-specific primers for human BMP7 (Hs_002333477_m1), glyceraldehyde-3-phosphate dehydrogenase (Hs_99999905_m1), E-cadherin (Hs_00170423_m1), vimentin (Hs_00185584_m1), and ß-actin (Hs_99999903_m1) (primer catalog, Applied Biosystems, Rotkreuz, Switzerland) on an ABI Prism 7700 Sequence Detection System (Applied Biosystems). The resulting values were normalized to glyceraldehyde-3-phosphate dehydrogenase or ß-actin. Semiquantitative PCR was performed on samples of experimentally induced bone metastases from PC-3M-Pro4 after intracardiac inoculation into nude mice using exon-specific and human-specific BMP primers.22,24

Cell Lines and Culture Conditions

The human prostate cancer cell lines PC-3 (American Type Culture Collection number CRL-1435; ATCC-LGC Promochem, Molsheim Cedex, France) and PC-3M-Pro4 were grown in Dulbecco??s modified Eagle??s medium plus 10% fetal calf serum (BioWhittaker, Verviers, Belgium). The PC-3M-Pro4 cells were generated from PC-3M cells by injecting PC-3M cells into athymic mouse prostates and selecting for variants with increasing metastatic potential by several rounds of reinjecting cells from xenograft tumors back into the mouse prostate.25,26 The human prostate cancer cell line LNCaP and LNCaP-derived cell lines C4-2 and C4-2 B427,28 were grown in T-Medium. Cells at 70 to 80% confluence were used for RNA extraction using an RNeasy Midi RNA extraction kit (Qiagen). Quantitative real-time PCR analysis was performed as described above.

Human PC-3M-Pro4 prostate cancer cells were stably transfected with a cytomegalovirus promoter-driven mammalian expression vector for luciferase, CMV-luc, and one clone with the highest expression of luciferase expression (PC-3M-Pro4/luc+) was successfully used for in vivo whole-body BLI.29-31 PC-3M-Pro4/luc+ were cultured in Dulbecco??s modified Eagle??s medium plus 10% fetal calf serum and 800 µg/ml geneticin/G-418 (Invitrogen, Breda, The Netherlands). Cells were regularly certified free of Mycoplasma contamination. For intraosseous or orthotopic inoculation, cell suspensions of PC-3M-Pro4/luc+ (1 x 105 cells/10 µl of phosphate-buffered saline) were prepared.29-31

Animals

Male nude (BALB/c nu/nu) mice were purchased from Charles River (L??Arbresle, France). Mice were housed in individual ventilated cages under sterile conditions according to the Dutch guidelines for the care and use of laboratory animals (DEC 4077). Surgical and analytical procedures were performed while mice were anesthetized.29,31,32

Intraosseous Inoculation of PC-3M-Pro4/luc+ Cells

A single cell suspension of PC-3M-Pro4/luc+ cells was injected into the right tibiae.29-31 The progression of cancer cell growth was monitored weekly by BLI. After the experimental period, the animals were sacrificed, and the tibiae were dissected and processed for further histomorphometrical and immunohistochemical analysis (see below).

Inoculation of PC-3M-Pro4/luc+ into the Mouse Prostate (Orthotopic Implantation)

A single-cell suspension of 1 x 105 PC-3M-Pro4/luc+ cells/10 µl of phosphate-buffered saline was surgically inoculated into the prostates of anesthetized 6-week-old male nude mice. The cutaneous wound was sutured. The progression of cancer cell growth was monitored weekly by BLI. The animals were subsequently treated daily with 100 µg/kg BMP7 or vehicle solution. After the experimental period, the animals were sacrificed. Directly after removal of the mouse prostate, BLI images of the animals were taken to establish the number of locoregional lymph node metastases. Tissues were dissected and processed for further histomorphometrical and immunohistochemical analysis (see below).

Induction of Systemic Metastases by Intracardiac Injection of PC-3M-Pro4/luc+ Cells

A single-cell suspension of 1 x 105 PC-3M-Pro4/luc+ cells/100 µl of phosphate-buffered saline was injected into the left cardiac ventricle, and cancer cell growth was monitored weekly by BLI and radiography.29-31 Daily BMP7 treatment started 2 days before inoculation of the cancer cells until the end of the experiment.

BMP7 Treatment and Prostate Cancer Growth in Vivo

Human BMP7 was obtained from Dr. S. Vukicevic (Laboratory of Mineralized Tissues, School of Medicine, Zagreb, Croatia). BMP7 was expressed, purified, and lyophilized.33 When used, BMP7 was freshly dissolved to a stock solution containing 1 mg/ml in 20 mmol/L acetate buffer with 5% mannitol, pH 4.5. BMP7 (20 µl) or vehicle solution (20 µl, 20 mmol/L acetate buffer with 5% mannitol) was administered daily into the tail veins of mice for 21 to 23 days (100 µg/kg human BMP7).

Whole-Body BLI and Quantification of the Bioluminescent Signal

BLI of tumors induced by the luciferase-expressing human prostate cancer cell lines was performed.29,31 Analyses for each metastatic site were performed after definition of the region of interest and then quantified.31 Values are expressed as relative light units.

Histomorphometry, Histochemistry, and Immunohistochemistry

Five-micrometer paraffin sections of patient-matched normal prostate and prostate cancer tissue were rehydrated, and a rabbit polyclonal antibody against BMP7 pro-domain was applied at a dilution of 1:100. Rabbit IgG (Jackson ImmunoResearch, La Roche, Switzerland) served as negative control. A species-specific biotinylated anti-IgG antibody followed by streptavidin/alkaline phosphatase conjugate (Amersham Biosciences, Roosendaal, The Netherlands) was used as the detection system. 3-Amino-9-ethyl-carbazole (Sigma, Buchs, Switzerland) served as chromogen (Sigma). Dissected tissues from animal studies were fixed in 4% paraformaldehyde (pH 6.8), decalcified (only bones) as described previously, and processed for paraffin embedding, sectioning, and staining32 with PS1 antibodies directed against phosphorylated Smad1.34,35

Immunoblotting

Crude cell lysates (10 µg/lane) were separated on sodium dodecyl sulfate-polyacrylamide gels and blotted on Hybond-P membranes (Amersham Biosciences). E-cadherin monoclonal antibody (5 µg/ml; Zymed Laboratories, Invitrogen), vimentin polyclonal rabbit antibody (ab7783; Abcam, Cambridge, UK), and a monoclonal mouse anti-actin antibody (1:5000; Transduction Laboratories, Lexington, KY) were used as primary antibodies. Binding was detected with a peroxidase-labeled anti-mouse secondary antibody (Amersham Biosciences) and the ECL Advanced chemiluminescence substrate (Amersham Biosciences) using the VersaDoc imaging system and QuantityOne imaging software (Bio-Rad, Veenedaal, The Netherlands).

Human BMP7 Enzyme-Linked Immunosorbent Assay

PC-3M-Pro4 cells and a clone stably transfected to overexpress rhBMP7 were seeded in a six-well plate (120,000 cells/ml), and conditioned medium was harvested and analyzed for BMP7 protein synthesis by enzyme-linked immunosorbent assay after 4 days of culture according to the manufacturer??s protocol (R&D Systems, Abingdon, UK).

Transient Transfections and Transcription Reporter Assays

Transient transfections and transcription reporter assays were performed and values expressed as luciferase intensity (relative light units).36 The experiments were performed in quadruplicate and repeated at least twice. Values are expressed as means ?? SEM. Incubation time with TGFß or BMP7 was 30 hours.

Luciferase Reporter Gene Constructs

For intracellular signaling of TGFß, the CAGA-luciferase construct, consisting of 12 Smad3/Smad4 binding sequences (CAGA boxes), and the luciferase-coding sequence was used. The CAGA boxes confer TGFß but not bone morphogenetic protein (BMP) stimulation to a heterologous promoter reporter construct.37

The BRE4-luciferase construct, which is based on the mouse Id1 promoter, was used to study the presence and functionality of BMP receptors.37 Signaling of BMP to stimulate the expression of BRE4-luciferase is transduced by BMP type I receptors and mediated by Smad1, Smad4, and Smad5, which form a complex with this reporter construct.38 The SBE4 construct consists of four Smad-binding elements (SBEs) in the promoter of the JunB gene, an immediate early gene that is potently induced by TGFß, activin, and bone morphogenetic proteins.39 The E-cadherin promoter-luciferase construct is based on the upstream fragment (positions C178 to +92). The pGL3-E-cadherin promoter plasmid40 was a kind gift from Dr. Antonio Garcia de Herreros (Institut Municipal d??Investigaci? Mdica, Barcelona, Spain).

Statistical Analysis

The paired t-test was performed for statistical evaluation of patient-matched mRNA expression, and the unpaired t-test was performed for comparison of mRNA expression in bulk and laser-captured tissue using GraphPad Prism Version 3 (GraphPad Software Inc.). Data are presented as means ?? SEM. Analysis of variance was performed for statistical evaluation of whole-body BLI data and reporter-luciferase assays. A P value equal or more than 0.05 was considered not significant.

Results

BMP7 mRNA Expression in Radical Prostatectomy Specimens

Samples of prostate cancer (PC) tissue and internal controls from the same patient (NP) were obtained after radical prostatectomy, and BMP7 mRNA was measured by real-time PCR. In bulk tissue BMP7 mRNA expression was generally lower than in whole tissue specimens of noncancerous prostate (11/16, mean decrease 26%), a nonsignificant difference (Figure 1, a and c ; P = 0.1). However, the multifocal nature of prostate cancer and the coexistence of epithelial and stromal tissues may cross-contaminate and dilute cell-specific gene expression analyzed in bulk tissue specimens. Therefore, we next collected pure populations of >1000 prostate cancer cells and patient-matched noncancerous prostate epithelial cells by LCM from 14 patients. In 93% of those cancer cell samples, BMP7 mRNA was strongly underexpressed when compared with patient-matched noncancerous epithelial cells (13/14, mean decrease 70%, P = 0.0019; Figure 1, b and d ). Comparison of the BMP7 mRNA expression in noncancerous prostate derived either from bulk tissue or from laser-captured epithelial cells revealed a 9.4 times higher expression in the epithelial compartment alone than in bulk tissue. BMP7 expression in laser-captured prostate cancer was also higher (2.8 times) when compared with bulk prostate cancer tissue, but these differences were less pronounced. These data strongly suggest that BMP7 is expressed mainly in the epithelial compartment of the prostate gland, particularly in the glandular epithelium.

Figure 1. BMP7 mRNA expression in noncancerous prostate epithelium (NP) and prostate cancer cells (PC) of radical prostatectomy specimens. a: BMP7 mRNA expression in patient-matched specimens consisting predominantly of NP and PC tissue (n = 16, P = 0.1, paired t-test), b: BMP7 mRNA expression in NP and neoplastic epithelial cells isolated by LCM (PC) (n = 14, P = 0.0019, paired t-test) c: Mean BMP7 mRNA expression values ?? SEM in bulk tissue in NP and PC both normalized to glyceraldehyde-3-phosphate dehydrogenase (P = 0.1, paired t-test). d: Mean BMP7 mRNA expression values ?? SEM in microdissected NP and neoplastic epithelium (PC), both normalized to glyceraldehyde-3-phosphate dehydrogenase (*P = 0.0019, paired t-test). BMP7 monoclonal antibody immunoreactivity and in clinical specimens of NP (e and f) and PC (g and h). e: Boxed area is magnified in f. BMP7 immunoreactivity was detected in the apical cytoplasm of the luminal cell layer of the normal acinar epithelium and parts of stroma (arrowhead), but weak or no staining of the basal cells was found. g and h: Prostate cancer cells lack BMP7 immunoreactivity, whereas BMP7 was still detectable in parts of the stroma (arrowhead). Scale bars = 100 µm.

Immunohistochemical analyses were performed for BMP7 on primary prostate cancer specimens and noncancerous tissue adjacent to the respective samples, which served as a histological control. Histological diagnosis and grading were performed by a certified pathologist (R.M.). BMP7 immunoreactivity was present in the cytoplasm of the luminal cell layer of the normal acinar epithelium in samples of noncancerous prostate tissue, although hardly any BMP7 could be detected in basal cells (Figure 1, e and f) . In addition, BMP7 protein expression was also found in the prostate stromal compartment and blood vessels. In contrast, prostate cancer cells showed lack of BMP7 immunoreactivity (Figure 1, g and h) .

Our LCM and immunohistochemistry data reveal that BMP7 expression was enriched in noncancerous prostate epithelium compared with prostate cancer cells in the same patients, suggesting that BMP7 down-regulation may facilitate prostate carcinogenesis, dissemination, and/or metastasis. Moreover, BMP7 expression (ratio prostate cancer/NP) appeared inversely correlated to Gleason patterns, but the distribution of patients for each pattern was too small to reach statistical significance (results not shown).

BMP7 mRNA Expression in Prostate Cancer Cell Lines

We further investigated whether BMP7 expression in prostate cancer cell lines was associated with tumorigenicity, metastatic potential, and EMT degree. The prostate cell lines examined have progressively greater tumorigenic and metastatic potential in the following order: LNCaP, C4-2, C4-2B4, PC-3, and PC-3M-Pro4. This arrangement of cell lines also parallels the gradual loss of expression of epithelial markers (E-cadherin epithelial marker) and acquisition of mesenchymal characteristics (Figure 2a) . Vimentin is a characteristic marker for the mesenchymal phenotype of cancer cells, and enhanced vimentin expression was previously found in motile prostate cell lines and in poorly differentiated and metastatic prostate carcinoma.41 In line with these observations, we found that vimentin expression increased with tumorigenicity and that E-cadherin/vimentin ratios decreased with augmented invasiveness and aggressiveness (Figure 2a) .

Figure 2. Real-time mRNA expression of E-cadherin, vimentin, and BMPs in prostate cancer cell lines with different tumorigenic potential in vitro and in experimentally induced bone metastases. a: E-cadherin/vimentin ratios in prostate cancer cell lines. b: BMP7 mRNA expression in prostate cancer cell lines. c: Correlation between E-cadherin/vimentin ratio and BMP7 mRNA expression in prostate cancer cell lines with different tumorigenic potential (R2 = 0.98, P = 0.0015). d: Relative expression levels of different BMP mRNA levels in an experimentally induced bone metastasis from PC-3M-Pro4 cells in nude mice. The mRNA levels (??SEM) were quantified by real-time reverse transcriptase-PCR relative to ß-actin endogenous control, except for d, where ß2-microglobulin was used for normalization as described previously.32

BMP7 mRNA levels in prostate cancer cell lines strongly correlated to E-cadherin/vimentin ratio (Figure 2, a and b) , whereas no association was observed for other BMPs (results not shown).24 The highly tumorigenic and metastatic human prostate cancer cell line PC-3 and its derivative PC-3M-Pro4 did not show detectable BMP7 mRNA expression (Figure 2b) and protein synthesis in vitro (human specific BMP7 enzyme-linked immunosorbent assay; all <10 pg/ml hBMP7, n = 4). The lack of BMP7 expression by PC-3M-Pro4 cells in vitro was confirmed in vivo in bone metastases (Figure 2d) , suggesting that the bone microenvironment is unable to induce BMP7 expression in metastatic cancer cells. In contrast, the LNCaP, C4-2, and C4-2B4 cell lines, characterized by a much lower tumorigenic and metastatic potential, showed substantial expression of BMP7 mRNA (Figure 2b) . E-cadherin/vimentin ratio together with BMP7 expression therefore decreased with increasing tumorigenic potential, and a strong correlation exists between BMP7 expression and E-cadherin/vimentin ratio (R2 = 0.98, P = 0.0015; Figure 2c ).

BMP7 Treatment of Bone Metastasis and Orthotopic Prostate Cancer Growth in Vivo

To test whether the observed decrease of BMP7 expression during prostate cancer progression may contribute to the acquisition of an invasive metastatic phenotype, we investigated whether systemic administration of rhBMP7 affected the growth of human prostate cancer cells in orthotopic tumor and bone metastasis animal models by whole-body BLI of human PC-3M-Pro4/luc+ cells. Three days after intrabone inoculation of PC-3M-Pro4/luc+ cells into the tibiae of BALB/c nu/nu mice, the animals were treated daily via tail intravenous injections of 100 µg/kg rhBMP7 or a vehicle solution, and growth of prostate cancer cells was monitored weekly by BLI and radiography. Administration of BMP7 for the duration of the experiment resulted in a significant and sustained inhibition of tumor cell growth and tumor-induced osteolysis in bone marrow (Figure 3, aCc) . Systemic administration of BMP7 acted directly on bone-residing PC-3M-Pro4/luc+ tumor cells as visualized by nuclear staining for phospho-Smad1, whereas this was significantly lower in vehicle-treated animals (Figure 3d) . Inoculation of PC-3M-Pro4/luc+ cells into the left cardiac ventricle of nude mice resulted in the formation of multiple bone metastases.29-31 Continuous daily treatment with BMP7 inhibited the growth of bone metastases (P < 0.05) despite a tendency toward a decrease. As a result, the average tumor burden per bone lesion was significantly decreased on BMP7 treatment (Figure 3e) .

Figure 3. Effects of daily systemic administration of 100 µg/kg rhBMP7 on the growth of human prostate cancer cells in bone marrow (aCd) and bone metastasis (e and f) in vivo after intraosseous transplantation or intracardiac inoculation of human PC-3M-Pro4/luc+ cells using whole body bioluminescent reporter imaging. a: BMP7 treatment of PC-3M-Pro4/luc+ cells growing in bone marrow of nude mice as monitored by BLI. b: Representative examples of bioluminescent photon emission in a vehicle- and BMP7-treated animal at day 21. c: Representative radiographs at day 21 after intraosseous transplantation of PC-3M-Pro4/luc+ cells growing in bone marrow (arrows indicate bone lesions in tibiae of nude mice). d: Immunohistochemical analyses of phosho-Smad1 localization at day 21 after intraosseous transplantation of PC-3M-Pro4/luc+ cells growing in bone marrow. The animals were either vehicle-treated (upper left panel) or treated with rhBMP7 (top right) 6 hours before explantation and fixation of the tibiae. The number of phospho-Smad1-positive cells in the BMP7-treated group is increased significantly. e: Effect of BMP7 treatment on the formation and growth of bone metastases by PC-3M-Pro4/luc+ in nude mice as monitored by BLI (tumor burden, number of bone metastases, representative examples of BLI at day 19). Relative light units (??SEM) emitted was quantified at least once weekly starting at day 0 after intraosseous implantation or intracardiac inoculation of the prostate cancer cells. n = 8 for each experimental group for intraosseous transplantation experiment (aCd); n = 10 for intracardiac inoculation (e). Values are expressed as means ?? SEM. *P < 0.05, **P < 0.001 (analysis of variance).

Our data therefore suggest that BMP7 inhibits the growth of bone metastases by human prostate cancer cells in nude mice. In contrast to its inhibitory effect on bone metastases, BMP7 treatment of orthotopically implanted PC-3M-Pro4/luc+ cells did not alter tumor growth (Figure 4a) . After removal of the prostate containing the orthotopically growing tumor, multiple loco-regional lymph nodes harbored bioluminescent PC-3M-Pro4/luc+ cells (Figure 4b) . Administration of BMP7, however, did not result in decreased formation of loco-regional lymph node metastases (Figure 4c) .

Figure 4. Effects of daily systemic administration of 100 µg/kg rhBMP7 on intraprostatic growth of human PC-3M-Pro4/luc+ cells and formation of loco-regional lymph node metastases nude mice. a: Tumor take of PC-3M-Pro4/luc+ cells growing in prostate on treatment with vehicle solution or daily 100 µg/kg rhBMP7 as monitored by BLI. b: The number of affected lymph nodes in nude mice after 21 days as detected by BLI. c: Representative example of bioluminescent photon emission of a mouse with multiple loco-regional lymph node metastases 21 days after intraprostatic inoculation of PC-3M-Pro4/luc+ cells. The mouse prostate was removed before bioluminescent reporter imaging to obtain sensitive measurements of affected lymph nodes (b). The presence of micrometastatic deposits in lymph nodes was confirmed by H&E staining. Relative light units emitted was quantified at least once weekly starting at day 0 after intraprostatic implantation of the prostate cancer cells. n = 8 for each experimental group. Values are expressed as means ?? SEM (analysis of variance).

BMP7 and Epithelial-to-Mesenchymal Transition in Prostate Cancer Cells

Next, we tested whether BMP7 acts on human PC-3M-Pro4 prostate cancer cells to inhibit the acquisition of an invasive mesenchymal phenotype by antagonizing Smad-dependent TGFß signaling. For this, we first studied the functionality of both TGFß and BMP receptors in PC-3M-Pro4 prostate cancer cells. Challenge of PC-3M-Pro4/luc+ cells with both BMP7 and TGFß, which were transiently transfected with SBE4-luciferase, resulted in a significant induction of reporter expression (Figure 5a) . This indicates that functional TGFß and BMP receptors are present in these cells, thus confirming our earlier observations.22,24 Exogenous addition of BMP7 but not of TGFß strongly stimulated BRE4-luciferase activity in PC-3M-Pro4 cells and indicated the presence of functioning, activated type 1 BMP-receptor complexes in these prostate cancer cells (Figure 5b) . Strikingly, when BMP7 and TGFß were given simultaneously, BRE4-luciferase activity was significantly stimulated in a synergistic manner (Figure 5b) .

Figure 5. The effects of BMP7 on gene promoter luciferase constructs in human PC-3M-Pro4 prostate cancer cells. a: Challenge of the SBE4 construct, consisting of four SBEs, demonstrates the presence of functional BMP7 and TGFß receptors (Smad-mediated signaling). *P 0.01 versus Co. b: BMP7, but not TGFß, induces BRE4-luciferase activity, indicating the presence and functionality of BMP receptors. Signaling of BMP to stimulate the expression of BRE4-luciferase is transduced by BMP type I receptors and mediated by Smad1, Smad4, and Smad5, which form a complex with this reporter construct. Strikingly, signaling of BMP7 to stimulate BRE4-luciferase is significantly enhanced when coincubated in the presence of TGFß. #P < 0.01 versus Co; *P < 0.01 versus BMP7. c and d: The presence of functionally active TGFß receptor complexes, particularly ALK5 (TGFß type 1 receptor), in PC-3M-Pro4 cells was demonstrated by the CAGA-luciferase reporter, whose activity depends on binding of activated Smad3/Smad4 transcription factor complexes. The CAGA boxes confer TGFß but not BMP stimulation to the promoter reporter constructs. The addition of BMP7 to TGFß-stimulated PC-3M-Pro4 cells dose-dependently inhibited TGFß-driven CAGA-luciferase activity. *P < 0.01 versus Co, **P < 0.01 versus TGFß alone (c); and *P < 0.01 versus TGFß alone, #P < 0.01 versus Co (d).

The presence of functionally active TGFß receptor complexes, particularly ALK5 (TGFß type 1 receptor), in PC-3M-Pro4 cells was demonstrated by the CAGA-luciferase reporter. The CAGA boxes confer TGFß but not BMP stimulation to the promoter-reporter constructs in PC-3M-Pro4 cells in a dose-dependent manner (Figure 5, c and d) . The addition of BMP7 to TGFß-stimulated PC-3M-Pro4 cells dose-dependently inhibited TGFß-driven CAGA-luciferase activity. This suggests that BMP7 counteracts TGFß-induced nuclear activation of activated Smad3/4 complexes (Figure 5, c and d) in human prostate cancer cells, whereas TGFß can stimulate BRE4-luciferase only in the presence of BMP7 (Figure 5b) .

The observed antitumor effects of BMP7 in vivo may therefore be caused by inhibition of TGFß-mediated growth-stimulatory responses in bone metastasis. This may be particularly true for the bone microenvironment, because bone is a storehouse for TGFß, and it has been shown that activated TGFß is released from bone matrix by bone-resorbing osteoclasts.29,31,32,42,43

The expression level of E-cadherin appears to be inversely related with prostate cancer grade, and its loss is considered a hallmark of EMT.23,44,45 Recombinant human BMP7 dose-dependently induced E-cadherin promoter activity (>150% at 1 µg/ml BMP7, P < 0.01), whereas TGFß had no effect (Figure 6a) . Strikingly, E-cadherin promoter activity was stimulated in a synergistic, dose-dependent, and significant manner by BMP7 in the presence of TGFß (Figure 6b , 350%). In line with this observation, Western blot analysis revealed that E-cadherin/vimentin ratio at the protein level was strongly up-regulated under these circumstances (Figure 6c) . These data therefore suggest that BMP7 can counteract the acquisition of an invasive metastatic phenotype by re-expressing the key epithelial marker E-cadherin, which is an absolute requirement for epithelial integrity, by decreasing cell motility, invasion, and migration.

Figure 6. The effects of BMP7 on E-cadherin promoter luciferase constructs and protein expression of E-cadherin/vimentin in human PC-3M-Pro4 prostate cancer cells. a and b: BMP7 induces a dose-dependent and significant increase in E-cadherin promoter-luciferase construct activity (P < 0.01 at >0.2 µg/ml BMP7). E-Cadherin promoter activity is significantly and synergistically enhanced when BMP7 and TGFß were coincubated (P < 0.05 versus BMP7 alone). *P 0.01 versus Co (a); *P < 0.05 versus Co, **P < 0.05 versus other conditions (b). c: E-cadherin and vimentin protein expression in PC-3M-Pro4 cells in the presence or absence of rhBMP7 and/or TGFß for 48 hours as determined by immunoblotting. Treatment of PC-3M-Pro4 cells with 0.5 µg/ml rhBMP7 in the presence of TGFß induces protein expression of E-cadherin. Values are expressed as E-cadherin/Vimentin ratio. d: Vimentin monoclonal antibody immunoreactivity of cultured PC-3M-Pro4 cells. In the presence of BMP7, the TGFß-induced vimentin expression and cellular redistribution was completely abolished. Values are expressed as means ?? SD (n = 4, analysis of variance).

When TGFß was added to PC-3M-Pro4/luc+ cells the intermediate filament and mesenchymal marker vimentin was redistributed in the prostate cancer cells (Figure 6d) . This TGFß-induced vimentin expression and cellular redistribution were completely abolished by BMP7 (Figure 6d) .

Discussion

In this report, novel evidence is provided for the importance of BMP7 (and TGFß) in the development of an epithelial cancer and metastatic behavior with prostate cancer as a model. Our clinical findings suggest that decreased BMP7 expression is implicated in human prostate cancer, and we postulate a role for BMP7 in controlling epithelial homeostasis of the prostate gland. In line with these clinical findings, experimental treatment of bone metastases from human prostate cancer in vivo positions BMP7 as a good candidate for the treatment of skeletal metastasis from prostate cancer.

Members of the TGFß superfamily play crucial roles in embryonic development, certain fibrotic diseases, inflammation, and cancer. Although TGFß inhibits the proliferation of normal prostate cells and functions as a tumor suppressor in early tumorigenesis, it acts as a tumor promoter in later stages of tumor progression and mediates oncogenic EMT.2,5,10,17,46-50 Prostate cancers often grow in a variety of growth patterns that are classified by Gleason histological grading. Prostate cancer classification according to this "gold standard" is predictive of disease status and also seems to parallel EMT. Laser capture microdissection of prostate epithelium/cancer was used because the epithelial compartment in the normal prostate represents only a small percentage (5%) of the entire organ. Furthermore, infiltrative prostate cancer patterns vary among individual cases. The latter can lead to misinterpretation of epithelium-specific gene expression among the various clinical specimens. By comparing pure populations of laser-captured cancer cells and noncancerous epithelial cells from the same patient, we observed significant and consistent BMP7 mRNA underexpression in invasive primary prostate carcinoma. In contrast, BMP7 mRNA expression in bulk tissue specimens of noncancerous prostate and prostate cancer tissue showed only a tendency toward underexpression in prostate cancer. Our observations support concerns that high-throughput RNA expression analysis performed on bulk tissue samples, either on primary tumors or its metastases, may not detect differences in relevant gene expression restricted to the epithelial or mesenchymal compartment.22 Clearly, LCM is preferred to establish putative changes in gene expression repertories in both epithelial and stromal compartments.

Emerging evidence from a variety of tumors suggests that the effects of BMPs (and TGFß) are cell-specific and could be either protumorigenic or antitumorigenic.48 For instance, BMP7 was shown to counteract EMT by inducing MET, to induce EMT, or to have no effect.2,4,10,49,51 In the noncancerous human prostate, expression of BMP7 protein was restricted mainly to the glandular epithelium, whereas primary prostate cancers show low or undetectable BMP7 protein expression. It also appeared that BMP7 expression is negatively correlated to Gleason score (data not shown), but the number of patients per Gleason pattern were too small for statistical analysis. Clearly more LCM and real-time PCR analyses are warranted to address this issue. In line with these clinical observations, the level of BMP7 expression in tested prostate cancer cell lines correlated with an epithelial phenotype and seems, therefore, inversely related to their tumorigenic and metastatic potential in vivo.

Daily systemic administration of recombinant BMP7 did not affect the growth of orthotopically implanted PC-3M-Pro4/luc+ prostate cancer cells and subsequent formation of lymph node metastasis in nude mice. However, the growth of micrometastatic deposits from human PC-3M-Pro4/luc+ prostate cancer cells in bone marrow was inhibited significantly, whereas the process of bone marrow colonization per se appeared unaffected. Our findings suggest, therefore, that the tumor microenvironment is an important determinant of the therapeutic response to BMP7.

Our in vitro studies reveal for the first time that BMP7 is a potent inhibitor of TGFß-induced EMT in PC-3M-Pro4 prostate cancer cells. In these cells, BMP7 counteracts TGFß-induced activation of activated Smad3/4 complexes. BMP7 induced a dose-dependent expression of E-cadherin that led to an overall increase in E-cadherin/vimentin ratio as an established indicator of less malignant and more epithelial phenotype. Surprisingly, we observed that these BMP7-mediated effects were synergistically enhanced by coincubation with TGFß at both transcriptional and protein levels. These actions of BMP7 may be of critical importance for explaining the successful experimental treatment of skeletal metastasis and intrabone growth. In this context, it is important to note that TGFß, which is highly concentrated as inactive form in bone matrix, can be released and activated by osteoclastic resorption. TGFß may act as paracrine growth factors for neighboring cancer cells that may have colonized the bone marrow.31,32,42,43 Compelling evidence suggests that the formation of micrometastatic deposits in bone marrow and subsequent development into clinically overt (macro)metastasis are critically dependent on the tumor-bone (marrow) interactions, in particular bone remodeling.31,43,52 Although little is known regarding the impact of bone (marrow) stromal cells in prostate cancer bone metastasis, it can be argued that the production and release of specific growth factors, colony stimulating factors, and cytokines like Wnts, which are critically important for the maintenance of the hematopoietic stem cell niche (and which regulate hematopoiesis), may also be essential for colonization, survival, and growth of cancer cells. Interestingly, TGFß/BMPs and other pathways like wingless/Wnt, Notch, and Hedgehog seem not only to be involved in bone development/homeostasis and hematopoiesis but also in prostate cancer initiation and progression.2,8,53-56 Our in vitro and in vivo data further support the notion that BMP7 can antagonize TGFß signaling routes in human prostate cancer cells that are metastatic to the skeleton.

Epithelial-to-mesenchymal transdifferentiation provides a mechanism for prostate epithelial cells to overcome physical constraints imposed on them by intercellular junctions and to adopt a motile phenotype. We hypothesize that the activation process of micrometastases in bone marrow may bear similarities to EMT that occurs at the primary site in various epithelial cancers and during ontogeny. TGFß/BMPs and their signaling molecules have been implicated in the cellular plasticity that occurs during organogenesis, tissue repair/remodeling, and carcinogenesis, where cells can shift between epithelial and mesenchymal phenotypes.1,2 EMT/MET can thus be viewed as a prime example of such cell plasticity. Considering the importance of the TGFß superfamily in bone metastasis,42 our findings suggest that the efficacy of BMP7 in the experimental treatment of bone metastases is interfering with recapitulation of such cell plasticity of bone marrow micrometastatic cancer cells, particularly the acquisition of an invasive phenotype by (perhaps TGFß-driven) EMT. Clinical studies also seem to underline the cellular plasticity of prostate cancer during dissemination and bone metastasis.57,58 The transition from a well-differentiated epithelial phenotype to an invasive mesenchymal phenotype may, of course, involve molecular mechanisms other than those described here.59 Other studies have provided evidence that motility and invasiveness can be enhanced without inducing a complete conversion of cellular identity.59 After migrating to new organ or tissue territories, metastatic prostate cancer cells can regain epithelial morphology by a phenomenon known as MET and re-establish E-cadherin expression and epithelial junctions.57-59 The observed reversion to an epithelial morphology in prostate cancer seems possible even at bone metastatic sites57-59 and is in full agreement with our in vitro and in vivo data presented here. Furthermore, previous studies demonstrated cellular plasticity of human prostate cancer cells (PC3 cells), particularly the EMT/MET interconversion induced by different culture conditions (monolayer versus three-dimensional cultures60 ).

In many prostate cancer patients, micrometastatic deposits may already exist in bone marrow after removal of the primary tumor that cannot be identified at the time of diagnosis (minimal residual disease). Whole-body bioluminescent reporter imaging of luciferase-expressing human prostate cancer cells revealed that growth of micrometastatic deposits in bone marrow is significantly inhibited by systemic administration of BMP7. Similar observations were made by us for other epithelial cancers, including breast cancer.50,61 Interestingly, recent data suggest that BMPs can inhibit the tumorigenic potential of human brain tumor-initiating cells, mediated via a significant reduction in the stem cell-like, tumor-initiating precursors.62 Accumulating evidence suggests that cancer stem cells with tumor-initiating potential exist in human prostate cancer.63 Although speculative at present, the observed therapeutic effects of BMP7 described here may be mediated, at least in part, by prostate cancer cells with tumor-initiating (and metastatic) potential. Studies are currently ongoing to address these important issues.

In conclusion, our data suggest that BMP7 controls the epithelial homeostasis in the human prostate gland by preserving the epithelial phenotype. Loss of BMP7 expression during prostate cancer progression could stimulate the TGFß-stimulated EMT and thus contribute to the acquisition of an invasive phenotype. However, exogenous BMP7 can still counteract the EMT process in the metastatic tumor, thus underscoring tumor plasticity, but its therapeutic response may be determined by the microenvironment. Therefore, BMP7 may represent a novel therapeutic molecule for repression of systemic prostate cancer progression.

【参考文献】
  Thiery JP: Epithelial-mesenchymal transitions in development and pathologies. Curr Opin Cell Biol 2003, 15:740-746

Prindull G: Hypothesis: cell plasticity, linking embryonal stem cells to adult stem cell reservoirs and metastatic cancer cells? Exp Hematol 2005, 33:738-746

Huber MA, Kraut N, Beug H: Molecular mechanisms for epithelial-mesenchymal transition during tumor progression. Curr Opin Cell Biol 2005, 17:548-558

Piek E, Moustakas A, Kurisaki A, Heldin CH, ten Dijke P: TGF-(ß) type I receptor/ALK-5 and Smad proteins mediate epithelial to mesenchymal transdifferentiation in NMuMG breast epithelial cells. J Cell Sci 1999, 112:4557-4568

Zavadil J, Bottinger EP: TGF-ß and epithelial-to-mesenchymal transitions. Oncogene 2005, 24:5764-5774

Kalluri R, Neilson EG: Epithelial-mesenchymal transition and its implications for fibrosis. J Clin Invest 2003, 112:1776-1784

Böttinger EP, Bitzer M: TGF-ß signaling in renal disease. J Am Soc Nephrol 2002, 13:2600-2610

Zavadil J, Cermak L, Soto-Nieves N, Bottinger EP: Integration of TGF-beta/Smad and Jagged1/Notch signalling in epithelial-to-mesenchymal transition. EMBO J 2004, 23:1155-1165

Zeisberg M, Kalluri R: The role of epithelial-to-mesenchymal transition in renal fibrosis. J Mol Med 2004, 82:175-181

Valcourt U, Kowanetz M, Niimi H, Heldin CH, Moustakas A: TGF-ß and the Smad signaling pathway support transcriptomic reprogramming during epithelial-mesenchymal cell transition. Mol Biol Cell 2005, 16:1987-2002

Vukicevic S, Latin V, Chen P, Batorsky R, Reddi AH, Sampath TK: Localization of osteogenic protein-1 (bone morphogenetic protein-7) during human embryonic development: high affinity binding to basement membranes. Biochem Biophys Res Commun 1994, 198:693-700

Jena N, Martin-Seisdedos C, McCue P, Croce CM: BMP7 null mutation in mice: developmental defects in skeleton, kidney, and eye. Exp Cell Res 1997, 230:28-37

Zeisberg M, Hanai J, Sugimoto H, Mammoto T, Charytan D, Strutz F, Kalluri R: BMP-7 counteracts TGF-ß1-induced epithelial-to-mesenchymal transition and reverses chronic renal injury. Nature Med 2003, 9:964-968

Zeisberg M, Shah AA, Kalluri R: Bone morphogenic protein-7 induces mesenchymal to epithelial transition in adult renal fibroblasts and facilitates regeneration of injured kidney. J Biol Chem 2005, 280:8094-8100

Dunsmuir WD, Gillett CE, Meyer LC, Young MP, Corbishley C, Eeles RA, Kirby RS: Molecular markers for predicting prostate cancer stage and survival. BJU Int 2000, 86:869-878

Richmond PJ, Karayiannakis AJ, Nagafuchi A, Kaisary AV, Pignatelli M: Aberrant E-cadherin and -catenin expression in prostate cancer: correlation with patient survival. Cancer Res 1997, 57:3189-3193

Bello-DeOcampo D, Tindall DJ: TGF-ßl/Smad signaling in prostate cancer. Curr Drug Targets 2003, 4:197-207

Horvath LG, Henshall SM, Kench JG, Turner JJ, Golovsky D, Brenner PC, O??Neill GF, Kooner R, Stricker PD, Grygiel JJ, Sutherland RL: Loss of BMP2, Smad8, and Smad4 expression in prostate cancer progression. Prostate 2004, 59:234-342

Wikström P, Stattin P, Franck-Lissbrant I, Damber JE, Bergh A: Transforming growth factor ß1 is associated with angiogenesis, metastasis, and poor clinical outcome in prostate cancer. Prostate 1998, 37:19-29

Dumont N, Arteaga CL: Targeting the TGF beta signaling network in human neoplasia. Cancer Cell 2003, 3:531-536

Yingling JM, Blanchard KL, Sawyer JS: Development of TGF-ß signalling inhibitors for cancer therapy. Nat Rev Drug Discov 2004, 3:1011-1022

Rentsch CA, Cecchini MG, Schwaninger R, Germann M, Markwalder R, Heller M, van der Pluijm G, Thalmann GN, Wetterwald A: Differential expression of TGFß-stimulated clone 22 in normal prostate and prostate cancer. Int J Cancer 2006, 118:899-906

Rubin MA, Mucci NR, Figurski J, Fecko A, Pienta KJ, Day ML: E-cadherin expression in prostate cancer: a broad survey using high-density tissue microarray technology. Hum Pathol 2001, 32:690-697

Schwaninger R, Rentsch CA, Wetterwald A, van der Horst G, van Bezooijen RL, van der Pluijm G, Lowik CW, Ackermann K, Pyerin W, Hamdy FC, Thalmann GN, Cecchini MG: Lack of noggin expression by cancer cells is a determinant of the osteoblast response in bone metastases. Am J Pathol 2007, 170:160-175

Sobel RE, Sadar MD: Cell lines in prostate cancer research: a compendium of old and new linesCpart I. J Urol 2005, 173:342-359

Pettaway CA, Pathak S, Greene G, Ramirez E, Wilson MR, Killion JJ, Fidler IJ: Selection of highly metastatic variants of different human prostate carcinomas using orthotopic implantation in nude mice. Clin Cancer Res 1996, 2:1627-1636

Wu TT, Sikes RA, Cui Q, Thalmann GN, Kao C, Murphy CF, Yang H, Zhau HE, Balian G, Chung LW: Establishing human prostate cancer cell xenografts in bone: induction of osteoblastic reaction by prostate-specific antigen-producing tumors in athymic and SCID/bg mice using LNCaP and lineage-derived metastatic sublines. Int J Cancer 1998, 77:887-894

Thalmann GN, Anezinis PE, Chang SM, Zhau HE, Kim EE, Hopwood VL, Pathak S, von Eschenbach AC, Chung LW: Androgen-independent cancer progression and bone metastasis in LNCaP model of human prostate cancer. Cancer Res 1994, 54:2577-2581

Wetterwald A, van der Pluijm G, Que I, Sijmons B, Buijs J, Karperien M, Lowik CW, Gautschi E, Thalmann GN, Cecchini MG: Optical imaging of cancer metastasis to bone marrow: a mouse model of minimal residual disease. Am J Pathol 2002, 160:1143-1153

Papapoulos SE, Hamdy NAT, van der Pluijm G: Bisphosphonates in the management of prostate carcinoma metastatic to the skeleton. Cancer 2000, 88:3047-3053

van der Pluijm G, Que I, Sijmons B, Buijs JT, Lowik CWGM, Wetterwald A, Thalmann GN, Papapoulos SE, Cecchini MG: Interference with the microenvironmental support impairs the de novo formation of bone metastases in vivo. Cancer Res 2005, 65:7682-7690

van der Pluijm G, Sijmons B, Vloedgraven H, Deckers M, Papapoulos S, Lowik C: Monitoring metastatic behavior of human tumor cells in mice with species-specific polymerase chain reaction: elevated expression of angiogenesis and bone resorption stimulators by breast cancer in bone metastases. J Bone Miner Res 2001, 16:1077-1091

Jones WK, Richmond EA, White K, Sasak H, Kusmik W, Smart J, Oppermann H, Rueger DC, Tucker RF: Osteogenic protein-1 (OP-1) expression and processing in Chinese hamster ovary cells: isolation of a soluble complex containing the mature and pro-domains of OP-1. Growth Factors 1994, 11:215-225

Persson U, Izumi H, Souchelnytskyi S, Itoh S, Grimsby S, Engstrom U, Heldin CH, Funa K, ten Dijke P: The L45 loop in type I receptors for TGF-ß family members is a critical determinant in specifying Smad isoform activation. FEBS Lett 1998, 434:83-87

Rosendahl A, Pardali E, Speletas M, ten Dijke P, Heldin CH, Sideras P: Activation of bone morphogenetic protein/Smad signaling in bronchial epithelial cells during airway inflammation. Am J Respir Cell Mol Biol 2002, 27:160-169

Goumans MJ, Valdimarsdottir G, Itoh S, Rosendahl A, Sideras P, ten Dijke P: Balancing the activation state of the endothelium via two distinct TGF-ß type I receptors. EMBO J 2002, 21:1743-1753

Dennler S, Itoh S, Vivien D, ten Dijke P, Huet S, Gauthier JM: Direct binding of Smad3 and Smad4 to critical TGFß-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene. EMBO J 1998, 17:3091-100

Korchynskyi O, ten Dijke P: Identification and functional characterization of distinct critically important bone morphogenetic protein-specific response elements in the Id1 promoter. J Biol Chem 2002, 277:4883-4891

Jonk LJ, Itoh S, Heldin CH, ten Dijke P, Kruijer W: Identification and functional characterization of a Smad binding element (SBE) in the JunB promoter that acts as a transforming growth factor-ß, activin, and bone morphogenetic protein-inducible enhancer. J Biol Chem 1998, 273:21145-21152

Batlle E, Sancho E, Franci C, Dominguez D, Monfar M, Baulida J, Garcia De Herreros A: The transcription factor snail is a repressor of E-cadherin gene expression in epithelial tumour cells. Nature Cell Biol 2000, 2:84-89

Lang SH, Hyde C, Reid IN, Hitchcock IS, Hart CA, Bryden AA, Villette JM, Stower MJ, Maitland NJ: Enhanced expression of vimentin in motile prostate cell lines and in poorly differentiated and metastatic prostate carcinoma. Prostate 2002, 52:253-263

Yin JJ, Selander K, Chirgwin JM, Dallas M, Grubbs BG, Wieser R, Massague J, Mundy GR, Guise TA: TGF-ß signaling blockade inhibits PTHrP secretion by breast cancer cells and bone metastases development. J Clin Invest 1999, 103:197-206

Mundy GR: Metastasis to bone: causes, consequences and therapeutic opportunities. Nature Rev Cancer 2002, 2:584-593

Umbas R, Isaacs WB, Bringuier PP, Schaafsma HE, Karthaus HF, Oosterhof GO, Debruyne FM, Schalken JA: Decreased E-cadherin expression is associated with poor prognosis in patients with prostate cancer. Cancer Res 1994, 54:3929-3933

Tomita K, van Bokhoven A, van Leenders GJ, Ruijter ET, Jansen CF, Bussemakers MJ, Schalken JA: Cadherin switching in human prostate cancer progression. Cancer Res 2000, 60:3650-36544

Tang B, de Castro K, Barnes HE, Parks WT, Stewart L, Bottinger EP, Danielpour D, Wakefield LM: Loss of responsiveness to transforming growth factor ß induces malignant transformation of nontumorigenic rat prostate epithelial cells. Cancer Res 1999, 59:4834-4842

Wakefield LM, Roberts AB: TGF-ß signaling: positive and negative effects on tumorigenesis. Curr Opin Genet Dev 2002, 12:22-29

Tu WH, Thomas TZ, Masumori N, Bhowmick NA, Gorska AE, Shyr Y, Kasper S, Case T, Roberts RL, Shappell SB, Moses HL, Matusik RJ: The loss of TGF-beta signaling promotes prostate cancer metastasis. Neoplasia 2003, 5:267-277

Yang S, Zhong C, Frenkel B, Reddi AH, Roy-Burman P: Diverse biological effect and Smad signaling of bone morphogenetic protein 7 in prostate tumor cells. Cancer Res 2005, 65:5769-5777

Buijs JT, Henriquez NV, van Overveld PGM, van der Horst G, Que I, Schwaninger R, Rentsch C, ten Dijke P, Cleton-Jansen A-M, Driouch K, Lidereau R, Vukicevic S, Cl?zardin P, Papapoulos SE, Cecchini MG, Löwik CWGM, van der Pluijm G: Bone morphogenetic protein 7 in the development and treatment of bone metastases from breast cancer. Cancer Res, in press

Danielpour D: Functions and regulation of transforming growth factor-beta (TGF-beta) in the prostate. Eur J Cancer 2005, 41:846-857

Schneider A, Kalikin LM, Mattos AC, Keller ET, Allen MJ, Pienta KJ, McCauley LK: Bone turnover mediates preferential localization of prostate cancer in the skeleton. Endocrinology 2005, 146:1727-1736

Chen G, Shukeir N, Potti A, Sircar K, Aprikian A, Goltzman D, Rabbani SA: Up-regulation of Wnt-1 and beta-catenin production in patients with advanced metastatic prostate carcinoma: potential pathogenetic and prognostic implications. Cancer 2004, 101:1345-1356

Huang HC, Klein PS: Interactions between BMP and Wnt signaling pathways in mammalian cancers. Cancer Biol Ther 2004, 3:676-678

Grego-Bessa J, Diez J, Timmerman L, de la Pompa JL: Notch and epithelial-mesenchyme transition in development and tumor progression: another turn of the screw. Cell Cycle 2004, 3:718-721

Karhadkar SS, Bova GS, Abdallah N, Dhara S, Gardner D, Maitra A, Isaacs JT, Berman DM, Beachy PA: Hedgehog signalling in prostate regeneration, neoplasia and metastasis. Nature 2004, 431:707-712

Bryden AA, Freemont AJ, Clarke NW, George NJ: Paradoxical expression of E-cadherin in prostatic bone metastases. BJU Int 1999, 84:1032-1034

Bryden AA, Hoyland JA, Freemont AJ, Clarke NW, Schembri Wismayer D, George NJ: E-cadherin and beta-catenin are down-regulated in prostatic bone metastases. BJU Int 2002, 89:400-403

Christiansen JJ, Rajasekaran AK: Reassessing epithelial to mesenchymal transition as a prerequisite for carcinoma invasion and metastasis. Cancer Res 2006, 66:8319-8326

Lang SH, Sharrard RM, Stark M, Villette JM, Maitland NJ: Prostate epithelial cell lines form spheroids with evidence of glandular differentiation in three-dimensional Matrigel cultures. Br J Cancer 2001, 85:590-599

Notting IC, Buijs JT, Mintardjo R, van der Horst G, Vukicevic S, Löwik CWGM, Schalij-Delfos N, Keunen J, van der Pluijm G: Bone morphogenetic protein 7 inhibits tumor growth of uveal melanoma. IOVS, in press

Piccirillo SGM, Reynolds BA, Zanetti N, Lamorte G, Binda E, Broggi G, Brem H, Olivi A, Dimeco A, Vescovi AL: Bone morphogenetic proteins inhibit the tumorigenic potential of brain tumour-initiating cells. Nature 2006, 444:761-765

Collins AT, Berry PA, Hyde C, Stower MJ, Maitland NJ: Prospective identification of tumorigenic prostate cancer stem cells. Cancer Res 2005, 65:10946-10951

作者单位：From the Departments of Urology,* Endocrinology, and Molecular Cell Biology, Leiden University Medical Center, Leiden, The Netherlands; the Department of Clinical Research and Department of Urology, University of Bern, Inselspital, and the Institute of Pathology,|| University of Bern, Bern, Switzerl