This article is mentioned in:

Introduction

The study of prostate carcinogenesis and tumor progression is difficult due to the lack of appropriate in vitro and in vivo models. Emerging evidence has shown that the tumor microenvironment plays a crucial role in prostate cancer (PCa) development/progression (1–4). Specifically, it has been observed that in PCa three disparate cellular outcomes predominate: i) the tumor remains well differentiated and clinically indolent; in this case the local stromal cells may act to restrain the growth of the cancer; ii) early in its genesis the tumor acquires a highly malignant phenotype, growing rapidly and displacing the original stromal population (often referred to as small-cell PCa); these less common aggressive tumors are relatively independent of the local microenvironment; and iii) the tumor co-opts the local stroma, taking on a classic stromagenic phenotype where interactions with the local microenvironment are critical to cancer growth. Therefore, tumor stromal cells can support tumor epithelial cell growth through the secretion of paracrine growth factors and/or providing a more favorable microenvironment for tumor growth. The stroma can elicit instructive, permissive, or inductive (reactive) effects on the parenchymal epithelium. For example, embryonic mesenchymal cells can ‘instruct’ epithelial cells to form functional, differentiated glands. By contrast, a ‘permissive’ stroma supports a previously-induced epithelial phenotype (5–8).

Molecular markers of stromal cell subpopulations have been poorly defined. Generally, fibroblasts are identified by their spindle-shaped morphology and the overlapping expression of various ‘indicators’ such as vimentin. Stroma cells are androgen receptor (AR)-negative. However, some cells retain the expression of AR (9,10). Stromal AR regulates epithelial proliferation, extracellular matrix (ECM) remodelling, neovasculature formation and immune cell infiltration through the secretion of pro-inflammatory cytokines/chemokines, whereas the loss of stromal AR leads to suppressed prostate tumorigenesis (9,11). Transforming growth factor-β (TGF-β) is a growth factor abundantly present in the stroma. This can induce or suppress differentiation and tumorigenesis in a dose- and context-dependant manner (4). Loss of the TGF-β type II receptor (TGFβR2) has been observed in the stroma of >60% of human PCa patients (12). Neurotrophin growth factor [nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3) and neurotrophin-4/5 (NT-4/5)] binding activates tyrosine kinase receptors encoded by the Trk family (Trk A, Trk B and Trk C), and the low-affinity receptor p75NTR. In the normal prostate, p75NTR and Trk A are found on benign prostate epithelial cells, whereas NGF-β is expressed by adjacent stromal cells (13–16). NGF-β stimulates the growth of prostate epithelial cells, which suggests the paracrine regulation of prostate cell growth by NGF-β or other neurotrophins. In contrast to benign prostate epithelial cells, p75NTR expression is decreased in cancer cells (13,14). A dramatic increase in p75 immunoreactivity has been observed in the stroma and this correlates with an increase in malignancy (16).

To better clarify the cellular interactions mediated by the reactive stroma, we characterized the stromal cell population present in benign prostatic hyperplasia (BPH) and PCa using a previously described method (17,18). To this end, we evaluated the expression of vimentin (in fibroblasts), vimentin and α-smooth muscle actin [α-SMA; in myofibroblasts (MFs)] or desmin (in muscle), as well as the expression of p75NTR, matrix metalloproteinase (MMP)-2 and MMP-9 and their tissue inhibitors (TIMPs). Stromal cells are a mixture of fibroblasts and MFs (19,20). The number of fibroblasts was higher in BPH- compared to PCa-derived cultures, whereas the number of MFs was higher in PCa- compared to BPH-derived cultures. The percentage MFs was higher in high-grade PCa-derived cultures compared to moderate- and low-grade PCa-derived cultures. p75NTR expression was also elevated in stromal cell cultures derived from PCa compared to those derived from BPH and in cultures derived from cases with Gleason scores ≥7 compared to those derived from cases with Gleason scores <7, as well as in cultures with a high concentration of MFs (vimentin/α-SMA/desmin-positive) compared to cultures with a high concentration of fibroblasts. The present study confirms a novel, malignant-dependent localization of p75 in smooth muscle cells SMCs/MFs in stromal cell cultures derived from human PCa. Therefore, p75 re-expression in stromal SMCs/MFs is a mechanism related to the general dedifferentiation of the stroma connected to neoplastic invasion.

Materials and methods

Reagents

All materials for tissue culture were purchased from HyClone (Cramlington, NE, USA). Plasticware was obtained from Nunc (Roskilde, Denmark). Antibodies, when not defined otherwise, were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Elisa kits for TGF-β1 and stromal cell-derived factor α (SDF1α) were purchased from RayBiotech (European distributor, Hölzel Diagnostika GmbH, Köln, Germany).

Stromal cell cultures from prostatic specimens

Tissue samples were collected by radical prostatectomy from patients attending the Urology Division of San Salvatore Hopital, L’Aquila, Italy. Our Institutional review board approved the protocol and a written informed consent was obtained from all human subjects. The tissue samples were enzymatically digested with collagenase type I (225 U/ml; Sigma, St. Louis, MA, USA) and hyaluronidase (125 U/ml; Sigma) in DMEM (with 10% FCS) at 37°C overnight. The supernatant containing the stromal cells was centrifuged at 250 × g for 5 min. The pellet was resuspended and plated in complete DMEM. After allowing the cells to grow to confluence, they were briefly trypsinized to release the stromal cells, which were then removed and expanded in DMEM with 5% FCS. After five passages, a homogeneous stromal cell population was established. Microscopy and immunocytochemistry were used to determine the purity of the stromal cell fraction. Epithelial cell exclusion assay was performed using a panel of different antibodies purchased from Sigma, when not specified. In particular, we used an anti-PAN cytokeratin antibody to evaluate the epithelial presence, anti-K18, clone CY90, anti-K8 and clone M20 to evaluate luminar cells and anti-K14, clone CKB1 and anti-K5 (N20) polyclonal antibody to recognize basal cells. Anti-desmin polyclonal antibody was used to determine smooth muscle cell presence. Anti α-SMA, clone 1A4, anti-vimentin and clone V9, were used to determine the desmin-negative MF and fibroblast cell populations. Cells were counted and expressed as a percentage of vimentin-positive cells per microscopic field at ×200 magnification. Five separate fields were considered and the percentage of positive cells was expressed as the mean ± SE.

Western blot analysis

Cells were grown to confluence in 100-mm dishes. The medium was then removed and the cells were washed with cold PBS and immediately lysed with 1 ml of lysis buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 30 mM p-nitrophenyl phosphate, 10 mM sodium pyrophosphate, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin and 10 μg/ml leupeptin). Lysates were centrifuged at 10,000 × g for 10 min. Equal amounts of protein (50 μg) were resolved using 7.5% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis, separated proteins were blotted onto a nitrocellulose membrane and analyzed by western blot analysis with the ECL method. Proteins were transferred from the gel onto a nitrocellulose filter paper. The membrane was washed with saline solution (Tris-HCl, NaCl and BSA at 1%) for 1 h and then incubated with different antibodies diluted at 1 μg/ml. After incubation the membrane was washed with Tris-NaCl plus 0.1% Tween-20 and incubated for 1 h with a second antibody conjugated with peroxidase (1:5,000). Protein bands were quantified by densitometry. Immunoreactive bands were visualized using enhanced chemiluminescence detection kit reagents (Amersham, Milan, Italy).

Gelatin zymography

The zymography assay (21) used gelatin as a substrate for MMP-2 and MMP-9. Gelatin at a concentration of 0.1% was incorporated into a 10% polyacrylamide gel containing 0.4% SDS. Electrophoresis under non-reducing conditions was performed using the Bio-Rad mini-gel system at 125 V for 90–120 min. For plasminogen activator analysis, SDS-polyacrilamide gels were polymerized with 0.1 mg/ml of lactose-free casein and 15 μg/ml of human plasminogen as previously described (22). After electrophoresis the gels were washed twice for 30 min in 2.5% Triton X-100 (v/v) to remove the SDS and then incubated overnight in the developing buffer [50 mM Tris-HCl (pH 7.6), 200 mM NaCl, 5 mM CaCl2, 0.2% (v/v) Brij-35] at 37°C. Digestion bands were quantified by the Java image processing program ‘ImageJ’ from the scanned gels. Reverse zymography for TIMPs was performed in concentrated supernatants containing 20 μg of proteins [conditioned medium concentrated 20-fold by Centricon (Amicon, Beverly, MA, USA) with 10 kDa molecular weight cut-off] analyzed by 14% SDS-PAGE containing 0.1 casein and 30% (v/v) serum-free supernatants from the MDA-MB-231 human breast cancer cell line. This cell line expresses a single 53 kDa band of EDTA-inhibitable caseinolytic activity in zymograms (MMP-3). Gels were processed as described above except for a 72-h incubation in collagenase buffer. Dark stained bands of 28-25 and 21 kDa molecular weight were identified as inhibitor bands of TIMP1/TIMP3 and TIMP2, respectively.

Statistics

The data are expressed as the mean and standard deviation (SD) and were compared using an unpaired Student’s t-test. Categorical data were analyzed by the exact Fisher’s test. A p-value <0.05 was considered to indicate a statistically significant difference.

Results

We characterized a total of 45 primary cultures derived from clinical samples, which, after the completion of the clinical work-up for each patient, were grouped according to the pathological diagnosis formulated through the microscopic analysis of the biopsies obtained in parallel with cultured tissue samples. We studied in primary culture 15 cases of BPH and 30 cases of PCa of various pathological grades.

Stromal cell characterization

Primary cell culture was characterized at each passage and a gradual decrease in the number of rounded cells with epithelial appearance and an increase in the number of elongated cells with fibroblastic aspects was observed, as demonstrated by the phase contrast micrographs (Fig. 1). In previous studies, it has been demonstrated that prostatic stromal cells are characterized by an increase in the expression of cytoskeleton proteins, such as vimentin and α-SMA, followed by a decrease in muscle markers, such as desmin (23,24). After five passages in our culture system, a cell population positively expressing vimentin and α-SMA, but negatively expressing cytokeratin 18 and weakly expressing desmin was obtained. These results indicate that our human pluripotent stem cell populations represent a combination of fibroblasts (vimentin-positive cells) and MFs (α-SMA-positive cells). Fig. 2 demonstrates the expression of cytoskeleton markers (α-SMA, vimentin and desmin) by western blot analysis. Fig. 3 demonstrates the expression of stromal markers, thymidine phosphorylase (TP), as well as that of the prostate markers, AR and prostate-specific antigen (PSA), or caveolin-1 and tenascin C as shown by western blot analysis in different cultures derived from BPH and PCa samples. The amount of these markers was quantified densitometrically as arbitrary densitometric units (ADUs) after normalization with β-actin. Vimentin expression was not statistically different in the BPH- compared to the PCa-derived cultures, both as the percentage of positve cells and adjusted densitometric units. Although the percentage of α-SMA-positive cells (Fig. 2B, C and G) was not statistically significant, the protein levels were significantly higher in the PCa - compared to the BPH-derived cultures (p<0.001, Fig. 2A, D and H). Similar results were obtained with desmin; although the percentage of desmin-positive cells (Fig. 2B, C and G) was not statistically significant, the protein levels were significantly higher in PCa- compared to BPH-derived cultures. Desmin expression was weakly and mainly associated with PCa stromal cultures (p<0.001, Fig. 2A, D and H). Although the number of samples was limited, we aimed to determine the correlation between the stromal cell population and histological grade (Gleason grade) in our samples. No statistically significant differences were observed for vimentin expression [2.1±1.0 (G2–4) ADU vs. 2.7±1.5 (G5–7) ADU and 4.3±1.8 (G8–10) ADU] and for α-SMA expression [4.7±1.2 (G2–4) ADU vs. 4.9±1.7 (G5–7) ADU and 5.7±2.1 (G8–10) ADU]. Desmin expression was significantly lower in cultures derived from cases with a low Gleason grade compared to those derived from cases with a high Gleason grade [0.4±0.2 (G2–4) ADU vs. 1.3±0.4 (G5–7) ADU and 2.4±0.6 in (G8–10), p<0.01]. The expression of K14 was observed in some stromal primary cultures and was restricted to PCa- compared to BPH-derived cultures (immuno-reactive bands for K14 were found in 1/15 BPH and 12/30 PCa cases; not significant) and in high-grade compared to moderate- and low-grade samples [1/7 (14.3%, G2–4), 5/15 (33.3%, G5–7) and 6/8 (75%, G8–10), p=0.043].

Figure 1 Phase contrast photomicrograph of stromal cell cutlures. (A–F) ×200 images. (A–C) Visible fibroblastic-like cells at different densities of growth ; (D–F) elongated cells with a final tendency to form fused cells (arrows, D and F) with several bright cytoplasmic organelles (arrows, E). The cells resemble myoepithelial/myofibroblast cells.

Figure 2 Western blot analyses of stromal cell markers, vimentin, desmin and α-muscle actin (α-SMA), in primary culture cell extracts. (A) Representative BPH-derived 10 cultures. (B) Percentage of vimentin (VMTN)-, desmin- and α-SMA-positive cells in BPH-derived 10 cultures analyzed by western blot analysis. (C) Mean percentage of VMTN, α-SMA and desmin in the total 15 cases of BPH. (D) Adjusted densitometric units obtained after normalization to actin levels of immunoreactive bands in western blot analysis in the 15 BPH-derived cultures. (E) Representative PCa-derived 12 cultures. (F) Percentage of VMTN-, desmin- and α-SMA-positive cells in BPH-derived 12 cultures analyzed by western blot analysis. (G) Adjusted densitometric units obtained after normalization to actin levels of immunoreactive bands in western blot analysis in the 30 PCa-derived cultures. (H) Mean percentage of VMTN, α-SMA and desmin in the 30 cases of PCa. Each lane was loaded with 40 μg of total protein.

Figure 3 Western blot analyses of thymidine phosphorylase (TP), androgen receptor (AR), prostate-specific antigen (PSA), caveolin-1, tenascin C and p75NTR in primary culture cell extracts harvested from (A) BPH- and (B) PCa-derived cultures. Each lane was loaded with 40 μg of total protein.

The levels of tenascin C verified by western blot analysis using the Santa Cruz Biotechnology (Scbt) antibody, N19, were significantly higher in cultures derived from PCa samples [1.2±0.9 ADU (mean ± SD) and 4.8±0.8 for BPH- and Ca-derived cultures, respectively, p<0.001] and in high-grade compared to moderate- and low-grade samples [3.2±0.4 (G2–4), 3.9±0.1 (G5–7) and 6.0±1 (G8–10), p<0.001].

The AR (Scbt N20) is expressed in a subset of prostatic stromal cells and functional stromal cell AR is required for normal prostate development and influences the growth of prostate tumors. Fibroblasts are negative for AR, whereas MFs express AR. We observed that AR expression was present only in some PCa-derived cultures (17/30, 56.7%, vs. 5/15, 33.3%, in PCa- and BPH-derived cultures, respectively) and in high-grade compared to moderate- and low-grade samples [3/10 (G2–4, 30%) vs. 7/15 (G5–7, 40%) and 7/8 (G8–10, 87.5%), p<0.005]. PSA secretion (Scbt A67-B/E3) was extremely low/absent in BPH- and PCa-derived cultures. The levels of p75NTR were significantly higher in PCa-derived (2.5±2.0 ADU) compared with BPH-derived cultures (0.8±0.7 ADU, p<0.005) and in high-grade compared to moderate- and low-grade samples [0.6±0.5 (G2–4) ADU vs. 2.1±1.4 (G5–7) ADU and 4.7±0.8 in (G8–10), p<0.001]. Since changes in the levels of stromal constituents, such as reduced levels or loss of caveolin-1 expression and increased TP (Imgenex, San Diego, CA) levels, were associated with PCa-derived cultures with high Gleason scores compared to BPH with low Gleason scores, we analyzed the expression of these proteins. We observed that caveolin-1 (Scbt H96) levels were significantly higher in BPH- compared to PCa-derived cultures [3.3±0.5 ADU for BPH vs. 1.5±0.9 ADU, respectively; p<0.001] and in low-grade compared to moderate- and high-grade samples [2.6±0.4 (G2–4) ADU vs. 1.2±0.2 (G5–7) ADU and 0.7±0.2 in (G8–10); p<0.005]. TP levels were significantly lower in BPH-compared to PCa-derived cultures [0.8±0.5 ADU for BPH vs. 3.5±0.7 ADU, respectively; p<0.001], whereas no statistically significant differences were observed for low-, moderate- and high-grade PCa-derived cultures [3.3±0.5 (G2–4) ADU vs. 4.1±0.3 (G5–7) ADU and 3.2±0.8 in (G8–10); not significant].

We then analyzed the secretion of two major stromal secreted cytokines (SDF1α and TGF-β) in conditioned media harvested from subconfluent BPH- and PCa-derived cultures. The results demonstrated that the levels of TGF-β1 and SDF1α were significantly higher in PCa-derived cultures (65.0±27.8 pg/ml for SDF1α and 158±64 pg/ml for TGF-β1) compared to those observed in BPH-derived cultures (12±4 pg/ml for SDF1α and 21±6 pg/ml for TGF-β1; p<0.001 for both cytokines). SDF1α and TGF-β1 secretion was higher and with respect to moderate-and low-grade samples [SDF1α: 40.3±12.0 (G2–4) ADU vs. 56.5±15.3 (G5–7) ADU and 72.7±31.2 (G8–10), p<0.005; TGF-β1: 84.6±37.0 (G2–4) ADU vs. 115.3±45.3 (G5–7) ADU and 184.8±51.2 (G8–10), p<0.001].

Gelatin and plasminogen-dependent zymography

Fig. 4 demonstrates that prostatic stromal cell cultures primarily secreted MMP-2, whereas MMP-9 secretion was low/absent in BPH-derived cultures (Fig. 4A) and restricted in some PCa-derived cultures (Fig. 4B; 12/30, 40%). Higher levels of the active form of MMP-2 were observed in PCa- compared to BPH-derived cultures.

Figure 4 Gelatinolytic activity (MMP-9 and MMP-2) of tissue inhibitors of metalloproteinase (TIMP1, TIMP2 and TIMP3) and plasminogen-dependent enzymes (tPA and uPA) present in conditioned media harvested after 48 h of subconfluent primary stromal culture cell cultures derived from (A) BPH and (B) PCa tissue samples. Each lane was loaded with 60 μl of conditioned medium. HMW, high molecular weight; LMW, low molecular weight.

TIMP1, TIMP2 and TIMP3 were also present in the conditioned media. BPH- and PCa-derived cultures (Fig. 4B) presented similar levels of TIMP1 (Fig. 4B), whereas the levels of TIMP2 were higher in PCa-derived cultures and TIMP3 was only expressed in BPH-derived cultures.

Low levels of urokinase plasminogen activator (uPA) and absent tissue levels of tissue plasminogen activator (tPA) were observed in conditioned media from BPH-derived cultures (Fig. 4A), whereas higher levels of uPA (both as high molecular weight and low molecular weight isoform) were observed in the conditioned media from PCa-derived cultures. tPA activity was observed in some (8/30, 26.7%) PCa-derived cultures. High levels of pro-MMP-9 and active MMP-2 as well as uPA and tPA were obserevd in the cultures with a higher content of MFs.

Discussion

The tumor-associated stroma is not simply a supporting element for cancer cells, but plays an important role in tumor growth, invasion and metastasis. Prostate stromal cells, and the ECM which they deposit, play a key role in restraining or promoting tumorigenesis through different mechanisms, including increased arachidonate metabolism and inflammation, promoting local tumor invasion. Eicosanoids modulate the interaction of tumor cells with various host components in cancer metastasis. Their synthesis involves the release of arachidonic acid (AA) from cellular phospholipids by phospholipase A2 (PLA2), followed by metabolism by cyclooxygenases and lipooxygenases (25). The specific nature of the human PCa stromal phenotype has been shown to be an independent clinical prognostic marker (26,27), underlining the importance of paracrine interactions in human disease. The prostate tumor microenvironment is complex, including cells of many different lineages (28–31). These include, but are not limited to, smooth muscle, various types of fibroblasts, senescent stromal cells, nerves and blood vessels and a wide variety of immune and inflammatory cell types. The diversity in the types of cells that compose the tumor stroma makes it difficult to study the contribution of each component. In addition, the lack of stromal cell lines that can retain the tumor-inducing properties shown by cancer-associated fibroblasts (CAFs) in vivo represents another important issue.

The results of the present study show differences in the percentage of stromal cell populations harvested from BPH and PCa primary stromal cultures. The two predominant cell types in the stroma of the prostate, are SMCs and fibroblasts. Fibroblasts represent the major stromal cell type, and a higher percentage of MFs are harvested from PCa. MFs are important components of the prostatic stroma. The mutual interaction through direct cell-cell contacts or paracrine signals between cancer cells and MFs is essential for invasive growth and is translated into a poor clinical prognosis. Tumor progression is deeply influenced by epigenetic changes induced by the tumor stroma. Reduced levels or loss of caveolin-1 expression and increased TP were observed in PCa-derived cultures with high Gleason scores compared to those observed in BPH-derived cultures with low Gleason scores. Caveolin-1 was strongly associated with the presence of fibroblasts. Although it has been demonstrated that the fibroblast expression of caveolin-1 favors directional migration and invasiveness of carcinoma cells in vitro as well as in vivo, stromal caveolin-1 remodels peri- and intratumoral microenvironments to facilitate tumor invasion, correlating with increased metastatic potency (32,33), current evidence indicates that the increased expression of caveolin-1 in prostate adenocarcinoma cells and the downregulation of the protein in prostate stroma, mediate progression to the castration-resistant phase of PCa (34). CAFs have been reported to promote epithelial-mesenchymal transition in cancer cells, thereby enhancing their aggressiveness and stem-like properties (7,35–38). We observed that PC3 xenograft-derived murine CAFs were enriched in caveolin-1-positive cells, suggesting that expeirmental in vivo models may be similar. The presence of a reactive stromal phenotype in PCa-derived stromal cells is associated with the retention of AR and p75NTR expression, TGF-β1, MMP-9 and VEGF secretion, as well as the overexpression of the chemokine, SDF1α/CXCL12. This phenotypic appearance can support the proliferation and invasion of PCa cell lines, such as PC3 and 22rv1. MF cultures secreted higher levels of active MMP-2 both in the pro-enzyme and active form. MMP-2 is a well-known MMP related to cancer cell invasion and metastasis; its overexpression can be induced by cytokines, growth factors and oncogenes (36–38). Some cultures of PCa-derived sromal cells express MMP-9 as a pro-enzyme (37–39). This in vitro study contributes to a better understanding of the cell composition of the reactive stromal compartment in PCa. The isolation of CAFs from aggressive or non-aggressive PCa cell xenografts may represent tool for the study of stroma-tumor cell interactions.

References