Background
Prostate cancer (PCa) is one of the most frequently diagnosed cancers in men and is the second leading cause of cancer deaths for men in the USA [
1]. The majority of deaths are due to the failures of diagnosis, current therapies, and subsequent development of the metastatic cancer, 80% of which is primarily metastasize to the bone [
2]. Thus, better strategies for the treatment of PCa will ultimately require a better understanding of the molecular mechanisms that trigger and drive cancer progression and metastatic processes.
Metastasis is a process that tumor cells leave a primary location, travel through circulation, and form a secondary tumor in the distant organ. Bone metastasis is a common complication associated with advanced cancers, including PCa, often causing acute pain and bone fractures, and the major reason for the cause of deaths in patients [
3]. The breakdown of the basement membrane surrounding the tumors as well as the increase in the abilities of proliferation, migration, and invasion are the hallmarks of metastasis [
4]. The well-studied protease linked with migration and invasion is the matrix metalloproteases (MMPs). MMP-2 and MMP-9, also known as gelatinases, play a key role for cleaving type I, IV collagen and contribute to the process of metastasis [
5]. In PCa, MMP-2 and MMP-9 are considered useful prognostic markers and are correlated to PCa progression [
6], indicating that the interaction between tumor cells and extracellular matrix is associated with tumorigenesis. Thus, targeting the factors that can regulate the activity of MMPs may develop as a strategy to suppress PCa progression.
Thrombospondin (TSP), a glycoprotein that belongs to a group of matricellular proteins, participates in cell-to-cell and cell-to-matrix communication [
7]. The human TSP protein family consists of five members, named TSP-1~5 [
8]. TSP-1 and TSP-2 are highly expressed during the tissue remodeling that is associated with cancer progression, whereas the roles of the other members of TSPs in tissue remodeling are less well understood. The structure of TSP-1 and TSP-2 is similar, but their expression patterns are temporally and spatially different during mouse development, suggesting that they may play different roles [
7‐
9]. In response to injury, the expression of TSP-2 is increased and is associated with tumor growth [
8]. In addition, TSP-2 might play a role in collagen fibrillogenesis in the skin and tendons, suggesting that TSP-2 modulates the cell surface properties of mesenchymal cells, and thus, regulates cell functions, such as adhesion and migration [
10]. Through binding to CD36 with their type I repeats, TSP-1 and TSP-2 have been reported to serve as an anti-angiogenic molecule [
11]. Additionally, the arginine-glycine-aspartic acid (RGD) sequence in TSP-2 has been reported that it binds to integrin α
νβ
3 and heparan sulfate proteoglycans which are associated with cell adhesion or binds to the low-density lipoprotein receptor-related protein (LRP) that modulates the concentration of TSP-2 in the pericellular environment by endocytosis and lysosomal degradation of the protein [
12]. Through interaction with LRP, TSP-2 binds to pro-MMP-2 and MMP-2 that regulates the extracellular levels of MMP-2 which is important for controlling several physiological processes, such as collagen fibrillogenesis, wound healing, and angiogenesis [
13,
14]. Moreover, TSP-1, TSP-2 was also demonstrated to play roles on anti-angiogenesis in the tumors [
11,
15]. However, the expression profile of TSP-2 is quite controversial, which is down-regulated in cervical cancer [
9] and ovarian cancer [
16], while it is overexpressed in oral cavity squamous cell carcinoma [
17], pulmonary adenocarcinoma [
18], and prostate cancer [
19], suggesting that TSP-2 may play another role rather than anti-angiogenesis. These studies indicate that the role of TSP-2 on tumorigenesis is still controversial, especially on metastasis.
MicroRNAs (miRNAs) are small non-coding RNA that regulates gene expression through binding to 3′ untranslated region (3′UTR) of mRNA, which with important functions in development, cell differentiation, regulation of cell cycle, and apoptosis [
20,
21]. Through up- or down-regulation of tumor suppressor genes, miRNAs may function in either an oncogenic or tumor suppressor role and they appear to play important and unique roles with respect to PCa progression [
22]. Several miRNAs have been illustrated to mediate metastasis, including miR-154, miR-376c, miR-377, miR-381, and miR-495 [
23]. It has been shown that miR-376c enhances proliferation, survival, and chemo-resistance by targeting activin receptor-like kinase 7 in ovarian cancer [
24]. Whereas, others showed that miR-376c inhibit cell proliferation and invasion by targeting the transforming growth factor-α in osteosarcoma [
25]. Thus, miR-376c may play “dual” roles on tumor progression. However, its role on PCa cells is largely uncertain. Herein, we showed for the first time that TSP-2 up-regulates MMP-2 expression and the subsequent cell motility in human PCa cells. Furthermore, these TSP-2’s effects are dependent on down-regulation of miR-376c expression.
Methods
Reagents
Recombinant human TSP-2 was purchased from R&D Systems (Minneapolis, MN, USA). The anti-rabbit TSP-2 was obtained from Abnova (Taipei, Taiwan); anti-rabbit p38, anti-mouse p-ERK, ERK, p-JNK, and p-p38 were obtained from Santa Cruz (CA, USA); anti-mouse MMP-2 was from R&D Systems (MN, USA). The inhibitors for p38 (SB203580), ERK (U0126), JNK (SP600125), and MMP-2 (MMP-2 inhibitor I) were purchased from Calbiochem (San Diego, CA, USA). The inhibitors for integrin α4β1 (sulfo-N-succinimidyl oleate (SSO)), CD36 (BIO1221), integrin αvβ3 (RGD), and the control peptide (RAD) were purchased from Torcis Bioscience (Ellisville, MO, USA). The Luciferase assay kit was purchased from Promega (Madison, WI, USA). TSP-2 shRNA plasmids were purchased from National RNAi Core Facility Platform (Taipei, Taiwan). The TSP-2 shRNA oligo sequences were 5′-CCGGCCCTCCTAAGACAAGGAACATCTCGAGATGTTCCTTGTCTTAGGAGGGTTTTTG-3′ which target to the sequence of TSP-2 is 5′-CCCTCCTAAGACAAGGAACAT-3′. MiR-376c mimic was purchased from Invitrogen (Carlsbad, CA, USA). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA).
Microarray data
The TSP-2 gene expression profile data was retrieved from the GEO database (
http://www.ncbi.nlm.nih.gov/geo/). A total of 182 normal prostate tissues, as well as benign, primary, and metastatic prostate tumors were obtained from the microarray data with the following accession numbers: GDS1439 and GDS2545. The TSP-2 gene expression profiles were retrieved from GEO data analysis tools. Further analysis of TSP-2 gene expression between normal, primary tumor, and metastatic tumor were examined by one-way ANOVA with Bonferroni’s multiple comparisons test and
p value ≤0.05 showed significance.
Cell culture
Human PCa cell lines (PC-3 and DU145) and human normal prostate epithelial cell lines (PZ-HPV7) were obtained from the American Type Culture Collection (ATCC). PC-3 and DU145 cells were grown in RPMI-1640 medium supplemented with 20 mM HEPES, 10% heat-inactivated fetal calf serum, 2 mM glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. PZ-HPV7 cells were grown in keratinocyte-SFM, containing bovine pituitary extract and recombinant epidermal growth factor. All cells were maintained in a humidified incubator at 37 °C, 5% CO2.
Transwell assay
The migration assay was performed using the transwell plates (Costar, NY, USA). The invasion assay was performed using the same transwell plate except for coating Corning® Matrigel® Matrix (Corning, NY, USA) in the lower chamber. Selection of invasive prostate cancer cells (PC-3-I3 and DU145-I5) was performed by using transwell invasion assay as described previously [
26]. Briefly, the transwell inserts were coated with Matrigel, prostate cancer cells were resuspended in 1% FBS containing media and seeded into the wells in the upper layer, with the lower layer supplied with 10% FBS containing media. After 48 h, the inserts were removed, and cells that had migrated through the membranes and become attached to the lower chamber compartments were trypsinized and expanded for second-round selection. The PC-3-I3 and DU145-I5 invasive prostate cancer cells were established after 3 and 5 rounds of selection, respectively.
Small interfering RNA (siRNA) transfection
Cells were transfected with siRNAs according to manufacturers’ recommendations on standard procedure [
26,
27]. The siRNAs (ON-TARGETplus SMARTpool) were purchased from GE Dharmacon (Lafayette, CO, USA). Cells were transfected with siRNA using Lipofectamine 2000 reagent. The mRNA knockdown efficiency was confirmed by real-time PCR as described in the following sections.
Reverse transcription (RT) and real-time PCR
Total RNA was extracted from PCa cells using a TRIzol kit as described previously [
26]. Briefly, the reverse transcription reaction was performed using the oligo (dT) primer. Real-time PCR analysis was carried out using SYBR with sequence-specific primers. The GAPDH mRNA expression was used as an internal control.
For miRNA detection, reverse transcription was performed using Mir-X™ miRNA First-Strand Synthesis and SYBR® RT-PCR with the specific forward primer of miR-376c (5′-AACATAGAGGAAATTCCACGT-3′). The U6 snRNA was used for normalization. The threshold was set above the non-template control background and within the linear phase of target gene amplification to calculate the cycle number at which the transcript was detected (denoted as CT).
Immunoblotting assay
Protein was isolated from PCa cells, and its concentration was then determined as described previously [
27]. Proteins were resolved by SDS-PAGE and transferred to Immobilon polyvinylidene fluoride membranes. After incubation with primary and secondary antibodies, the membranes were visualized by enhanced chemiluminescence using Kodak X-OMAT LS film.
Zymography analysis
The supernatants of the indicated condition of cells were mixed with sample buffer without reducing agent and heating. The sample was performed with SDS-PAGE containing gelatin (1 mg/ml). Afterwards, the gel was washed with 2.5% Triton X-100 to remove SDS, rinsed with 50 mM Tris–HCl, pH 7.5, and then incubated overnight at room temperature with the developing buffer (50 mM Tris–HCl, pH 7.5, 5 mM CaCl2, 1 μM ZnCl2, 0.02% thimerosal, 1% Triton X-100). The zymographic activities were revealed by staining with 1% Coomassie blue. The same samples were performed with SDS-PAGE without gelatin and staining with 1% Coomassie blue as loading control.
Immunohistochemical (IHC) staining
The protein expression was determined on tissue slides using IHC staining as described previously [
26]. Human PCa tissue array (T195b and PR956) was purchased from Biomax (MD, USA) in the form of 5 μm sections of paraffin-embedded tissue on glass slides. The tissue slides were incubated with human TSP-2 and MMP-2 antibodies, followed by counterstaining with hematoxylin.
Plasmid construction and luciferase activity assay
The 3′UTRs of the human MMP-2 gene were amplified by PCR using the following primer: Forward primer, 5′-GAGTTTAAACCCTCTTTAAGTCTGTTTCTTC-3′, Reverse primer, 5′-GCGCTAGCCAACTAATAATGGCCTTTTT-3′. The 3′UTRs of MMP-2 were cloned downstream of the reporter gene in the pGL2-Control vector. The predicted MMP-2 binding site for miRNA was identified by the miRDB (
http://mirdb.org/miRDB). Mutant plasmids that attenuate the interaction between MMP-2 3′UTR and miRNA were generated using a QuikChange Site-Directed Mutagenesis kit (Stratagene, Cedar Creek, TX, USA). Mutagenesis of miR-376c targeting seed region of MMP-2 3′UTR were amplified by PCR using the following primers: Forward primer, 5′-CAATTAATAGAGTGCTTTCTGGGTGCAAGGCACTTTTCACG-3, Reverse primer, 5′-CGTGAAAAGTGCCTTGCACCCAGAAAGCACTCTATTAATTG′-3”. These plasmids with 3′UTR of MMP-2 and β-galactosidase as control were transfected into cells using lipofectamine 2000. Following transfection, these cells were incubated with the indicated agents. Cell extracts were prepared and used for measuring the luciferase and β-galactosidase activities as the manufacturer’s recommendations. Activities of luciferase and β-galactosidase were then measured by Luciferase Assay System (Promega, WI, USA).
PC-3 constitutively expressed pLenti CMV V5-Luc cells were transfected with TSP-2 shRNA plasmids and selected by puromycin. PC-3 stable cells (2 × 106) which containing control vector or TSP-2 shRNA-expressing vector were intratibially injected with 200 μl Matrigel into the right tibia of nude mice (4-week old). Bone metastasis was monitored using an in vivo imaging system (Xenogen IVIS imaging system). The mice were humanely sacrificed after 4 weeks. The tumor mass was removed from the right tibia bone to detect its weight.
Microcomputed tomography (micro-CT) analysis
After the mice were sacrificed, the tibia of tumor growth were dissected and fixed in 4% paraformaldehyde for micro-CT analysis. The tibia for micro-CT scanning was assessed by using a micro-CT scanner (Skyscan 1176; Bruker, Kontich, Belgium). Reconstruction of sections was carried out with GPU-based scanner software (NRecon, Bruker). In addition, the grayscale was based on the Hounsfield unit, and the validated calcium standards were scanned as the density reference. The three-dimensional microstructural volumes from the micro-CT scans were analyzed using Skyscan software (CTAn, Bruker). Bone volume was used to assess the bone resorption area of the bone metastasis.
The 499 prostate adenocarcinoma datasets were retrieved from the TCGA. The datasets included mRNA, microRNA expression levels, and clinical data. Total 50 paired adjacent normal and tumor specimens were used to analyze expression level of TSP-2 and miR-376c. The paired t test was performed, and p value ≤0.05 showed significance. The TSP-2 and MMP-2 expression levels were correlated with Gleason score of the tumor specimens. The one-way ANOVA with Bonferroni’s multiple comparisons test was performed, and p value ≤0.05 showed significance. The correlation between TSP-2 and MMP-2 was analyzed by all mRNA sequencing results in the dataset.
Statistical analysis
All data presented as mean ± standard error of the mean (SEM). Statistical analysis between the two samples was performed using the Student t test. p < 0.05 was considered significant.
Discussion
TSP-2 has been demonstrated to play an anti-angiogenic role on tumor cells [
11,
15]. However, its expression profile is controversial, being down-regulated in some tumors [
9,
16], while being overexpressed in others [
17‐
19]. Thus, the biological functions of TSP-2 are still uncertain, especially in PCa. Our IHC results indicated that TSP-2 expression was positively correlated with PCa progression, and this result is consistent with the TSP-2 expression pattern in the other datasets of PCa. Moreover, the TCGA dataset of PCa consolidates our finding in this study. The TSP-2 expression is higher in PCa patients and correlated with tumor progression.
TSP-2 exhibit an extensive modular domains named exosites, which have been implicated in interaction with various cell surface, matrix, and proteolytic proteins. Moreover, the properdin-like type 1 repeats (TSR) of TSP-2 were found to interact with MMP-2 and endocytosis of complexes by LRP. This mechanism reveals the modulation of MMP-2 secretion from cells by TSP-2 expression [
14]. Although TSP-2 binds to MMP-2 that regulates the extracellular levels of MMP-2 during collagen fibrillogenesis, wound healing, and angiogenesis [
13], little is known about the roles of TSP-2 on PCa progression. To investigate the biological roles of TSP-2, PCa cell lines, PC-3, DU145, and LNCaP were applied. We revealed that TSP-2 induces MMP-2 expression and the subsequent migration and invasion in PC-3, DU145, and LNCaP cells. Moreover, the migratory ability of PC-3 is significantly higher than that in DU145 cells. This result is consistent to other studies showing the higher motility in PC-3 cells [
38]. It may have resulted from the higher expression of TSP-2 in PC-3, thus inducing more MMP-2 expression. Furthermore, the higher TSP-2 expression level showed the higher MMP-2 expression, migratory, and invasive abilities which were also observed in PC-3-I3 and DU145-I5, compared to PC-3 and DU145, respectively. Meanwhile, the knockdown of TSP-2 also suppressed the migratory and invasive abilities in PC-3 cells through attenuating MMP-2 expression. These results confirmed the biological roles of TSP-2 in human PCa cells. Furthermore, we also showed that metastatic PCa has a higher TSP-2 and MMP-2 expression when compared with normal prostate tissue, suggesting that TSP-2 might be associated with advanced PCa. Moreover, bone metastasis is obviously inhibited while TSP-2 is knocked down when compared to wild-type mice in vivo. Taken together, these results reveal that the TSP-2 expression level is associated with tumor metastasis, suggesting that TSP-2 could be used as a biomarker of PCa progression.
The development of an antiangiogenic target has been a common strategy for cancer therapy for a long time, but substantial benefits remain unrealized because tumors elicit evasive resistance [
39]. Interestingly, the inhibition of VEGFR/PDGFR by applying their kinase inhibitors sunitinib/SU11248 can accelerate metastatic tumor growth and decrease overall survival in mice, but cannot function as an anti-cancer strategy [
40]. TSP-1 and TSP-2 were considered as an anti-cancer molecular by inhibiting angiogenesis through antagonizing VEGF expression or suppressing metastasis by manipulating MMPs activity in several cancers, including PCa [
41,
42]. Surprisingly, under hypoxia conditions, TSP-1 might trigger cell migration in advanced PCa cells [
43], suggesting that the anti-angiogenic molecules may switch to play a positive role for tumor cell migration. Herein, we showed for the first time that the expression TSP-2 is correlated to the PCa progression, especially on the metastasis. Since, TSP-2 is the homology to TSP-1, we considered that TSP-2 may share the similar mechanism to TSP-1, showing the promotion of the migratory effects on tumor progression. Further experiments are needed for clarifying this issue in the future.
TSP-1 and TSP-2 have been known as the anti-angiogenic molecules which may bind to CD36 [
30,
41]. Interestingly, the inhibition between TSP-1 and CD36 by the TSP1 antibody A4.1 blocks the TSP-1’s effects on anti-angiogenesis [
44] and also suppresses the migration of C4-2 cells [
43], suggesting that the CD36 is responsible for the TSP-1-induced migration. This result is similar to our study showing that CD36 participated in the TSP-2-induced MMP-2 expression and the subsequent migration and invasion. In addition to CD36, TSP-2 also regulates angiogenesis and cell motility, through integrins, including integrin α
4β
1 [
31] and α
Vβ
3 [
32]. Integrins are a large family of cell-surface glycoproteins, which bind to several extracellular matrix components and regulate cytoskeletal organization and facilitate cell motility. Overexpression of integrin α
Vβ
3 has been found in many cancers, including melanoma, prostate, and breast cancer and is associated with their malignancy and responsible for their bone metastasis [
45,
46]. Herein, we also showed for the first time that the TSP-2’s effects on PCa cell migration and invasion are mediated through integrin α
Vβ
3, but not integrin α
4β
1 in human PCa cells.
It has been shown that miRNAs are involved in multiple biological processes and are also tightly correlated with tumor progression, including PCa [
20]. The miR-376c belongs to the miR-376 cluster gene family, containing miR-376a, miR-376a*, and miR-376b. The lower miR-376c level is correlated with a higher PSA and Gleason score [
10], suggesting that miR-376c is a negative regulator for PCa progression. Furthermore, miR-376c is shown to serve as an important regulator for androgen signaling by targeting the 3′UTR of UDP-glucuronosyltransferase 2B15 and UGT2B17 in PCa cells [
47]. These studies are similar to our results showing that the miR-376c expression level is negatively associated with cell migration, invasion, and PCa progression. Meanwhile, we also showed the first evidence that miR-376c is essential for the TSP-2-induced migration in PCa cells. Thus, the modulation of miR-376c expression may develop as a novel strategy for PCa therapy.
To understand the physiological role of TSP-2 in vivo, we established a PC-3 cells with stably knockdown of its TSP-2 (PC-3/shTSP-2-Luc). We found that the expression of MMP-2, as well as the abilities of migration and invasion was significantly reduced in PC-3/shTSP-2-Luc cells, whereas its miR-376c is higher than normal PC-3/Luc cells. Through monitoring by bioluminescence imaging, we showed that the knockdown of TSP-2 dramatically suppresses bone metastasis and osteolytic abilities of prostate tumor. Interestingly, the tumor with PC-3/shTSP-2-Luc is smaller than that with PC-3/Luc, although the cell proliferation rates between these cells are almost the same. We considered that TSP-2 might regulate osteoclastogenesis and bone remodeling in vivo. This issue needs further investigation in the future.
The concept of “vicious cycle” has been well established in bone metastasis. The tumor-bone interaction promotes both bone destruction and tumor growth during bone metastasis [
48]. In the past, the role of MMPs in tumor progression is focused on invasion and metastasis. However, the crucial role of MMPs in vicious cycle of bone metastasis has been discussed recently. MMPs improve the tumor growth not only through bone destruction but also by regulating bioactivity of the vicious cycle-related factors such as PTHrP, RANKL, and TGFβ [
49]. Here, our findings reveal that TSP-2/MMP-2 axis in bone metastasis may be regulated through complicated mechanism and remains to be discussed in the future.
Acknowledgements
Our gratitude goes to Michael Burton, Asia University for the English editing.