Monocyte Chemoattractant Protein-1 promotes cancer cell migration via c-Raf/MAPK/AP-1 pathway and MMP-9 production in osteosarcoma

All shRNA plasmids with specific targets, and control shRNA plasmid were purchased from National RNAi Core Facility (Academia Sinica, Taipei, Taiwan). Protein G beads; all secondary antibodies against rabbit or mouse (IgG-conjugated horseradish peroxidase); rabbit polyclonal antibodies against β-actin, MMP-2, MMP-3, MMP-9, MMP-12, MMP-13, c-Raf, p-MEK, MEK, p-ERK, ERK, p-JNK, JNK, p-p38, p38, p-c-Jun, c-Jun were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). All siRNAs which target specific genes (si-c-Jun: sc-29,223; si-MMP9: sc-29,400; si-CCR2: sc-270,220; si-control: sc-44,231) were purchased from Santa Cruz Biotechnology. p-c-Raf (Ser338) was purchased from Cell Signaling (Danvers, MA, USA). U0126 and PD98059 were purchased from Calbiochem (San Diego, CA, USA).

The dominant-negative mutants specific for MEK1, ERK2, JNK and p38 were kindly gifted from Dr. W. M. Fu (National Taiwan University, Taipei, Taiwan), Dr. M. Cobb, (University of Texas Southwestern Medical Center, Dallas, TX, USA), Dr. M Karin (University of California, San Diego, CA, USA), and Dr. J. Han (University of Texas Southwestern Medical Center, Dallas, TX, USA), respectively. Recombinant human MCP-1/CCL2 was purchased from PeproTech (Rocky Hill, NJ, USA). AP-1 luciferase plasmid was purchased from Stratagene (La Jolla, CA, USA). The remaining chemicals used were purchased from Sigma–Aldrich (St. Louis, MO, USA).

All osteosarcoma cell lines (U2OS, HOS, and MG-63) and normal osteoblast cell line (hFOB 1.19) used in the current study were obtained from American Type Cell Culture Collection (Manassas, VA, USA). U2OS cells were maintained in McCoy’s 5A medium. MG63, HOS, and hFOB 1.19 cells were maintained in DMEM medium. To make a complete medium, the other components were added according to ATCC’s description. Furthermore, the cells were cultured in the presence of antibiotics (100 U/mL of penicillin and 100 μg/mL of streptomycin) at 37 °C and 5% CO₂.

Transwell migration assay was conducted with Transwell plate (Costar, NY; pore size, 8 μm). Briefly, 300 μL of the serum-free medium was added in the lower chamber with different concentrations of MCP-1. Meanwhile, 100 μL of serum-free medium contained 1 × 10⁴ cells were added in the upper chamber. The cells were incubated at 37 °C in 5% CO₂ for 24 h. After 24 h later, the transwell inserts were fixed in 3.7% formaldehyde for 15 min. Then, 0.05% crystal violet dissolved in PBS was added to stain the cells for 15 min. The Transwell inserts were washed with PBS, and the cells in the upper chamber were removed using cotton swabs. The cells which migrated to the lower side of the Transwell inserts were further observed and counted using a microscope. For each experimental condition, a minimum of three experiments were conducted in triplicate.

The MG63 (M10, M20 and M30) migration-prone subclones were established by using Transwell inserts (6 wells plate with 8 μm pore size). MG63 osteosarcoma cells (1 × 10⁴) suspended in 100 μL of serum-free medium were seeded in the upper chamber, while 300 μL growth medium contained 10% FBS was loaded into the lower compartment. After 24 h later, the cell which migrated across the Transwell insert to the bottom of plate were detached by trypsin and cultured as MG63 (M1). The cells were cultured for 2 days for a second round of selection. The MG63 migration-prone subclone was continued migration seletion for 10, 20, 30 rounds to generate MG63 (M10), MG63 (M20) and MG63 (M30), respectively.

Each of the 12 wells was filled with 1 × 10⁵ cells, and the plate was incubated for 24 h. The confluent monolayer of cultured cells was scratched using a fine pipette tip. The rate of wound closure was recorded and calculated through microscopic observation. For each experimental condition, a minimum of three experiments were conducted in triplicate.

Protein expression was examined using a Thermo Scientific Pierce BCA Protein Assay Kit (Thermo Fisher Scientific Inc., Waltham, USA). Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis was conducted to resolve proteins in cell lysate, followed by transferred to Immobilon polyvinylidene difluoride (PVDF) membranes. Next, to block the blots, PVDF membranes were incubated in 4% BSA at room temperature for 1 h. The PVDF membranes were further incubated with primary antibodies for 1 h at room temperature, then washed three times with TBST. The membranes were later incubated with HRP-conjugated secondary antibodies for another 1 h at room temperature follow by washed with TBST for three times. The membranes were then detected by chemiluminescent substrate (Amersham™ ECL™ Western Blotting Detection Reagents; GE Healthcare Life Sciences, Marlborough, MA, USA) and monitored by using a charge-coupled device camera-based detection system (UVP Inc., Upland, CA, USA). The data were quantitatized by using ImageJ software (National Institute of Health, USA).

The total RNA was extracted by using the TRIzol kit (MDBio Inc., Taipei, Taiwan) according to the manufacture’s protocol. Next, for reverse transcription into cDNA, 2 μg of total RNA was reacted with an oligo(dT) primer. TaqMan® one-step PCR Master Mix (Applied Biosystems, Foster City, CA, USA) was used for analysis by using quantitative real-time polymerase chain reaction (qPCR). After reverse transcription, the cDNA (100 ng/25 μL per-reaction) was further added with primers and TaqMan® probes for detecting specific sequences, as well as TaqMan Universal PCR Master Mix according to the manufacturer’s instructions. The polymerase activation cycle was 10 min at 95 °C, 15 s at 95 °C for 40 cycles, and finally 60 s at 60 °C. For each experimental condition, a minimum of three experiments were conducted in triplicate with a StepOnePlus sequence detection system.

The cells with 80 % confluency were co-transfected with AP-1-luciferase vector (0.8 μg) and a β-galactosidase expression vector (0.4 μg) for 24 h by using Lipofectamine 2000 (LF2000; Invitrogen, Carlsbad, CA, USA). The plasmids and LF2000 were mixed and incubated for 20 min and subsequently added to the cells for 24 h, followed by incubated with the indicated agents for another 24 h. Before the cells were washed with cold PBS, the media were first removed. To lyse the cells, 100 μL of reporter lysis buffer (Promega, Madison, WI, USA) was added, and the supernatant was collected. In addition, luciferase substrate was added to 20 μL of lysates with 20–30 μg protein. Luminescence was recorded using a microplate luminometer. The luciferase activity was evaluated after normalizing the cells with the cotransfected β-galactosidase expression vector. For each experimental condition, a minimum of three experiments were conducted in triplicate.

The cells grown in 10-cm dishes were treated with MCP-1 for the indicated conditions and the cells were washed two times with ice-cold PBS, followed by scraped in PBS and collected cells by centrifuged for 15 mins. Subsequently, ice-cold buffer 1 was added (10 mM HEPES, pH 7.9; 10 mM KCl; 0.1 mM EDTA, pH 8; 0.1 mM EGTA, pH 8; 10 mM PMSF, 10 mM DTT, 10 mM NaF, 10 mM Na3VO4). The cells were sheared mechanically with a syringe and needle, then samples were centrifuged and the supernatant, i.e., the cytoplasmic fraction, was collected. The remaining pellet was washed 3 times with buffer 1, and then re-suspended in buffer 2 (20 mM HEPES, pH 7.9; 400 mM NaCl; 1 mM EDTA, pH 8; 1 mM EGTA, pH 8; 10 mM PMSF, 10 mM DTT, 10 mM NaF, 10 mM Na3VO4) and centrifuged; the supernatant was collected as the nuclear protein fraction.

Chromatin immunoprecipitation analysis was conducted as previously described [32]. DNA was immunoprecipitated using anti-c-Jun mAb, and it was further purified. The DNA was extracted by adding phenol–chloroform. The purified DNA pellet was used for PCR. After the PCR reaction, products were resolved using 1.5% agarose gel electrophoresis. Next, UV light was used for visualization. The primers 5′- ATCCTGCTTCAAAGAGCCTG-3′ and 5′-GTCTGAAGGCCCTGAGTGGT-3′ were used for amplification across the human MMP-9 promoter region (− 547 to − 327).

The MCP-1 and control shRNA lentiviral constructs (pLKO.1) were obtained from the National RNAi Core Facility (Academia Sinica, Taipei, Taiwan). The control shRNA (sh-control: ASN0000000004, target sequence: CCTAAGGTTAAGTCGCCCTCG) and the shRNAs that targeted MCP-1 (sh-MCP-1-A: TRCN0000006279, target sequence: GATGTAAACATTATGCCTTA; sh-MCP-1-B: TRCN0000006283, target sequence: CCCAGTCACCTGCTGTTATAA; and sh-MCP-1-C: TRCN0000381382, target sequence: TCATAGCAGCCACCTTCATTC) were purchased from the National RNAi Core Facility (Taipei, Taiwan).

The HEK293T cells were used to prepare lentivirus. Briefly, the shRNA plasmid mixed with packaging vectors pCMV and pMDG were co-transfected into HEK293T cells. After 24 h and 48 h post-transfection, the cell culture supernatants were collected and stored at -80 °C. To establish knockdown stable clones, the MG63 cells were transduced with cell culture supernatants describe above in the presence of 8 μg/ml of polybrene (Sigma–Aldrich). After 48 h post-transduction, the cells which expressed shRNA vectors were selected by the culture medium contained puromycin (10 μg/ml). Finally, the shRNA expression stable clones were generated after 2 weeks of selection with puromycin. In addition, all experiments used at least 2 distinct shRNA.

All in vivo experiments were conducted according to Guidelines for Animal Care of the Institutional Animal Care and Use Committee of Shin-Kong Wu Ho-Su Memorial Hospital (Taipei, Taiwan) (Ethical approval No: Most1040002). The mice were purchased from the Lasco (Taipei, Taiwan). The 5-weeks male CB17-SCID mice were intravenous tail injected with osteosarcoma cells (2 × 10⁶ / 100 μL). Six weeks after tumor implantation, the mice were sacrificed and lungs were collected to analyze metastatic nodules. The lung tissues were fixed by 10% formalin and then embedded in paraffin and subsquently performed with hematoxylin and eosin (HE) staining.

Human osteosarcoma tissue microarrays (BO244, T261, T262, T262A, T263, and OS804b) were obtained from Biomax (Rockville, MD, USA). All tissue microarray contained normal bone tissue (11 cases), stage I osteosarcoma (7 cases), stage II osteosarcoma (49 cases), and stage III osteosarcoma (7 cases). The paraffin-embedded tissues (5-μm thick) were rehydrated and incubated in 3% hydrogen peroxide to suppress endogenous peroxidase activity. Next, 3% bovine serum albumin (BSA) was prepared and used to block the samples, and it was subsequently replaced with phosphate-buffered saline (PBS) for incubation. The samples were further incubated at 4 °C with a primary mouse polyclonal antihuman antibody. After overnight incubation, the samples were washed with PBS. After three washes, the samples were incubated with a secondary antibody labeled with biotin. An ABC Kit purchased from Vector Laboratories (Burlingame, CA, USA) was used to detect the bound antibodies. Next, the samples were stained with chromogen diaminobenzidine. After another wash, the samples were stained with Delafield’s hematoxylin. Finally, the samples were dehydrated, mounted, and observed at five different degrees (0 (negative), 1 (very weak), 2 (weak), 3 (moderate), 4 (strong), and 5 (very strong)) of independent and blinded observations. The total intensity score was obtained from five immunohistochemistry (IHC) scores.

The values are represented as means ± the standard deviation (SD). Significant differences between the experimental groups and controls were assessed using the Student’s t test. Overall survival analysis was performed through the Fisher LSD post hoc tests. The differences in overall survival of the two groups were compared using the log-rank test; p < 0.05 was considered statistically significant.

MCP-1 has been shown to increase cell migration and metastasis in various human cancer cells. To understand the effect of MCP-1 on osteosarcoma cells, we first selected and cultured an osteosarcoma cell line, MG63, with different degrees of migratory ability including 10, 20, and 30 generations and compared their migratory efficiency (Fig. 1a). The higher the generation was, the more the cells could migrate. Consequently, we detected the MCP-1 protein (Fig. 1b) and mRNA (Fig. 1c) expression among different selected cells. MCP-1 protein and mRNA production both increased the most during the 30 generation MG63 cells. Meanwhile, the association between MCP-1 and osteosarcoma cell migration potential was confirmed in osteosarcoma cell lines including MG-63, U2OS, HOS as well as normal osteoblast cell line hFOB 1.19 (Fig. 1d-f), which was in agreement with our findings in migration-prone cells above. Of the different concentrations of MCP-1, the MG63, U2OS and HOS cells stimulated with 10 ng/mL of MCP-1 exhibited the highest migratory degrees (Fig. 1g). In the HOS cells, the highest migratory ability was observed for stimulation with less than 5 ng/mL of MCP-1. In the wound healing ability test, 10 ng/mL of MCP-1 triggered the highest degrees of migration in the three osteosarcoma cell lines (Fig. 1h and i). When two different concentrations of MCP-1 antibody were used in the MG63 cells, the original migratory effect could be significantly reduced (p < 0.05) (Fig. 1j). Therefore, MCP-1 production was highly correlated with osteosarcoma cell migration in vitro.

Studies have revealed that MMPs including MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-12, and MMP-13 are significantly related to osteosarcoma metastasis and poor prognosis [19, 33‐38]. To identify the mediator of MCP-1-promoted osteosarcoma migration, we further examined the expression of MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-12, and MMP-13 mRNA under MCP-1 stimulation (Fig. 2a). The results revealed that substantial amounts of only MMP-9 mRNA were produced after MCP-1 treatment. In addition, MMP-9 mRNA was upregulated in a dose-independent manner (Fig. 2b). Western blotting demonstrated that among the expression levels of MMP-2, MMP-3, MMP-9, MMP-12, and MMP-13 only that of MMP-9 increased in a dose-dependent manner (Fig. 2c and Fig. S2). The expression levels of MMP-9 among migration-prone subclone, osteosarcoma cell lines and normal osteoblasts were also evaluated (Fig. S1, upper panel), and our results indicated that MCP-1 expression was positively associated with MMP-9 expression. When osteosarcoma cells were treated with MMP-9 monoclonal antibody (mAb), the degree of cell migration significantly decreased (Fig. 2d). Moreover, with MMP-9 inhibitor (SB-3CT), less osteosarcoma cell migration was demonstrated (Fig. 2e). In addition to pharmacological inhibition, we also used MMP-9 small interfering RNA (siRNA) to confirm the migratory effect (Fig. 2f-g). MCP-1-promoted osteosarcoma cell migratory ability remarkably decreased after MMP-9 siRNA pretreatment. Therefore, MCP-1 triggered osteosarcoma metastasis through the regulation of MMP-9 expression.

CCR2, one of the most important MCP-1 receptors, regulates MCP-1 effects more strictly than MCP-1 itself [39]. A study of 27 patient samples with osteosarcoma revealed CCR2 expression in every sample [40]. However, in another study, CCR4 was found in 21 out of 22 osteosarcoma samples [40]. Consequently, we examined whether MCP-1 induced MG63 osteosarcoma cell migration through the CCR2 or CCR4 receptor by using pharmacological inhibitors. In the migration assay, osteosarcoma cell migration was remarkably reduced by the CCR2 inhibitor but not by the CCR4 inhibitor (Fig. 3a). MMP-9 mRNA expression was also substantially inhibited by the CCR2 inhibitor but not by the CCR4 inhibitor after MCP-1 stimulation (Fig. 3b). In addition, MMP-9 protein production was reduced with CCR2 inhibitor treatment (Fig. 3c and Fig. S3). We further used CCR2 mAb to examine the MCP-1-mediated MG63 cell migratory effect (Fig. 3d). With CCR2 mAb, MCP-1-induced migration was drastically reversed. Moreover, pretreatment with CCR2 mAb demonstrated a similar effect as the CCR2 inhibitor on MMP-9 expression level in response to MCP-1 stimulation (Fig. 3e-f). In addition to the CCR2 inhibitor and mAb, we used CCR2 siRNA to confirm the involvement of CCR2 in MCP-1-promoted osteosarcoma cell migration. CCR2 siRNA could reverse both the cell migration degree (Fig. 3g) and the production of MCP-1-mediated MMP-9 mRNA (Fig. 3h). Finally, we defined the expression level of CCR2 among osteosarcoma cell lines and normal osteoblast cell line, which may be associated with migration ability in these cell lines (Fig. 3i). As expected, the positive correlation was found between CCR2 expression level and migration ability among these cell lines (Fig. 3j), suggesting that CCR2 was closely related to MCP-1-induced osteosarcoma cell metastasis as a key receptor.

Previous studies have demonstrated that c-Raf is important in the regulation of epithelial to mesenchymal transition, which is an essential process of tumor invasion and metastasis in numerous types of cancer, including osteosarcoma and chondrosarcoma [32, 41‐43]. Under MCP-1 stimulation, western blotting demonstrated that phosphate c-Raf was activated in a time-dependent manner (Fig. 4a). Pretreatment with c-Raf inhibitor (GW5074) for 30 min markedly inhibited MCP-1-induced phosphorylation of c-Raf (Fig. 4b). Pharmacologically, the MG63 cells treated with GW5074 demonstrated a drastically decreased migratory ability (Fig. 4c and Fig. S3) as well as MMP-9 expression (Fig. 4d and e). Finally, to further examine whether c-Raf was involved in MCP-1-mediated osteosarcoma migration, we used the c-Raf short hairpin RNA (shRNA) (Fig. 4f). When the MG63 cells were transfected with c-Raf shRNA, both cell migration (Fig. 4g) and MMP-9 mRNA expression (Fig. 4h) were remarkably reversed. According to these results, MCP-1 induced MMP-9 expression and cell migration through the c-Raf signaling pathway in osteosarcoma.

The c-Raf/MEK/ERK signaling cascade has been reported to be related to tumorigenesis and metastasis [32, 43, 44]. Studies have revealed that MEK, ERK, JNK, and p38 MAPK play crucial roles in MMP-9 regulation and osteosarcoma metastasis [45]. Therefore, we examined whether MEK, ERK, JNK, and p38 MAPK were also important mediators in MCP-1-mediated MG63 cell migration. We used different pharmacological antagonists, including MEK inhibitors (U0126 and PD98059), p38 MAPK inhibitor (SB203580), and JNK inhibitor (SP600125) to demonstrate the influences on MCP-1-promoted MG63 cell migration (Fig. 5a). MMP-9 production and MCP-1-mediated MG63 cell migratory ability were drastically reduced by the antagonists (Fig. 5b and Fig. S3). Western blotting revealed that MMP-9 mRNA expression was decreased, and MMP-9 protein production was inhibited (Fig. 5c). To determine whether MEK, ERK, JNK, and p38 MAPK were activated during MCP-1-triggered cell migration, each protein phosphorylation was measured using western blotting (Fig. 5d). In addition, we used MEK, ERK, JNK, and p38 MAPK mutants to examine the cell migratory effect (Fig. 5e) and MMP-9 mRNA regulation (Fig. 5f). With different methods of antagonists, MEK, ERK, JNK, and p38 MAPK mutants also remarkably reversed the effects of MCP-1 upregulation. Pretreatment with MEK inhibitors (U0126 and PD98059), p38 MAPK inhibitor (SB203580), and JNK inhibitor (SP600125) for 30 min markedly inhibited MCP-1-induced phosphorylation of ERK, JNK, and p38 MAPK (Fig. 5g-i). Therefore, MAPK is crucial in osteosarcoma metastasis promoted by MCP-1 upregulation.

AP-1, an important binding site of c-Jun and c-Fos, has been reported to regulate numerous types of gene transcription, such as MMPs [46]. Studies have revealed the critical role of c-Jun in osteosarcoma progression and metastasis [47, 48]. Therefore, we hypothesized that AP-1 was also involved in MCP-1-mediated migration. Through the pharmacological blocking of the c-Jun function (curcumin and tanshinone IIA), substantially fewer MG63 cells migrated than through MCP-1 stimulation only (Fig. 6a and Fig. S3). Additionally, both curcumin and tanshinone IIA remarkably reduced MMP-9 expression in qPCR (Fig. 6b) and the Western blotting assay (Fig. 6c). Next, we examined c-Jun activation and transportation during the MCP-1-induced migration in osteosarcoma cells. Western blotting demonstrated that c-Jun was phosphorylated in the cytosol only, and the total expression of c-Jun was detected in the nucleus (Fig. 6d). In addition to pharmacological antagonists, c-Jun functional siRNA was used. c-Jun siRNA did not exert a noticeable effect on MG63 cell migration; however, it could drastically reverse the effects of MCP-1 on MG63 cell (Fig. 6e-f). As presented in Fig. 6g, MMP-9 mRNA expression was also remarkably reduced even after MCP-1 stimulation. Therefore, AP-1 was an essential mediator in MCP-1-promoted MMP-9 upregulation and osteosarcoma migration.

AP-1 activation has been reported to regulate MMP-9 transcription and cancer invasion [46, 49]. Figure 6g and h present the AP-1 luciferase activity. MCP-1 activated AP-1 in a dose-dependent manner (Fig. 6h). The CCR2 inhibitor, GW5074, U0126, PD938059, SB203580, and vSP600125 drastically reduced the MCP-1-mediated migration in the MG63 cells (Fig. 6i). In addition, western blotting demonstrated that the CCR2 and c-Raf antagonists reduced c-Jun phosphorylation (Fig. 6j). The chromatin immunoprecipitation assay revealed that the binding of c-Jun to the AP-1 segment regulating the MMP-9 expression was increased after treatment with MCP-1 (Fig. 6k). By contrast, pretreatment with the CCR2 inhibitor and c-Raf inhibitor resulted in less AP-1 binding and MMP-9 transcription (Fig. 6l). U0126, PD98059, SB203580, and SP600125 also attenuated MCP-1-promoted AP-1 binding effects (Fig. 6l). Taken together, MCP-1-promoted AP-1 activation and MMP-9 expression through the CCR2, c-Raf, MAPK, and c-Jun pathways.

The in vitro results in the current study showed that MCP-1 contributes to cancer cell migration by inducing MMP-9 expression in osteosarcoma cells. Therefore, we conducted the osteosarcoma cells (MG63) which stably expressed MCP-1 or MCP-1 shRNA, which were further subjected to confirm MCP-1 expression levels and migration potential (Fig. 7a-d). Also, the in vivo animal study was performed with MG63 cells stably expressed MCP-1 shRNA or control vector by intravenous injection into the tail vein. Twenty eight days after implantation with cancer cells, the mice were sacrificed and lung nodules were monitored to evaluate metastatic potential of stable clones. The data indicated that lung nodules were dramatically abolished in MG-63/MCP-1-shRNA stable clone (Fig. 7e and g). Meanwhile, in hematoxylin and eosin (H&E) staining and IHC results, the lung morphology in mice implanted with MG-63/MCP-1-shRNA stable clone exhibited decreased degrees of lung metastatic nodules, while control group mice were infiltrated with metastases (Fig. 7f). These evidence indicated that MCP-1 promotes metastasis of osteosarcoma in vivo.

Our findings indicate that MCP-1 induces MMP-9 expression in osteosarcoma cells. To confirm these findings, we examined the expressions of MCP-1 and MMP-9 in different-stage osteosarcoma cells as well as in normal human bone tissue. The tissue array results revealed that higher concentrations of MCP-1 and MMP-9 were produced as the high grade of osteosarcoma increased (Fig. 7h). Compared with the normal bone tissue, stage II osteosarcoma cells exhibited 2–3 times higher MCP-1 expression, and stage III osteosarcoma cells exhibited 3–4 times higher MCP-1 production (Fig. 7i). The MMP-9 expression in stage II and stage III osteosarcoma cells was 2–3 and 3–4 times, respectively, higher than that in normal tissue (Fig. 7j). Furthermore, we investigated the correlation between MCP-1 and MMP-9 upregulation. The correlation coefficient of MCP-1 and MMP-9 expression was 0.69 (Fig. 7k). These results indicate that MCP-1 expression is correlated with MMP-9 expression, tumor progression, and prognosis in patients with osteosarcoma.