Overexpression of macrophage inhibitory cytokine-1 induces metastasis of human prostate cancer cells through the FAK–RhoA signaling pathway

Abstract

An elevated level of macrophage inhibitory cytokine-1 (MIC-1) is reported in the sera of patients with metastatic prostate cancer compared with that of benign diseases and healthy adults. We investigated the mechanistic role of MIC-1 overexpression in the metastasis of prostate cancer cells. Our study showed a progressive increase in secretory MIC-1 production correlated with the increase in the metastatic potential of PC-3 and LNPCa prostate cancer metastatic variants. Further, the in vitro studies using ‘loss-’ and ‘gain’-of-function approaches showed that ectopic overexpression of MIC-1 (PC-3-MIC-1) and forced downregulation of MIC-1(PC-3M-siMIC-1) enhanced and reduced the motility and invasiveness of these cells, respectively. Supporting our in vitro observations, all the mice orthotopically implanted with PC-3-MIC-1 cells developed metastasis compared with none in the PC-3-vector group. Our results showed that MIC-1 overexpression was associated with apparent changes in actin organization. In addition, an enhanced phosphorylation of focal adhesion kinase (FAK) and guanosine-5′-triphosphate (GTP)-bound RhoA was also seen; however, no significant change was observed in total FAK and RhoA levels in the PC-3-MIC-1 cells. Altogether, our findings show that MIC-1 has a role in prostate cancer metastasis, in part, by promoting the motility of these cells. Activation of the FAK–RhoA signaling pathway is involved in MIC-1-mediated actin reorganization, and thus, leads to an increase in the motility of prostate cancer cells.

Introduction

Prostate cancer (PCa) is the most common malignancy in the male population of the Western world. It is the second leading cause of cancer-related deaths in the United States, representing nearly 10% of cancer deaths in men (Jemal et al., 2008). Despite the fact that early diagnosis is made in most cases, the incidence and mortality rate of this cancer are still increasing steadily. Metastasis in PCa is a critical determinant for the survival of the patients. Locally treated PCa reoccurs in 40% of men (D’Amico et al., 1998). Metastatic PCa is highly resistant to therapeutic intervention. Therefore, the development of potential novel targeted therapeutics to inhibit PCa metastasis may have a significant impact on the incidence of PCa mortality. These therapeutics may be important for the ‘watchful waiting’ group of PCa patients, who are diagnosed with low-stage PCa and are under observation for signs of high PCa progression.

The causes of metastasis of PCa are not yet precisely understood. Many gene products, including numerous growth factors and cytokines and/or their receptors, are overexpressed or downregulated during the development of this hyperproliferative disease. Macrophage inhibitory cytokine-1 (MIC-1), a member of the TGF-β superfamily, has been observed to be frequently over-expressed during the progression of PCa in androgen-independent and recurrent diseases (Karan et al., 2002, 2003). Owing to its expression and function in various organs, MIC-1 is also known by different names (Hromas et al., 1997; Lawton et al., 1997; Paralkar et al., 1998; Bottner et al., 1999; Baek et al., 2001). The nascent MIC-1 protein comprises a signal peptide, propeptide and mature peptide. Following intracellular cleavage, the mature protein is secreted as a homodimer linked by a disulfide bond (Bootcov et al., 1997). Importantly, we and others have reported that the high expression of MIC-1 may be associated with the acquisition of an androgen-independent phenotype as well as a poor outcome for patients diagnosed with PCa (Chen et al., 2007; Selander et al., 2007; Wakchoure et al., 2009). In addition to PCa, elevated levels of MIC-1 are also reported in the serum of patients with metastatic breast and colorectal cancers compared with that of normal adults (Welsh et al., 2003). Furthermore, MIC-1 induces invasiveness of gastric cancer cells by the upregulation of the urokinase-type plasminogen activator system (Lee et al., 2003). Thus, these observations suggest a possible role of MIC-1 in PCa progression.

The studies on the functional role of MIC-1 in PCa cells are contradictory. Various experimental evidences suggest androgen-mediated regulation of MIC-1 in PCa cells (Paralkar et al., 1998; Karan et al., 2003; Chen et al., 2007). In an attempt to investigate the role of MIC-1 in PCa, the exogenous expression of MIC-1 in AR-negative, MIC-1 null DU145 cells, showed a reduced cell adhesion by decreasing RhoE and catenin δ1 expression (Liu et al., 2003). Recently, however, it has been reported that in AR-positive LNPCa PCa cells, MIC-1 promotes cell proliferation through an autocrine/paracrine manner (Chen et al., 2007). The MIC-1-mediated proliferation is carried out through an alternative TGF-β pathway, involving ERK1/2 and p90RSK (Chen et al., 2007). In fact, MIC-1 has been shown to perform diverse biological functions, depending on the context of the cells (Paralkar et al., 1998; Fairlie et al., 1999; Lee et al., 2003). Hence, at this time, the role and mechanisms of action of MIC-1, which appear to be dependent on cancer cell types, remain to be precisely established (Karan et al., 2009). Importantly, due to the effect of the prostate microenvironment on the ultimate role of a protein like TGF-β (Barrack, 1997), more studies on appropriate animal models are warranted. In spite of some knowledge regarding the role of MIC-1 in PCa growth, the clinical observations showing that MIC-1 expression is generally increased in metastatic PCa patients (Selander et al., 2007), and that MIC-1 expression correlates to prognosis, remain to be understood.

Previous studies have shown that a PC-3 cell line variant (saPC-3), which grows in suspension and has a more metastatic property, expresses a higher level of MIC-1 than parental PC-3 cells (Patrikainen et al., 2007). In addition, very low endogenous MIC-1 expression is detected in PC-3 cells compared with AR-positive LNPCa-C81 or LNPCa-LN3 cells (Karan et al., 2003). In this study, using well-established highly metastatic variant PCa cell line models (Kozlowski et al., 1984; Pettaway et al., 1996), we showed a progressive increase in secreted MIC-1 production with an increase in the metastatic potential of the cells. To further investigate the mechanistic role of MIC-1 overexpression in PCa metastasis, we adopted gain- and loss-of-function approaches. Both overexpression of MIC-1 in PC-3 cells and downregulation of MIC-1 in PC-3M cells showed that MIC-1 has a significant role in PCa metastasis. We also showed that the MIC-1-mediated enhancement in the metastatic property of PC-3 cells might be partly through the increase in motility of PC-3 cells. Furthermore, MIC-1 activated the focal adhesion kinase (FAK)–RhoA signaling pathway to increase the motility of PC-3 cells. Finally, we showed that at least in a subpopulation of PCa cells, MIC-1 has a role in increasing the metastatic property of those cells.

Results

Metastatic variants of PCa cells produce a high level of secretory MIC-1

Clinical evidence has shown that patients with metastatic PCa have increased blood levels of MIC-1. Therefore, the expression pattern of secreted MIC-1 was investigated in PC-3, LNPCa-C33 and their metastatic variants (PC-3M, PC-3MLN4 and LNPCa-LN3). For the detection of secreted MIC-1 (mature form), we used our well-characterized anti-MIC-1 antibody, which was produced to the C-terminal region of the MIC-1 protein (Chen et al., 2007). Similar to clinical observations, in these two different metastatic cell line models, along with the increase in the metastatic potential of PCa cells, a high level of secretory MIC-1 was detected (Figures 1a and b).

Figure 1.

Analysis of macrophage inhibitory cytokine-1 (MIC-1) expression in metastatic variants of prostate cancer (PCa) cells and PC-3-derived sublines. MIC-1 expression was analyzed in (a) parental PC-3 and two of its metastatic variant cells (PC-3M and PC-3M-LN4) by immunoblot analysis, (b) LNPCa-C33 and its metastatic variant LNPCa-LN3 cells and (c) a pooled population of PC-3-vector and PC-3-MIC-1 cells. A total of 20 µg protein from enriched cell culture media was resolved by electrophoresis (15% SDS–polyacrylamide gel electrophoresis (PAGE)), transferred to polyvinylidene difluoride (PVDF) and then probed with anti-MIC-1 antibody. After incubation with horseradish peroxidase (HRP)-conjugated anti-rabbit antibody, the signal was detected using an electrochemiluminescence reagent kit. Coomassie blue-stained protein band on gel was used as a loading control. The metastatic variants of PC-3 and LNPCa-33cells produce a high level of secretory MIC-1 with increasing metastatic potential.

Characterization of PC-3-derived MIC-1-transfected sublines

To determine the significance of MIC-1 in PCa progression, we transfected MIC-1 overexpression construct and empty vector (to serve as a control) in PC-3 cells. Multiple stable clones were selected on growing the cells in puromycin-containing media. MIC-1 expression and secretion in the derived sublines were examined by western blot analysis. The selected sublines were monitored over a period of 2–3 months for the stable overexpression of MIC-1. A pooled population of the sublines that showed an approximately threefold increase in MIC-1 expression was selected and used for all the in vitro and in vivo studies (Figure 1c). Likewise, a pooled population of cells transfected with empty vector was used as a control in all the functional studies. Importantly, in terms of cell behavior in the cell culture, we observed similar results for individual selected clones and pooled cells. Therefore, to simplify the presentation only results with pooled cells are shown.

MIC-1 promotes motility and invasiveness of PC-3 and PC-3M cells

Different studies have shown that the invasive and metastatic potential of cancer cells is strongly related to a variety of phenotypic characteristics. Among these characteristics, the motility of cells highly influences the metastatic property of cells. In PCa patients, increasing serum MIC-1 levels are associated with the progression of metastatic PCa (Selander et al., 2007; Wakchoure et al., 2009). Furthermore, as mentioned previously, in this study, we found that in multiple PCa cell lines, the metastatic variants of PCa cells express a higher level of MIC-1 than their corresponding parental cells (Figures 1a and b). Therefore, cell motility may be a key factor, which gets enhanced due to MIC-1 over-expression. The number of MIC-1-overexpressing PC-3 cells (PC-3-MIC-1) migrating to the lower surface of the porous membrane was significantly higher than that of the vector control (PC-3-vector) cells (Figure 2a; **P<0.0005). Likewise, motility of PC-3M control cells (PC-3M-Con) was significantly less than PC-3M cells with a transient knockdown of MIC-1 (PC-3M-siMIC-1) (Supplementary Figure 1; Figure 2a; **P<0.0005). To further examine if the increased cell motility also correlated with the invasive potential of PCa cells, we performed an in vitro Matrigel invasion assay. We observed that significantly more PC-3-MIC-1 cells invade the Matrigel in comparison to control PC-3-vector cells (Figure 2b; *P<0.05). As expected, a significantly smaller number of PC-3M-siMIC-1 cells invade the Matrigel than PC-3M-Con cells (Figure 2b; *P<0.05). However, in both the motility and invasion assays, there was no significant difference between PC-3-MIC-1 and PC-3M-Con cells (Figures 2a and b; P>0.05, not significant (NS)). To elucidate whether secreted MIC-1 can exert effect on neighboring cells through the paracrine loop and increase the motility of those cells, a wound-healing assay was carried out (Figure 2c; Supplementary Materials and Methods). Normal PC-3 cells incubated with media collected from PC-3-MIC-1 cells promote a significantly faster wound healing compared with cells incubated with media collected from PC-3-vector cells (Figure 2c; **P < 0.0005). However, media collected from PC-3M cells with transient knockdown of MIC-1 (Supplementary Materials and Methods; Supplementary Figure 1) resulted in an opposite outcome. Further, media from PC-3-MIC-1 and PC-3M cells have a similar effect on the extent of wound healing (P>0.05, NS).

Figure 2.

Effect of macrophage inhibitory cytokine-1 (MIC-1) overexpression on cell motility and invasion. Cells were seeded on noncoated or Matrigel-coated membranes for motility (a) and invasion (b) assays, respectively, and incubated for 22 h. Medium containing 10% fetal bovine serum in the lower chamber was used as a chemoattractant. The cells that migrate through the membrane were stained and photographed under bright-field microscopy (magnification, × 10). The number of cells that migrated throughout the membrane was determined by averaging 10 random fields of view. The data are expressed as the number of cells per field of view and is the average of three independent experiments. Error bars indicate s.e. of the average (**P < 0.0005; *P< 0.05). (c) To show the paracrine effect of MIC-1 on the motility of other neighbor cells, we carried out a wound-healing assay, using conditioned media collected from PC-3-MIC-1, PC-3-vector, PC-3M-Con and PC-3M-siMIC-1 cells. In this experiment, cells in culture were monitored for their ability to migrate into a site of mechanical injury. Introduced wounds were photographed before and 10 h after incubation of cells in conditioned media. The results show that normal PC-3 cells incubated with conditioned media collected from PC-3-MIC-1 and PC-M-Con cells heal the wound quicker (**P < 0.0005) than cells incubated with media from PC-3-vectro or PC-3-siMIC-1 cells. The wound area covered after 10 h of incubation was determined as described in Supplementary Materials and Methods.

For all the aforementioned experiments, non-transfected PC-3 and PC-M cells were also taken as additional controls. The results for these cells were almost the same as the corresponding empty-vector-transfected PC-3 cells, and control-siRNA-transfected PC-3M cells (data not shown). Therefore, for further studies, only transfected control cells were used.

MIC-1 has no effect on the incidence of tumor formation and overall growth of PC-3 cells

The exact role of MIC-1 in PCa cell survival seems to be very ambiguous. To date, most of the studies have been carried out in vitro or in xenograft model. But, the prostate has a unique microenvironment, and the actual role of a molecule in PCa progression may be micro-environment dependent. Therefore, we evaluated the effect of MIC-1 on the growth of tumors by orthotopic implantation. MIC-1- and empty-vector-transfected PC-3 cells were injected orthotopically into the prostate gland of athymic nude mice. After 40 days, an incidence of tumor was found in both the groups. There was no significant difference in the tumor weight (Figure 3a; Supplementary Table 1; P=0.2). Using immunohisto-chemical analysis, we carried out the expression of MIC-1 at the primary tumor in both groups. As expected, we observed a more intense staining of MIC-1 in primary tumors obtained from PC-3-MIC-1 cells than PC-3-vector cells (Figure 3b). In parallel, animals injected with PC-3M cells showed tumors with a significantly higher weight than PC-3-vector and PC-3-MIC-1 cells (Supplementary Figure 2A; Supplementary Table 1; P=0.04 and 0.0004). Tumors obtained from PC-3M cells have a similar intensity of MIC-1 expression as PC-3-MIC-1 tumors (Supplementary Figure 2C).

Figure 3.

Macroscopic and microscopic examination of orthotopic prostate tumors and organs with metastasis lesions. (a) Macroscopic appearance of selected urogenital organs from animals implanted with PC-3-MIC-1 and PC-3-vector cells. Seminal vesicles and prostate tumor were dissected from 40-day-old animals from both groups. There was no difference in the gross appearance of tumors derived from both groups. But, animals injected with PC-3-MIC-1 cells were found to be very week, emaciated and having less body weight than animals injected with PC-3-vctor control cells. (b) Immunohistochemical analysis of primary tumors developed in immunodeficient mice by orthotopic implantation of PC-3-MIC-1 and PC-3-vector cells. Tumor sections were stained for macrophage inhibitory cytokine-1 (MIC-1)-positive cells using MIC-1-specific affinity-purified anti-rabbit polyclonal antibody. The two panels represent × 10 and × 20 magnification. Tissue derived from PC-3-MIC-1 tumors shows high signal intensity. (c) Histological appearance of liver, lung and lymph node showing prominent metastasis in the mice implanted with PC-3-MIC-1 cells. Normal (N) and cancer cells (Ca) in different organs are separately indicated, but in the lymph node, the normal tissues are completely occupied by tumor cells. It was found that MIC-1 promotes metastasis in the orthotopic prostate cancer model.

Overexpression of MIC-1 results in metastasis of PC-3 cells to distant organs and mice carrying the tumors originated from PC-3-MIC-1 cells lose normal body weight

In spite of the fact that patients with metastatic PCa have a high serum concentration of MIC-1, the actual role of MIC-1 in PCa metastasis has yet to be defined. Therefore, in our orthotopic animal model, we checked the incidence of metastasis in both animal groups. During autopsy of those mice, we found that only animals of PC-3-MIC-1 groups developed massive metastasis to different organs, and no metastasis was observed in animals form the PC-3-vector group. Most of the gross metastasis were identified in the lymph nodes, the liver and kidneys, and were confirmed by light microscopic examination (Figure 3c). Furthermore, all the animals injected with PC-3M cells showed an incidence of metastasis and had a similar pattern (same organ and number) of metastasis as PC-3-MIC-1 animals (Table 1).

Table 1.

Incidence of metastasis in animals injected with PC-3-MIC-1 and PC-3-vector cells

Type of animals	Number of animals	Number of animals with different organ-specific metastasis					P-value for incidence of metastasis
		Liver	Kidney	Lymph node	Lungs	Bone
PC-3-MIC-1	6	6/6	6/6	6/6	0/6	0/6
PC-3-vector	6	0/6	0/6	0/6	0/6	0/6	*P < 0.05
PC-3M	6	6/6	6/6	6/6	0/6	0/6

^*Fisher’s exact test.

Cancer-related weight loss is a major cause of morbidity and mortality in cancer patients, and is mostly caused by cytokines. Earlier studies have also shown the role of MIC-1 in tumor-induced anorexia and weight loss (Johnen et al., 2007). Therefore, we examined the effect of MIC-1 overexpression on overall body weight change. For this purpose, the body weight of all the animals of both groups was measured on the date of tumor cell injection, and on the date of killing. The difference of mean body weight at these two different time points was significantly less (P = 0.0008) in the case of animals injected with PC-3-MIC-1 cells compared with animals injected with PC-3-vector cells (Supplementary Table 2; Figure 3a). Animals harboring PC-3-M cells also have a significantly smaller body weight than animals injected with PC-3-vector cells. The difference was not significant as compared with animals injected with PC-3-MIC-1 cells (data not shown).

MIC-1 alters actin rearrangement and activates the FAK–RhoA signaling pathway, which may promote cell motility

Cell motility mostly depends on a rearrangement of the actin cytoskeleton, which is typically associated with the coordinated assembly and disassembly of the cortical actin network. Polymerization of globular actin (G-actin) leads to the formation of long fibrous actin (F-actin), which helps to form the cellular outgrowths responsible for cell motility. As we observed an increase in cell motility in PC-3-MIC-1 cells as compared to PC-3-vector control cells, we wanted to examine the actin organization in these cells. Localization and distribution of F-actin was analyzed by phalloidin staining. Confocal microscopy of the phalloidin-stained cells showed more intense F-actin staining in PC-3-MIC-1 cells than in the PC-3-vector cells (Figure 4a). The actin reorganization leads to the formation of more microspikes, lamellipodia and filopodia-like cellular projections in PC-3-MIC-1 cells (**P<0.0005). Various studies have shown the role of FAK in actin rearrangement, and FAK-mediated actin rearrangement is an important regulator of cell migration. Our biochemical analysis showed an increased level of activated FAK (pY925) level in PC-3-MIC-1 cells over PC-3-vector cells (Figure 4b). The level of total FAK, however, remained unchanged in both of the cell types, suggesting that MIC-1 promotes activation of FAK, which may be responsible for the greater motility observed in MIC-1 transfected cells (Figure 4b). Further, it has also been shown that FAK-mediated actin assembly is mostly carried out by the Rho–GTPase pathway (Sieg et al., 2000). In addition, studies have also shown that in highly invasive PC-3 cells, activation of RhoA contributes to an increase in the motile property of the cells (Hodge et al., 2003). Therefore, we examined the RhoA activation status in both of the PC-3-derived cell lines. For this purpose, we performed Rhotekin-RBD bead pulldown assays. The activity of RhoA was found to be increased more than twofold in highly invasive PC-3-MIC-1 cells compared with PC-3-vector low invasive cells (Figure 4b). However, no difference was observed in total RhoA expression level (Figure 4b). Thus, the FAK–RhoA signaling pathway is activated in MIC-1-overexpressing cells.

Figure 4.

Effect of macrophage inhibitory cytokine-1 (MIC-1) overexpression on actin rearrangement, expression and activity of RhoA and FAK. (a) Actin organization: the PC-3-MIC-1 cells and control PC-3-vector cells were grown on coverslips, fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) and permeabilized with 0.1% Triton X-100 in PBS. After that F-actin staining was carried out by using rhodamine phalloidin (R415). The mounted glass slides were observed under Zeiss confocal laser-scanning microscope. Increased filopodia structures (indicated by arrows) were observed in MIC-1-transfected cells compared with the PC-3-vector cells. Filopodias longer than 1.5µm were counted on 10 individual cells per group, and the average per cell was calculated + s.e. (**P < 0.0005). (b) Western blot analysis showing a similar level of total FAK and RhoA protein in PC-3-MIC-1 and PC-3-vector cells, but increased expression of activated FAK and RhoA protein in PC-3-MIC-1 cells compared with PC-3-vector cells. RhoA activity was determined by using Rhotekin assay. Relative levels of all the proteins are indicated in each figure normalized to the internal control β-actin.

Discussion

Reports on the detection of elevated levels of MIC-1 in metastatic PCa patients led us to investigate the biological function of MIC-1 in PCa progression so that further therapeutic studies can be designed. Corroborating the clinical findings, this study clearly shows an increase in MIC-1 production with the increase in metastatic potential of PCa cells. We have also shown a role of MIC-1 in potentiating the invasiveness of PCa cells, in vitro and in vivo. In addition, our study has shown that MIC-1 promotes actin reorganization in PC-3 cells, by the activation of the FAK–RhoA signaling pathway.

Among the available PCa cell line models, the metastatic variants of PC-3 and LNPCa cells generated by multiple selection cycles of orthotopic injection (Kozlowski et al., 1984; Pettaway et al., 1996) are powerful tools to understand the molecular mechanisms behind PCa metastasis. In these models, PC-3M-LN4 cells are more metastatic than PC-3M cells and PC-3M cells are more metastatic than the original parental PC-3 cells (Kozlowski et al., 1984; Pettaway et al., 1996). Likewise, LNPCa-LN3 cells are more metastatic than the parental LNPCa cells. In this study, a high level of secretory MIC-1 production with increasing metastatic potential of PC-3 and LNPCa cells clearly indicates that MIC-1, which is overexpressed in metastatic PCa, has a possible role in PCa metastasis.

Distant metastasis is the major cause of human cancer-related death (Sporn, 1996). Successful metastasis spread depends both on intrinsic properties of tumor cells, and on tumor microenvironment-derived factors. Various cytokines are known to augment the ability of a tumor cell to metastasize by helping in several steps of the cancer metastasis process. A number of studies have shown that cell motility is an essential factor for PCa metastasis, and the motility of PCa cells depends on several known and unknown factors. Upregulation of growth factor synthesis, secretion and/or altered expression of growth factor receptors has the potential to enhance PCa cell motility (Banyard and Zetter, 1998). Likewise, additional growth factors, including autocrine motility factor, TGF-β and so on, also modulate PCa cell motility (Banyard and Zetter, 1998). Therefore, there is a need to identify and characterize genes that are responsible for generating aggressive PCa. We present here in vitro and in vivo evidence that an increase in MIC-1 expression leads to a rearrangement of the actin cytoskeleton and promotes cell motility, invasion and metastasis. These MIC-1 functions are consistent with the classic model of metastasis, which assumes that metastases arise from rare cells, and those cells along with the early oncogenic alteration afterward gather mutations that promote metastasis (Robinson et al., 2004). The role of MIC-1 PCa metastasis further supports the significance of the TGF-β superfamily in cancer metastasis (Padua and Massague, 2009).

Tumor cell invasion through the extracellular matrix and tissue barriers requires the combined effects of increased cell motility and proteolytic degradation of extracellular matrix. Our in vitro motility assay showed that MIC-1 promotes the motility of PC-3 and PC-3M cells. Further, we showed that, like other known paracrine effects of MIC-1 (Chen et al., 2007), MIC-1 exert it effect both in autocrine and paracrine manners to increase PCa cell motility in our cell line models. We have also shown that MIC-1 promotes invasion of PC-3 and PC-3M cells through extracellular matrix, and in vitro, it enhances the incidence of PC-3 cells metastasis. Therefore, the high incidence of metastasis in animals injected with PC-3-MIC-1 cells rather than PC-3-vector may be partly due to an increase in cell motility. A similar type of metastasis pattern between PC-3-MIC-1 and PC-3M groups further indicates that MIC-1 may be the key factor in providing the metastasis property of PC-3M cells. Cell motility mostly depends on rearrangement of the actin cytoskeleton, which is typically associated with the coordinated assembly and disassembly of the cortical actin network. Polymerization of G-actin leads to the formation of long F-actin, which helps to form the cellular outgrowths responsible for cell motility. The localization and distribution of F-actin was analyzed by phalloidin staining in both MIC-1-overexpressing cells (PC-3-MIC-1) and control cells (PC-3-vector). Confocal microscopy of the phalloidin-stained cells showed more intense F-actin staining in PC-3-MIC-1 cells than in the PC-3-vector cells. Further, our biochemical analysis showed an increased level of activated FAK (pY925) level in PC-3-MIC-1 cells over PC-3-vector cells. The level of total FAK, however, remained unchanged in the cell types, suggesting that MIC-1 promotes activation of FAK, which may be responsible for the greater motility, observed in MIC-1-transfected cells. During migration of cells, the cell adhesion dynamics are closely associated to control actin assembly and disassembly, and FAK contributes to both (Sieg et al., 2000). The FAK-mediated actin assembly is mostly carried out by the Rho–GTPase pathway (McLean et al., 2005). RhoA, Rac1 and Cdc42 Rho-GTPase regulate cell motility through stimulation of actin cytoskeletal rearrangement. Specifically, RhoA facilitates the development of focal adhesions and stress fibers. Here, we revealed that the expression of MIC-1 promotes activation of RhoA-GTPase, which is similar to a previous report that highly invasive PC-3 cells have more activated RhoA-GTPase (Hodge et al., 2003). The Rho family GTPase cycles between an active GTP-bound state and an inactive GDP-bound state, through the activity of guanine nucleotide exchange factors (GEF) and GTPase-activating proteins. Studies have shown that FAK-mediated phosphorylation of the RHO guanine nucleotide exchange factor (p190RHOGEF) and PDZ-domain-containing RHO-GEF (PDZRHOGEF) correlates with higher RhoA activity (Chikumi et al., 2002; Zhai et al., 2003; Schlaepfer et al., 2004). Experimental evidence has also shown that activated RhoA can activate FAK through an unknown mechanism (Chikumi et al., 2002). Therefore, it is possible that MIC-1 activates FAK and/or RhoA through an unknown mechanism, which may further activate each other and thus can regulate actin rearrangement (Figure 5).

Figure 5.

Proposed model for macrophage inhibitory cytokine-1 (MIC-1)-mediated signaling events that promote actin cytoskeleton rearrangement. Including integrins, FAK transduces stimuli from a variety of cell-surface-associated receptors like growth factor receptors and G-protein-linked receptors. But, due to a lack of clear knowledge regarding the MIC-1 receptor, how MIC-1 activates FAK is yet to be elucidated. However, MIC-1-mediated phosphorylation of FAK may promote phosphorylation of the RhoA-specific GDP/GTP exchange factor (RhoA-GEF), and thus, enhance activation of RhoA. It is also possible that MIC-1 may activate RhoA-GEF through an unknown mechanism, which further produces more RhoA-GTP. Later, activated RhoA may enhance phosphorylation of FAK. In this way, a positive feedback loop may regulate the activation of both FAK and RhoA, which regulate actin reorganization and cell motility. Mechanisms that are not fully elucidated are indicated by dotted arrows or question marks.

MIC-1 has been implicated to have a role in cancer. Both antiapoptotic and proapoptotic effects have been described in a variety of tumor cells (Karan et al., 2009). Experimental studies using human prostate LNPCa cells support the role of MIC-1 in cell proliferation (Chen et al., 2007). However, in the androgen-independent PC-3 PCa cell line, ectopic expression of MIC-1 was shown to reduce the growth of subcutaneously grown tumors (Lambert et al., 2006). Likewise, when androgen-insensitive DU145 cells were treated with recombinant MIC-1, loss of cell adhesion and induction of apoptosis was observed (Liu et al., 2003). In our in vivo experiment, we have found that MIC-1 has no impact on overall primary tumor growth, which is not in accordance with previously reported observations (Lambert et al., 2006). This difference in results may be due to the different experimental models (orthotopic vs subcutaneous) and the context-dependent function of MIC-1. In comparison with previous experiments, the present orthotopic implantation of PCa cells mimics the progression of PCa in its natural microenvironment. Importantly, recent studies suggest tumor stage, tissue of origin and a microenvironment-dependent function of MIC-1. In addition, studies have also shown that MIC-1 bioavailability is tumor microenvironment dependent (Bauskin et al., 2005). Therefore, in the prostate microenvironment, the role of MIC-1 as an antitumor factor may be altered or compensated by other host factors. Furthermore, a significantly higher tumor weight in PC-3M groups than in PC-3-vector and PC-3-MIC-1 groups may be due to the accumulation of many tumorigenic genetic alterations in these cells than parental PC-3 or PC-3-MIC-1 cells. In addition, no overall effect of MIC-1 overexpression on tumor growth (derived from PC-3 cells) has also raised an important question regarding the regulation of MIC-1 expression in different stages of PCa progression. It was envisioned that the induction of MIC-1 might be an early response due to inflammation, infection or injury in the prostate gland (Karan et al., 2009). The exact regulatory mechanism for MIC-1 expression in the early inflammatory and late metastatic stages of PCa still needs to be checked. Further studies on the role of androgen in the regulation of MIC-1 expression in PCa cells will be helpful to clarify the controlled expression of MIC-1 in PCa. This information will help to design effective therapeutics for PCa.

In addition, loss of body weight in animals bearing tumors generated by PC-3-MIC-1 and PC-3M cells supports the previous report regarding the role of MIC-1 in tumor-induced weight loss (Johnen et al., 2007). This effect of MIC-1 may be due to its role in the central regulation of appetite and weight (Johnen et al., 2007).

Our data have shown an overexpression of MIC-1 in metastatic variants of PCa cells, and importantly, its functional role in the promotion of PCa cells metastasis. The MIC-1-mediated increase in the invasive or metastatic property of PCa cells is partly due to actin rearrangement, which may be due to the activation of both FAK and RhoA in MIC-1-overexpressing cells. Further investigations are needed to delineate the detailed signaling cascade involved in MIC-1-mediated actin reorganization. In addition to the role of MIC-1 in PCa metastasis, we also elucidate that MIC-1 over-expression has no overall effect on tumor growth in the prostate microenvironment. This finding indicates a microenvironment-dependent role of MIC-1 in PCa progression. Taken together, the data in this study show that MIC-1 is a potential target for the prevention of PCa metastasis.

Materials and methods

Cell culture and transfection

The PC-3 cell line was purchased from the American Type Culture Collection (Manassas, VA, USA). The PC-3M, PC-3M-LN4 and LNPCa-LN3 cells were obtained from Dr Fidler. LNPCa-C33 cells were obtained from Dr Lin (UNMC, Omaha, NE, USA). All cell lines were grown in regular phenol red-free RPMI 1640 culture medium (Gibco, Grand Island, NY, USA) supplemented with 5% fetal bovine serum (FBS), 1% glutamine and 1% penicillin–streptomycin. For overexpression of MIC-1 in PC-3 cells, the MIC-1 overexpression construct (pMSCV.puro-MIC-1) was generated (Supplementary Materials and Methods), and was stably transfected in PC-3 cells. For a control, cells were transfected with empty vector (pMSCV.puro). Stable clones were selected and maintained in a medium containing puromycin (3 µg/ml; InvivoGen, San Diego, CA, USA). The selection medium was replaced with a complete medium with antibiotic supplement for at least 5 days before any analysis.

For transient knockdown of MIC-1 in PC-3M cells, the stealth RNAi-mediated approach was adopted (Supplementary Materials and Methods; Supplementary Figure 1).

Confocal immunofluorescence microscopy

For phalloidin staining, cells were grown at low density on sterile coverslips for 20 h. After washing with 0.1 m HEPES containing Hanks buffer, the cells were fixed in 3.7% formaldehyde solution in phosphate-buffered saline (PBS) for 10 min at room temperature. Then cells were washed twice with PBS and permeabilized with 0.1% Triton X-100 in PBS for 3–5 min. This was followed by washing with PBS. Nonspecific blocking was carried out using 1% bovine serum albumin for 30 min. For F-actin staining, cells were incubated with rhodamine phalloidin (R415; Molecular Probes, Eugene, OR, USA) for 20 min at room temperature. Then cells were washed 3–4 times with PBS, and finally slides were washed twice with PBS and mounted on glass slides in antifade Vectashield mounting medium (Vector Laboratories, Burlin-game, CA, USA). The slides were observed under a Zeiss confocal laser-scanning microscope, and photographs were taken digitally using LSM 510 software.

Motility and invasion assays

For motility assays, 1 × 10⁶ cells suspended in serum-free medium were plated in the top chamber of polyethylene teraphthalate membranes (six-well insert, pore size 8 µm; Becton Dickinson, San Jose, CA, USA). The lower chamber of the well was filled with 2.0 ml of 10% serum-containing medium and the cells were allowed to migrate for 22 h under chemotactic drive. After incubation, the cells that did not migrate through the pores in the membrane were removed by scraping the membrane with a cotton swab. The migrated cells on the lower side of the membrane were stained with a Diff-Quick cell stain kit (Dade Behring, Newark, DE, USA) and photographed in 10 random fields of view at × 10 magnification. Cell numbers were counted and expressed as the average number of cells per field of view. For invasion assay, cells (5.0 × 10⁵) were seeded on Matrigel-coated membrane inserts (BD Biosciences, Bedford, MA, USA). The bottom chamber contained 2.0 ml of serum-supplemented medium as a chemoattractant. After incubation for 22 h at 37 °C, the cells that invaded through the Matrigel-coated membrane were fixed, stained and counted as the motility assay. For both the experiments, the data were represented as the average of the three independent experiments with the standard error of the average indicated.

in vivo tumor growth and metastasis

To test the effect of the overexpression of MIC-1 on the tumorigenicity and metastatic property of PC-3 cells, we carried out orthotopic implantation. In parallel, to compare the tumorigenicity and metastatic property of PC-3M cells with PC-3-vector and PC-3-MIC-1 cells, PC-3M cells were also orthotopically implanted. Immunodeficient male mice (10- to 12-week old) were purchased from the animal production area of the National Cancer Institute, Frederick Cancer Research and Development Center (Frederick, MD, USA). The mice were treated in accordance with the Institutional Animal Care and Use Committee guidelines. The PC-3-vector, PC-3-MIC-1 and PC-3M cells were harvested from subconfluent cultures by a brief exposure to 0.25% trypsin and 0.02% EDTA. After neutralizing the effect of trypsin with 10% FBS, the cells were washed once in PBS. Cells were resuspended in PBS at a concentration of 10⁶cells per 50 µl. Single-cell suspensions of > 90% viability were used for the injections. Animals (six animals per group) were anesthetized with intraperitoneal injection of ketamine and xylazine mixture (4:1). The abdomen was cleaned with iodine solution and a 1 cm midline incision was made to expose the prostate gland. One million cells suspended in 50 µl of PBS were injected into a dorsal prostatic lobe using a 30-guage needle (Hamilton syringe, Reno, NY, USA). The abdominal wound was closed in two layers with catgut and wound clips. Animals were monitored twice weekly.

To evaluate the difference in body weight gain or loss, we measured the body weight of individual animals on the day of tumor cell injection, and on the day of animal killing. The mean body weights at these two different time points were compared. To determine the tumor growth and metastasis, we killed the mice by CO₂ asphyxiation and autopsied on the 40th day after the implantation of the tumor cells. After inspection of macroscopical tumor growth, the prostate and seminal vesicles were taken out as one unit and the weight of the resected organs was assessed in milligrams (mg). Regional and distant lymph nodes, lung, liver as well as other organs suspected for harboring metastasis were routinely formalin-fixed, embedded, sectioned, and stained with hematoxylin and eosin using standard techniques for microscopic examination. Before post-mortem examination of the animals, the total body weight of the animals was measured. Later on, the mean difference in body weight (body weight on day 1/ body weight on day 40) was compared between the two groups. Further statistical analysis was carried out as mentioned in the Supplementary Materials and Methods.

For antibodies, plasmid construct, siRNAs transfection, western blot, wound-healing assay, filopodia measurements and statistical analysis, a detailed description is given in the Supplementary Information.