Introduction
Prostate cancer is the second leading type of cancer in men in United States. In 2010, new cases of prostate cancer were estimated at 217,730, resulting in 32,050 deaths in [
1]. The major cause of death is bone metastasis. Metastasis is a very complicated process during which cancer cells go through a series of steps: (i) cell dissociation from the primary tumor environment, (ii) cell adhesion to the endothelial surface at the target, (iii) cell invasion through the endothelial surface, (iv) cell invasion into new environment, and (v) cell proliferation.
In our previous study, we found that SLUG, a zinc-finger transcription factor, was elevated in mouse prostate tumors and human prostate cancer cell lines [
2]. SLUG belongs to the Slug/Snail superfamily [
3,
4], and it regulates epithelial-mesenchymal transition (EMT) in a variety of cancers [
5]. EMT is a dynamic process that promotes cell motility with decreased adhesive ability, and thus is thought to be a major starting point for cancer metastasis [
6]. SLUG plays a major role in EMT during embryonic development and metastasis of breast cancers, through partial inhibition of E-cadherin [
7,
8,
3].
In the tumor microenvironment, a complex network of chemokines and receptors affects metastasis. The CXCL12/CXCR4 pathway was originally discovered in the immune system to play an important role in cancer cell metastasis [
9‐
12]. Mice deficient of either CXCR4 or CXCL12 had abnormal development in the central nervous system [
13]. CXCL12 belongs to chemokine family of small peptides with 8 to 12 kDA size that control cell activation, differentiation, and trafficking [
14,
15]. CXCL12 is expressed by several organs: lung, liver, skeletal muscle, brain, heart, kidney, skin, and bone marrow; its secretion is related to tissue damage [
16]. The CXCR4/CXCL12 axis can coordinate metastasis of a variety of cancers, such as bladder [
17], breast [
18], head and neck [
19], ovarian [
20], renal cell [
21], and prostate [
22,
23]. Interestingly, SLUG is required for transcriptional and functional regulation of CXCL12 during bone tissue remodeling [
24].
Although the role of SLUG in cancer metastasis has been documented in other cancers besides prostate cancer, its molecular mechanism remains elusive. In this study, we examined the regulation of the Slug-CXC4R/CXCL12-metastasis triangle in an in vitro cell culture model of human prostate cancer cells. We used gain- and loss-of-function approaches to study (i) how SLUG regulates the CXCR4/CXCL12 axis, and (ii) the functional role of CXCL12 in SLUG-induced migration and invasion of human prostate cancer cell lines. We found that forced expression of SLUG significantly upregulated both CXCL12 and CXCR4 expression and their downstream target MMP9. Knockdown of SLUG decreased CXCL12 and CXCR4 expression in prostate cancer cells. Furthermore, we showed that downregulation of CXCL12/CXCR4 axis via CXCL12 knockdown impaired SLUG-mediated MMP9 expression, migration and invasion. Lastly, we provide evidence that CXCL12 and SLUG regulate migration and invasion of prostate cancer cells independent of cell growth. Our findings suggest that prostate cancer cells can gain invasive characteristics through upregulation of autocrine CXCL12.
Discussion
Metastasis is the spread of a disease from one organ or tissue to another non-adjacent organ or tissue; and thus, it is regulated by numerous signaling pathways in both the cancer cells and microenvironment. CXCR4/CXCL12 axis plays role in cancer cell metastasis and proliferation; the importance of the CXC4/CXCL12 axis may differ in different types of cancer cells, due to their discrete expression. For example, CXCR4 expression is lower in gastrointestinal tumors than breast cancer [
32]. Overexpression of CXCR4 in prostate cancer cells accelerated prostate tumor metastasis, prostate tumor vascularization, and tumor growth in vivo [
33]. CXCL12 stimulates chemotaxis of metastatic prostate cancer cells expressing a high level of CXCR4 and accelerates their migration [
34]. Conversely, blockade of CXCR4/CXCL12 interaction in prostate cancer cells via CXCR4 knockdown significantly inhibits bone metastasis in vivo [
35]. Androgens promote migration of prostate cancer cells via KLF5-mediated upregulation of CXCR4 expression [
36].
In this study, we used gain- and loss-of-function approaches to determine that SLUG positively regulated both CXCL12 and CXCR4 at the RNA and protein level. Because SLUG is a zinc-finger transcription factor and mainly functions as a transcription repressor when it is tethered to promoters of target genes [
4,
7], we therefore assumed that SLUG regulates CXCL12 and CXCR4 in an indirect manner, i.e., by suppressing expression of one or more inhibitors of these two molecules. It was recently reported that MiR-886-3p directly targets CXCL12 and decreases its expression [
37]. In future studies, we will examine if SLUG directly downregulates MiR-886-3p in prostate cancer cells. Interestingly, CXCL12 can increase the RNA and protein level of the CXCR4 receptor in basal cell carcinoma and PC3 cells [
38,
39]. Therefore, it is possible that SLUG upregulates CXCR4 in a CXCL12-dependent manner. It has been heavily documented that CXCL12 is expressed in the bone microenvironment and creates migration and invasion paths for the tumor cells with CXCR4 expression [
40]. Our current findings indicate that CXCL12 is expressed in prostate cancer cells and was induced by SLUG. Notably, it was recently shown that Slug is required for transcriptional and functional regulation of CXCL12 during the remodeling of bone tissue [
24].
Elevated SLUG expression in tumors is correlated with tumor metastasis in many types of tumors [
41,
25,
42], and forced expression of SLUG promotes metastasis of breast cancer in mouse models through partial inhibition of E-cadherin [
43]. In this study, we found that SLUG overexpression upregulated endogenous CXCL12 and increased prostate cancer cell migration and invasion, but reduced adhesion (data nor shown). In contrast, knockdown of endogenous CXCL12 expression impaired SLUG-mediated MMP9 expression, and migration and invasion in PC3 cells. Thus, our new findings that CXCL12/CXCR4 is a mediator of SLUG-induced migration and invasion of prostate cancer cells provide insight into the molecular mechanisms by which SLUG promotes tumor cell metastasis in vivo. Neutralizing CXCL12 with specific antibodies in NOD/SCID mice resulted in reduced metastasis to the lungs, adrenal glands, and liver [
21]. Therefore, it would be worthwhile to use mouse models to test whether CXCL12 is a key mediator of SLUG-induced metastasis of prostate cancer in vivo.
It has been suggested that CXCL12 promotes tumor invasion by inducing MMP9 [
44], which degrades extracellular matrix components. MMP9 is expressed and secreted from both prostate cancer cells and their microenvironment [
30,
45]. In addition, high expression of SLUG and MMP9 is found in pancreatic cancer tissues [
25]. It remains to be determined whether MMP9 is upregulated by SLUG. Here, we showed that SLUG upregulated both CXCL12 and its downstream target MMP9 expression, and that MMP9 is regulated by SLUG through CXCL12. In the future, it needs to be determined if MMP9 is critical for SLUG-induced invasion of prostate cancer cells.
Overall, our data indicate that CXCL12 is a key mediator for SLUG-induced migration and invasion of human prostate cancer cell lines in vitro; thereby suggesting that autocrine CXCL12 is an important factor for tumor metastasis.
Materials and methods
Cell Culture
PC3, 22RV1, LNCaP, and DU-145 cells were obtained from American Type Culture Collection (ATCC, Manassas, VA). These cells were maintained in culture medium, according to the manufacturer's instructions.
Plasmids
pMig-Slug was constructed by cloning human SLUG gene into pMIGR1 retroviral vector. pLKO.1-Slug shRNA1 (target sequence: 5'-CAGCTGTAAATACTGTGACAA-3'), pLKO.1-Slug shRNA2 (target sequence: 5'- CCAAATCATTTCAACTGAAA-3'), pLKO.1-CXCL12 shRNA1 (target sequence: 5'-TGTGCATTGACCCGAAGCTAA), and pLKO.1-CXCL12 shRNA2 (target sequence: 5'-GCCAACGTCAAGCATCTCAAA-3') were obtained from Open Biosystem (Huntville, AL). pLKO.1 control shRNA (containing non-target scramble shRNA, Addgene plasmid #1864) were purchased from Addgene (Cambridge, MA).
Viral Production and Infection
293T cells were seeded at 3 × 105 cells per well in a 6-well plate. The next day, a mixture of plasmid DNA was transfected separately into 293T cells using Superfect transfection reagent (Qiagen, Valencia, CA). For retrovirus production, pCL-Ampho (packaging plasmid) was mixed with pMig-based retroviral vectors. To generate the lentiviruses, the packaging plasmids (pCMV-VSVG and psPAX2) were co-transfected with pLKO.1-Slug shRNA or pLKO.1-control shRNA (containing non-target shRNA). The viruses were collected 24 hr after transfection. For viral infection, PC3, 22RV1, or DU-145 cells were seeded at 50% confluence in 6-well plates. The next day, the virus-containing supernatants from 293T cultures were mixed with polybrene (Sigma, St. Louis, MO) at a final concentration of 4 mg/ml, and added to the cells in each well. The plate was centrifuged at 2,000 rpm for 1 hr at 35°C, and returned to the cell culture incubator. PC3, 22RV1, DU145, and LNCaP cells were infected with retroviruses (pMig-Slug or pMig vector) for 3 times to achieve 100% transduction in these cells. Cells infected with pLKO.1 lentiviruses were selected with puromycin (1 μg/ml), starting at 48 hr after infection.
RNA Isolation, cDNA Synthesis, RT-PCR, and qPCR Analysis
Total RNA extraction from cultured cells was accomplished by using RNeasy Plus mini kit (Qiagen, Valencia, CA). cDNA was synthesized by random priming from 1 μg of total RNA with the SuperScript III First-Strand Synthesis Super Mix kit (Invitrogen, San Diego, CA), according to the manufacturer's protocol. Primers used for the RT-PCR and qPCR analysis were synthesized by Integrated DNA Technologies (Coralville, IA). RT-PCR was performed by using the Hotstar Taq DNA polymerase kit (McLab, San Francisco, CA), and qPCR was performed by using the Perfecta SYBR Green FastMix (Quanta Bioscience, CA), according to the manufacturer's protocol. Data were analyzed by using the comparative CT method; CT refers to the ''threshold cycle,'' and is determined for each experiment using MyiQ software. Quantities of gene specific mRNA expression were determined by the CT method. Amplification of GAPDH was performed for each reverse-transcribed sample as an endogenous quantification standard. The fold-difference in gene expression was determined by 2_ΔΔCT. ΔΔCT is equal to (ΔCT of experimental conditions -ΔCT of control conditions). ΔCT is equal to (gene-specific CT -GAPDH CT). The primers are as following: SLUG, 5'-CTTCCTGGTCAAGAAGCA-3' and 5'-GGGAAATAATCACTGTATGTGTG-3'; CXCR4, 5'-ATATACACTTCAGATAACTACACCGAG-3' and 5'-TCAGTTTCTTCTGGTAACCCATGACCA-3'; CXCL12, 5'-ACCGCGCTCTGCCTCAGCGACGGGAAG-3' and 5' TGTTGTTCTTCAGCCGGGCTACAATCTG-3'; MMP9,5'-AGCGGGCGGCGCCTCTGGAGGTTCGA-3' and 5' CCTGGCAGAAATAGGCTTTCTCTCGGT-3'; GAPDH, 5' ATTGACCTCAACTACATGGTTTACATG-3' and 5'-TTGGAGGGATCTCGCTCCTGGAAG-3'.
Enzyme-linked Immunosorbent Assay (ELISA)
Conditioned cell culture medium was centrifuged and an SDF1-α immunoassay kit (R&D Systems Inc. Minneapolis, MN) was used for CXCL12 detection. 100 μl of sample or control (or standard) was added into each well, according to the manufacturer's protocol. The optical density of each well was measured within 30 min, using a microplate reader set to 450 nm.
Western Blot Analysis
The cells were lysed in the protein lysis buffer (20 mM Tris, 100 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 1 mM beta glycerophosphate, 1 mM sodium orthovanadate), supplemented with 1 ml protease inhibitor cocktail (Sigma, St. Louis, MO). The protein samples were analyzed by Western blot analysis using an ECL kit (Pierce, Rockford, IL) with antibodies against following antigens: Slug (ANASPEC, Fremont, CA), CXCR4 (Abcam, Boston, MA), GAPDH (Bethyl Laborotaries, Montgomery, TX).
Zymographic Analysis of MMP activity
Cells overexpressing pMig or pMig-Slug (70-80% confluence) were washed twice with PBS, and the medium was changed to serum free cell culture medium. After 48 hr, the conditioned medium was collected and centrifuged for 5 min at 400 × g. A 500 μl aliquot was concentrated to < 100 ul in a Microcon concentrator (Millipore, Billerica, MA) at 6500 × g at 4°C. Protein concentration was determined using BCA assay (Thermo Scientific, Rockford, IL), and 20 μg of the protein from each sample was electrophoresed on a 10% zymography gel containing 0.1% gelatin (Invitrogen, San Diego, CA). MMP activity was detected by incubating the gel in 1× Zymogram Renaturing Buffer for 30 min at room temperature and then equilibrating the gel for 30 min at room temperature with gentle agitation. The gel was incubated with fresh 1× Zymogram Developing Buffer overnight, followed by staining with Coomassie Blue for 30 min. Contrast was adjusted by destaining with Coomassie destaining solution (Methanol: Acetic acid: Water (50: 10: 40). The staining gels were then air-dried in cellophane mounts and images were captured.
Wound Healing Assay
The cells were seeded in a 12-well plate (15 × 104). After the cells formed a confluent mono layer, scratches were performed using a 100 ul tip. The culture medium was replaced with fresh complete medium. The closure of scratch was analyzed under the microscope and images were captured at 18 - 24 hr after incubation.
Invasion Assay
The cells were seeded at 6 × 104 cells per well into the 96-well plate of an Oris™ Cell Invasion Assay Kit (Platypus, Madison, WI). The plate was incubated for ~16 hr at 37°C. The stoppers were then removed. Collagen I Overlay was added to create a 3-D ECM environment for invasion and incubated for 1 hr at 37°C. Cell culture medium was added and the cells were allowed to invade for 72 hr, and were stained with DAPI before images were captured.
Statistical Analysis
qPCR data and cell growth data were analyzed by the Student's t-test (one-tailed). P < 0.05 was used to define statistically significant differences.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BU and WSW designed the experiments, participated in discussion of the data and draft of the manuscript. BU conducted experiments. All authors read and approved the final manuscript.